fe-safe 2017 - abaqus-docs.mit.edu · pdf file4.3 master s-n curve f ... systemes simulia...

VERITY MODULE IN fe-safefe-safe 2017

Contents

Copyright © 2016 Dassault Systemes Simulia Corp. Verity® in fe-safe V2 User Manual 1

Issue: 12.0 Date: 08.09.2016

Contents

1 INTRODUCTION TO VERITY IN FE-SAFE .................................................................................. 7

1.1 ABOUT FE-SAFE ..................................................................................................................... 7

1.2 ABOUT VERITY IN FE-SAFE ...................................................................................................... 7

1.3 HOW TO USE THIS MANUAL .................................................................................................... 8

2 USING VERITY IN FE-SAFE .................................................................................................. 9

2.1 OPENING AN FEA MODEL ...................................................................................................... 9

2.1.1 Selecting the Datasets .......................................................................................................... 9

2.1.2 Defining the Welds ............................................................................................................. 10

2.1.3 Evaluating the structural stresses ....................................................................................... 10

2.2 ANALYSING THE MODEL ....................................................................................................... 11

2.2.1 Defining the Loading ........................................................................................................... 12

2.2.2 Defining the Materials ......................................................................................................... 12

2.2.3 Performing the Analysis ...................................................................................................... 13

3 INTERFACE REFERENCE ................................................................................................... 17

3.1 WELD DEFINITIONS .............................................................................................................. 17

3.1.1 The Weld Tree .................................................................................................................... 17

3.1.2 The automatic weld finder ................................................................................................... 19

3.1.3 Spot Weld Parameters ........................................................................................................ 31

3.1.4 Legacy Line Definition Parameters ..................................................................................... 33

3.1.5 Group selection and editing ................................................................................................ 41

3.1.6 Model Units ........................................................................................................................ 42

3.1.7 Other Options ..................................................................................................................... 42

3.2 DIAGNOSTICS ...................................................................................................................... 43

3.2.1 Structural Stresses Table.................................................................................................... 43

3.2.2 Worst cycle mean stress and damage parameter contour................................................... 45

3.2.3 Log for Items ...................................................................................................................... 46

3.2.4 Histories for Items ............................................................................................................... 46

3.3 WELD MASTER S-N CURVES ................................................................................................. 48

4 THE VERITYSSM METHOD ............................................................................................... 51

4.1 APPROACHES TO EVALUATION OF WELDED JOINTS ................................................................. 51

4.2 THE STRUCTURAL STRESS METHOD ...................................................................................... 53

4.2.1 Shell/Plate Element Procedures ......................................................................................... 54

4.2.2 Calculation Examples ......................................................................................................... 55

4.3 MASTER S-N CURVE FORMULATION ...................................................................................... 57

4.3.1 Structural Stress Based K Estimation ................................................................................. 57

4.3.2 A Two-Stage Growth Model ................................................................................................ 59

4.3.3 Equivalent Structural Stress Parameter .............................................................................. 60

4.3.4 Initial Crack Size Effects ..................................................................................................... 62

4.4 S-N DATA CORRELATION ..................................................................................................... 63

4.5 A SIMPLE EXAMPLE OF THE STRUCTURAL STRESS METHOD .................................................... 66

Contents

2 Verity® in fe-safe V2 User Manual Copyright © 2016 Dassault Systemes Simulia Corp.

Issue: 12.0 Date: 08.09.2016

4.6 PROPORTIONAL COMBINATION OF NORMAL AND SHEAR EQUIVALENT STRUCTURAL STRESSERROR! BOOKMARK NOT

DEFINED.

4.7 MULTIAXIAL COMBINATION OF NORMAL AND SHEAR EQUIVALENT STRUCTURAL STRESSERROR! BOOKMARK NOT

DEFINED.

4.8 MULTIAXIAL S-N CURVE NORMALISATION .................................. ERROR! BOOKMARK NOT DEFINED.

4.9 MULTIAXIAL PARAMETER CONFIGURATION ................................ ERROR! BOOKMARK NOT DEFINED.

5 PREPARING FE MODELS FOR USE WITH VERITY® ................................................................ 69

5.1 SUPPORTED FE PROGRAMS ................................................................................................. 69

5.1.1 Abaqus ............................................................................................................................... 70

5.1.2 Nastran ............................................................................................................................... 75

5.1.3 I-DEAS ............................................................................................................................... 79

5.1.4 Ansys ................................................................................................................................. 81

5.2 MODELLING CONSIDERATIONS AND APPLICATIONS .................................................................. 82

5.2.1 Full Penetration Welds ........................................................................................................ 82

5.2.2 Partial Penetration Weld and Root Failures......................................................................... 83

5.2.3 Lap Fillet Weld and Failure Modes ...................................................................................... 89

5.2.4 Fillet Weld between Two Plates with Dissimilar Thicknesses .............................................. 91

5.2.5 Weld Line Curvature Modelling ........................................................................................... 93

5.2.6 Calculations without Modelling Weld ................................................................................... 96

6 TUTORIAL A: SINGLE WELD LINE (USING ABAQUS) ............................................................ 101

6.1 INTRODUCTION................................................................................................................... 101

6.2 PREPARATION .................................................................................................................... 101

6.3 OPEN THE MODEL .............................................................................................................. 102

6.3.1 Select the Datasets ........................................................................................................... 102

6.3.2 Define the Weld ................................................................................................................ 106

6.3.3 Evaluate the structural stresses ........................................................................................ 113

6.4 ANALYSE THE MODEL ......................................................................................................... 114

6.4.1 Define the Loading ............................................................................................................ 114

6.4.2 Define the Materials .......................................................................................................... 115

6.4.3 Perform the Analysis ......................................................................................................... 116

7 TUTORIAL B: MULTIPLE WELD LINES (USING ANSYS) ......................................................... 119

7.1 INTRODUCTION................................................................................................................... 119

7.2 PREPARATION .................................................................................................................... 119

7.3 OPEN THE MODEL .............................................................................................................. 120


7.3.2 Define the Welds .............................................................................................................. 122






Contents


Issue: 12.0 Date: 08.09.2016

8 TUTORIAL C: SHELL WELD LINES (USING NASTRAN) .......................................................... 131

8.1 INTRODUCTION .................................................................................................................. 131

8.2 PREPARATION ................................................................................................................... 131

8.3 OPEN THE MODEL ............................................................................................................. 132

8.3.1 Select the datasets ........................................................................................................... 132

8.3.2 Define the materials .......................................................................................................... 135

8.3.3 Define the welds ............................................................................................................... 135


8.4 ANALYSE THE MODEL ........................................................................................................ 138

8.4.1 Define the loading ............................................................................................................. 138

8.4.2 Define the algorithm ......................................................................................................... 139

8.4.3 Define the Weld material .................................................................................................. 140

8.4.4 Perform the analysis ......................................................................................................... 141

9 TUTORIAL D: SPOT WELDS (DETECT ALL CONNECTING ELEMENTS) ...................................... 143

9.1 INTRODUCTION .................................................................................................................. 143

9.2 PREPARATION ................................................................................................................... 144

9.3 OPEN THE MODEL ............................................................................................................. 144

9.3.1 Select the Datasets .......................................................................................................... 144

9.3.2 Define the Welds .............................................................................................................. 148



9.4.1 Define the Loading ........................................................................................................... 150


9.4.3 Perform the Analysis......................................................................................................... 153

10 TUTORIAL E: SPOT WELDS (DETECT CONNECTING ELEMENT GROUP) ................................... 155

10.1 INTRODUCTION .................................................................................................................. 155

10.2 PREPARATION ................................................................................................................... 156

10.3 OPEN THE MODEL ............................................................................................................. 157


10.3.2 Define the Welds .............................................................................................................. 160



10.4.1 Define the Loading ........................................................................................................... 162


10.4.3 Perform the Analysis......................................................................................................... 165

11 TUTORIAL F: AUTOMATIC WELD FINDER FOR SOLID WELDS (USING ABAQUS) ......................... 167

11.1 INTRODUCTION .................................................................................................................. 167

11.2 PREPARATION ................................................................................................................... 168

11.3 OPEN THE MODEL ............................................................................................................. 169


11.4 FILLET GROUP .................................................................................................................. 174

11.5 DEFINE THE WELD ............................................................................................................. 175

Contents


Issue: 12.0 Date: 08.09.2016



11.6.2 Limit analysis to the verity weld toe groups ....................................................................... 182


12 TUTORIAL G: AUTOMATIC WELD FINDER FOR SOLID WELDS (USING ANSYS) .......................... 185

12.1 INTRODUCTION................................................................................................................... 185

12.2 PREPARATION .................................................................................................................... 186

12.3 OPEN THE MODEL .............................................................................................................. 187


12.4 FILLET GROUP ................................................................................................................... 189

12.5 DEFINE THE WELDS ............................................................................................................ 190



12.6.2 Limit analysis to the Verity groups ..................................................................................... 199


13 TUTORIAL H: AUTOMATIC WELD FINDER FOR SHELL WELDS (USING NASTRAN) ..................... 203

13.1 INTRODUCTION................................................................................................................... 203

13.2 PREPARATION .................................................................................................................... 204

13.3 OPEN THE MODEL .............................................................................................................. 204

13.3.1 Select the datasets ........................................................................................................... 204

13.3.2 Define the welds ............................................................................................................... 207


13.4.1 Define the loading ............................................................................................................. 210

13.4.2 Define the algorithm .......................................................................................................... 211



13.5 SPECIFIC LINE ANALYSIS ..................................................................................................... 212

13.5.1 Define the line welds ......................................................................................................... 213

13.5.2 Analyse the Model ............................................................................................................ 215



14 REFERENCES ............................................................................................................... 219


Issue: 12.0 Date: 08.09.2016

Trademarks

fe-safe, Abaqus, Isight, Tosca, the 3DS logo, and SIMULIA are commercial trademarks or registered

trademarks of Dassault Systèmes or its subsidiaries in the United States and/or other countries. Use of

any Dassault Systèmes or its subsidiaries trademarks is subject to their express written approval.

Other company, product, and service names may be trademarks or service marks of their respective

owners.

Legal Notices

fe-safe and this documentation may be used or reproduced only in accordance with the terms of the

software license agreement signed by the customer, or, absent such an agreement, the then current

software license agreement to which the documentation relates.

This documentation and the software described in this documentation are subject to change without

prior notice.

Dassault Systèmes and its subsidiaries shall not be responsible for the consequences of any errors or

omissions that may appear in this documentation.

© Dassault Systèmes Simulia Corp, 2016.


Issue: 12.0 Date: 08.09.2016

Third-Party Copyright Notices

Certain portions of fe-safe contain elements subject to copyright owned by the entities listed below.

© Battelle

© Endurica LLC

© Amec Foster Wheeler Nuclear UK Limited

fe-safe Licensed Programs may include open source software components. Source code for these

components is available if required by the license.

The open source software components are grouped under the applicable licensing terms. Where

required, links to common license terms are included below.

IP Asset Name IP Asset

Version

Copyright Notice

Under BSD 2-Clause

UnZip (from

Info-ZIP)

2.4 Copyright (c) 1990-2009 Info-ZIP. All rights

reserved.

Under BSD 3-Clause

Qt Solutions 2.6 Copyright (c) 2014 Digia Plc and/or its

subsidiary(-ies)

All rights reserved.

http://opensource.org/licenses/BSD-2-Clause

http://opensource.org/licenses/BSD-3-Clause

Introduction to Verity® in fe-safe


Issue: 12.0 Date: 08.09.2016

1 Introduction to Verity in fe-safe

1.1 About fe-safe

fe-safe is a powerful, comprehensive and easy-to-use suite of fatigue analysis software for Finite

Element models. It is used alongside commercial FEA software to calculate:

where fatigue cracks will occur;

when fatigue cracks will initiate;

the factors of safety on working stresses (for rapid optimisation);

the probability of survival at different service lives (the ‘warranty claim’ curve).

Results are presented as contour plots which can be plotted using standard FE viewers.

fe-safe has direct interfaces to the leading FEA suites.

1.2 About Verity in fe-safe

Verity in fe-safe is an add-on module for use with fe-safe fatigue analysis software, that computes

equivalent structural stresses based on the Battelle Structural Stress Method, and uses these stresses

to calculate fatigue damage and corresponding fatigue lives. Users of Verity in fe-safe are assumed to

have a working knowledge of fe-safe, including such techniques as configuring a fatigue analysis and

setting properties for different parts of the model, defining the fatigue loading, running an analysis and

exporting fatigue results. The use and application of fe-safe is described in the fe-safe User guide,

which should be referred to alongside the Verity in fe-safe manual.

Features of Verity in fe-safe include:

Support for Finite Element solutions from Abaqus FIL and ODB, Nastran OP2, I-DEAS UNV and

Ansys RST files.Automatic identification of weld line and through-thickness connectivity.

Automatic detection of spot weld nuggets and their associated sheets.

Support for a wide range of FE dimensionalities. The FE mesh in the vicinity of the weld line may be

comprised of 3D solids (either hexahedral/pentahedral or tetrahedral) or 3D mid-surface shells.

Beam and 1D solid elements may be used in the mesh, but must not be included in the weld

element domain that contributes nodal forces to the nodes along the weld line and is used to derive

the weld line connectivity.

Consistent structural stress calculations regardless of mesh sizes, element types, or the FE

software used (a critical step to realising 6 sigma in CAE applications).

Structural stress calculations for both normal and Mode III (in plane) shear forces.

Use of a single master S-N curve, eliminating the need to determine ‘weld classification’ (required

by other weld fatigue approaches).

Support for element types of any dimensionality that feature linear or quadratic order.

The ability to process multiple load cases from multiple results files.

Introduction to Verity® in fe-safe


Issue: 12.0 Date: 08.09.2016

The ability to process load cases from linear or nonlinear static analyses.

User-specified rotations of the weld line local coordinate system (to accommodate structural stress

computations on through-thickness sections that are not normal to the plate surfaces).

If the reference surface is offset from the mid-surface during an analysis involving 3D shells, this

offset may then be input so that the structural stress calculations will account for the offset.

User-specified symmetry at either end of the weld line that will de-activate virtual node weld end

correction.

Support for further improving the accuracy of fatigue life predictions by allowing the user to

formulate their own custom, master S-N curve based on proprietary data.

Compatibility with tools in PreProcessors such as Ansa that automatically identifies and generates

the required weld definitions.

1.3 How to Use This Manual

Chapter 2 (Using Verity in fe-safe) provides an overview of the actions required in order to perform

Verity or LBF analysis in fe-safe, while chapter 3 (Interface Reference) offers further details of the

available options. The Verity Method itself is discussed in chapter 0. The prerequisites for Verity

analysis are explained in chapter 0 (Preparing FE Models for use with Verity®). Worked examples of

the process are provided in several tutorials.

Users new to fe-safe

Because this manual assumes some familiarity with fe-safe, it will be necessary to learn a little about

the main program first. Work through some of the tutorials in the fe-safe User guide, including at least

one demonstrating the use of files from your preferred FEA software, then return here.

fe-safe users new to Verity®

Some understanding of the Verity method is required to make the best use of the software. Chapter 0

provides a description for those unfamiliar with the method.

Work through one or more tutorials including at least one demonstrating the use of files from your

preferred FEA software . Then read chapter 0 to find out how to prepare your own FE models for use

with Verity®. Once you have done this, follow the procedure described in chapter 2 with your own

models, referring to chapter 3 as necessary.

Experienced users of Verity in fe-safe

Experienced users are most likely to refer to chapter 3, which provides a detailed interface reference,

including descriptions of infrequently-used parameters.

Preparing FE Models for use with Verity®


Issue: 12.0 Date: 08.09.2016

2 Using Verity in fe-safe

2.1 Opening an FEA Model

To perform a fatigue analysis using this technique, an FE model must be opened using the File menu

item Open Finite Element Model....

The model to be read is then selected in an open file dialog box.

2.1.1 Selecting the Datasets

If the model is being read for the first time, fe-safe pre-scans the model to see what it contains. A pre-

scan dialog box is displayed allowing the selection of the stresses and forces to read:

Figure 2-1

To perform a Verity based analysis some datasets containing Forces for Verity must be selected in

addition to datasets containing Stress.

Select the required forces and stresses and click OK.

fe-safe will read the selected data and echo to the Message Log a summary of the datasets that have

been read.



Issue: 12.0 Date: 08.09.2016

2.1.2 Defining the Welds

Once the stresses and forces have been read the weld locations should be defined prior to evaluation

of the structural stresses. This can be done using the Weld Preparation >> Define and Analyse Weld

Geometry menu and a new dialog box will be displayed:

Figure 2-2

Remove any unwanted weld definitions using the Delete… or Delete All… buttons, then add welds as

necessary using the Add… button.

For further options, see section 3.1.

2.1.3 Evaluating the structural stresses

Once all weld definitions are complete click the Save and analyse button to start the calculation of the

structural stresses. fe-safe will locate the weld line domain and identify the local coordinate system at

each weld line node, and then run the Verity module to calculate structural stresses and echo the Verity

outputs to the message log, as well as the following information:



Issue: 12.0 Date: 08.09.2016

Processing foces dataset 1.1 at time 2.22e-016.

Saving structural stresses to FED file ...

Verity stress generation and import completed.

This tells you that fe-safe is importing the structural stresses calculated by Verity®. When subsequently

running the analysis if a node/element is part of a Verity group, the stress dataset(s) defined in the

loading are used to look up the correct structural stress dataset(s).

Once the import is completed the Current FE Models window should show one or more groups

beginning with WELD (one for each failure mode) which contain the elements for the weld lines. A

summary of the Verity analysis run can be found in the verity-diagnostics.log file (Weld

preparation >> View Structural Stress Diagnostics Log menu item) and a table of the structural

stresses calculated can be found in the verity-ss.txt file (Weld preparation>> View Structural

Stress File menu item).

2.2 Analysing the Model

After the import process has completed the WELD groups will appear in the Group Parameters section

in the Fatigue from FEA window. These groups have limited configuration options, the algorithm for

these groups can be a structural stress method weld analysis for either normal or shear stresses, or Do

not analyse. If the algorithm cell for a WELD group is selected then the following dialog box is

displayed, from which this selection is made via the respective radio button.



Issue: 12.0 Date: 08.09.2016

2.2.1 Defining the Loading

To define the type of loading for the analysis, select the Loading Settings tab on the Fatigue from FEA

dialog box.

Figure 2-3

Remove any unwanted loadings by right clicking on the tree and selecting Clear all loadings, then add

the desired loadings using the Add… button. When finished click OK.

2.2.2 Defining the Materials

Materials from the verity.dbase are provided for use with the structural stress algorithms. This

currently includes 8 materials:

Steel Weld (50%)

Steel Weld –2SD (2.2%)

Steel Weld –3SD (0.12%)

Steel Weld/Throat (50%)

Steel Weld/Throat –2SD (2.2%)

Steel Weld/Throat –3SD (0.12%)

Al alloy Weld (50%)

Al alloy Weld –2SD (2.2%)



Issue: 12.0 Date: 08.09.2016

The material will default to the steel mean life curve each time a new Verity structural stress calculation

is performed. Note that in fe-safe 6.5-03 and 2016 versions, there were separate materials for Steel

Weld shear. In fe-safe 2017 the Verity materials contain two S-N curves instead: the normal S-N curve

and a new attribute storing the T-N curve for the shear response. For the Steel Weld (50%), Steel Weld

-2SD and -3SD materials these are equivalent to the S-N curves of the previous respective Steel Weld

Shear materials. Fe-safe automatically picks the appropriate curve for the selected algorithm. For other

material classes, (Root or Throat failures and aluminium alloy welds) Battelle have not published the

shear response data. The T-N curves used in the shear damage calculation have therefore been

calculated from the normal S-N curve assuming a Von Mises assumption (as recommended by Dong

and Hong in [1]). So in effect in the absence of a published T-N curve for shear, the shear algorithm

multiplies the shear amplitude by √ and then uses the normal S-N curve.

2.2.3 Performing the Analysis

Once the WELD group is configured, click the Analyse button. The FEA Fatigue Analysis Summary

dialog box is displayed, showing the parameters that will be used for the analysis:

Figure 2-4



Issue: 12.0 Date: 08.09.2016

Click Continue to begin analysis. As the analysis is being performed the worst life calculated so far will

be displayed in the Message Log window.

Once the analysis is complete, a results file will be produced by default containing the log10 (life)

contours. If a node is included in the analysis of more than one weld, the result exported for the node is

the worst-case result if no failure mode had been specified, although separate results can be

maintained where the mode weld line domains are defined on different elements (see 3.1.4) and failure

mode is specified. For example the fusion domain can be defined on the fillet elements of the weld line,

and the toe domain on the parent plate side of the line. Note that if the automatic weld finding capability

is used then this will automatically happen (but see compact contour mode settings for root mode in

3.1.2). Additional contours export options can be configured in the FEA Fatigue >> Analysis Options

dialog box, Export tab:

Figure 2-5



Issue: 12.0 Date: 08.09.2016

All nodes on an element use element’s worst value: all nodes on an element will be assigned the

worst contour value of that element.

All nodes on an element use element’s averaged value: all nodes on an element will be assigned

the average contour value of that element.

Skipped nodes use element’s worst contour value [Default]: nodes not analysed (e.g. those of weld

elements which are not part of the weld line) will be assigned the worst contour value of the

remaining nodes on element.

Skipped nodes use element’s averaged contour value: nodes not analysed (e.g. those of weld

elements which are not part of the weld line) will be assigned the average contour value of the

remaining nodes on element.

Skipped nodes use default contour value: nodes not analysed (e.g. those of weld elements which

are not part of the weld line) will be assigned the default contour value:

Contour Default value

Life / Log of Life Infinite life

FOS Max FOS

FRF 10

Dang Van - radial 10

Dang Van - vertical 10

Dang Van - survived 1

Probabilities -100

SRP PP 0

SRP PC 1

SRP CP 2

SRP CC 3

All other contours 0

Skipped nodes are not exported: nodes not analysed (e.g. those of weld elements which are not

part of the weld line) will not be assigned any value.

More detailed description of all configuration options can be found in Chapter 5 of the fe-safe User

Guide.



Issue: 12.0 Date: 08.09.2016



Issue: 12.0 Date: 08.09.2016

3 Interface Reference

Note that only those interface elements of particular importance for Verity Analysis are described here.

3.1 Weld Definitions

The weld definition dialog box can be displayed after the datasets have been read by using the Weld

Preparation >> Define and Analyse Weld Geometry menu:

Figure 3-1

3.1.1 The Weld Tree

The weld tree is used to define all the welds and the parameters associated with each of the welds. The

root tree items with the icon are line welds (either individual lines or groups of automatically located

lines defined via a fillet group), while those with the icon represent a set of spot welds. The name of

a weld can be edited by selecting the weld and then clicking the Edit button or double-clicking the weld

name. The weld names must be unique.



Issue: 12.0 Date: 08.09.2016

Prior to fe-safe 6.5-02, each weld line and failure mode combination was defined individually as a

separate weld in the tree, and required a domain group and set of line nodes to be explicitly specified.

This method is still available, but is now referred to as a Legacy Line Definition. The alternative is to

define a set of lines and/or failure modes in a single weld definition in the tree, by providing a weld fillet

group, and much of the previously required data is automatically derived by fe-safe (see 3.1.2).

For each weld there are a series of children defining the parameters for a weld. Those with a icon

are required. The list of the required parameters for a weld is a function of the type of elements used

within the FEA model.

If the checkbox Show all configuration options for each weld is checked then a series of optional

parameters are also shown. These are marked with a icon. The list of optional parameters is affected

by the status of the open weld parameter as well as the element type.

Buttons to the right of the weld tree allow welds to be modified, stored and retrieved:

Add… creates a new weld definition using default settings. The type (Line Welds (From Weld

Fillet Elements), Spot (Detect All Connecting Elements), Spot (Detect Connecting Element

Group) or Legacy Line Definition) is selected from a popup menu.

Note: Line Welds has replaced the Solid Weld option prior to fe-safe version 2016, and now

includes Shell elements.

The new Legacy Line Definition now performs the functionality that was previously termed Line

Welds prior to fe-safe version 2016.

Delete removes the currently selected weld.

Duplicate creates a copy of the currently selected weld, which can then be edited. This is

particularly useful when there are welds of the same type but with different nodes (for example,

the two sides of a weld).

Delete All… removes all welds.

Edit… allows the currently selected aspect of a weld

Open… replaces the current tree with a set of weld definitions stored on disk.

Append… preserves the current tree, but adds the weld definitions from a file. This will only

succeed if there are no duplicate definitions.

Save All… writes the current tree to a file.

Save One… writes the currently selected weld to a file.

All the possible parameters that can be defined for a weld are documented below. To start editing a

parameter select it and then click the Edit button. Text items can be edited in place in the tree control.

Items that have a pre-defined list of options will display a dialog box containing the list to select from.



Issue: 12.0 Date: 08.09.2016

3.1.2 The automatic weld finder

Weld lines on either solid or shell elements can be defined automatically as an alternative to the legacy

weld line method (see Section 3.1.4). The intention is to provide a simpler way of defining the weld

domain and weld lines, especially when multiple weld lines are to be analysed, or several different

failure modes (Toe, Fusion, Throat, Root). Instead of inputting the weld domain and detailed information

on the weld line nodes for each line and failure mode, the fillet group is defined instead. The fillet group

may already be available in fe-safe, for example as a Set, Material, or Property group defined in the FE

model or it may be a user-defined group in fe-safe. The various weld lines and weld domains are then

automatically computed by fe-safe, although it is still necessary to provide a domain group for root

failures on solid element welds. Multiple weld fillets in the model can be supplied as a single fillet group,

including both solid and shell elements (though a mix of shell and solid elements is not allowed in the

same weld), and the software will automatically locate the separate lines in the various fillets, although it

is possible that some complex geometries involving intersecting lines in branching fillets may contain

some ambiguities, which would require definition of separate fillet groups. fe-safe locates potential weld

lines by identifying nodes connected by edges lying along the interface between fillet and non-fillet

elements. The root lines are also located, either because these are formed from internal surface nodes

in fully penetrative welds, or lie on a fillet-parent material interface where there are two faces with

significant change in normal direction.

To invoke this option, click the Add button on the main Weld Definition dialog and select Weld Lines

(From Weld Fillet Elements) from the resulting popup menu. This will result in the weld definition dialog

displayed below. The only compulsory input is the fillet group. Note this group specifies all weld fillet

elements, not the weld domain. For toe, fusion, or throat failure modes, the weld domains can be

automatically extracted.

For root failures the weld domain can be detected automatically for shell welds, but for solid welds a

root domain group must be specified. If multiple root lines are to be analysed, this root domain group

covers the combined domain of all of them. Note that the specified root domain may be a superset of

the actual weld domain; the weld finder will determine which elements from it actually lie on the failure

plane.



Issue: 12.0 Date: 08.09.2016

Figure 3-3

Double clicking on the value field of the Fillet Group parameter invokes the Group definition dialog (see

section 3.1.5), and allows selection of a loaded group or user-defined group in the fe-safe Group

Manager. By default only Toe and Fusion failures are automatically analysed; use the optional

parameters (Auto-detect Fusion failures, etc) to configure automatic extraction of Toe, Fusion, or Throat

modes. Alternatively explicit domains can be defined for any of these. Note that if a Toe or Fusion

domain is specified which relates only to one line of a fillet, but auto-extraction is on for that mode, then

the other line will still have its domain automatically extracted. If detailed diagnostics are specified on

the Weld Definition dialog (see section 3.1.7), then a comprehensive set of groups will be created in the

Group Manager (in the list of unused groups) specifying the nodes and elements of each line, and the

located domain.



Issue: 12.0 Date: 08.09.2016

Recall that if the checkbox Show all configuration options for each weld is checked then a series of

optional parameters are also shown. These are marked with a icon. Descriptions of each parameter

are included below.

Fillet Group

This is a list of elements defining the weld fillet material, (i.e., the filler material added in the process of

creating a line weld).

Auto-detect Fusion Failures

This option can be turned On or Off (default is On), and automatically identifies fusion plane failure

modes.

Auto-Detect Shell Root Failures

This option can be turned On or Off (default is Off), and automatically identifies weld root failure modes

(for shell welds only).

Auto-detect Throat Failures

This option can be turned On or Off (default is Off), and automatically identifies weld throat failure

modes.

Auto-detect Toe Failures

This option can be turned On or Off (default is On), and automatically identifies weld toe failure modes.

Compact contour mode

This option can be turned On or Off (default is On). It applies to sub-surface line elements on solid

welds, namely root failures, or throat failures analysed from the root. When On, this results in fatigue

results for the root line solid elements being projected onto the surface line(s) for easier viewing in the

Output file – i.e. there is no need to view/cut the results, in order to see the life contour. If throat failure

mode is analysed, then the root results will be projected onto the surface throat failure line; otherwise

they are projected onto both surface weld lines at the edges of the fillet. Note that the same element

can have results from several failure modes (eg. a fusion result and a projected compact contour result

from root failure), and fe-safe will pick the worst for the final result. If throat failures are analysed from

both surface and root using automatic domain identification, and compact contour mode is on, then both

results are displayed on different sides of the line (e.g. the domains discussed below in Throat Failure

Sense). However if the domain had been explicitly specified then only that one side is available, so only

one set of throat results will be displayed which will be the worst of the two.



Issue: 12.0 Date: 08.09.2016

Crack Length

As in 3.1.4 the default value of this parameter is set at crack length = plate thickness. There is a slight

difference relating to the default value, due to differences in the plate thickness used as a default for

this parameter. In Automatic weld finder for solid welds, the plate thickness parameter may be extracted

from the model rather than pre-configured. If this happens the plate thickness may vary somewhat

along the length of the weld line, so the maximum plate thickness value is used as default.

Domain Constraint

This is used in automatic domain location for solid welds, and is ignored when an explicitly specified

domain is in use. If left at the default Unconstrained, then the search for through nodes in the weld

domain will continue all the way through the model until a surface is reached, or until no further node or

element edge is found aligning with the failure plane (which would result in diagnostic warnings).

Alternatively use the popup menu after double-clicking on this value field to select Constrained by Plate

Thickness, which will terminate the domain search at a maximum distance from the weld line, specified

in the Plate Thickness option. Note that in Solid Welds, this plate thickness may not be exactly what is

used in the Verity structural stress calculations (see Structual Stress Plate Thickness option below).

Failure Plane Co-planarity Tolerance

This is used to specify a tolerance in degrees for the angle between the calculated failure plane normal,

and that of a candidate element face. This is used to test potential domain solid elements if no

candidate nodes are located for that element within the nodal tolerance cone (see also Through

Thickness Cone Tolerance). The domain location algorithm would then attempt to find aligned points

close to edge intersections which are within the cone tolerance distance of the edge, but only if the

edge lies on an element face within this parameter’s tolerance of the failure plane. The nodal forces are

then interpolated (potentially between a corner and a mid-side node on the edge). Note that solid

elements should ideally be meshed so that the weld domain contains a set of elements with faces

aligning along the failure plane. However automatic tet meshers are not guaranteed to produce this.

This parameter is intended to allow small violations of this co-planarity assumption. However this should

not be used to set large tolerance values so that arbitrary meshes can be processed in Verity, as testing

indicates that meaningful results can be obtained only if the through-thickness mesh is regular.

Fusion Domain Group

This is used to override the automatic fusion domain extraction by explicitly specifying the domain. The

automatic algorithm will extract a domain on the fillet side of the failure plane. This can be used to place

the domain on the parent material side, or to vary the plate thickness along the line. If a domain is

specified but some elements are never included, then an unused group is created which consists of all

domain group elements which have not been placed in the weld domain (for example their nodes, or

faces may not align, see Failure Plane Co-Planarity Tolerance and Through Thickness Cone

Tolerance).



Issue: 12.0 Date: 08.09.2016

I(r) Function

As part of the fatigue life calculation procedure, Verity uses a dimensionless polynomial function of the

bending ratio, (see Section 4.3.3). The following equations are available to the user for analysing

normal stresses:

which have been designated as “Structural Joints / Displacement Control” and “Small Details / Load

Control”, respectively. The “Structural Joints / Displacement Control” equation is the default option in

the Verity GUI. This ensures the best possible correlation for the engineering applications that are

usually analysed in Verity (as opposed to the less common assessment of test specimens). Note that

for backward compatibility with previous versions of Verity the user must set this parameter to “Small

Details / Load Control”. Note that when shear stresses are analysed then this selection is ignored, and

instead a fixed shear polynomial I(r) function is used given in section 4.3.5.

Number of smoothing iterations

Sometimes a line weld configuration can yield large structural stress fluctuations. To address this issue

a structural stress smoothing algorithm is available, which provides the ability to smooth the resultant

structural stress profile. This algorithm is cyclic in nature and up to 10 iterations can be applied. The

default value is zero, i.e. the smoothing function is disabled.

Plate Thickness

The plate thickness may be specified, in the same units as those selected for the FEA model distance

units using Loaded FEA Model Properties dialog box accessed from the Current FE Models window in

the fe-safe GUI. With shell welds for FE models that do not contain shell thickness this parameter must

be specified, for solid welds this is always optional.

For solid welds, if a domain has been set it is not necessary to specify this, nor if the desired domain

goes all the way through the model to a surface. If set, then this parameter has two potential uses,

depending on the selections of the Domain Constrained and Stuctural Stress Plate Thickness settings.

The former relates to its use in determining the maximum thickness of the weld domain anywhere along

the weld line, whereas the latter determines whether it is used in the Verity structural stress calculations

as a constant along the line (rather than the actual distance determined by the local domain thickness).

See also Domain Constrained and Stuctural Stress Plate Thickness. Note that if the plate thickness is

set, then normally one or both of these options should be changed, as the default will be to ignore this

parameter.



Issue: 12.0 Date: 08.09.2016

Reference Length

The reference length value defines the parameter L used in the virtual node weld end correction, and is

used only when this correction is being applied. The reference length is the parameter L in the

derivation of the weld end correction. If the specified reference length is not greater than or equal to 2.5

times the crack length, it will be reset to this value. If the weld line is closed or if the symmetry flag has

been set at both ends, the reference length is not needed.

Root Domain Group

If root failures are to be analysed then the root domain (same group for all root lines) must be specified.

Note that the selected elements for solid welds must be the ones on the fillet side of the continuation of

the parent/fillet interface plane (i.e. directly under the back of the fillet). Although in principle the

elements on the other side could be used in the Verity calculations, and the same results should be

obtained (if forces are validly balanced), in this context that would create an ambiguity in the algorithm

used to determine the failure plane. The root failure plane is aligned to an element’s face in the root

domain that has an edge on the root line and is adjacent to an element that is neither in the root domain

nor the weld fillet. Note that this requires the root domain to be directly under the fillet, not under the

continuation of the fusion plane on the parent material side (as the latter would have two such faces,

one in each fusion plane, rendering it ambiguous). . Note that the specified root domain may be a

superset of the actual weld domain; the weld finder will determine which elements from it actually lie on

the failure plane.

Rotate LCS After Solve

This flag indicates whether the nodal forces and moments along the weld line and through-thickness (if

necessary) will be rotated into the local co-ordinate system before solution of the weld line system of

equations, or the resultant line forces and moments will be rotated into the local co-ordinate system

after system solution.

By default, Verity will rotate the nodal forces and moments accumulated at the weld line and through-

thickness nodes (if necessary) before the solution of the weld line system of equations and application

of the weld end correction. In this case, only the local x and y line forces and moments will be computed

during system solution and subjected to the weld end correction (if necessary). If the user specifies that

the rotation be performed after the system solution, the global line forces and moments of all

components will be computed from the accumulated global nodal forces during system solution. After

the system solution and end correction (if necessary), the resultant line forces and moments will be

rotated into the local co-ordinate system. In either case, only the computed line forces and moments in

the local x and y directions are used to compute the normal and in-plane structural stresses.



Issue: 12.0 Date: 08.09.2016

Start Element or Node (for Symmetry Start)

This option is used to resolve ambiguities in weld line sense when Weld Symmetry has been set to

Start or End. In the case that weld symmetry is used, Verity needs to know which end to apply the

symmetry setting to (see Weld Symmetry). This can only be used when the fillet group contains a single

fillet, although there can be multiple (connected) lines induced by large angle changes (in excess of 45

degrees). Either of the surface lines or root line can be used, and the other lines will be automatically

aligned in the same sense. Specify either the start element or node corresponding to the end of the

weld line corresponding to the Start, for use in applying the Weld Symmetry setting. If the direction

changes by more then 45 degrees between two nodes, then a new line is created, but the direction

sense will be continued into connected split lines from the line containing the specified start item.

However weld ends internal to a series of weld lines, induced by such splitting will always have

symmetry on at the junction (and likewise if the end is embedded in the parent material). So for

example if a weld line goes from nodes 1,2,…50 followed by a large angle change and a new line from

nodes 51 to 100, then specifying Weld Symmetry as Start and start element 1, will result in symmetry

on at nodes 1,50 and 51, but not 100. This is the same as specifying Weld Symmetry to End, and

setting the Start Node to 100 instead. However note that in automatically determined throat failures this

can result in swapping over the domain to the other side of the failure plane, as the choice of element

parity depends on the line direction sense, and the lines will be reordered so that they start in the sense

of the specified start item.

The start item value is first interpreted as an element ID. If no such element exists, or if no line has an

end at such an element, then the software tries to reinterpret it as a node ID. If no such node exists

either, or it is not the start (or end) of a surface weld line, then an error occurs. If the line consists of a

single element then the value must be a node.

Note that if Weld Symmetry is Both or None, then there is no need to set this parameter.

Structural Stress Plate Thickness

This option is used to select what plate thickness is used in the Verity structural stress calculations. By

default the determined domain geometry is used to calculate a node-by-node thickness which may vary

along the weld line (Derived from Domain Geometry). This popup menu can be used to select the

alternative From Configured Plate Thickness option, which will result in Verity using the value specified

in the Plate Thickness option for every line node. This is mainly intended as a backwards compatability

option with weld lines specified as per section 3.1.4.



Issue: 12.0 Date: 08.09.2016

Throat Domain Group

If throat failures are to be analysed then the throat domain may optionally be specified. For solid welds

the default surface throat line(s) contains central nodes on the fillet surface which are equidistant from

two paired surface weld lines (to within a small tolerance). The failure plane direction is from this line to

the root line. There should be a sequence of element faces which are aligned with this, so mid-nodes

cannot be used, and there should be an even number of elements across the fillet surface for the

automatic method. Alternatively if the geometric centre of the fillet is not the desired failure plane, or

there is an odd number of elements across the fillet, then the domain can be explicitly specified, and

the failure plane will be the set of faces closest to the automatic central plane. Use the popup menu to

select the group, if needed.

For shell welds the analysis of the throat should occur at both sides of the fillet where the weld line

interfaces with the parent material.

Throat Failure Sense

Solid Weld throat failures can be analysed from either the surface line to the root or vice versa, or in

both senses. Use the popup menu to select the sense. The selection of this sense will determine which

throat domain is identified (see Figure 3-6 below). The local Verity coordinate system (Vx, Vy, Vz) as

discussed below in Throat Wrap Around Line Angle determines this domain because Vz will be inverted,

and Vx must conform to a right-hand-rule.

The default setting for this parameter is From Surface. See also Compact Contour Mode for display of

root line results projected onto the surface if From Root or Both is used.

Throat Wrap Around Line Angle

This angle is used to determine what additional solid elements are included in the domain at the base of

a throat failure of a solid weld. This is only used if automatic throat failure location is on and no explicit

throat domain group was specified. The normal criteria for including elements in the domain is that they

should be on the correct side of the failure plane, which in automatic mode means the side pointed to

by the local Verity x-axis (Vx) in a right handed coordinate system where the Verity z-axis (Vz) is the

through thickness direction, and the Verity y axis (Vy) is the local line direction. Vz is the red vector on

Figure 3-5 below, if the failure sense is from surface to root. However at the root nodes which lie on the

fillet/parent plane interface, this rotation can be used to further rotate the local x-axis so it points into the

root parent plate. Additional elements sharing the root node which also lie below the corresponding

rotated plane are included in the weld domain.



Issue: 12.0 Date: 08.09.2016

Note that a negative angle will act to reduce the domain. The positive rotation sense will be selected by

the software to be either clockwise or anti-clockwise depending on whether (Vx,w) is a left or right

handed system respectively, where w is the vector from root to surface (either the Vz axis or its

opposite depending on failure sense). A small tolerance is applied so that some element nodes can be

marginally on the “wrong” side of the plane, as long as the element centre is on the correct side, and no

node is incorrectly positioned by more than 5% of the element scale (mean distance from corners to

centroid).

Take for instance this example specimen. Vz is the red vector on Figure 3-5 below:

Figure 3-4

Figure 3-5



Issue: 12.0 Date: 08.09.2016

There are two potential weld throat domains on either side of the theoretical crack path that has been

modelled from the middle of the weld throat to the weld root (the red arrow in Figure 3-5 above) where

parent material 1 (blue) and parent material 2 (green) meet weld material (grey). Assuming partial

penetration, there are no shared nodes between parent material 1 (blue) and parent material 2 (green).

The possible domains are shown in side view:

Figure 3-6

The weld domain on the left must contain two elements on the parent material 1, whilst the weld domain

on the right can only contain one element in the parent material 2. This is because of connectivity in the

model and shape/size of each element in the base material. Let’s investigate this further:

Figure 3-7



Issue: 12.0 Date: 08.09.2016

Note that while there is one shared (also referenced as common node or equivalenced) node between

the three sets shown above at left, there are still four nodes on the plane between the two parent

material groups whose nodes are not shared (unequivalenced nodes with no interaction). The plane

containing the four nodes at right is a ‘free’ surface inside the mesh, and should not be crossed by the

weld domain. In order to maintain this convention, the weld domain in Figure 3-6 at right has an angle of

-45 while the weld domain in Figure 3-6 at left has an angle of 45 about the weld line. Note that a

positive sign always acts to increase the set of elements in the domain, and a negative sign always

serves to reduce it, so positive means a rotation away from the domain side of the failure plane,

whereas negative means a rotation towards the domain (thus further restricting it). fe-safe automatically

calculates whether this requires a clockwise or anti-clockwise rotation, depending on whether (Vx,w) is

a left or right handed system respectively, where w is the vector from root to surface (either the Vz axis

or its opposite depending on failure sense). Because there is a small tolerance of 5% of element scale it

is not necessary to be absolutely precise on the angle, but for positive rotations with some mesh

irregularities, or weld line curvature, it may be better to use a value slightly larger than what is strictly

required (e.g. +50 rather than +45) to be sure that the rotated axis fully includes all the required

elements. Finally note that in the above example at right the rotation of -45 is not strictly necessary, as

the failure plane bisects the next parent material blue element along its diagonal, and so this would not

be included even with zero angle. However in a mesh including wedge elements or tetrahedral

elements, where the blue element diagonal might form part of an element face, then the -45 rotation

would be necessary to exclude them.

Through Thickness Tolerance Cone

This parameter applies to solid welds and is specified as a proportional displacement (default 0.05)

which defines a cone around a weld line node, centred on the through thickness direction vector

(normal to the local weld line tangent lying in the failure plane). If the height of the through thickness

vector to the current node is h, then the tolerance cone is defined as a search through an angle theta

(Θ) where theta is the inverse tangent of the height times the parameter value:

Θ=arctan(0.05*h) in the case of the default.

This is used to search for through nodes to include in the structural stress computations. Some tet

meshes may require a parameter as high as 0.1 in curved sections of the weld line in order to include

all the necessary nodes, but increasing this too high may result in spurious extra nodes being included

from neighbouring elements, and more accurate results would be achieved by improving the regularity

of the mesh. Note that this parameter is used for all solid weld mesh types, but typically tet meshes may

need a larger value, as they tend to be less regular than hexahedral (brick) meshes.



Issue: 12.0 Date: 08.09.2016

If nodes do not line up with the weld line normal, then virtual nodes will be inserted where the weld line

normal passes within this cone tolerance of an element edge, and the forces will be interpolated

between the nearest nodes on the edge. See also Failure Plane Co-planarity tolerance above. On

curved sections for which high cone tolerances are required because element edges do not quite align

with the local normal, one problem can be that additional unwanted through nodes can be acquired

from neighbouring elements. Therefore a secondary pruning stage is applied, and where subsets of

through nodes are found at a similar projected distance from the weld line (i.e. projected onto the

normal), then only the one closest to the normal is retained. The proximity parameter used for grouping

the set to be pruned nodes whose inter-node projected distance is below 25% of the median inter-node

projected distance in the through node set for the line node.

Toe Domain Group

If toe failures are to be analysed then the toe domain may optionally be specified. The default toe failure

plane is normal to the weld toe face (i.e. the parent material face adjacent to the weld line), and the

domain parity selection is in the sense of above the “top” of the fillet, or right of the the “bottom” of the

fillet (See also Figure 11-9 of Tutorial F below, or Figure 12-9 of Tutorial G below). The automatic

search would continue along this plane until the model is traversed and a surface face is reached, or

possibly the plate thickness is reached (see Domain Constraint above).

Use this option to specify the domain explicitly, which must be done if the failure plane is significantly

not normal to the parent material toe face. This also allows for variable domain thickness along the weld

line, or for the element parity selection sense to be inverted (note this latter would result in the line

results not being visible in the FE model results file without using a view cut or removing the obscuring

fillet elements).

Weld Symmetry

Same as Weld Symmetry in 3.1.4 below. Note that if Start or End is selected then it is necessary to

specify a start element or node as discussed above in Start Element or Node (for Symmetry Start). Start

or End is not available for fillet groups containing multiple sub-fillets, as the application of the start

sense then becomes ambiguous. If weld symmetry is required then multiple solid weld definitions must

be supplied, one per fillet (with separate fillet groups). Also note that any line end which is not on an

open surface will automatically have symmetry on at the appropriate end regardless of the parameter

setting. This also includes lines which end because of a large angle change, or if the end of the line is

embedded in a parent material surface, or if the line is a closed loop. Symmetry on means that no weld

end correction is applied there.



Issue: 12.0 Date: 08.09.2016

Wrap Around Angle

This parameter is used to determine what additional elements are included in the domain at the end of

the weld line and has a range of [-90,90] degrees, where the zero angle refers to the continuation of the

weld line beyond its end (which at the start would be 180 degrees from the weld line direction). The

rotation axis is about the through thickness direction vector, with positive sense towards the weld

domain. On Toe failure modes, a zero value would include one element beyond the line end; -90 would

curtail the domain by excluding this element; whereas +90 results in a wrap around of the domain as

illustrated in Figure 3-8 below. An element is included if all of its nodes are on the domain side of the

rotated line, or if the centre of the element is so located, and no node is more than 10% of the mean

element edge length on the wrong side. Default value is 90 degrees, in line with Verity modelling

guidelines. For hexahedral meshes it would be usual to set this to either 0, 90 or -90; but intermediate

settings such as 45 degrees can result in partial wrap around in tet meshes. Note that 45 degrees

would effectively be the same as 0 degrees in a typical hexahedral mesh, as at least one corner of the

partially included element would still be outside.

Figure 3-8

3.1.3 Spot Weld Parameters

A spot nugget may be modelled in two ways:

as a beam (in Abaqus or Nastran);

as an Abaqus (discrete or point-based) fastener where the weld’saxial direction is represented

by the z-axis of the element’s local coordinate system.



Issue: 12.0 Date: 08.09.2016

as a Nastran CWELD element where the weld’s axial direction is represented by the x-axis of

the element’s local coordinate system. Note that Nastran CWELD elements defined as point-to-

point or point-to-patch are supported for spot weld analysis in fe-safe, but patch-to-patch

CWELD element types are not supported

It is assumed that nuggets are represented exclusively by beams or fasteners.

For a beam or point-based fastener the connection to each sheet must occur at a single node. For a

discrete fastener the connection can occur anywhere on the sheet. For a beam the nodal forces are

extracted from the associated shell elements. For a fastener the forces are extracted from the

associated connector element. Irrespective of nugget representation structural stresses will be

calculated for each sheet and each intermediate nugget segment. However, note that only sheet failure

is considered during the fe-safe analysis phase as the equivalant structural stress calculations are only

applicable to sheet failure.

A weld tree item with the icon designates a spot weld definition. The icon can be added by clicking

the Add… button then selecting one of two spot weld definition options, namely Spot (Detect All

Connecting Elements) and Spot (Detect Connecting Element Group).

The purpose of Spot (Detect All Connecting Elements) is to allow the user to create a single weld

definition for every spot weld in the associated model. This option assumes that the nugget element

under consideration, i.e. a beam or fastener, is used exclusively as a nugget element in the FE model.

After creating the definition (and after clicking the Save and analyse button):

every spot weld nugget is detected;

every spot weld sheet is detected;

a single nugget diameter value is applied to every spot weld.

These extracted values are then used to calculate structural stresses, per spot weld. Only one Spot

(Detect All Connecting Elements) definition can be created by the user.

The Spot (Detect Connecting Element Group) option creates a definition that instructs fe-safe to find

and then process a group (or groups) of spot weld elements. Given a suitably defined group, i.e. a

group of beams or fasteners; the resultant procedure ensures that:

every spot weld sheet is detected;

a single nugget diameter value is applied to every spot weld.

The extracted values are then used to calculate structural stresses, per selected group of spot welds.

Multiple Spot (Detect All Connecting Group) definitions can be created by the user.

Irrespective of definition option the following parameters will be observed under the spot weld definition

icon:



Issue: 12.0 Date: 08.09.2016

Connecting Elements

Specifies the group of spot nugget elements to be assessed. For the Spot (Detect All Connecting

Elements) option no group is required, i.e. every spot weld will be detected and then assessed, so the

option is pre-set to All. For the Spot (Detect Connecting Element Group) option the user selects from a

list of groups (as defined).

Nugget Diameter

The diameter of every spot weld nugget.

Nugget Diameter Estimate

A sheet-thickness dependent nugget diameter estimate that is implemented during the calculation

phase (the results of which are displayed in a Verity output file). When this is turned-on the Nugget

Diameter option is disabled.

I(r) Function

As described above.

3.1.4 Legacy Line Definition Parameters

Element Type

The FE mesh away from the weld line may be comprised of elements of any type. The elements in the

weld element domain that contribute nodal forces to the nodes along the weld line and are used to

derive the weld line connectivity must be of the same dimensionality. Editing the element type will

display a dialog box allowing the selection of one of the valid element types.

Figure 3-2



Issue: 12.0 Date: 08.09.2016

Shells

This category refers to 3D mid-surface shell elements that have no through-thickness dimension. At this

time 3D continuum shell elements that have an explicitly modelled through-thickness dimension are not

supported. If the FE program allows, the user may model the surface of the shell element as being

offset from the mid-surface and then specify this offset (see Specify Mid-Surface Offset). It is assumed

that orientation of all shell elements along the weld line remains consistent.

Solids

This category refers to hexahedral (brick) and pentahedral (wedge) 3D solid elements. Tetrahedral

elements require special handling and belong to a separate category. The only restrictions on solid

mesh topology is that only rectangular element faces may lie upon the through-thickness cut from the

weld line and that the mesh be regular on the through-thickness cut. By regular it is meant that corner

nodes line up with corner nodes, midside nodes line up with midside nodes, and all nodes lie upon the

through thickness direction which is normal to the weld line in the failure plane to the weld toe surface.

With respect to the input required by shell meshes, solid meshes require only one more command: to

specify the reference normal to the weld toe surface at the start node.

Tetrahedrons

Tetrahedral elements are 3D solids but because they are often generated by automatic meshers their

through-thickness topology may not be regular. The possible irregularity requires special handling.

Since through-thickness face nodes may lie anywhere on the through-thickness face, in order to

determine the weld line connectivity all nodes lying on the through-thickness cut from the weld line,

along the full length of the weld line, must be specified (see Specify Through-Thickness Face Nodes).

The locations at which nodal forces are extracted on the through-thickness line is determined by

specifying stations on the through-thickness line or by identifying intersections between the through-

thickness line and element edges on the through-thickness face (see Through-Thickness Station

Points). As with solid meshes, the reference normal at the start node must be specified.

Although tetrahedral meshes that are irregular on the through-thickness face are allowed, testing has

indicated that meaningful answers can only be obtained if the through-thickness mesh is regular. This is

because nodal forces extrapolated to locations other than the nodes are no longer in equilibrium and

because if forces from corner nodes and midside nodes are mixed then the equations for processing

structural stresses from nodal forces are no longer valid. If tetrahedral elements are employed, the user

is strongly encouraged to manage the mesh so that it is regular on the through-thickness face.

Plate Thickness

The plate thickness must be specified. The specified plate thickness is constant along the weld line and

is in the same units as those selected for the FEA model distance units.



Issue: 12.0 Date: 08.09.2016

Weld line nodes

This is a list of nodes defining the line of the weld. The nodes may be input in any order. In the tree the

weld line is picked from a list of existing node groups. See section 3.1.5 for how this is performed. The

structural stresses will be calculated along the line of the weld.

Weld reference elements

This is a list of elements defining the weld line and identifying the elements on through thickness plane

of the weld, (i.e., for a weld toe the section where the base metal joins the weld). An element domain at

the toe is used for weld toe failure analyses and a similar weld element domain at the throat is used for

weld throat failure analyses, etc.

For shell elements, a single element defines the through thickness plane, but for solid and tetrahedral

elements, the weld element domain must include all of the elements through the thickness, on the plane

of the the through-thickness cut on which to analyse weld fatigue. All rows through the thickness are

required for solids in order to calculate the moment at the weld line.

If a weld line is open (i.e. not a closed circular weld), it is required that up to two additional elements

from the Finite Element mesh at the end of the weld line be included in the domain (and all

corresponding elements through the thickness as well), if they exist in the mesh. These two elements

are not the Start Element (as the start element is on the weld line and contains two nodes in the node

domain). The additional elements share a node with the start element (and at the other end of the weld

line the end element) and contribute to the force balance at the start or end node.

The requirement for weld element domain preparation on solid mesh topology is that only rectangular

element faces may lie upon the through-thickness cut from the weld line and that the mesh is regular on

the through-thickness cut. By regular it is meant that corner nodes line up with corner nodes, midside

nodes line up with midside nodes, and all nodes lie upon the the through thickness direction which is

normal to the weld line in the failure plane.

Start Node and Element

The start node must always be specified. The start node is the first node to be processed along a weld

line. If the weld line is closed the choice of the start node is arbitrary; if the weld line is open then the

start node must be one of the end nodes.

The start element gets the process of automatic determination of the weld line connectivity going by

indicating the initial positive direction along the weld line. To be the start element, the element must be

connected to the start node, must share the weld toe surface, and must share an edge with the weld

line.

For 3D solid meshes it is possible to have an element that is connected to the start node but is not in

the weld toe surface and/or does not share an edge with the weld line. For 3D shell meshes it is

possible for an element to connect to the start node but not to share and edge with the weld line. All of

these elements are not valid start elements.



Issue: 12.0 Date: 08.09.2016

Shell layer

Shell orientation can be configured for the weld line elements. The shell layer option is only required for

meshes comprised of 3D shell elements.

This specifies which face of the shell is to be considered for structural stress calculations: Top (default),

Bottom or Both layers can be specified. It is assumed that orientation of all shell elements along the

weld line remains consistent. If a node is included in the analysis of more than one weld, the result

exported for the node is the worst-case result.

Reference Normal

The reference normal is only required for meshes comprised of 3D solids or tet elements. The reference

normal is the normal to the weld toe surface at the start node, in the form of a vector.

The weld line nodes define the toe of the weld and lie in the weld toe surface. The start node is the

user-specified first node along the weld line. The weld toe surface is considered the top surface; the

bottom surface is through-thickness. The reference normal should point outward from the weld toe

surface, away from the through-thickness direction.

If the mesh is comprised of shell elements, the normal to any element along the weld line is well defined

(Note that the user is responsible for ensuring that the direction of the shell element normals around the

weld line are consistent). All shell element faces by definition lie in the weld toe surface, with the top

(toe) surface section being defined as the shell element section on “top” in the direction of the normal.

The reference normal is required for 3D solids and tets because fe-safe has no way of knowing what is

the weld toe surface – that is, which faces of the elements connected to the weld line constitute the

weld toe surface. By supplying the reference normal, fe-safe knows how to determine what is the weld

toe surface face of connected elements. To get the reference normal, all the user has to do is create a

line perpendicular to the weld toe surface and away from the through-thickness direction in the FE

modeller- if the through-thickness direction is parallel to the normal direction then the nodes at the top

and bottom surfaces will determine this line. The vector from point 1 to point 2 of this line defines the

reference normal.

The reference normal is updated automatically by fe-safe as the weld line comprised of 3D solids and

tets is traversed; all the user has to supply is the reference normal at the start node to get the process

started.

Through thickness face nodes

The face nodes on the through-thickness section along the entire length of a weld line comprised of 3D

solid tetrahedral elements are specified.

Through thickness face nodes need only be specified for meshes comprised of 3D solid tetrahedral

elements, due to the possible irregularity of tet meshes through the thickness.

In the tree the through thickness nodes are picked from a list of existing node groups. See section 3.1.5

for how this is performed.



Issue: 12.0 Date: 08.09.2016

Weld type

Is the weld open or closed?

A closed weld is one that has no ends.

Crack Length

The default value of this parameter is set at crack length = plate thickness. This is based on the failure

criterion of the first appearance of a through-thickness crack. The crack length at failure is used only

when the virtual node weld end correction is being applied. The crack length is the parameter L1 in the

derivation of the weld end correction. If the weld line is closed or if the symmetry flag has been set at

both ends, the crack length is not needed.

Failure Mode

The weld line can have its associated failure mode specified (Toe, Fusion, Throat or Root), or this can

be left at Unspecified. This selection does not affect the structural stress calculations but can affect the

subsequent group processing during the fe-safe fatigue calculations, if different failure modes (e.g. Toe

and Fusion) are specified for the same nodes. For example the same node can appear in two line

specifications: one for a Toe failure and one for Fusion. If no Failure Mode is specified on the weld

definition, then only one value will be stored for the node in the structural stress database, which will be

the worst of the analysed modes. Later when the weld groups are analysed by the fe-safe fatigue

calculation, the associated elements of the node will both have the same fatigue life, even though the

intention might have been to have one element having the fusion result and the other having the toe

result. In order to achieve individual mode-specific results, it is necessary to specify the failure mode.

This will also result in the creation of separate analysis groups for each specified mode.

Reference Length

Same as this parameter in 3.1.2 above.



Issue: 12.0 Date: 08.09.2016

Through thickness node search co-planarity tolerance angle

When searching into the solid domain for the through thickness nodes that are to be associated with

each solid weld line node, the legacy fe-safe algorithm uses element edges for hexahedral or

pentahedral elements; note tetrahedral meshes use a different approach to find through nodes aligned

within a tolerance cone. At fe-safe 6.5 some additional checks were put in place to ensure that the

element edges remain aligned with the weld line normal. It was possible for the original algorithm to

deviate onto incorrectly aligned nodes under some circumstances. Firstly the element face containing

the putative through node edge to be traversed must be approximately co-planar with the failure plane

computed at the line node. This parameter specifies the tolerance threshold on this co-planarity in

degrees. If the element face normal deviates by more than this (default 5 degrees) from the weld line

normal to the failure plane, then the through node search will terminate (for that line node). If this

happens then a warning message will be displayed at the end of the analysis. The diagnostic log will

contain further information about the element(s) at which this happened. The weld will still be analysed,

but this is likely to result in some inaccuracies as it probably means that some through nodes are

missing. This may occur in weld lines with complex curvature. This tolerance parameter may be

increased up to a maximum of 30 degrees. However a finer mesh may be a more accurate solution.

Through thickness node validation cone tolerance

Secondly (following on from the previous parameter), a final validation check is performed on all

through nodes from solid hexahedral or pentahedral meshes, to confirm that all through nodes lie within

a cone centred on the through thickness normal whose semi-angle is defined via this parameter. The

parameter defines the tangent of the cone semi-angle (i.e. allowed deviation distance off the normal as

a proportion of the projected distance of the through node from the line along the normal). This is for

solid non-tetrahedral meshes similar to the Tet mesh through thickness connectivity parameter defined

below, but it is used for final validation rather than location of the through nodes. If located through

nodes fail this validation check then the weld line will not be processed and an error message will result.

The default value is 0.1 (equivalent to 5.7 degrees). The tolerance parameter may need to be

increased for weld lines with complex curvature, but a finer mesh may be a more accurate solution, as

there is an implicit assumption in Verity that through nodes lie along the weld line normal. Note that this

parameter is not used for tetrahedral meshes, because these have their own cone tolerance parameter

which was used to locate their through nodes in the first place.

Mid surface offset

The offset from a shell element mid-surface to the surface on which the shell is modelled is set.

The mid-surface offset is defined as a fraction of the plate thickness. A negative sign indicates that the

offset is in the direction opposite to the element normal. For instance, if the shell element is modelled on

the top surface of the plate, the offset is 0.5. If the shell element is modelled on the bottom surface of

the plate, the offset is -0.5. The mid-surface offset is used to modify the line moments along the weld

line since the computed nodal forces are with respect to the modelled shell surface.



Issue: 12.0 Date: 08.09.2016

Weld symmetry

Symmetry end conditions are specified for the weld line.

The specification of symmetry has no effect on a closed weld line. If the weld line is open, the end for

which symmetry conditions are specified is treated as closed and no virtual node weld end correction is

applied. The start end is where the start node is located. The keyword Both causes symmetry

conditions to be applied at both ends. If symmetry conditions are applied at both ends then no virtual

node weld end correction is applied at either end and there is no need to specify the crack length at

failure or the reference length.

Through thickness station points

The number of station points on a line through-thickness from the top surface to the bottom surface at

which to interpolate nodal forces so that structural stresses may be computed from a mesh comprised

of 3D solid tetrahedral elements.

There are two algorithms used to identify through-thickness locations at which nodal forces are

interpolated for meshes comprised of 3D tetrahedral elements:

Station Points

The user specifies the number of station points. At this number of equally spaced stations through the

thickness from a weld line node, nodal forces will be interpolated from surrounding nodes. The

interpolated nodal forces will be used to compute line forces and moments at the weld line node. If the

through-thickness mesh is regular (see Element Type section above), then if the number of station

points equals the number of elements through-thickness the stations will line up with actual nodes and

no interpolation is necessary.

Edge Intersections

The user specifies zero station points. The default is zero, so omission of the station points command

will lead to use of the edge intersection algorithm. For this algorithm, a line is extended through-

thickness from the top surface to the bottom surface at a weld line node. Where this line intersects

element edges, station points will be established to which nodal forces will be interpolated. As before,

the interpolated nodal forces will be used to compute line forces and moments at the weld line node. If

the through-thickness mesh is regular, the edge intersections will occur at actual nodes and no

interpolation is necessary.

Meaningful results can be computed from 3D solid meshes comprised of tetrahedral elements only if

the through-thickness mesh is regular. As such it is most straightforward to use the station point

algorithm by specifying the number of station points to be equal to the number of elements through-

thickness.



Issue: 12.0 Date: 08.09.2016

Rotate weld line co-ord system (deg.)

The angle in degrees by which to rotate the automatically computed weld line local co-ordinate system

about the weld line to arrive at the LCS for use in structural stress computations is specified. Specifying

a rotation angle for the LCS only affects results for 3D shell elements and should be used to alter the

angle of the through thickness plane. For all other elements types the through thickness plane is

defined by the element domain.

The positive x-axis of the LCS is perpendicular to the weld line, in the weld toe surface, and points into

the weld element domain. The positive y-axis of the LCS points along the weld line and proceeds from

the start node in the direction of the start element to the end of the weld line. The positive z-axis is the

cross product of the x-axis with the y-axis. The positive z-axis can be reversed by reversing the

direction in which the weld line is traversed (i.e., change the start element for closed weld lines or the

start node and element for open weld lines).

In the vast majority of cases, the direction of the cut through the thickness of the plate on which to base

structural stress computations is parallel to the weld toe surface normal. As such, the normal is used to

identify the weld toe surface and through-thickness faces along the weld line. The local z-axis is then

also parallel with the normal, although it may point in the opposite direction. The algorithm to identify the

weld line connectivity will work even if the normal and the through-thickness face are roughly parallel

(see Tolerances section), although the structural stress computations will proceed on the basis that the

two and the local z-axis are parallel.

There may be cases where the desired through-thickness direction and the normal are so out of

alignment that fe-safe fails to correctly identify the weld line connectivity or that the accuracy of the

structural stress computations suffers. In these cases the user may specify an angle by which to rotate

the automatically computed weld line local x-axis about the weld line (the local y-axis) so that the cross

product produces a local z-axis that is parallel to the through-thickness face. This is the angle between

the weld toe surface normal and the through-thickness line from the bottom surface to the top surface.

During the determination of the weld line connectivity fe-safe will take this angle into account so that it

will be able to correctly identify the through-thickness faces. The resulting structural stresses will then

be relative to the angled through-thickness face, not the line through-thickness that is parallel to the

weld toe surface normal. The specified plate thickness should be the angled distance from the top

surface to the bottom surface.

It is valid to specify a rotation angle for meshes comprised of 3D shell elements.

Rotation into weld coord system

Same as Rotate LCS After Solve in 3.1.2 above.



Issue: 12.0 Date: 08.09.2016

Tet Mesh through thickness connectivity tolerance

The tolerance for determining tet mesh through-thickness connectivity is used to evaluate when a

through-thickness line intersects an element face edge or the location on a through-thickness face that

coincides with a station point. The tolerance is given as a fraction of the plate thickness. The more

refined the mesh, the less the required tolerance. The default tet mesh tolerance is 0.001. For coarse

meshes of, say, two quadratic elements through the thickness, it may be necessary to set the tet mesh

tolerance as high as 0.075 if there is significant curvature in the through-thickness face along the weld

line.

Since the only way to produce meaningful structural stresses from a tet mesh is to force the through-

thickness mesh to be regular, either the station point or the edge intersection algorithm will produce

stations through-thickness at which to extract nodal forces that coincide with nodes (thus eliminating the

need for interpolation of nodal forces). It is thus fairly straightforward to examine the debug output file to

determine that the appropriate through-thickness nodes are being found. If they are not being found,

successively increase the tolerance slightly until all are being found.

Number of smoothing iterations


I(r) Function


3.1.5 Group selection and editing

The group selection dialog box will be displayed whenever a parameter requiring an element or node

group is edited. Note that the title depends upon which parameter is being edited:

Figure 3-9



Issue: 12.0 Date: 08.09.2016

To pick one of the existing groups, either loaded from the source model or predefined earlier select it

and click OK.

The Import…, Export..., New..., and Edit... buttons in the group selection dialog box are legacy methods

for defining and editing groups of nodes or elements to be used in weld definitions. It is now

recommended to use the group manager (FEA Fatigue >> Manage Groups... menu) for all operations

related to element and node groups instead. See section 5 of the fe-safe User Guide for more details.

3.1.6 Model Units

The units of the FE model are required for two purposes:

1. To evaluate the plate depth in mm – this is used to convert the structural stresses to equivalent

structural stresses. This is done as part of the Open Finite Element Model… process, or in the

Current FE Models window as described in section 5 of the fe-safe User guide.

2. To convert the stresses to SI units for use with the master weld S-N curves. This is done when the

fatigue analysis is performed

Drop down boxes provide a list of commonly used units, but any conversion factor may be applied.

3.1.7 Other Options

Output Extensive Diagnostics

Verity diagnostics will be saved to the verity-diagnostics.log file, at the end of the analysis

process the user will be asked if they wish to view the debug log. If this checkbox is ticked additional

diagnostics relating to the structural stress calculation will be requested. Note that selecting this option

can generate a lot of additional data, and increase the overall analysis time

Merge All Groups Into ‘WELD’ Group

This option allows to have multiple WELD groups in the Group Parameters control, so that each weld

can have individual properties assigned:

Figure 3-10



Issue: 12.0 Date: 08.09.2016

or to merge all welds into a single WELD group, as below:

Figure 3-11

Note that in Line Welds (3.1.2) the WELD group is broken up into groups for different failure modes

(e.g. WELD[Toe], WELD[Fusion]. If the node is associated with different element for different modes

(e.g. fillet side for fusion, parent plate side for toe), then this sub-typing of the nodal structural stress

data means that the separate failure mode results are preserved. If there are multiple results arising

from the same node mapped to the same element (possibly after projection from the root to a surface

line in compact contour mode), then all such WELD groups are analysed and the final result would then

be the worst life arising from any group.

3.2 Diagnostics

3.2.1 Structural Stresses Table

The calculated structural stresses are exported to the verity-ss.txt file. A sample output

(truncated) is shown below.

Weld line: elems1_nodes1

Results File:

Load case:

Step: 1

Increment: 1

Time: 1.000000

Node WL Distance NSS

Membrane

NSS

Bending

NSS Total top

NSS Total

btm

TSS

Membrane

TSS

Bending

TSS Total

top

TSS Total

btm

---- --------- ---------- ---------- ---------- ---------- ---------- ---------- ---------- ----------

50 0.136E+03 0.621E+03 -0.163E+06 0.163E+06 -0.162E+06 0.253E+03 0.203E+05 0.205E+05 -0.201E+05

49 0.459E+01 0.584E+03 -0.218E+06 0.219E+06 -0.218E+06 -0.186E+02 -0.571E+02 -0.757E+02 0.385E+02

51 0.918E+01 0.595E+03 -0.163E+06 0.163E+06 -0.162E+06 -0.242E+03 -0.202E+05 -0.205E+05 0.201E+05

52 0.235E+01 0.621E+03 -0.184E+06 0.184E+06 -0.183E+06 -0.287E+03 -0.231E+05 -0.235E+05 0.229E+05

54 0.114E+01 0.548E+03 -0.165E+06 0.165E+06 -0.164E+06 -0.266E+03 -0.189E+05 -0.192E+05 0.188E+05



Issue: 12.0 Date: 08.09.2016

Where:

WL Distance - Distance along the weld line from the start node. For a closed weld the value

of WL Distance for the start node is the total length of the weld line, rather

than the distance along the weld line.

NSS Membrane - Membrane structural stress normal to the through-thickness plane (σm)

NSS Bending - Bending structural stress normal to the through-thickness plane (σb)

NSS Total Top - Total structural stress at the top surface normal to the through-thickness

plane (σT(top))

NSS Total Bottom - Total structural stress at the bottom surface normal to the through-thickness

plane (σT(bottom))

TSS Membrane - Membrane shear stress on the through-thickness plane (τm)

TSS Bending - Bending shear stress on the through-thickness plane (τb)

TSS Total Top - Total shear stress at the top surface on the through-thickness plane (τT(top))

TSS Total Bottom - Total shear stress at the bottom surface on the through-thickness plane

(τT(bottom))

Figure 3-12

S – structural stress normal to the through-thickness plane:

2

6

t

m

t

fx'y'

bms

S – Shear stress on through-thickness plane:

2

6

t

m

t

f y'x'

bms

For a more detailed description please refer to Chapter 4.2.

through-thickness plane

Weld

t

S

S



Issue: 12.0 Date: 08.09.2016

3.2.2 Worst cycle mean stress and damage parameter contour

For fe-safe analysis algorithms that base the fatigue calculation on extracting cycles, the most

damaging cycle seen at a node can have its damage parameter exported. The meaning of this contour

depends on the selected algorithm, and for Verity the exported contours are the equivalent structural

stress (see 4.3.3) mean and amplitude, either the normal or shear form depending on the respective

structural stress analysis algorithm (see 2.2). The export is selected via a checkbox (Worst cycle mean

stress and damage parameter) on the Contours tab invoked via the Exports... button on the Fatigue

from FEA dialog box, as in Figure 3-13 below. See also section 22.1 of the fe-safe user guide.

Figure 3-13



Issue: 12.0 Date: 08.09.2016

3.2.3 Log for Items

The equivalent structural stress cycles can be written to the log file for a list of specified items. Access

to this control is from the Exports ... button on the Fatigue from FEA dialog box. The ids defining the

nodes to export must be defined on the List of Items tab to enable the export options controls.

On the Log for Items tab there is a checkbox to Cycle-by-cycle life table for critical plane. If the ids

specified relate to a node within the WELD group then equivalent structural stresses values for each

cycle are exported to the log file. The log contains a table with a row for each cycle in each block giving

the cycle minimum and maximum equivalent structural stress, the respective point indices of these

within the load history, and the life which would be obtained just from that cycle. The format is per the

example below:

WORST PLANE CYCLE LIFE TABLE for Element e1033.6 Block 1 0 degs

(Maximum of 100 most damaging cycles shown, 1 is the first point)

nf (Repeats) Cycle Pt 1 Pt 2

MPa MPa

1.26e+06 113.141 -113.141 1 8

3.50e+06 81.208 -82.269 3 2

2.92e+08 66.999 27.098 5 4

4.10e+08 -13.549 -49.361 7 6

3.2.4 Histories for Items

The equivalent structural stress history for the top and bottom of the plate can be exported for a list of

specified nodes.

Access to this control is from the Exports ... button on the Fatigue from FEA dialog box. The ids defining

the nodes to export must be defined on the List of Items tab to enable the export options controls.

On the Histories for Items tab there is a checkbox to export Load histories. If the ids specified relate to a

node within the WELD group then equivalent structural stresses are exported. It should be noted that

the ids will be element ids unless nodal averaged stresses are used then they will be nodal ids. You can

check what the stresses in your model are by expanding a dataset in the Current FE Models window.



Issue: 12.0 Date: 08.09.2016

The tree item will indicate if the stresses are nodal or elemental:

Figure 3-14

When the analysis is performed a history plot file will be created for each node. These are listed

towards the bottom of the analysis .log file. These files can be opened using the File menu item Open

Data File…, and then plotted using one of the options on the View menu or toolbar.

Figure 3-15



Issue: 12.0 Date: 08.09.2016

3.3 Weld master S-N curves

The Verity database contains the master S-N curves that are provided for Verity analyses.

When defining these curves it should be noted that Stress Range versus Life (rather than Stress

Amplitude) is required.

The S-N curves can be plotted using the Generate Material Plot Data option on the Material menu. The

option to plot Stress Range rather than Stress Amplitude is provided on this dialog box.

The master S-N curves for steel and aluminium alloy weldments are based on systematic analysis of

test data spanning many industries and many welded joint configurations.

Where users have their own test data, it is recommended that they assess the validity of these master

curves against their own test results.

The database in fe-safe allows users to enter their own fatigue curves in addition to the master curves,

should they wish to do so.



Issue: 12.0 Date: 08.09.2016

Note that from fe-safe 2017 the shear T-N curves are also specified in the same database, so each

material contains an S-N curve for normal stress and a T-N curve for shear stress. If there is no shear

test data available then the default is to assume a Von Mises response, which means that the shear

stress will be scaled by √ and then the normal S-N curve is used. As Battelle have only provided shear

curves for Steel Welds for Toe failures [2], the default scaling has been used for other standard Verity

materials (aluminium alloys and root/throat failures). Note that earllier versions of fe-safe provided

separate “shear” materials; the new approach is to provide one material with both S-N and T-N curves,

and then use the curve appropriate to the selected algorithm.



Issue: 12.0 Date: 08.09.2016



Issue: 12.0 Date: 08.09.2016

4 The VeritySSM Method

The VeritySSM method is based on the mapping of the balanced nodal forces/moments along an

arbitrary weld line (available from a typical Finite Element run) into the work-equivalent tractions (or line

forces/moments). A complex stress state due to notch effects can then be represented in the form of a

simple stress state in structural mechanics, in terms of through-thickness membrane and bending

components at each nodal location. The resulting structural stress calculations are mesh-insensitive

(assuming the overall geometry of a component is reasonably represented in the Finite Element model),

regardless of element size, element type and integration order used.

4.1 Approaches to Evaluation of Welded Joints

Stress concentration in welded joints (and notched structures) dominates the fatigue behaviour of

welded structures. Traditional Finite Element methods are not capable of consistently capturing the

stress concentration effects, due to their mesh-sensitivity at welds (resulting from notch stress

singularity). In practice, use of an artificial and subjective radius generates large variability in results.

Fatigue design and evaluation of welded joints are typically carried out by a weld classification

approach in which a (theoretically infinite) family of parallel nominal stress based S-N curves are used

(based upon joint types and loading modes). Extrapolation-based hot spot stress methods offer the

potential to reduce the required number of the S-N curves, and these have become increasingly popular

in offshore and marine applications. An example of hot spot stress based SCF using three extrapolation

techniques shows the variability in results:

Figure 4-1



Issue: 12.0 Date: 08.09.2016

One issue with using any extrapolation-based hot spot stress procedures is that the stress gradients are

more localised in plate structures than in tubular structures. In Figure 4-2, the surface stresses normal

to weld (indicated by arrows) are normalised by the respective nominal bending stresses. The surface

stress gradients shown become increasingly localised as the joint type changes from tube-to-tube

through tube-to-plate to plate-to-plate joints. As a result, extrapolations using 0.5t/1t, 0.5t/1.5t or nodal

value at 0.5t yield a unity in SCF, i.e., the nominal stress. If the Finite Element mesh is not refined

enough or not yet converged, extrapolated hot spot stresses tend to vary with the element sizes, types,

joint types, and loading mode (as shown in Figure 4-1).

Figure 4-2

In all global based stress analysis procedures for fatigue evaluation, the ultimate goal is to identify an

appropriate global stress parameter which:

a. can be calculated consistently with a minimum mesh-sensitivity (mesh sizes, element shapes,

element types, etc.) at a fatigue prone location such as at weld toe; and

b. is demonstrably capable of correlating different fatigue behaviours (such as S-N data) observed

in various joint types, loading modes, plate thicknesses, etc.



Issue: 12.0 Date: 08.09.2016

Nominal stress definition (if applicable for some joint configurations) satisfies (a) since it can be

calculated by simple mesh-insensitive formulae. It does not satisfy (b) since it cannot be used to

correlate S-N data from various joint types and loading modes. This is why the IIW recommended the

use of a very large family of (essentially parallel) S-N curves with respect to the nominal stress

parameter:

Figure 4-3

The fact that the IIW S-N curves are essentially parallel to one another suggests that a scaling

parameter which correctly measures the stress concentration in various welded joint types and load

modes should be able to collapse all the parallel S-N curves into a single master S-N curve. The

VeritySSM method uses an effective global stress parameter to establish such a curve.

4.2 The Structural Stress Method

The structural stress method is used to analyse stress concentration at a fatigue prone location such as

a weld toe, as shown in (a) below:

Figure 4-4



Issue: 12.0 Date: 08.09.2016

It was postulated that this can be represented by an equilibrium-equivalent simple stress state (b) and a

self-equilibrium stress state (c). The former describes a stress state corresponding to an equivalent far

field stress state in fracture mechanics context (or more simply, a generalised nominal stress state at

the same location), while the latter can be estimated by introducing a characteristic depth t1 shown by

the dashed lines in (c).

For displacement-based Finite Element methods, the balanced nodal forces and moments within each

element automatically satisfy the equilibrium conditions at every nodal position. Therefore, the

equilibrium-equivalent structural stress state (in the form of membrane and bending) can be calculated

by using the nodal forces/moments at a location of concern.

4.2.1 Shell/Plate Element Procedures

In order to calculate the structural stresses in terms of membrane and bending components, line forces

and moments must be properly formulated by introducing work-equivalent arguments. For an arbitrarily

curved, closed weld line (i.e., one with two ends overlapping each other, such as a tubular joint), the

nodal forces can be related to line forces along the weld as:

1

3

2

1

1221

3322

2211

1111

1

3

2

1

.

.

3

)(

60

6

...

6

...

3

)(

0

60

0

063

)(

6

60

63

)(

.

.

nnnnn

nn

n f

f

f

f

llll

llll

llll

llll

F

F

F

F

(1)

The weld start node is the same as the weld end, i.e., Fn = F1 and fn = f1, where f1, f2, …, fn–1 are line

forces along y’. In the matrix on the left, li (i =1, 2, …, n–1) represents the element edge length from the

ith element projected onto the weld toe line. The corresponding line moments can be calculated in an

identical manner by replacing the balanced nodal forces F1, F2, …, Fn–1 in local y’ direction with

balanced nodal moments M1, M2, …, Mn–1, with respect to x’ in Eq (1).



Issue: 12.0 Date: 08.09.2016

Note that the nodal force Fi in Eq. (1) represents the summation of the nodal forces at node i from the

adjoining weld toe elements situated on the positive side of the y’ axis. Before Eq. (1) can be

constructed, coordinate transformation for the nodal forces and nodal moments from the global x-y-z

into the local x’-y’-z’ system must be performed, with x’ travelling along the weld line and y’ being

perpendicular to the weld line. The linear system of equations described by Eq. (1) can be solved

simultaneously to obtain line forces for all nodes along the line connecting all weld toe nodes.

Substituting the corresponding nodal moments into Eq. (1), one obtains line moments in the same

manner. Then, the structural stress shown in Figure 4-4 (b) at each node along the weld (such as weld

toe) can be calculated as:

2

6

t

m

t

fx'y'

bms (2)

For parabolic plate or shell elements, Eq. (1) can be formulated in an identical fashion, as can in-plane

shear.

4.2.2 Calculation Examples

Tubular T-Joint

A round robin study on fracture assessment collected detailed strain gauge measurements on a tubular

T-joint (for deriving hot spot stress based stress concentration at the saddle positions shown). To

demonstrate the effectiveness of the present structural stress procedures, four shell element models

with drastically different element sizes near the tube-to-tube weld are shown, at approximately

0.25tx0.25t, 0.5tx0.5t, 1tx1t and 2tx2t:



Issue: 12.0 Date: 08.09.2016

Note that the weld was not modelled at the tube-to-tube intersection to simplify mesh generation. The

four shell models produced the following structural stresses along the weld toe on the chord side:

Since the structural stresses along the weld possess the quarter symmetry, only the results for a

quarter of the weld length measured from the saddle point are shown. The maximum structural stress

concentration occurs at the saddle position. Within the angular span of 90o along the 3D curved weld

from saddle to crown positions, the 2tx2t mesh represents the weld line with only three nodal positions

(or about two and half linear elements) as shown by the triangle symbols. Therefore, the difference in

the structural stress calculations from the 2tx2t mesh is mainly due to geometric changes at the weld

line (tube to tube intersection) resulting from the large linear element sizes used. However, the

structural stress based SCF at the saddle position is still within about 5% of the fine mesh case

(.25tx.25t). Excluding the 2tx2t mesh, the SCF variations in the other three models are all within 2% of

each other.

Plate To I-Beam Joint

In these models, the box fillet weld was modelled as simple nodal connections between attachment

plate edge and I beam. The weld line in this instance is considered as being open-ended. The virtual

node method is automatically activated in constructing Eq. (1):



Issue: 12.0 Date: 08.09.2016

Four drastically different element sizes ranging from 0.5tx0.5t to 4tx4t are used in the mesh designs,

resulting in the following structural stress distributions (normalised by the nominal bending stress)

calculated along the weld toe on the attachment plate side:

It can be seen that the variation in the structural stress calculated at the weld end positions is within

about 1% for all four cases. The validity of the SCF was demonstrated using models with the fillet weld

being properly represented by a row of inclined shell elements. Note that in all four models, the weld is

not modelled for simplicity. As the element sizes change at the beam to attachment intersection, the

geometric representation remains the same.

4.3 Master S-N Curve Formulation

A stress-based scaling parameter to collapse the multiple S-N curves of Figure 4-3 can only be found if

fracture mechanics principles apply, so that crack propagation dominates fatigue lives in welded joints.

A large amount of S-N test data supports this.

4.3.1 Structural Stress Based K Estimation

The simplest candidate fracture mechanics parameter which can be considered is the stress intensity

factor K, but generalised K solutions are not available for welded joints. Fortunately, the structural

stress definition of Figure 4-3 (b) is consistent with the far-field stress definition ( ∞) in fracture

mechanics. Therefore, the structural stress calculation process can be viewed as a stress

transformation process from an actual complex joint in a structure under arbitrary loading to a simple

fracture specimen, in which the complex loading and geometry effects are captured in the form of

membrane and bending:



Issue: 12.0 Date: 08.09.2016

Figure 4-5

As a result, K for any crack size along the weld can be estimated by using the existing K solution for a

simple plate fracture mechanics specimen subjected to both membrane tension and bending, by

considering either an edge crack or a surface elliptical crack.

Figure 4-6 shows the validation for considering an edge crack in a T fillet weld:

Figure 4-6

The case corresponding to “W/O notch stress” was obtained by directly plugging the structural stress

components (membrane and bending) calculated using the present structural stress method into the

existing K solution for an edge notch specimen under remote tension and bending, respectively. The

case “W/ notch stress” refers to the use of the self-equilibrating part of the stress state shown in Figure

4-4 (c) which is analytically estimated.



Issue: 12.0 Date: 08.09.2016

It can be seen that without considering the notch stress (or self-equilibrating part of the stress state), the

current solution provides an accurate K estimation for crack size a/t larger than about 0.1. With the

notch stress effects, K can be calculated for any given infinitesimally small a/t. The current solution in

Figure 4-6 is higher for small a/t than the weight function solution from Glinka, since Glinka introduced a

small weld toe radius in performing the Finite Element stress calculation to avoid mesh-sensitivity. In

the present calculations, the weld toe radius was assumed to be zero, i.e., simulating a sharp notch at

the weld toe.

4.3.2 A Two-Stage Growth Model

The non-monotonic K behaviour as a function of a/t shown in Figure 4-6 is characteristic of all joint

types investigated. As the crack size a/t becomes smaller than about a/t ~ 0.1, the elevated K is

attributed to the dominance of the notch stresses at weld toe. It can then be postulated that both the

short crack and long crack growth processes may be characterised by the two distinct stages of the K

behaviour as a crack propagates from a/t < 0.1 to a/t > 0.1. Along this line, it can be argued that two

stages of stress intensity solutions in the form of the notch-stress dominated Ka/t<0.1 and far-field stress

dominated Ka/t<0.1 can be separated to characterise the full range crack growth behaviour from

0 < a/t < 0.1 (“small crack”) to 0.1 ≤ a/t ≤ 1 (“long crack”). Here, the term K refers to the stress intensity

factor range corresponding to remote stress range. It then follows:

1.0/21.0/1 )()( tata KfKfCdN

da (3)

By introducing a stress intensity magnification factor Mkn in dimensionless form and assuming a power-

law form of the two stage crack growths corresponding to f1(K)a/t≤0.1 and f2(K)a/t>0.1, respectively, Eq.

(3) can be re-written as:

m

n

n

kn KMCdNda )()( (4)

The terms Mkn and Kn are defined below:

) and thickness throughon (based

effects) notch local with(

bmtt

n

knK

KM

(5)

signifying the notch-induced magnification of the stress intensity factors as a/t approaches zero. n

represents the constant crack growth exponent for the first stage of the crack growth and m represents

the conventional Paris law exponent, both of which are to be determined by experimental crack growth

rate data.



Issue: 12.0 Date: 08.09.2016

The validation of the two-stage growth model is shown in Figure 4-7. The crack growth data were taken

from well-known short crack growth data by Tanaka and Nakai and Shin and Smith. Without relying on

any crack closure arguments, all the so called anomalous crack growth data in (a) are collapsed into

single straight data band in (b), with a unified slope of m=3.6. Note that the first exponent in Eq. (4) was

empirically determined as n=2.

Figure 4-7

4.3.3 Equivalent Structural Stress Parameter

The two-stage crack (Eq. 4) with two stage growth exponents being n = 2 and m = 3.6 can be integrated

as:

faa

a

mn

kn KMC

daN

0)()(

(6)

An extensive investigation of Mkn for various joint types showed that it can be approximated by a single

curve as a function of a/t for all joint types once the denominator in Eq. (5) is formulated using the

mesh-insensitive structural stress.



Issue: 12.0 Date: 08.09.2016

Note that the integral in Eq. (6) is not very sensitive to the final crack size af and can be written in a

relative crack length form as:

)()(1

)()(

)/(2

11/

0/

rItCKMC

tatdN m

s

mta

ta

mn

kni

(7)

where I(r) is a dimensionless function of bending r (r = ∆b / ∆s) after performing the following

integration for a given m:

1/

0/)()()()(

)/()(

ta

ta

m

bmm

n

kn

i

t

af

t

afr

t

afM

tadrI

Then, Eq. (7) can be expressed in terms of N once the dimensionless I(r) function is known:

mmm

m

ms NrItC

11

2

21

)(

(8)

Eq. (8) uniquely describes a family of an infinite number of structural stress based S-N curves (∆s – N)

as a function of thickness effects (t), and bending ratio effects r. If Eq. (8) provides a good

representation of the fatigue behaviour of welded joints, an equivalent structural stress parameter can

be defined by normalising the structural stress range ∆s with the two variables expressed in terms of t

and r on the right hand side of Eq. (8):

mm

m

s

s

rIt

S1

2

2

)(

(9)

where the thickness term t(2–m)/2m becomes unity for t =1 (unit thickness), so the thickness t can be

interpreted as a ratio of actual thickness t to a unit thickness, rendering the term dimensionless. With

this interpretation, the equivalent ∆Ss retains a stress unit. It is worth noting that the equivalent structural

stress parameter described by Eq. (9) captures the stress concentration effects (∆s), thickness effects

(t), and loading mode effects (r) on fatigue behaviour.



Issue: 12.0 Date: 08.09.2016

4.3.4 Initial Crack Size Effects

Before Eq. (9) can be used to construct a single master S-N curve for welded joints, assumptions in

performing the integration and the effects of Mkn on I(r) must be quantified. It is well known that initial

crack size (ai / t) can make a significant difference to the final life prediction based on fracture

mechanics as described in Eqs. (6-8).

Figure 4-8 shows the effects of a series of assumed initial crack sizes, using the edge crack based K

solution. Note that I(r) is presented as I(r)1/m after considering the exponent 1/m in Eq. (12) to facilitate

the comparison between the two graphs. Without considering the local notch effects, i.e., Mkn, different

initial crack sizes ai / t produce significantly different I(r)1/m curves as a function of r. Once Mkn is

considered, the dependency of the I(r) on initial crack size ai / t becomes insignificant, particularly for

the two cases with small initial crack size (ai / t), as shown in Figure 4-8 (b). The increase in I(r)1/m is

about 8.5% as r increases from r = 0 (pure membrane) to r = 1 (pure bending) under load controlled

conditions. This implies that with a strong notch effects in typical welded joints characterised by Mkn, the

usual initial crack size effects on life predictions observed in typical fracture mechanics specimens

without stress riser are significantly diminished in welded joints. Based on Figure 4-8 (b), ai / t = 0.001

will be used in the rest of this chapter.

Figure 4-8

Hong and Forte [2] showed that a similar approach can be used in shear dominated loadings, with

some modifications. The parameter m is modified to . A different function is used:



Issue: 12.0 Date: 08.09.2016

And the bending ratio term r is replaced by its shear equivalent

| |

| | | |

Furthermore the equivalent structural stress is scaled by √ .

We then have an equivalent structural stress for shear dominated loading given by

mm

m

s

rIt

S1

2

2

)(

3

(10)

4.4 S-N Data Correlation

The ultimate test for proving if the equivalent stress parameter (Eq. 9) is valid is to demonstrate if a

large amount of experimental S-N data can be collapsed into a single narrow band. In doing so, a

massive amount of S-N data (over 800 fatigue tests) from drastically different joint geometries, plate

thicknesses, and loading modes were collected from literature and published reports, as highlighted

below:

Figure 4-9

For each set of specimens and fatigue tests, the structural stress calculations were performed under

given loading conditions and failure criteria. The results are summarised in Figure 4-10 in terms of both

nominal stress range and equivalent structural stress range as given in Eq. (9). A wide scatter can be

seen in (a), which is expected. Once the equivalent structural stress range is used according to Eq. (9),

the all the S-N data are collapsed into a narrower band, regardless of the diverse joint types, plate

thicknesses, and load modes under which the fatigue tests were conducted over about 40 years. The

same methodology has been proven with the same effectiveness in correlating tubular and pipe/vessel

joints.

t

Joint G’ (t=12.7mm)

Joint G’: Maddox, S.J., 1982, “Influence of Tensile Residual Stresses on the Fatigue

Behavior of Welded Joints in Steel,” Residual Stress Effects in Fatigue,

ASTM STP 776, pp. 63-96, ASTM.

t

Joint Gb (t=20mm)

Joint Gb: Huther I., Lieurade, H.P., Sayhi, N., and Buisson, R., 1998, “Fatigue

Strength of Longitudinal Non-Load-Carrying Welded Joints,” Welding in the World,

Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

Joint B(t=12.7mm), Joint B(Kihl)(6.35mm),

13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t

Joint B: Mantaghi, S. and Maddox, S.J., 1993, ”The Application of Fatigue Design Rules to Large

Welded Structures,”

Joint B(Kihl): Kihl, D.P., and Sarkani, S., 1997, “Thickness Effects on the fatigue Strength of

Welded Steel Cruciforms,” International Journal of Fatigue, Vol.19 Supp. No.1,

pp.S311-S316.

13/10/8AW, 50/50/16AW, 50/50/16AW(DW), 100/50/16AW(QT Steel): Maddox, S.J.,

1987, ”The Effect of Plate Thickness on the Fatigue Strength of Fillet Welded Joints,”

TWI.

t Joint C(t=12.7mm)

Joint C: Mantaghi, S. and Maddox, S.J., 1993, ”The Application of Fatigue Design Rules to Large



13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)




13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)




13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)


Welded Structures,”t Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

t Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

tJoint D(t=12.7mm)

Joint D: Mantaghi, S. and Maddox, S.J., 1993, ”The Application of Fatigue Design Rules to Large


tJoint D(t=12.7mm)



tJoint E (t=12.7mm)

Joint E: Mantaghi, S. and Maddox, S.J., 1993, ”The Application of Fatigue Design Rules to Large


tJoint F (t=12.7mm),

Joint F(Rorup)(12.5mm)

Joint F: Mantaghi, S. and Maddox, S.J., 1993, ”The Application of Fatigue Design Rules to Large


Joint F(Rorup): Rorup, J., and Petershagen, H., 2000, “The Effect of Compression Mean Stress

on the fatigue Strength of Welded Structures,” Welding in the World, Vol., 44(5), pp.20-35.

tJoint E (t=12.7mm)









tJ o in t E (t= 1 2 .7 m m )

J o in t E : M a nt ag hi , S . an d M ad do x , S . J ., 1 9 9 3 , ” T h e Ap p lica tio n o f F a tig u e D esig n R u les to L a rg e

W eld ed S tru ctu res ,”

tJ o in t F (t= 1 2 .7 m m ),

J o in t F (R o ru p)(1 2 .5 m m )

J o in t F : M a nta g hi , S . a n d M a d do x, S .J ., 1 9 9 3 , ” T h e Ap p lica tio n o f F a tig u e D e sig n R u les to L a rg e


J o in t F (R o r u p ): R o ru p , J . , a n d P et ers h ag en , H ., 2 0 0 0 , “ T he E ff ect o f C o m pre ssio n M e an S tre ss

o n th e f a tigu e S tre n gth o f W eld ed S tru ctu re s ,” W eld in g in th e W o rld , V o l. , 4 4 (5 ), pp .2 0 -3 5 .

tJoint E (t=12.7mm)









t Bell (t=16mm)

Bell: Bell, R., and Vosikovsky, O., 1993, “Fatigue Life Prediction o f Welded Joints

for Offshore Structures under Various Amplitude Loading,”

Journal of Offshore Mechanics and Arctic Engineering , Vol.115, pp.123-130.

t LS1 (t=1.5mm)

LS 1: Zhang, J., Dong, P. and Gao, Y., 2001, “Evaluation of Stress Intensity Factor-Based Predictive

Technique for Fatigue Life of Resistance Spot Welds, SAE paper 2 001-01-0830, SAE

t Bell (t=16mm)

Bell: Bell, R., and Vosikovsky, O., 1993, “Fatigue Life Prediction of Welded Joints



t LS1 (t=1.5mm)


Technique for Fatigue Life of Resistance Spot Welds, SAE paper 2001-01-0830, SAE

t Bell (t=16mm)



Journal of Offshore Mechanics and Arctic Engineering, Vol.115, pp.123-130.

t LS1 (t=1.5mm)



t Bell (t=16mm)




t LS1 (t=1.5mm)


Technique for Fatigue Life of Resistance Spot Welds, SAE paper 2001-01-0830, SAE t

CP1 (t=1.5mm)

CP 1: Zhang, J., Dong, P.and Gao, Y., 2001, “Evaluation of Stress Intensity Factor-Based Predictive


t Double Gussets (t=90mm, crack=2mm)

Double Gussets(Niemi): Wagner, M., 1998, “Fatigue Strength of Structural Members

with In-Plane Notches,” IIW Doc. XIII-1730-98, IIW.

t

CP1 (t=1.5mm)






Double Edge Gusset (90mm)

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.

t





t

Joint Gb (t=20mm)



Vol. 41, pp.298-313.


13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)




13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)




13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)




13/10/8AW(13mm), 50/50/16AW(50mm),

50/50/16AW(DW)(50mm),100/50/16AW(100mm),

100/50/16AW(QT Steel)(100mm)

t





pp.S311-S316.



TWI.

t Joint C(t=12.7mm)


Welded Structures,”t Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

t Joint-Cb(Booth)(t=38mm), Joint-Cb(Pook)(38mm)

tJoint D(t=12.7mm)



tJoint D(t=12.7mm)



tJoint E (t=12.7mm)









tJoint E (t=12.7mm)









tJ o in t E (t= 1 2 .7 m m )

J o in t E : M a nt ag hi , S . an d M ad do x , S . J ., 1 9 9 3 , ” T h e Ap p lica tio n o f F a tig u e D esig n R u les to L a rg e


tJ o in t F (t= 1 2 .7 m m ),

J o in t F (R o ru p)(1 2 .5 m m )

J o in t F : M a nta g hi , S . a n d M a d do x, S .J ., 1 9 9 3 , ” T h e Ap p lica tio n o f F a tig u e D e sig n R u les to L a rg e


J o in t F (R o r u p ): R o ru p , J . , a n d P et ers h ag en , H ., 2 0 0 0 , “ T he E ff ect o f C o m pre ssio n M e an S tre ss

o n th e f a tigu e S tre n gth o f W eld ed S tru ctu re s ,” W eld in g in th e W o rld , V o l. , 4 4 (5 ), pp .2 0 -3 5 .

tJoint E (t=12.7mm)









t Bell (t=16mm)

Bell: Bell, R., and Vosikovsky, O., 1993, “Fatigue Life Prediction o f Welded Joints



t LS1 (t=1.5mm)



t Bell (t=16mm)




t LS1 (t=1.5mm)



t Bell (t=16mm)



Journal of Offshore Mechanics and Arctic Engineering, Vol.115, pp.123-130.

t LS1 (t=1.5mm)



t Bell (t=16mm)




t LS1 (t=1.5mm)


Technique for Fatigue Life of Resistance Spot Welds, SAE paper 2001-01-0830, SAE t

CP1 (t=1.5mm)






t

CP1 (t=1.5mm)






Double Edge Gusset (90mm)

Fig. 3: Corner Joints Tested by Yagi (1992)

Fig. 2: “Type b)” Crack Scenarios at Welds on Plate Edges

t

CP1 (t=1.5mm)






t = 5-80mm



Issue: 12.0 Date: 08.09.2016

Figure 4-10

By regression analysis, the mean line of (b) can be represented in the form of:

333.3

1

'

1

1

2

216308

)(

NCN

rIt

m

mm

m

s (11)

with the stress unit in MPa and thickness in mm and m = 3.6. In Eq. (11), 1/m' represents the negative

slope of the master S-N curve in (b). Note that although in various publications, m = m' (=3 for steel

welds) is often assumed. In this investigation, it is found that they are close, but not necessarily the

same. For simple fatigue test specimens, nominal stresses are often well-defined, then,

nsss SSCF

in which SCFss signifies the structural stress based SCF and ∆Sn the typical nominal stress range

definition.



Issue: 12.0 Date: 08.09.2016

For a given simple joint specimen of interest, the structural stress based SCF can be calculated using

the present structural stress procedure. Then the nominal stress based S-N behaviour can be predicted

by using the master S-N curve from Eq. (11) as:

'

11

2

2

)(m

ss

mm

m

n NSCF

rItCS

(12)

Both C and m' are given in Eq. (11) based on the master S-N database generated in this investigation.

Two examples are given:

Figure 4-11

The mean nominal stress S-N curves for various plate thicknesses are predicted using Eq. (12) for the

cruciform joint tests (under remote tension loading) by Maddox. A good correlation between the test

data and those predicted by the master S-N (Eq. 11) curve is evident.

Hong and Forte [2] showed that a similar master S-N curve can be derived for shear dominated

loadings using Eq. (10) for the shear equivalent structural stress, but the slope is modified to ⁄ .

Further details are given in [2]. The mean shear-dominant S-N curve is:

⁄



Issue: 12.0 Date: 08.09.2016

4.5 A Simple Example of the Structural Stress Method

As widely documented in many text books on basic Finite Element theory, displacement based Finite

Element methods (commonly used in industrial day to day applications) tend to over-estimate the

stiffness of a structure. An adequate mesh refinement is always required in order to achieve a

converged solution of the displacement field predicted by the structural mechanics theory. This

becomes particularly important when dealing with bending due to the potential presence of shear

locking, for example discussed by Cook et al.

To demonstrate how the equilibrium-based structural stress procedure discussed above can be applied

to a simple problem on which a hand calculation using the above equation can be carried out, a

cantilever bending example was directly taken out from an elementary Finite Element text book by Cook

et al, as shown:

Figure 4-12

Figure 4-13



Issue: 12.0 Date: 08.09.2016

Two linear plane stress elements were used to model the simple beam, as shown. Such a model is

deemed too coarse to provide any meaningful stress results in the Finite Element calculations. Indeed,

as shown in Figure 4-15, the stresses calculated at the mid-length of the beam (Section A-A) are

±117.6MPa from the element on the right, in contrast to the beam theory solution of 750MPa. If

averaged stresses between the two elements (from both right and left sides of A-A section) at the nodal

positions were used, the calculated stresses become 235.3MPa. Due to its equilibrium considerations,

the structural stress procedure using nodal forces should provide improved estimate of the stresses.

The nodal forces from the Finite Element calculations are also shown. By applying Eq. (2), after

substituting the nodal forces in x-direction, the line forces at Node 2 (at beam bottom) and Node 5 (at

beam top) becomes, respectively:

N/mm750]250)250(2[2

2

)2(2

N/mm750)]250(2502[2

2

)2(2

255

522

xx

xx

FFl

f

FFl

f

(12)

Note that the beam thickness (t) is 1mm, the structural stresses are solely determined by the membrane

component, for example, at Node 2 (bottom):

MPa750mm1

N/mm750 ms (13)

The structural stress procedure for this case provides exactly the same solution as the beam theory.

This is expected since the beam is under statically determined loading conditions. The equilibrium

arguments underlying the structural stress procedure require that the structural stress so calculated

should satisfy the equilibrium conditions. It should be cautioned that this example is only intended to

illustrate the basic concept of the structural stress method and does not imply that the method can be

used without limitation.



Issue: 12.0 Date: 08.09.2016



Issue: 12.0 Date: 08.09.2016

5 Preparing FE Models for use with Verity®

5.1 Supported FE Programs

For line weld assessment the following Finite Element programs are supported by Verity in fe-safe:

Abaqus

Nastran

I-DEAS

Ansys

For spot weld assessment only Abaqus and Nastran packages are supported.

Irrespective of the type of weld under consideration Verity in fe-safe requires element nodal forces and

the model geometry. The FE results must include the element nodal forces for each element that

contributes to the nodal forces acting along each defined weld, saved in the global coordinate system.

The elements comprising a weld definition can be 3D shells or 3D solids. 3D beam and 1D solid

elements are also recognised, to allow their use in the FE model without generating error messages

when the weld connectivity is processed.

Supported Finite Elements types are listed for each FE program below, with the appropriate element

identifier and a brief description of the formulation. Unsupported element types will be ignored —a

warning message will be issued stating that for correct results the element must not be used in any

weld line definition.

Where element thickness has been read, the Current FE Models Assembly item will have a sub-item

labelled Element Thickness. If this is not present for a shell model the Plate Thickness parameter must

be set, see 3.1.2.

Numbering Schemes

Node numbering around each of the supported elements can be found in the relevant FE program

documentation. Based on this node numbering, an edge, face, and face edge numbering scheme is

established that is used in the algorithms to identify the weld line connectivity. Numbering scheme

tables are included below for each FE program.

In each table the node positions are the positions in the node numbering array, also known as the

element incidences. Each table is arranged in terms of the element shape. The element shape

encompasses both linear and quadratic element orders, so that the specified midside nodal positions

are only relevant for quadratic elements.

For the face numbering schemes, the nodal positions are arranged such that the normal to the face is

outward from the element. The first midside node position occurs between the first and second corner

node positions, the second midside node between the second and third corner nodes, etc., until the last

midside node occurs between the last and first corner nodes.



Issue: 12.0 Date: 08.09.2016

For the edge connectivity and face edge connectivity, the positive edge direction points in the direction

from the first corner node to the second corner node.

In the debug output concerning the weld line node connectivity, edges or face edges and faces are

identified by numbers that correspond to the tables. If the edge has a negative number, it means that

the direction pointing from a weld line node to the next weld line node is in the negative edge direction.

The element face on the weld toe surface for 3D shell elements is always 1 by default since there is

only one face.

5.1.1 Abaqus

Supported versions: 6.10 and above.

File type: Sequential access FIL file (extension .fil).

Output database ODB file (extension .odb)

Output method: Element nodal forces are output using the NFORC option in the *EL

FILE or *ELEMENT OUTPUT commands

Reading of element thickness in ODB models is supported.

More fe-safe details: fe-safe Documentation, Appendix G.

3D shell

connectivity:

Element Shape Edge Edge Node Positions

1st corner 2nd corner midside

quadrilateral

1 1 2 5

2 2 3 6

3 3 4 7

4 4 1 8

triangle

1 1 2 4

2 2 3 5

3 3 1 6



Issue: 12.0 Date: 08.09.2016

3D solid face connectivity:

Element

Shape Face

Face Node Positions

corner midside

1st 2nd 3rd 4th 1st 2nd 3rd 4th

brick

1 1 4 3 2 12 11 10 9

2 5 6 7 8 13 14 15 16

3 1 2 6 5 9 18 13 17

4 2 3 7 6 10 19 14 18

5 3 4 8 7 11 20 15 19

6 1 5 8 4 17 16 20 12

wedge

1 1 3 2 9 8 7

2 4 5 6 10 11 12

3 1 2 5 4 7 14 10 13

4 2 3 6 5 8 15 11 14

5 3 1 4 6 9 13 12 15

tetrahedron

1 1 3 2 7 6 5

2 1 2 4 5 9 8

3 2 3 4 6 10 9

4 1 4 3 8 10 7



Issue: 12.0 Date: 08.09.2016

3D solid face edge

connectivity:

Element Shape Face Face Edge Face Edge Node Positions


brick

1

1 1 4 12

2 4 3 11

3 3 2 10

4 2 1 9

2

1 5 6 13

2 6 7 14

3 7 8 15

4 8 5 16

3

1 1 2 9

2 2 6 18

3 6 5 13

4 5 1 17

4

1 2 3 10

2 3 7 19

3 7 6 14

4 6 2 18

5

1 3 4 11

2 4 8 20

3 8 7 15

4 7 3 19

6

1 1 5 17

2 5 8 16

3 8 4 20

4 4 1 12

wedge

1

1 1 3 9

2 3 2 8

3 2 1 7

2

1 4 5 10

2 5 6 11

3 6 4 12

3

1 1 2 7

2 2 5 14

3 5 4 10

4 4 1 13

4

1 2 3 8

2 3 6 15

3 6 5 11



Issue: 12.0 Date: 08.09.2016



4 5 2 14

5

1 3 1 9

2 1 4 13

3 4 6 12

4 6 3 15


tetrahedron

1

1 1 3 7

2 3 2 6

3 2 1 5

2

1 1 2 5

2 2 4 9

3 4 1 8

3

1 2 3 6

2 3 4 10

3 4 2 9

4

1 1 4 8

2 4 3 10

3 3 1 7



Issue: 12.0 Date: 08.09.2016

Supported elements:

Element Type Element Name1 Brief Description

2

3D Shells

S4 4-node linear general-purpose full ∫ quadrilateral

S4R 4-node linear general-purpose reduced ∫ quadrilateral

S8R 8-node quadratic thick reduced ∫ quadrilateral

S3 = S3R 3-node linear general-purpose reduced ∫ triangle

3D Solids

C3D4(H) 4-node linear reduced ∫ tetrahedron

C3D6(H) 6-node linear reduced ∫ wedge

C3D8(H,I,IH) 8-node linear full ∫ brick

C3D8R(H) 8-node linear reduced ∫ brick

C3D10(H,M,MH) 10-node quadratic reduced ∫ tetrahedron

C3D15(H) 15-node quadratic reduced ∫ wedge

C3D20(H) 20-node quadratic full ∫ brick

C3D20R(H) 20-node quadratic reduced ∫ brick

3D Beams

B31(H) 2-node linear reduced ∫

B31OS(H) 2-node linear reduced ∫ warped

B32(H) 3-node quadratic reduced ∫

B32OS(H) 3-node quadratic reduced ∫ warped

B33(H),B34 2-node cubic full ∫

PIPE31(H) 2-node linear reduced ∫ pipe

PIPE32(H) 3-node quadratic reduced ∫ pipe

1 ( ) denotes optional extensions: (I)=incompatible modes, (H)=hybrid w/ pressure, (M)=modified

2 See Abaqus documentation for a more complete element description.

Note: Elements using five degrees of freedom (S4R5, STRI65, S8R5, S9R5) are not supported as for

those elements moments corresponding to the in-surface rotations are not available for output.



Issue: 12.0 Date: 08.09.2016

5.1.2 Nastran

Supported versions: 70.7 and above.

File type: Sequential access OP2 (OUTPUT2) file (extension .op2).

Output method: Element nodal forces are output using the GPFORCE command.

The OP2 file is requested by issuing the bulk data commands

PARAM,POST,-1 or PARAM,POST,-2. No other values of POST are

supported. For POST equal to –1 or –2, the bulk data command

PARAM,POSTEXT,YES must be issued as well in order for datablocks

required to be written to the database. Note that the geometry must be

included in the file, that is PARAM,OGEOM,NO command must not be

used.

If the NASTRAN bulk data parameter POST is assigned the value –1,

then a labelled results database with an .op2 extension will be written. If

POST is assigned the value –2 then an unlabeled results database with

a .op2 extension will be written. If the NASTRAN-level parameter

OP2NEW is set to YES and POST is set to –1 or –2, then the

datablocks will be written in post-2001 format. If OP2NEW is instead set

to NO with POST equal to –1 or –2, then the datablocks will be written

in pre-2001 format. The NASTRAN default for OP2NEW is NO.

Reading of element thickness in OP2 models is supported.




Issue: 12.0 Date: 08.09.2016

3D shell and

connectivity:

See Abaqus table.

3D solid face connectivity:

Element

Shape Face

Face Node Positions

corner midside

1st 2nd 3rd 4th 1st 2nd 3rd 4th

brick

1 1 4 3 2 12 11 10 9

2 5 6 7 8 17 18 19 20

3 1 2 6 5 9 14 17 13

4 2 3 7 6 10 15 18 14

5 3 4 8 7 11 16 19 15

6 1 5 8 4 13 20 16 12

wedge

1 1 3 2 9 8 7

2 4 5 6 13 14 15

3 1 2 5 4 7 11 13 10

4 2 3 6 5 8 12 14 11

5 3 1 4 6 9 10 15 12

tetrahedron

1 1 3 2 7 6 5

2 1 2 4 5 9 8

3 2 3 4 6 10 9

4 1 4 3 8 10 7



Issue: 12.0 Date: 08.09.2016

3D solid face edge

connectivity:



brick

1

1 1 4 12

2 4 3 11

3 3 2 10

4 2 1 9

2

1 5 6 17

2 6 7 18

3 7 8 19

4 8 5 20

3

1 1 2 9

2 2 6 14

3 6 5 17

4 5 1 13

4

1 2 3 10

2 3 7 15

3 7 6 18

4 6 2 14

5

1 3 4 11

2 4 8 16

3 8 7 19

4 7 3 15

6

1 1 5 13

2 5 8 20

3 8 4 16

4 4 1 12

wedge

1

1 1 3 9

2 3 2 8

3 2 1 7

2

1 4 5 13

2 5 6 14

3 6 4 15

3

1 1 2 7

2 2 5 11

3 5 4 13

4 4 1 10

4

1 2 3 8

2 3 6 12

3 6 5 14



Issue: 12.0 Date: 08.09.2016



4 5 2 11

5

1 3 1 9

2 1 4 10

3 4 6 15

4 6 3 12

tetrahedron

1

1 1 3 7

2 3 2 6

3 2 1 5

2

1 1 2 5

2 2 4 9

3 4 1 8

3

1 2 3 6

2 3 4 10

3 4 2 9

4

1 1 4 8

2 4 3 10

3 3 1 7

Supported elements:

Element Type Element Name Brief Description

3D Shells

CQUAD4, CQUADR 4-node linear quadrilateral (3D shell)

CQUAD8, CQUAD4, CQUAD4FD,

CQUAD9FD, CQUADX, CQDX4FD,

CQDX9FD

8-node quadratic quadrilateral (3D

shell)

CTRIAR, CTRIX3FD 3-node linear triangle (3D shell)

CTRIA6, CTRIA6FD, CTRIA3FD,

CTRIAX, CTRIAX6, CTRIAX6FD 6-node quadratic triangle (3D shell)

3D Solids

CHEXP 8-node linear brick

CHEXA, CHEXA20F, CHEXAFD,

CHEXAL, CHEXPR 20-node quadratic brick

CPENP 6-node linear wedge

CPENTA, CPENPR, CPENT15F,

CPENT6FD 15-node quadratic wedge

CTETP 4-node linear tetrahedron

CTETRA, CTETPR, CTETR10F,

CTETR4FD 10-node quadratic tetrahedron

3D Beams

1D Solids

CBAR 2-node linear beam or solid

CBEAM 2-node linear warped beam

CROD 3-node quadratic beam or solid



Issue: 12.0 Date: 08.09.2016

Note: CTRIA3 elements do not support stress output to element-nodes, only to centroids, which causes

a data mismatch if such elements were to be used for Verity analysis – nodal force data is stored at

nodes and stress data is stored at the centroid. As this may cause the results of fatigue analysis to be

incorrect and dependant on the element orientation along the weld line the CTRIA3 element type is not

supported for Verity analysis; it is recommended to use the CTRIAR elements instead.

Note: It is not recommended to use a mix of element types (e.g. CQUADR and CQUAD4) along the

weld line.

5.1.3 I-DEAS

Supported versions: Tested with NX series version 11.

File type: Sequential access UNV (universal) file (extension .unv).

Output method: Element nodal forces are output using the ELEMENT FORCE option.


3D shell and connectivity: See Abaqus table and the note on node numbering below.

3D solid face connectivity: See Abaqus table and the note on node numbering below.

3D solid face edge

connectivity:

See Abaqus table and the note on node numbering below.

Node numbering in the UNV file is in an inconvenient form for processing since the corner nodes are

not given first for higher order elements. As a consequence, the program modifies the I-DEAS node

numbering to match that of Abaqus:

Element Type I-DEAS Element Incidence in Abaqus order

6-node triangle 1 3 5 2 4 6

8-node quad 1 3 5 7 2 4 6 8

10-node tet 1 3 5 10 2 4 6 7 9

15-node wedge 1 3 5 10 12 14 2 4 6 11 13 15 7 8 9

20-node brick 1 3 5 7 13 15 17 19 2 4 6 8 14 16 18 20 9 10 12 11



Issue: 12.0 Date: 08.09.2016

Supported elements:

Element Type Descriptor ID Brief Description

3D Shells

94 4-node linear quadrilateral

95 8-node quadratic quadrilateral

91 3-node linear triangle

92 6-node quadratic triangle

3D Solids

115 8-node linear brick

116 20-node quadratic brick

112 6-node linear wedge

113 15-node quadratic wedge

111 4-node linear tetrahedron

118 10-node quadratic tetrahedron



Issue: 12.0 Date: 08.09.2016

5.1.4 Ansys

Supported versions: 5.7 through 13.x.

File type: Direct access RST file (extension .rst).

Output method: Element nodal forces are output using the command OUTRES,NLOAD.

Reading of element thickness in RST models is supported.


3D shell

connectivity:

See Abaqus table.

3D solid face connectivity: See Abaqus table.

3D solid face edge

connectivity:

See Abaqus table.

Supported elements:

Element Type Element Number Brief Description

3D Shells 28, 41, 43, 57, 63, 143, 181

1 4-node linear quadrilateral

91, 93, 99, 2811 8-node quadratic quadrilateral

3D Solids

5, 45, 46, 62, 64, 65, 69, 70, 73, 96,

97, 122, 164, 1852

8-node linear brick

90, 95, 117, 128, 147, 186, 191, 226,

2312

20-node quadratic brick

87, 92, 98, 123, 127, 148, 168, 187,

227, 232 10-node quadratic tetrahedron

1 triangle if linear element degenerated to 3 nodes or quadratic degenerated to 6 nodes

2 tetrahedron if linear element degenerated to 4 nodes or quadratic element degenerated to 10 nodes, wedge if

linear element degenerated to 6 nodes or quadratic element degenerated to 15 nodes

3 See Ansys manual for a more complete element description



Issue: 12.0 Date: 08.09.2016

5.2 Modelling Considerations And Applications

The structural stress calculation procedure is based on equilibrium considerations, using nodal forces

and moments in shell and plate element models. Therefore, a reasonable representation of the overall

load path to a joint and joint stiffness is all that is required in modelling a structure. As a result, the

modelling requirements for the presence of welds can be significantly relaxed from those in typical

stress-based Finite Element computations. Some detailed parametric analyses are presented below to

illustrate how typical joint types can be effectively modelled, and where special care is required when

using Verity®.

5.2.1 Full Penetration Welds

A full penetration fillet weld in a plate T-fillet joint can be modelled using an inclined weld element

forming a triangle with the two base plates:

Figure 5-1

In this type of joint configuration with full penetration, weld toe failure into the base plate is often the

dominant fatigue failure mode (as shown). The weld element thickness tw = l/√2 has been used for all

reported results cited here, where l is the fillet weld leg length. Due to the rigidity of the triangle

formation, the structural stresses calculated for weld toe cracking are not sensitive to variations in

assumed weld element thickness.



Issue: 12.0 Date: 08.09.2016

A 3D T-joint example was employed:

Figure 5-2

A solid element model (a) was used to establish the reference structural stress. A series of plate

element models (b) with various weld element thickness assumptions were used to compare the

structural stress calculations at the same weld toe location as that in the solid model. The results are

summarised in (c).

The structural stress for the case with tw = l√2 (overall weld throat depth) was slightly over-estimated,

comparing the results from the solid model in which the weld size and shape was accurately

represented. The over-estimation is primarily due to the asymmetrically increased stiffness as the weld

element thickness was increased. If the vertical element in the triangle formation is proportionally

increased, the increase in the structural stress results becomes much less noticeable.

5.2.2 Partial Penetration Weld and Root Failures

The same fillet-welded T-joint can be used to illustrate the applications of the structural stress

procedures in partial penetration welds:



Issue: 12.0 Date: 08.09.2016

Figure 5-3 Weld representation in a plate element model and various possible failure modes in T-fillet weld with

partial penetration

In addition to the weld toe failure modes (Toe 1 and Toe 2) in (a), a weld failure originating from the

weld root can be approximately modelled using a plate element model as shown in (b). In reality, three

weld throat failure modes through the weld metal are possible, indicated by (1)-(3) in (b). In addition,

failure path (4) corresponds to a crack originating from the weld root and propagating through the base

plate. The structural stresses relevant to the four weld root failure paths can be approximated as normal

stresses to the respective arrow lines (1) through (4), as shown in (c) with respect to the plate element

representation of the partial penetration weld. For failure mode (2), two weld elements are needed so

that the nodal forces from the middle node can be used as weld line nodes (c). This was not considered

in the current investigation, since the corresponding structural stresses can be approximated by taking

the algebraic sum of those corresponding to failure paths (1) and (3).



Issue: 12.0 Date: 08.09.2016

A parametric analysis on the effects of fillet weld leg size on structural stress behaviour for the same

fillet-welded T-joint discussed earlier, but with partial penetration this time, is shown below:

Figure 5-4 Parametric analysis of structural stress as a function of weld size and failure mode (reference SS values

are calculated using solid model at l =1t)

The structural stresses with respect to various potential failure modes are summarised in (b) as a

function of the fillet weld throat size tw in terms of the fillet weld leg length l. In this parametric

investigation, the structural stress corresponding to weld throat failure mode (2) was also calculated

using a solid element model for verification purposes at l=t, represented as “Throat (2) – Ref” in (b).

With the same solid model, the reference structural stress values for failure modes “Toe 2 – Ref” and

“Throat (4) – Ref” were calculated for comparison with the plate element model results for l=t.



Issue: 12.0 Date: 08.09.2016

As shown in (b), if the fillet weld leg length is small, the structural stresses corresponding to the two

weld throat failure modes (1) and (3) dominate the fatigue behaviour, since the structural stresses

corresponding to the two weld toe failure modes (“Toe 1” and “Toe 2”) are relatively low. As the weld leg

length increases up to about l=t, there is a rapid reduction in the magnitude of the structural stresses

characterising weld throat failures through the weld metal. Both weld root failure mode (4) into the plate

and weld toe failure modes become dominant. Note that for weld root failure mode (4), as the weld leg

length increases, the failure position with respect to the horizontal plate also changes due to the

limitations of this particular Finite Element model. Therefore, the failure mode modelled in the plate

element model no long represents failure mode (4) in (b).

With l=t, a good agreement between the structural stress results from both plate element model and

solid element models can be seen for failure mode “Toe 2” and “Throat (4)” in (b). For failure mode

“Throat (2)”, the structural stress calculated from the solid element model takes an average of those

corresponding to “Throat (1)” and “Throat (3)” calculated from plate element models. Consequently, the

results suggest that the simple weld representation scheme shown in (c) for partial penetration welds is

adequate to capture some of the major weld throat failure modes.

The procedures discussed above for partial penetration welds can be used to determine a minimum

weld leg size beyond which weld throat failure modes can be effectively suppressed. From Figure 5-4, it

can be inferred that the minimum leg size should be at least l=t for this T-joint. Such an observation can

be related to practical experiences with weld toe failures in partial penetration welds. This also

highlights the potential variability in predicting weld throat failures in partial penetration welds, since any

stress analysis procedure requires a precise weld throat depth (i.e., weld element thickness tw) known

before hand. In practice, the weld throat depth can vary significantly even within the same weld.



Issue: 12.0 Date: 08.09.2016

A more complex example is shown below for a tube-to-tube joint with partial penetration fillet welds with

a leg size of about 2t:

Figure 5-5

The dominant failure mode is weld toe failure into the horizontal tube (or chord) wall under the given in-

plane bending conditions. Two possible weld representation schemes were examined (b). Under a fixed

tw = 3mm, Figure 5-6a shows the structural stress results as a function of the vertical element thickness

in the triangle representation of the fillet weld. For tr=0, a gap is assumed as shown in Figure 5-5b. By

varying both tw and tr (see Figure 5-6b for details), the difference in the structural stress calculations is

not noticeable.



Issue: 12.0 Date: 08.09.2016

In view of these parametric investigations, for weld toe failure the weld representation for partial

penetration welds can use either a triangle formation or single inclined weld element shown in Figure

5-5b. The weld element thickness tw can be simply calculated based on nominal weld leg size. Any

additional variations have no significant effects on the structural stress calculations. However, it must be

noted that when using the triangle formation in Figure 5-5b, the ability to calculate the structural

stresses with respect to the weld element (tr) for characterising weld throat failures is lost, unless tr is

made negligible (such as tr ≈ 0.1t). If weld throat cracking is of interest, the weld should be represented

as a row of inclined weld elements with element thickness tw being a nominal or a minimum weld throat

depth specified as a weld, which can vary significantly in practice. For this reason, the S-N data

representing weld throat failures typically exhibit a much wider scatter than those representing weld toe

failures.

Figure 5-6 structural stress results using two weld presentation schemes for partial penetration weld and various

weld element thickness definition (see Figure 5-5b), assuming weld toe cracking into the chord tube wall.



Issue: 12.0 Date: 08.09.2016

5.2.3 Lap Fillet Weld and Failure Modes

Non-Load Carrying Lap Fillet Weld

A cover plate fillet weld specimen is shown below with two weld presentation schemes. For weld toe

failure into the base plate, the two schemes yield the same structural stress based SCF, supporting the

findings discussed earlier. In such non-load carrying lap fillet weld, weld throat cracking is unlikely.

Figure 5-7



Issue: 12.0 Date: 08.09.2016

Load Carrying Lap Fillet Weld

Load-carrying lap fillets are used frequently in many automotive applications. One simple joint

configuration in this category is shown below, with the potential failure modes:

Figure 5-8

As discussed in section 5.2.2, the weld leg size is critically important in determining which failure mode

is dominant for this type of joint. For convenience, the weld throat depth is again measured in terms of

the weld leg size on the bottom plate in (b). In the shell element representation (b), failure Mode (A) can

be directly modelled as shown. Mode (B) and Mode (C) are approximate representations of the weld

failures at the top and bottom edges (or fusion lines) of the weld. As discussed earlier, it can be shown

that the stress states from Mode (B) and Mode (C) in (b) can be used to represent mode (D) in (a) by

taking the algebraic sum of the two.



Issue: 12.0 Date: 08.09.2016

(c) summarises a parametric study on the weld size effects on structural stress based SCF for each of

the potential failure modes in (b). A solid model with a precise representation of the partial penetration

weld geometry (leg length l=t) and the two well-defined failure orientations (“Weld Toe – Ref” and “A –

Ref”) was used to calculate their respective structural stresses as the reference values for comparison

purposes with the simplified weld presentation in the shell model. The structural stress based SCFs are

shown in (c) as the symbols and for the two failure modes, which are essentially the same as the

shell model results at l/t=1, validating the adequacy of the shell element representation of the fillet weld

in the shell model.

When the weld leg size (l/t) is smaller than unity, the weld throat failures (A) and (C) in (b) dominate the

failure mode in this specimen. The algebraic summation of the structural stresses from (A) and (C) is

near zero for all weld leg sizes investigated, implying that the weld throat failure through the middle of

the weld metal (D) in (a) is unlikely. Furthermore, failures corresponding to (A) and the weld toe on the

horizontal plate become more dominant. Again, the structural stress based SCFs for both mode (A) and

the weld toe in the bottom plate are not sensitive to the variation in weld size investigated.

5.2.4 Fillet Weld between Two Plates with Dissimilar Thicknesses

The earlier fillet weld examples were analysed by assuming that two plates joined by a fillet weld are of

similar thicknesses, and that the fillet weld size is of the order of the plate thickness. In such situations,

the weld element can be easily defined by connecting the mid-thickness planes between the two plates,

such as that shown in Figure 5-7a. If the thicknesses of the two plates are significantly different, or if the

fillet weld is significantly smaller than one of the plate thickness, the resulting additional local bending

effects not captured using the previous weld presentation can be significant. The previous weld

representation discussed in the earlier sections is shown as a direction connection between the inclined

weld elements with the mid-thickness nodes in the two adjacent plates:

Figure 5-9



Issue: 12.0 Date: 08.09.2016

Without losing generality, (b) illustrates a typical situation in which the local bending effects can be

approximated by placing the weld element nodes through the mid-positions of the two fillet weld leg

lengths. This issue was first identified when comparing the structural stress results between 3D solid

element models and shell element models for a padding plate joint, see Figure 5-10. As shown in

Figure 5-10a, the difference in the two plate thicknesses is 5mm. In addition, the fillet weld has a leg

length of 5mm. The simplest weld representation (“Scenario 1” in Figure 5-10b) does not consider the

localised load transfer through the weld. According to the discussions with respect to Figure 5-9b, an

improved modelling scheme (“Scenario 2” in Figure 5-10c) is applicable to such problems.

The overall structural stress distributions along the weld toe in the base plate are compared in Figure

5-11a. The increase in structural stress based SCF using Scenario 2 is due to the additional bending

effects captured by using the weld representation scheme shown in Figure 5-10c. The actual structural

stress based SCF can be accurately calculated by using 3D solid element models in which the small

fillet weld can be correctly modelled. As shown in Figure 5-11b, the results from the shell element

models with scenario 2 are very close to those from the 3D solid element models. The shell element

models with the weld representation described as Scenario 1 (Figure 5-10b) under-estimate the

structural stress based SCF.

Figure 5-10



Issue: 12.0 Date: 08.09.2016

Figure 5-11

5.2.5 Weld Line Curvature Modelling

Weld line curvature requires some additional considerations when using Verity®. A weld line is defined

as a continuous curve in space, described by a series of inter-connected nodes with unique nodal

numbers in a Finite Element model. For example, consider a curved weld line in a linear shell element

model:

Figure 5-12



Issue: 12.0 Date: 08.09.2016

The normal vector direction (y’) at Node 3 (N3) between the Elements 2 and 3 (E2 and E3) is

determined by the vector summation of y’E2 and y’E3. If the intersection angle () is too large, mesh-size

sensitivity can develop, and the structural stress calculated at this node may not correctly represent the

actual structural stress state for fatigue evaluation purposes. Analyses performed for various curved

welds suggest that in general if the intersecting angle is smaller than ~30, the structural stress results

are not noticeably affected.

This criterion can often be relaxed for a particular application, either through an analyst’s experience or

by performing a series of parametric analyses. For example, consider a longitudinal stiffener joint

modelled by a shell element model with half symmetry assumed:

Figure 5-13 A longitudinal stiffener and a curved weld line definition for structural stress calculation for weld toe

crack with respect to the base plate (note: the weld end on the base plate in txt mesh is simulated with 4 linear

elements over the 180 angular span)

The weld line definition for weld toe cracking into the base plate is shown by the arrow line encircling

the entire weld toe, along which Verity calculates structural stresses in a simultaneous manner. In the

mesh shown, four linear elements were used to model the half circle around weld end (=45 in this

case). If more elements are used in the .5t mesh, the change in the structural stresses is still not

significant, but this cannot be generalised for broad applications. It is always desirable to perform a few

parametric analyses to establish a criterion for representing weld line curvature, if a coarse mesh is to

be used.



Issue: 12.0 Date: 08.09.2016

The above discussions also imply that a 90 sharp turn along a weld line in a Finite Element model

should be avoided. For the same longitudinal stiffener joint:

Figure 5-14 Representation producing a sharp corner at weld end on base plate (not recommended)

One single weld line definition encircling the weld toe contour can no longer be correctly presented as a

continuous curve. As a result, structural stresses calculated at the weld end and corner are mesh-

sensitive and lacking physical meaning. If there are no other choices, the weld toe cracking (into the

base plate thickness) at the weld end may be analysed by considering an alternate weld definition

(straight line). Obviously, the SCF at the weld end is much higher than that shown in Figure 5-13. Weld

toe cracking along the two straight sides of the weld can be simulated by using a straight weld line

definition on each side of the weld toe (dashed lines in Figure 5-14). At the weld end, the structural

stresses normal to the weld line (x’ direction) will be calculated in Verity with the virtual node method

automatically. However, the structural stress information for failure locations situated within the curved

weld end in Figure 5-13 will not be available from the straight weld line definition.

It should be noted that the weld line definition along the weld toe on attachment plate side is not

affected in Figure 5-14. The resulting structural stresses are reasonably accurate, comparing with those

with the weld being properly modelled in Figure 5-13. Further discussions on this (using a different

example) follow.



Issue: 12.0 Date: 08.09.2016

5.2.6 Calculations without Modelling Weld

An extreme case in simplified weld representations is shown in Figure 5-15a for the same joint type as

Figure 5-13 and Figure 5-14, in which the fillet weld is only represented as a line of nodes connecting

the two plates:

Figure 5-15 An extreme case in which the fillet weld is not modelled –an alternative weld line definition for

calculating the structural stress SCF at the weld end in the direction of loading (SS calculated using nodal forces at

N1 and N2 from the one element shown)



Issue: 12.0 Date: 08.09.2016

Special caution must be taken in interpreting the analysis results since some of the load transfer

mechanisms may not be correctly represented when using the oversimplified weld representation

procedure. Under the loading direction given in (a), the structural stress normal to the weld end (in the

direction of loading or negative y direction) can be calculated with a weld line definition as shown (solid

blue line). However, the virtual node position in this instance should be determined as the actual weld

toe position measured from the mid-plane of the attachment plate (l1=t in this example). Once Verity is

used with respect to the weld line, the corresponding structural stresses at the weld (in the negative y

direction) should be very close to those shown in (b).

The structural stress based SCFs were calculated manually before the virtual node method was

implemented in Verity®. The required nodal forces in the negative y direction were retrieved for each of

the highlighted elements in each of the four models with different element sizes. The line force f1 at the

weld end was then calculated using l1=t. The line moment m1 was calculated in the same way by

retrieving the nodal moments with respect to x from the same elements. Given the over-simplified weld

representation scheme used in Figure 5-15, it is surprising to see that the structural stress based SCFs

are close to those calculated using the curved weld line definition in Figure 5-13. Furthermore, as the

element size increases, the SCF remains essentially the same up to 10t10t. This proves the

effectiveness of the virtual node method. However, the surprising consistency of the structural stress

results from the four meshes is attributed to the fact that in the plate element model in Figure 5-15, the

overall geometry remains the same as the element size increases, while the 180 arc at the weld end in

Figure 5-13 deviates increasingly from the smoothness requirements (Section 5.2.5) in the linear

element models.

For this particularly example the structural stresses at the weld end can be calculated with reasonable

accuracy using the specific calculation procedures as described, but if the loading is applied in the x

direction on one of the two longer edges of the base plate, the structural stress based SCFs in the x

direction become unity since there are no interactions between the base and longitudinal attachment

plates without modelling the presence of the weld.



Issue: 12.0 Date: 08.09.2016

With the same considerations, if FPSO Detail 5 is analysed without modelling the weld it does not

introduce any significant error in the structural stress calculations for the weld toe on the base plate:

Figure 5-16 Comparison of the structural stress results with element sizes without modelling the presence of the

weld



Issue: 12.0 Date: 08.09.2016

Figure 5-17 Comparison of the structural stress results with element sizes with a proper weld representation

The structural stress based SCF is at about 1.6 (Figure 5-16) without modelling the weld, versus about 1.5 with

modelling the weld (Figure 5-17). Note that at the time of these calculations, the virtual node position was

placed at 10mm (or l1 = 10mm). In the earlier discussions regarding S-N data interpretation for the same detail,

l1 = 20mm was used to be consistent with the actual failure criterion reported

Tutorial A: Single Weld Line


Issue: 12.0 Date: 08.09.2016



Issue: 12.0 Date: 08.09.2016

6 Tutorial A: Single Weld Line (Using Abaqus)

6.1 Introduction

This tutorial outlines how to analyse welded structures using Verity®.

This analysis is based upon a solid hexahedral model with one row of elements through the thickness.

The model was created for the SAE Fatigue Challenge in 2003 and 2004. Figure 6-1 shows the model;

the plot was created using an FEA pre-processor:

Figure 6-1

The brown area shows the weld itself. For this analysis the weld line to be analysed is the one between

the weld (brown) and the marked weld toe (dark blue). In practice most likely two weld lines, one at

each side of the weld, would be analysed.

The remainder of the model will be analysed using a local strain approach and the material Manten

steel.

The model is provided in all the supported formats. This tutorial will use an Abaqus model; you can use

one of the other FEA results format if you prefer, but if so the element and node numbers in the weld

line and weld toe will need to be modified. A full weld definition configuration file is supplied for each

format.

6.2 Preparation

Start the program by selecting fe-safe from the Windows “Start” menu (Windows) or by running the

script fe-safe (Linux) – (see section 5.2 in the fe-safe User Guide).

When starting fe-safe, the Configure fe-safe Project Directory dialog appears. The path shown in the

Project Directory field is where the fe-safe project directory will reside. The user can either accept the

default name and location or enter a more convenient name and location.



Issue: 12.0 Date: 08.09.2016

If an existing project directory is chosen, existing data in that project will be retained. This includes

model data, load history files, analysis settings, and more.

Reset any existing analysis options to their defaults in the Clear Data and Settings dialog box [Tools >>

Clear Data and Settings...] by selecting the Check all option and clicking OK.

6.3 Open the Model

6.3.1 Select the Datasets

To open the model, select File >> FEA Solutions >> Open Finite Element Model... On the file selection

dialog box select the file <DataDir>\weld_fatigue\solid-tubess3_xx.odb and click Open.

This will display the Select Datasets to Read pre-scan dialog box:

Figure 6-2



Issue: 12.0 Date: 08.09.2016

The dialog box shows that there is a dataset containing Stresses (used by fe-safe) and Nodal Forces

(used by Verity®). Ensure that both are selected so that the weld can be analysed using Verity and the

rest of the model can be analysed using fe-safe, then click OK.

The stresses will be extracted from the model and information similar to that shown below will be

displayed in the Message Log window. This information is also written to the file

<ProjectDir>\Model\reader.log.

odb_io version 6.3-02:9994

Reads Abaqus 6 ODB databases

Copyright Dassault Systemes UK Limited 1996-2012

Reading : C:\data\verity\solid-tubess3_610.odb

Writing : Pre-scan

Strains : Read all (Pre-scanning)

ODB interface exe exited with code 62097

Pre-scan complete.

odb_io completed

NOTE : Storing current session macro commands to

D:\Data\Users\fe-safe.version.6.3\output\current.macro




Reading : C:\Program Files (x86)\Safe_Technology\fe-

safe\version.6.3\data\verity\solid-tubess3_610.odb pre-scan

Writing : D:\Data\Users\fe-safe.version.6.3\projects\project_01\model\FESAFE.FED

Selected position is Element Nodal

Strains : Selected datasets

Dataset : S

From Step : 1

Iteration : 1

Time : 2.22e-16

Title :

Direct Min/Max : -33612.7 50056.2

Shear Min/Max : -30004.6 19650.8

No. Elements : 4412



Issue: 12.0 Date: 08.09.2016

When the model has finished loading, the Loaded FEA Models Properties dialog box dialog box

appears, as shown in Figure 6-3.

Figure 6-3



Issue: 12.0 Date: 08.09.2016

If the dialog box dialog box does not appear automatically, then it can be displayed by right-clicking on

the icon in the Current FE Models window and selecting Properties:

Figure 6-4

Ensure that the units are as shown in Figure 6-3, then click OK.

A dialog box will show prompting to edit element groups loaded from the model, click No.

A summary of the open model appears in the Current FE Models window, showing the loaded datasets

and element group information – see Figure 6-5.

Figure 6-5

Note: If the window does not appear exactly as shown in Figure 6-5, then expand the tree view to show

more details.



Issue: 12.0 Date: 08.09.2016

When a new model is loaded, the output filename automatically defaults to:

<ResultsDir>\{source_file_name}Results.{source_file_extension}

which in this example is:

<ResultsDir>\solid-tubess3_xxResults.odb

The output filename can be modified either by manually editing the Output File field in the Fatigue from

FEA dialog box, or by clicking the adjacent browse button: .

6.3.2 Define the Weld

After the model has been opened the weld locations should be defined prior to evaluation of the

structural stresses. This can be done using the Weld Preparation>> Define and Analyse Weld

Geometry… menu. A new dialog box will be displayed:

Figure 6-6

If this is the first time Verity has been used, the weld tree area will most likely be blank. If it is not then

use the Delete All… button to clear the tree.



Issue: 12.0 Date: 08.09.2016

To create a weld, click the Add… button and select Legacy Line Definition as the type of the weld to be

defined. This will add a new weld definition to the tree with the name New Line Weld:

Figure 6-7

Note: If the New Weld Line does not appear as shown above, then expand the tree view to show more

details.

To edit parameters in the weld tree, select each parameter and click on the Edit button, or double-click

the parameter or value. Define each required weld parameter as discussed below:

Name

Rename the weld line by or double-clicking on the New Line Weld value. This will allow the item to be

edited in the tree. Change the weld name to TV_to_LG_tube__in_TV_tube and press the enter key

() to accept it.



Issue: 12.0 Date: 08.09.2016

Element Type

To indicate that this weld domain contains Solid hexahedral elements (as opposed to Tetrahedrons or

Shells), edit the Element Type=Shells parameter:

Figure 6-8

Use the drop-down box to select Solids. The weld tree will be rebuilt to include the parameters required

for a weld in a solid model.

Plate Thickness

Edit the plate thickness value and specify a thickness as 0.312 inches. All measurement defined in the

Weld Definition dialog box match those of the consistent units system specified in the Current FE

Models window. The FE model contains distances in inches and loads in lbf, so the definition must

match these. The plate thickness needs to be constant for the length of the weld.



Issue: 12.0 Date: 08.09.2016

Reference Normal

Next, specify the Reference Normal. This is only required for Solid (hexahedral) or Tetrahedral

elements. This parameter is the normal to the free surface of the start element. Edit this parameter and

set the normal to (0,0,–1) in the displayed dialog box:

Figure 6-9

Click OK.

Start Element & Start Node

The start node and element are required to tell fe-safe where one end of the weld starts. From Figure

6-10 and Figure 6-12 it can be seen that they are element 76 and node 354. Define the two parameters:

Start Element =76

Start Node =354

in the tree for the weld definition.



Issue: 12.0 Date: 08.09.2016

Weld Line Nodes

The position of the weld is defined using a line of nodes along the weld and in the weld toe surface.

The top left section of the weld (including node numbers) is shown below. Following the weld line

(between brown and dark blue sections) the sequence of nodes beginning 354, 353, 352 can be seen

to define the weld line.

Figure 6-10

Edit the parameter Weld line nodes. This will pop up the weld line nodes group selector dialog box. In

this example a nodal set with all weld line nodes has been already defined in the FE model with a name

of NTT, so select NTT_nodal (or PART-1-1_NTT_nodal in the case that instance support is turned on

in fe-safe) from the list and click OK to confirm:

Figure 6-11



Issue: 12.0 Date: 08.09.2016

Note: If the finite element model does not include a node group containing the weld line nodes, a user

defined group can be configured as discussed in section 5 of the fe-safe User Guide. Groups defined

for Verity need not contain node numbers defined in order.

Weld Reference Elements

The weld reference elements define a theoretical crack path along the weld toe through the base

material for this example. Elements in the base material along the weld toe are used for weld toe failure

analyses and elements in the weld fillet material is used for weld throat failure analyses.

The toe defines the section where the base metal joins the weld. The requirement for weld toe element

preparation on solid mesh topology is that only rectangular element faces may lie upon the through-

thickness cut from the weld line and that the mesh is regular on the through-thickness cut. By regular it

is meant that corner nodes line up with corner nodes, midside nodes line up with midside nodes, and all

nodes lie upon the line normal to the weld toe surface.

The top half of the weld including element numbers is shown below:

Figure 6-12



Issue: 12.0 Date: 08.09.2016

Edit the parameter Weld reference elements, which will pop up the weld reference elements group

selector dialog box. Again for this example an elements set with all weld toe elements has been defined

in the FE model with a name of TTOE. Select TTOE (or PART-1-1_TTOE in the case that instance

support is turned on in fe-safe) from the list and click OK to confirm:

Figure 6-13

Click OK.

Note: As with nodes, a user defined group can be configured for elements. Groups defined for Verity

need not contain element numbers defined in order.

Final Configuration

This tutorial analyses a single weld only, additional welds can be defined in the same manner.

The Just Save button can save this definition to disk for later use. The configuration you have just

defined has been included in your installation directory in the file

<DataDir>\weld_fatigue\solid-tubess3.def. You can load this definition using the Open…

button if you have had problems defining the analysis.

Now the configuration is complete it should look similar to:



Issue: 12.0 Date: 08.09.2016

Figure 6-14

Note: If support for instances was turned on (FEA Fatigue>>Abaqus ODB Interface Options with a

checkbox engaged next to Allow ODB files with multiple parts and instances) at the time of ODB

reading, the groups will appear as shown above with the instance (PART-1-1).

6.3.3 Evaluate the structural stresses

Now click Save and analyse and Verity will evaluate the structural stresses. After the structural stresses

are calculated Information similar to that shown below will be displayed in the Message Log window.

This information is also written to the file <ProjectDir>\Model\verity-diagnostics.log.

Running Verity module on 1 line welds...

Verity structural stress file backed up to C:\temp\model\verity-ss.arc.

Processing weld New_Line_Weld.

Processing forces dataset 1.1 at time 2.22e-016.

Processing forces for weld New_Line_Weld.



Analysis of welds took 0.141 seconds.



Issue: 12.0 Date: 08.09.2016

The structural stresses evaluated by Verity have been merged in with the stresses read by fe-safe and

a new group named WELD has been created. A summary of the Verity analysis run can be found in the

verity-diagnostics.log file (Weld Preparation>> View Structural Stress Diagnostics Log menu)

and a table of the structural stresses calculated can be found in the verity-ss.txt file (Weld

Preparation >> View Structural Stress File menu).

Note: a warning containing the information that an item in the definition that has no instance defined,

may be displayed. fe-safe will try to accommodate for the lack of instance, which should allow the

tutorial to continue without issue.

6.4 Analyse the Model

After evaluating the structural stresses using Verity we should have a group named WELD in the

Fatigue from FEA window. The material will have defaulted to the Steel Weld 50% S-N curve, and the

algorithm will have defaulted to SSM-Normal.

In this tutorial the mean life curve will be used for the weld. The remainder of the model will be analysed

using a local strain approach and the material Manten steel.

6.4.1 Define the Loading

The loading will be constant amplitude zero mean loading (this is sometimes referred to as R=-1). We

will cycle between the stresses in the dataset multiplied by 1 and multiplied by –1.

On the Fatigue from FEA dialog box select the Loading Settings tab to switch to the loading tree. Right

mouse click on the loadings tree and select Clear all loadings to remove any existing loading.

Select the dataset named Dataset 1: (1.1) S : Stress from the Current FE Models window. Click the

Add... button and then select the A user-defined LOAD * dataset. Specify 1 and –1 in the Loading Scale

box:

Figure 6-15



Issue: 12.0 Date: 08.09.2016

Click OK. This will add an item to the loading as shown:

Figure 6-16

Select the Analysis Settings tab from the Fatigue from FEA dialog box to switch to the main settings.

6.4.2 Define the Materials

In the Group Parameters table double-click the Algorithm column and in the displayed dialog box select

Analyse with Material’s default Algorithm and click OK. This will configure all applicable groups to use

the algorithm best suited to the selected material.

Select the material SAE_950C-Manten in the local.dbase file in your Material Databases window

and double-click the column titled Material. You will be asked if you wish to set the material, click Yes.

The material for the WELD group should then be reverted to a material from the verity.dbase.

The default life curve can easily be changed from the mean curve to a –2 or –3 standard deviations life

curve. To do this select the required life curve in the verity.dbase file in the Material Databases

window and then click on the cell defining the WELD group material in the Group Parameters table on

the Fatigue from FEA dialog box. Figure 6-17 shows the current list of S-N curves in the Verity

database.

Figure 6-17



Issue: 12.0 Date: 08.09.2016

6.4.3 Perform the Analysis

The Fatigue from FEA dialog box should now look similar to:

Figure 6-18

The results filename will have defaulted to a new .odb file.

To start the analysis, click the Analyse button. A summary of what you have defined is displayed. Click

Continue and the analysis will start. As the analysis is being performed the worst life so far is displayed

in the Message Log window.

Summary

=======

Worst Life-Repeats : 19172.797

at Element [0]1398.4

Analysis time : 0:00:01.150

Fatigue Analysis Completed.

The analysis log shows that the worst-case life for the whole model is:

19172.797 repeats of the loading, at element 1398, node 4.



Issue: 12.0 Date: 08.09.2016

A copy of the original .odb file was created, onto which a new step containing the fatigue results was

appended.

The exported fatigue results in the file contain the log10(life) contour of fatigue lives, which should look

similar to:

Figure 6-19

Note that element 1398 is not at the weld, and comes from analysis on the base materials using SAE-

950.

Tutorial B: Multiple Weld Lines


Issue: 12.0 Date: 08.09.2016



Issue: 12.0 Date: 08.09.2016

7 Tutorial B: Multiple Weld Lines (Using Ansys)

7.1 Introduction

This tutorial outlines how to analyse welded structures in an Ansys .rst file using Verity®.

Figure 7-1 shows the solid hexahedral model with one row of elements through the thickness. The

model was created for the SAE Fatigue Challenge in 2003 and 2004.

Figure 7-1

All of the weld definitions are merged to one group for configurations with the nominal master S-N curve

for weld fatigue at toes. The remainder of the model will be configured for no analysis.

The tutorial uses an Ansys RST model. However, the same techniques can be applied to all FE formats

supported for Verity-in-fe-safe, which currently include:

Ansys results (*.rst) files

Abaqus ODB files

Nastran OP2 files

7.2 Preparation





Issue: 12.0 Date: 08.09.2016






Ensure that the following basic settings are correctly configured. If fe-safe is being run for the first time

then it is possible to skip this preparatory section.



7.3 Open the Model


To open the model, use the File >> FEA Solutions >> Open Finite Element Model... menu, select the

file <DataDir>\weld_fatigue\t-tube_sld_xx.rst and click Open. This will display the Select

Datasets to Read pre-scan dialog box:

Figure 7-2



Issue: 12.0 Date: 08.09.2016

The dialog box shows that there is a dataset containing stresses, strains, temperatures and forces. For

this tutorial make sure that Forces and Stress datasets are selected so that the weld can be analysed

using Verity and the rest of the model can be analysed using fe-safe, then click OK.

The selected datasets will be extracted from the model. Once this process is complete, the Loaded FEA

Models Properties dialog box dialog box appears, as shown in Figure 7-3.

Figure 7-3



Issue: 12.0 Date: 08.09.2016



Figure 7-4

Ensure that the units are as shown in Figure 7-3, then click OK. A dialog box dialog box will appear

prompting to edit element groups loaded from the model, click No.

7.4 Define the Welds


structural stresses. This can be done using the Weld Preparation >> Define and Analyse Weld

Geometry menu. A new dialog box dialog box will be displayed. Initially this may contain the weld

definitions that were used in the previous Verity run (Tutorial A):

Figure 7-5



Issue: 12.0 Date: 08.09.2016

These welds will not be needed, unless you are re-running the same job. Use the Delete All… button to

clear the tree, and click OK when prompted.

Tutorial A covers the details of defining a weld. For this tutorial, click the Open… button. The Open

Verity Weld Definition dialog box dialog box will appear:

Figure 7-6

Use the drop-down list to sort the directory by the weld definition of type Weld Definition *.def. select the

file <DataDir>\weld_fatigue\t-tube_sld1.def, and click OK. This will load the correct set of

welds into the weld definition tree. In this case there will be four defined welds:

Figure 7-7



Issue: 12.0 Date: 08.09.2016

The weld definition and configuration should be complete and look similar to this:

Figure 7-8

Expand the defined welds to view configuration details. The lists of nodes and elements are defined in

the weld definition file as opposed to the finite element nodel. These can be accessed by editing Weld

line nodes or Weld reference elements (as discussed in Tutorial A).



Issue: 12.0 Date: 08.09.2016


Now click Save and analyse and Verity will evaluate the structural stresses. An information similar to

that shown will be displayed in the Message Log window after the structural stresses have been

calculated. This information is also written to the file <ProjectDir>\Model\verity-

diagnostics.log:


Verity structural stress file backed up to C:\temp\model\verity-ss.arc.

Processing weld tvtb_to_lgtb_in_tvtb_neg_z.

Processing weld tvtb_to_lgtb_in_tvtb_pos_z.

Processing weld tvtb_to_lgtb_in_lgtb_neg_z.

Processing weld tvtb_to_lgtb_in_lgtb_pos_z.

Processing forces dataset 1.1 at time 1.

Processing forces for weld tvtb_to_lgtb_in_tvtb_neg_z.

Processing forces for weld tvtb_to_lgtb_in_tvtb_pos_z.

Processing forces for weld tvtb_to_lgtb_in_lgtb_neg_z.

Processing forces for weld tvtb_to_lgtb_in_lgtb_pos_z.





a new group named WELD has been created. A summary of the Verity analysis run can be found in the

verity-diagnostics.log file (Weld Preparation >> View Structural Stress Diagnostic Log menu)

and a table of the structural stresses calculated can be found in the verity-ss.txt file (Weld

Preparation >> View Structural Stress file menu).


After evaluating the structural stresses using Verity we should have a group named WELD in the

Fatigue from FEA window. The material will have defaulted to the Steel Weld 50% S-N curve.

In this tutorial the mean life curve will be used for the weld. The remainder of the model will be

configured for no analysis.




On the Fatigue from FEA dialog box dialog box select the Loading Settings tab to switch to the loading

tree. Right mouse click on the loadings tree and select Clear all loadings to remove any existing

loading.



Issue: 12.0 Date: 08.09.2016



box:

Figure 7-9


Figure 7-10

Select the Analysis Settings tab from the Fatigue from FEA dialog box dialog box to switch to the main

settings.



Issue: 12.0 Date: 08.09.2016


For this analysis, we will turn off the analysis for the entire model, then turn on just the analysis of the

WELD group of elements. In the Group Parameters table double-click on the cell in the Algorithm

column and WELD row and select Normal Structral Stress:

Figure 7-11

Click OK.



Issue: 12.0 Date: 08.09.2016

This tutorial uses the default material, Steel Weld (50%) for the WELD group. Select the algorithm cells

of the other groups (Material1,Material2 and Default), then right click and Select Edit, then change the

algorithm selection in the resulting dialog to Do Not Analyse. The Fatigue from FEA dialog box dialog

box should now look similar to the following:

Figure 7-12

If you have some options configured in Exports from a previous analysis, click Exports…, select the List

of Items tab, click None, and then click OK.

The results filename will have defaulted to a new .rst file.



Issue: 12.0 Date: 08.09.2016


To start the analysis, click the Analyse button. A summary of what you have defined is displayed:

Figure 7-13

Click Continue and the analysis will start. As the analysis is being performed the worst life so far is

displayed in the Message Log window.

Summary

=======


at Element 2809.3




355807 repeats of the loading, at element 2809, node 3.



Issue: 12.0 Date: 08.09.2016

A copy of the original .rst file excluding the original steps was created, onto which a new step

containing the fatigue results was written.

The exported fatigue results are written to the Ansys variable SX and contain the log10(life) contour of

fatigue lives, which should look similar to:

Figure 7-14

Tutorial C: Shell Weld Line


Issue: 12.0 Date: 08.09.2016

8 Tutorial C: Shell Weld Lines (Using Nastran)

8.1 Introduction

This tutorial outlines how to analyse welded structures in a Nastran .op2 file using Verity®.

Figure 8-1 shows a linear shell model which is based on models created for the SAE Fatigue Challenge

in 2003 and 2004.

Figure 8-1

The weld definitions will be generated as separate groups for configurations with different master S-N

curves for weld fatigue at toes. The remainder of the model (except the weld material shown in yellow)

will be analysed using a local strain approach and the material Manten steel.

The tutorial uses a Nastran OP2 model. However, the same techniques can be applied to all FE formats



Abaqus ODB files

Nastran OP2 files

8.2 Preparation





Issue: 12.0 Date: 08.09.2016








Reset any existing analysis options to their defaults in the Clear Data and Settings dialog box dialog

box [Tools >> Clear Data and Settings...] by selecting the Check all option and clicking OK.

8.3 Open the Model

8.3.1 Select the datasets


file <DataDir>\weld_fatigue\t-tube_2k4_pn2.op2 and click Open. This will display the Select

Datasets to Read pre-scan dialogue:

Figure 8-2



Issue: 12.0 Date: 08.09.2016






Figure 8-3




Issue: 12.0 Date: 08.09.2016

A dialog box will appear prompting to edit element groups loaded from the model, click Yes.

Figure 8-4

Use the button to move the group associated with Property 3 (property definition 3 in the

Nastran input file) which represents the elements in the weld fillet, and the group associated with

Matierial 1 which represents all elements in the whole model with material definition 1. The Default

group will be used to limit analysis in the weld elements which are not part of the remaining Property 1

and Property 2 groups. Click OK.

If the Current FE Models Properties dialog box dialog box does not appear automatically, then it can be

displayed by right-clicking on the icon in the Current FE Models window and selecting Properties.

If the Edit Group List prompt does not appear, the Select Groups to Analyse dialog box can be opened

using FEA Fatigue >> Manage Groups… menu selection.



Issue: 12.0 Date: 08.09.2016

8.3.2 Define the materials

From the Material Databases window in the local.dbase file select the material SAE_950C-Manten

material and in the Group Parameters table double-click the column titled Material. You will be asked if

you wish to set the material, click Yes.

8.3.3 Define the welds



Geometry menu. A new dialog box will be displayed. Initially this may contain the weld definitions that

were used in a previous Verity run (e.g. a previous Tutorial):

Figure 8-5





Issue: 12.0 Date: 08.09.2016

Tutorial A covers the details of defining a weld. For this tutorial, click the Open… button. The Open

Verity Weld Definition dialog box will appear:

Figure 8-6

Use the drop-down list to sort the directory by the weld definition of type Weld Definition *.def. select the

file <DataDir>\weld_fatigue\t-tube_2k4_pn2.def, and click OK. This will load the correct set

of welds into the weld definition tree. In this case there will be two defined welds:

Figure 8-7



Issue: 12.0 Date: 08.09.2016

Expand the defined welds to view configuration details. The lists of nodes and elements are defined in

the weld definition file as opposed to the finite element nodel. These can be accessed by editing Weld

line nodes or Weld reference elements (as discussed in Tutorial A).

Figure 8-8

Deselect the checkbox to Merge all groups into ‘WELD’ group. This allows the two weld lines to be

configured differently in the Groups Parameters settings (e.g. materials and algorithms).


Now click Save and analyse and Verity will evaluate the structural stresses. Information similar to that

shown below will be displayed in the Message Log window after the structural stresses have been

calculated. This information is also written to the file <ProjectDir>\Model\verity-

diagnostics.log:



Issue: 12.0 Date: 08.09.2016


Verity structural stress file backed up to C:\Users\E5Q\Documents\fe-

safe.version.6.5\projects\project_01\model\verity-ss.arc.

Processing weld TV_to_LG_tube_in_TV_tube.

Processing weld TV_to_LG_tube_in_LG_tube_open.

Processing forces dataset 1.1.

Processing forces for weld TV_to_LG_tube_in_TV_tube.

Processing forces for weld TV_to_LG_tube_in_LG_tube_open.





two groups starting with WELD have been created, representing each of the weld definitions. A

summary of the Verity analysis run can be found in the verity-diagnostics.log file (Weld

Preparation >> View Structural Stress Diagnostic Log menu) and a table of the structural stresses

calculated can be found in the verity-ss.txt file (Weld Preparation >> View Structural Stress File

menu).


After evaluating the structural stresses using Verity we should have two groups starting with WELD in

the Fatigue from FEA window. The algorithm will have defaulted to SSM-Normal, and the material will

have defaulted to the 50% failure equivalent structural stress S-N curve for both groups. This will suffice

for the toe of the TV tube, whilst the toe on the LG tube will need to be represented by the 2.2% failure

equivalent structural stress S-N curve.

The remainder of the base materials will be configured for analysis with Manten steel.

8.4.1 Define the loading







Issue: 12.0 Date: 08.09.2016



box:

Figure 8-9


Figure 8-10


8.4.2 Define the algorithm






Issue: 12.0 Date: 08.09.2016

Double-click the cell defining the Default group in the Algorithm column and select Do not analyse and

click OK.

Figure 8-11

Check the Algorithm column to make sure that each Algorithm-related cell is configured as above

8.4.3 Define the Weld material

Figure 8-12 shows the current list of S-N curves in the Verity database. This tutorial uses the default

material, Steel Weld (50%) for the WELD_TV_to_LG_tube_in_TV_tube group and Steel Weld -2SD

(2.2%) for the WELD_TV_to_LG_tube_in_LG_tube_open group.

Figure 8-12



Issue: 12.0 Date: 08.09.2016

From the Material Databases window in the verity.dbase file select the Steel Weld -2SD (2.2%)

material and in the Group Parameters table double-click on the cell defining the

WELD_TV_to_LG_tube_in_LG_tube_open group in the column titled Material. Click Yes.

Check the other options to make sure no more adjustments are required; the Fatigue from FEA dialog


Figure 8-13

If you have some options configured in Exports from a previous analysis, click Exports…, select the List

of Items tab, click None, and then click OK.

The results filename will have defaulted to a new .op2 file. Change the Output File name and extension

for a third-party post-processing format if needed.

8.4.4 Perform the analysis

To start the analysis, click the Analyse button. A summary of what you have defined is displayed.



Summary

=======


at Element 1262.4:1




Issue: 12.0 Date: 08.09.2016



94530 repeats of the loading, at element 1262, node 1, bottom layer.

A copy of the original .op2 file excluding the original steps was created, onto which a new step


Tutorial D: Spot Welds by Detect All


Issue: 12.0 Date: 08.09.2016

9 Tutorial D: Spot Welds (Detect all connecting elements)

9.1 Introduction

This tutorial outlines how to analyse spot welds in an Abaqus .odb file by using Verity®.

Figure 9-1 shows the model which consists of two sheet metal strips joined by four spot welds. Each

spot is modelled using a beam (B31) element. The FEA loading applied is lap shearing.

Figure 9-1



Issue: 12.0 Date: 08.09.2016

9.2 Preparation












9.3 Open the Model



file <DataDir>\weld_fatigue\four_beams_lap_shear_2017.odb and click Open. This will

display the Select Datasets to Read pre-scan dialogue:



Issue: 12.0 Date: 08.09.2016

Figure 9-2

The dialog box shows that there are 5 increments containing Stresses and Nodal Forces. Use the

Quick-select options to select the Stress and Force datasets for the Last increment only, and click on

Apply to Dataset List. This will result in the selections shown in Figure 9-2. Once this is done click OK.


appears. Ensure that the units are as shown in Figure 9-3, then click OK.

A dialog box will appear prompting to edit element groups loaded from the model, click No.



Issue: 12.0 Date: 08.09.2016

Figure 9-3

If the dialog box dialog box shown in Figure 9-3 does not appear automatically, then it can be displayed

by double-clicking on the icon in the Current FE Models window and selecting Properties (see

Figure 9-4).



Issue: 12.0 Date: 08.09.2016

A summary of the open model appears in the Current FE Models window, showing Stress and Force

datasets and element groups - see Figure 9-4.

Figure 9-4

Note: If the window does not appear exactly as shown in Figure 9-4, then expand the tree view to show

more details.




<ResultsDir>\four_beams_lap_shear_613Results.odb


FEA dialogue, or by clicking the adjacent browse button: .



Issue: 12.0 Date: 08.09.2016

9.3.2 Define the Welds

After the stresses and forces have been read-in the user needs to define the spot welds that are to be

evaluated. This can be done by using the Weld Preparation >> Define and Analyse Weld Geometry

menu. A new dialog box will appear. If there are weld definitions on display use the Delete All… button

to create a blank screen.

Figure 9-5

To create a weld definition click the Add… button. A pop-up menu is then displayed (see Figure 9-5).

Here, the user has two spot weld evaluation options. By selecting Spot (Detect All Connecting

Elements) the resultant definition will ensure that every spot weld in the model will be detected and then

evaluated by Verity®. This is based on the assumption that every connecting element represents a spot

weld (as stipulated in Section 3.1.3). The alternative option, Spot (Detect Connecting Element Group),

allows the user to select a group of spot welds for evaluation purposes.



Issue: 12.0 Date: 08.09.2016

In this example there are four beam elements in the entire model with each representing a spot weld.

The intention is to evaluate these four welds. The Spot (Detect All Connecting Elements) option should

therefore be selected. The resultant weld definition is displayed in Figure 9-6:

Figure 9-6

Note: If the New Spot Welds item does not appear as shown above, then expand the tree view to show

more details.

Connecting Elements

Whenever the Spot (Detect All Connecting Elements) option is selected the Connecting Elements

parameter is set to All (which is non-editable).

Nugget Diameter

The four spot weld nuggets under investigation have a diameter of 5mm. The default value will

therefore suffice.

Note that the two optional parameters (that can be viewed by selecting the Show all configuration

options checkbox) do not need to be modified, i.e. there is no need to calculate an estimate for the

nugget diameter values and the “Structural Joints / Displacement Control” I(r) function will be sufficient

for the problem at hand.



Issue: 12.0 Date: 08.09.2016


Click Save and analyse and Verity will evaluate the structural stresses. During this calculation phase a

series of calculation-related messages will be displayed in the Message Log. This information is also

written to the Verity diagnostics file in the Project Directory, such as:

<ProjectDir>\Model\verity-diagnostics.log. The log will show the detection of each beam,

and the processing of each spot weld, for the one Force dataset:

Running Verity module on all detected spot welds...


4 beam elements detected.

4 spot welds have been prepared for assessment.

4 spot welds have been processed.




a new group named WELD has been created. A summary of the Verity analysis can be found in the

verity-diagnostics.log file (select the menu option Weld Preparation >> View Structural Stress

Diagnostics Log) and a table of the structural stresses calculated can be found in the verity-ss.txt

file (select the menu option Weld Preparation >> View Structural Stress File menu).


After evaluating the structural stresses we should have a group named WELD in the Fatigue from FEA

window. The material will have defaulted to the 50% failure equivalent structural stress S-N curve.

In this tutorial the mean life curve will be used for the weld.


The loading will be constant amplitude loading with a minimum load of zero (this is sometimes referred

to as R=0). We will cycle between the stresses in the dataset multiplied by 10 and multiplied by 0.





Issue: 12.0 Date: 08.09.2016


Add... button and then select the A user-defined LOAD * dataset. Specify 10 and 0 in the Loading Scale

box:

Figure 9-7


Figure 9-8




Issue: 12.0 Date: 08.09.2016



WELD group of elements. In the Group Parameters table double-click the Algorithm column and in the

displayed dialog box select Do not analyse and click OK. This will stop analysis for elements in all

groups. Next, double-click on the cell in the Algorithm column and WELD row and select Normal

Structural Stress:

Figure 9-9

Click OK.



Issue: 12.0 Date: 08.09.2016

This tutorial uses the default material, Steel Weld (50%). Check the other options to make sure no more

adjustments are required; the Fatigue from FEA dialog box should now look similar to the following:

Figure 9-10




Figure 9-11



Issue: 12.0 Date: 08.09.2016



Summary

=======


at Element [0]187.3:1





A copy of the original .odb file excluding the original steps was created, onto which a new step



similar to:

Figure 9-12

Tutorial E: Spot Welds by Group


Issue: 12.0 Date: 08.09.2016

10 Tutorial E: Spot Welds (Detect connecting element group)

10.1 Introduction

This tutorial outlines how to analyse spot welds in an Abaqus .odb file by using Verity®.

Figure 10-1 shows the model which consists of two sheet metal strips joined by four spot welds. Each

spot is modelled using a beam (B31) element. The FEA loading applied is lap shearing.

Figure 10-1



Issue: 12.0 Date: 08.09.2016

10.2 Preparation














Issue: 12.0 Date: 08.09.2016

10.3 Open the Model



file <DataDir>\weld_fatigue\four_beams_lap_shear_613.odb and click Open. This will

display the Select Datasets to Read pre-scan dialogue:

Figure 10-2

The dialog box shows that there are 5 increments containing Stresses and Nodal Forces. Use the

Quick-select options to select the Stress and Force datasets for the Last increment only, and click on

Apply to Dataset List. This will result in the selections shown in Figure 10-2. Once this is done click OK.


appears. Ensure that the units are as shown in Figure 10-3, then click OK.

A dialog box will appear prompting to edit element groups loaded from the model, click No.



Issue: 12.0 Date: 08.09.2016

Figure 10-3

If the dialog box dialog box shown in Figure 10-3 does not appear automatically, then it can be

displayed by double-clicking on the icon in the Current FE Models window and selecting Properties

(see Figure 10-4).



Issue: 12.0 Date: 08.09.2016

A summary of the open model appears in the Current FE Models window, showing Stress and Force

datasets and element groups – see Figure 10-4.

Figure 10-4

Note: If the window does not appear exactly as shown in Figure 10-4, then expand the tree view to

show more details.




<ResultsDir>\four_beams_lap_shear_613Results.odb


FEA dialogue, or by clicking the adjacent browse button: .



Issue: 12.0 Date: 08.09.2016

10.3.2 Define the Welds

After the stresses and forces have been read-in the user needs to define the spot welds that are to be

evaluated. This can be done by using the Weld Preparation >> Define and Analyse Weld Geometry

menu. A new dialog box will appear. If there are weld definitions on display use the Delete All… button

to create a blank screen.

Figure 10-5

To create a weld definition click the Add… button. A pop-up menu is then displayed (see Figure 10-5).

Here, the user has two spot weld evaluation options. By selecting Spot (Detect All Connecting

Elements) the resultant definition will ensure that every spot weld in the model will be detected and then

evaluated by Verity®. This is based on the assumption that every connecting element represents a spot

weld (as stipulated in Section 3.1.3). The alternative option, Spot (Detect Connecting Element Group),

allows the user to select a group of spot welds for evaluation purposes.

In this example there are four beam elements in the entire model with each representing a spot weld.

The intention is to evaluate one of the four weld groups. The Spot (Detect Connecting Element Group)

option should therefore be selected. The two required parameters will be shown.



Issue: 12.0 Date: 08.09.2016

Connecting Elements

Whenever the Spot (Detect Connecting Element Group) option is selected the Connecting Elements

parameter is used to select the group containing the connecting elements of interest.

Edit the Connecting Elements parameter. Select the group SPOTWELD-BEAM4.

Nugget Diameter

The spot weld nugget under investigation has a diameter of 5mm. The default value will therefore

suffice.

Note that the two optional parameters (that can be viewed by selecting the Show all configuration

options checkbox) do not need to be modified, i.e. there is no need to calculate an estimate for the

nugget diameter values and the “Structural Joints / Displacement Control” I(r) function will be sufficient

for the problem at hand.

The resultant weld definition is displayed in Figure 10-6:

Figure 10-6

Note: If the New Spot Welds item does not appear as shown above, then expand the tree view to show

more details.



Issue: 12.0 Date: 08.09.2016


Click Save and analyse and Verity will evaluate the structural stresses. During this calculation phase a

series of calculation-related messages will be displayed in the Message Log. This information is also

written to the Verity diagnostics file in the Project Directory, such as:

<ProjectDir>\Model\verity-diagnostics.log. The log will show the detection of the beam,

and the processing of the spot weld, for the one Force dataset:

Running Verity module on all detected spot welds...


1 connecting element detected.

1 spot weld has been prepared for assessment.

1 spot weld has been processed.




a new group named WELD has been created. A summary of the Verity analysis can be found in the

verity-diagnostics.log file (select the menu option Weld Preparation >> View Structural Stress

Diagnostics Log) and a table of the structural stresses calculated can be found in the verity-ss.txt

file (select the menu option Weld Preparation>> View Structural Stress File menu).


After evaluating the structural stresses we should have a group named WELD in the Fatigue from FEA

window. The material will have defaulted to the 50% failure equivalent structural stress S-N curve.

In this tutorial the mean life curve will be used for the weld.


The loading will be constant amplitude loading with a minimum load of zero (this is sometimes referred

to as R=0). We will cycle between the stresses in the dataset multiplied by 10 and multiplied by 0.





Issue: 12.0 Date: 08.09.2016


Add... button and then select the A user-defined LOAD * dataset. Specify 10 and 0 in the Loading Scale

box:

Figure 10-7


Figure 10-8




Issue: 12.0 Date: 08.09.2016



WELD group of elements. In the Group Parameters table double-click the Algorithm column and in the

displayed dialog box select Do not analyse and click OK. This will stop analysis for elements in all

groups. Next, double-click on the cell in the Algorithm column and WELD row and select Normal

Structural Stress:

Figure 10-9

Click OK.



Issue: 12.0 Date: 08.09.2016


adjustments are required; the Fatigue from FEA dialog box should now look similar to the following:

Figure 10-10




Figure 10-11



Issue: 12.0 Date: 08.09.2016



Summary

=======


at Element 172.3:1





A copy of the original .odb file excluding the original steps was created, onto which a new step



similar to:

Figure 10-12

Tutorial F: Multiple Solid Weld Lines using Abaqus


Issue: 12.0 Date: 08.09.2016

11 Tutorial F: Automatic weld finder for Solid Welds (using Abaqus)

11.1 Introduction

This tutorial outlines an alternative way of analysing solid weld lines in an Abaqus .odb file using

Verity®.

The intention is to provide a simpler way of defining the weld domain and weld lines, especially when

multiple weld lines are to be analysed, or several different failure modes (Toe, Fusion, Throat, Root).

Figure 11-1 shows the solid hexahedral model with one row of elements through the thickness.. Note

that this is the same model as was used in Tutorial A (See section 7 above). The difference here is that

instead of inputting the weld domain and detailed information on the weld line nodes for each line and

failure mode, the weld fillet material is defined as a group is defined instead. The various weld lines and

weld domains are then automatically computed by fe-safe, although it is still necessary to provide a

domain group for root failures.

Note that as well as the fillet displayed on the underside of the intersecting tubes, there is a similar fillet

on the upper side (occluded in Figure 11-1). Both weld fillets can be supplied as a single fillet group and

the software will automatically locate the separate lines in the two fillets. fe-safe locates potential weld

lines by connecting up nodes with edges lying along the interface between fillet and non-fillet elements.

Note that this will typically produce some spurious short potential weld lines on the edges of the weld-

ends. However these are rejected, because they form closed loops including a fillet edge, and will be

labelled as Weld_End lines in the fe-safe message logs. The root lines are also located, either because

these are formed from internal surface nodes in fully penetrative welds, or lie on a fillet-parent material

interface where there are two faces with significant change in normal direction.

This analysis is based upon a solid hexahedral model with one row of elements through the thickness.

Figure 6-1 shows the model:

Figure 11-1



Issue: 12.0 Date: 08.09.2016

The brown area shows the weld fillet. The dark blue elements are the one used in the weld line

previously analysed in Tutorial A, but the automatic solid weld finder software used in this tutorial

locates three other toe domains: one on the other side of the displayed fillet, and a similar pair of toe

domains on the other side (occluded in Figure 11-1).

The tutorial uses an Abaqus ODB model. However, the same techniques can be applied to all FE

formats supported for Verity-in-fe-safe, which currently include:


Abaqus ODB files

Nastran OP2 files

Only the definition of the lines (and the number of lines) differs from the “classic” Verity method of

Tutorial A; once the software has located the weld lines and extracted the Verity local coordinate

system, the Verity method proceeds as before (although there can be some differences in exactly how

the though thickness domain has been assigned, and also the assumed plate thickness will default to

being determined by the extracted geometry, not the specified nominal thickness). Note that minor

differences may also be observed in Verity results due to slight differences in the way that surface

normals, and hence the Verity local coordinate system have been extracted. These differences would

probably appear on curved sections when using quadratic elements.

11.2 Preparation












Issue: 12.0 Date: 08.09.2016

11.3 Open the Model


To open the model, select File >> FEA Solutions >> Open Finite Element Model... On the file selection

dialog box select the file <DataDir>\weld_fatigue\solid-tubess3_xx.odb and click Open.

When prompted ‘Do you want to pre-scan solid-tubes3*.odb’ click Yes and the Select Datasets to Read

pre-scan dialog box will appear:

Figure 11-2

The dialog box shows that the last increment in the step contains a dataset containing Stresses (used

to generate loading for Verity in fe-safe) and Nodal Forces (used by Verity®). Ensure that both are

selected and click OK.

The stresses will be extracted from the model and information similar to that shown below will be

displayed in the Message Log window. This information is also written to the file

<ProjectDir>\Model\reader.log.



Issue: 12.0 Date: 08.09.2016




Reading : C:\data\verity\solid-tubess3_610.odb

Writing : Pre-scan

Strains : Read all (Pre-scanning)

ODB interface exe exited with code 62097

Pre-scan complete.

odb_io completed

NOTE : Storing current session macro commands to

D:\Data\Users\fe-safe.version.6.3\output\current.macro




Reading : C:\Program Files (x86)\Safe_Technology\fe-

safe\version.6.3\data\verity\solid-tubess3_610.odb pre-scan

Writing : D:\Data\Users\fe-safe.version.6.3\projects\project_01\model\FESAFE.FED

Selected position is Element Nodal

Strains : Selected datasets

Dataset : S

From Step : 1

Iteration : 1

Time : 2.22e-16

Title :

Direct Min/Max : -33612.7 50056.2

Shear Min/Max : -30004.6 19650.8

No. Elements : 4412



Issue: 12.0 Date: 08.09.2016


appears, as shown in Figure 11-3.

Figure 11-3

Set the Loaded FEA Models Properties to Stress Units of psi and Force Units of lbf and Distance Units

of in as shown. Ensure that the units are as shown in Figure 11-3, then click OK.



Issue: 12.0 Date: 08.09.2016



Figure 11-4

A dialog box will show prompting to edit element groups loaded from the model, click No.



Issue: 12.0 Date: 08.09.2016

A summary of the open model appears in the Current FE Models window, showing the loaded datasets

and element group information – see Figure 11-5.

Figure 11-5

Note: If the window does not appear exactly as shown in Figure 11-5, then expand the tree view to

show more details.




<ResultsDir>\solid-tubess3_xxResults.odb


FEA dialog box, or by clicking the adjacent browse button: .



Issue: 12.0 Date: 08.09.2016

11.4 Fillet Group

The only compulsory input to this style of solid weld definition, is that a group must be defined in fe-safe

for the weld fillet material (which may comprise in fact several fillets). Sometimes this group may have

been loaded automatically with the Finite Element model, for example as one of the existing element

set, material or property groups. If so then this group can be selected, but here we assume that the

required group has not been identified in the pre-processor and needs to be defined as an elemental

group in the fe-safe Group Manager. Using the Group Manager this list of numbers can be used to

define the group in order that it be available to the automatic weld line extraction. In this case the weld

fillet elements are in the range 4293-4320 and 2083-2110 (inclusively). This group will not be configured

for analysis in the list of Group Parameters, but it will be used below to define the weld.

Select FEA Fatigue >> Manage Groups… to display the Select Groups to Analyse dialog box. Define a

new (elemental) group in the in the Advanced Group Creation area of the dialog box with a Name:

Weld_Fillet. Set the value of the group to the series of IDs: 4293-4320,2083-2110 then click the Create

Button. The Weld_Fillet (56 elements) group will appear in the list of Unused Groups. Click OK to

dismiss the dialog box.

Alternatively, a file definition of this group has been provided, in a text file according to the format for

User-Defined ASCII group files in fe-safe (see Appendix E of the fe-safe documentation for details).

This file is in the data directory of your installation along with the model file for this exercise. To load the

group file, select the Load… button from the Group Manager dialog box, and select the file

<DataDir>\weld_fatigue\solid_tubes3_weld_fillet_group.txt. You may need to

change the file type to Text files (*.txt) to locate your required group definition file. Click Open. After

selecting the fillet group definition, select Load As Elemental and click OK when prompted, and then

click OK to dismiss the Group Manager dialog box when the Weld_Fillet (56 elements) group appears

in the list of Unused Groups.

Note that if the fillet was already defined in the FEA model file by an existing group, then there would be

no need to create a new one



Issue: 12.0 Date: 08.09.2016

11.5 Define the Weld



Geometry… menu. A new dialog box will be displayed:

Figure 11-6


use the Delete All… button to clear the tree. Click Yes when prompted to delete.

To create a weld, click the Add… button and select Line Welds (From Weld Fillet Elements) as the type

of the weld to be defined. This will add a new weld definition to the tree with the name New Solid

Weld:

Note: If the New Weld Line does not appear as shown in Figure 11-7, then expand the tree view to

show more details, and click the checkbox to show all configuration options.



Issue: 12.0 Date: 08.09.2016

Figure 11-7



Name

Rename the weld line by or double-clicking on the New Line Weld value. This will allow the item to be

edited in the tree. Change the weld name to Toes_LG_tube__in_TV_tube and press the enter key

() to accept it.



Issue: 12.0 Date: 08.09.2016

Fillet Group

Double click on the blank Fillet Group value and then select the previously created Weld_Fillet group in

the resulting list. Click OK.

Figure 11-8

Failure Modes

In this analysis we only consider toe failures, so double click on the Auto-detect Fusion Failures and

change the setting to Off.

The analysis definition is now complete.

The analysis definition can also be obtained by using the Open button in the Weld Definition dialog and

loading <DataDir>\weld_fatigue\solid_tubes3_solid_weld_toe.wdf (you may need to

change the file type filter from .def to .wdf).



Issue: 12.0 Date: 08.09.2016

The updated weld definition dialog box should now look like Figure 11-9 below.

Figure 11-9



Issue: 12.0 Date: 08.09.2016

Evaluate the structural stresses

Now click Save and analyse and the automatic weld finder for solid welds will identify 4 weld lines along

the interface edges between the fillet and the parent tube material. The weld toe element and node

domains will be automatically defined along with the associated through thickness nodes (when

necessary) for each weld line node. Based on the extracted domains and line geometry, Verity will

evaluate the structural stresses. After the structural stresses are calculated, information similar to that

shown below will be displayed in the Message Log window. This information is also written to the file

<ProjectDir>\Model\verity-diagnostics.log.

Number of nodes found on line 0 is 57

Start and end nodes are: n2062 , n354

Classification of line 0 is Weld_Line_Surface



Classification of line 1 is Weld_End







…



Classification of line 17 is Fillet_Edge

Processing weld Toe_Welds_Tube_Toe_1.




Processing forces dataset 1.1 at time 2.22e-016.

Processing forces for weld Toe_Welds_Tube_Toe_1.






Issue: 12.0 Date: 08.09.2016





a new group named WELD[Toe] has been created, which appears in the Group Parameters area of the

Analysis Settings tab.

A summary of the Verity analysis run can be found in the verity-diagnostics.log file (Weld

Preparation>> ViewStructural Stress Diagnostics Log menu) and a table of the structural stresses

calculated can be found in the verity-ss.txt file (Weld Preparation>> View Structural Stress File

menu).

The automatically located toe weld domains are displayed in red in Figure 11.10. Note that the Wrap

Around Angle parameter offers further control of exactly which elements are included around the ends

of the weld line (see 3.1.2).

Figure 11-10



Issue: 12.0 Date: 08.09.2016

Set the Wrap Around Angle parameter to 90 to achieve the correct weld domain for the topmost and

bottommost weld lines shown.

Note: a warning containing the information that an item in the definition that has no instance defined,

may be displayed. fe-safe will try to accommodate for the lack of instance, which should allow the

tutorial to continue without issue.


After evaluating the structural stresses using Verity we should have a group named WELD[Toe] in the

Group Parameters area of the Fatigue from FEA window. The material will have defaulted to the mean

weld master S-N curve called Steel Weld 50% The Algorithm will have defaulted to SSM-Normal:-

None. In this tutorial the mean life curve will be used for the weld. The remainder of the model will not

be analysed as we are only interested in the weld in this context (see Tutorial A if the entire model is to

be analysed).


The loading will be a fully reversed load with R=-1. We will cycle between the stresses in the dataset

multiplied by 1 and multiplied by –1. This is the normal default loading in fe-safe, so some of the

following steps may not be necessary.

dialog box In the Current FE Models window, select the stress dataset corresponding to step 1

increment 1 of the loaded FE model: Dataset 1: (1.1) S : Stress.


mouse click on the loadings tree and select Clear all loadings to remove any existing loading. Click Yes

when prompted to Delete Loadings.

Click the Add... button and then select the A user-defined LOAD * dataset. Click OK when prompted

with the Automatic Block Creation dialog box.

In the Dataset Embedded Load History dialog box, specify 1 and –1 in the Loading Scale box:

Figure 11-11



Issue: 12.0 Date: 08.09.2016

Click OK. This will add an item to the newly created Elastic Block as shown:

Figure 11-12


11.6.2 Limit analysis to the weld toe groups

In the Group Parameters table, select all of the groups except the WELD[Toe] group and right-click the

Algorithm column and select Edit…. In the displayed dialog box select Do Not Analyse and click OK.

Figure

11-13


adjustments are required.



Issue: 12.0 Date: 08.09.2016


The Fatigue from FEA dialog box should now look similar to Figure 11-14:

Figure 11-14

The Output File field will have defaulted to a new .odb file.




Summary

=======


at Element 15.5



The analysis log shows that the worst-case life for the four toe domains is:

305139.906 repeats of the loading, at element 15, node 5.

A copy of the original .odb file was created, onto which a new step containing the fatigue results was

appended.



Issue: 12.0 Date: 08.09.2016


similar to:

Figure 11-15

Tutorial H: Multiple Solid Weld Lines using Ansys


Issue: 12.0 Date: 08.09.2016

12 Tutorial G: Automatic weld finder for Solid Welds (Using Ansys)

12.1 Introduction

This tutorial outlines an alternative way of analysing solid weld lines in an Ansys .rst file using Verity®.

The intention is to provide a simpler way of defining the weld domain and weld lines, especially when

multiple weld lines are to be analysed, or several different failure modes (Toe, Fusion, Throat, Root).

Figure 12-1 shows the solid hexahedral model with two rows of elements through the thickness. Note

that this is the same model as was used in Tutorial B (See section 8 above). The difference here is that

instead of inputting the weld domain and detailed information on the weld line nodes for each line and

failure mode, the weld fillet material is defined as a group instead. The various weld lines and weld

domains are then automatically computed by fe-safe, although it is still necessary to provide a domain

group for root failures.

Note that as well as the fillet displayed on the underside of the intersecting tubes, there is a similar fillet

on the upper side (occluded in Figure 12-1). Both weld fillets can be supplied as a single fillet group and

the software will automatically locate the separate lines in the two fillets. fe-safe locates potential weld

lines by connecting up nodes with edges lying along the interface between fillet and non-fillet elements.

Note that this will typically produce some spurious short potential weld lines on the edges of the weld-

ends. However these are rejected, because they form closed loops including a fillet edge, and will be

labelled as Weld_End lines in the fe-safe message logs. The root lines are also located, either because

these are formed from internal surface nodes in fully penetrative welds, or lie on a fillet-parent material

interface where there are two faces with significant change in normal direction.

This analysis is based upon a solid hexahedral model with two rows of elements through the thickness.

Figure 6-1 shows the model:



Issue: 12.0 Date: 08.09.2016

Figure 12-1

The tutorial uses an Ansys RST model. However, the same techniques can be applied to all FE formats



Abaqus ODB files

Nastran OP2 files

Only the definition of the lines differs from the “classic” Verity method of Tutorial B; once the software

has located the weld lines and extracted the Verity local coordinate system, the Verity method proceeds

as before (although there can be some differences in exactly how the though thickness domain has

been assigned, and also the assumed plate thickness will default to being determined by the extracted

geometry, not the specified nominal thickness). Note that minor differences may also be observed in

Verity results due to slight differences in the way that surface normals, and hence the Verity local

coordinate system have been extracted. These differences would probably appear on curved sections

when using quadratic elements.

12.2 Preparation








Issue: 12.0 Date: 08.09.2016





12.3 Open the Model



file <DataDir>\weld_fatigue\t-tube_sld_v12.rst and click Open. When prompted ‘Do you

want to pre-scan t-tube_sld_v12.rst’ click Yes and the Select Datasets to Read pre-scan dialog box will

appear:

Figure 12-2

The dialog box shows that there is a dataset containing stresses, strains, temperatures and forces.

Ensure that Forces and Stress datasets are selected and click OK.



Issue: 12.0 Date: 08.09.2016



Figure 12-3

Set the Loaded FEA Models Properties to Stress Units of psi and Force Units of lbf and Distance Units

of in as shown. Ensure that the units are as shown in Figure 12-3, then click OK.



Issue: 12.0 Date: 08.09.2016



Figure 12-4

Ensure that the units are as shown in Figure 12-3, then click OK. A dialog box will appear prompting to

edit element groups loaded from the model, click No.

12.4 Fillet Group

The only compulsory input to this style of solid weld definition, is that a group must be defined in fe-safe

for the weld fillet material (which may comprise in fact several fillets). Sometimes this group may have

been loaded automatically with the Finite Element model, for example as one of the existing material

groups. If so then this group can be selected, but here we assume that the required group has been

defined in a text editor but needs to be loaded into the fe-safe Group Manager, in order that it be

available to the automatic weld line extraction. The file should contain a list of element identifiers for the

fillet material (NOT the weld domain!), in a text file according to the format for User-Defined ASCII

group files in fe-safe (see Appendix E of the fe-safe documentation for details). A file is provided in the

data directory of your installation along with the model file for this exercise. The group contains the

following elements representing the weld fillet material:

GROUP both_fillets

1667-1711, 5089-5133

3378-3422, 6800-6844

END



Issue: 12.0 Date: 08.09.2016

Select FEA Fatigue >> Manage Groups… to display the Select Groups to Analyse dialog box. To load

the User-Defined ASCII group file, select the Load… button from the Group Manager dialog box, and

select the file <DataDir>\weld_fatigue\t-tube_sld_both_fillets_group.txt. You may

need to change the file type to Text files (*.txt) to locate your required group definition file. click Open.

After selecting the fillet group definition, select Load As Elemental and click OK when prompted.

The group both_fillets (180 elements) will appear in the list of Unused Groups. Click OK to dismiss the

dialog box.

Note that if the fillet was already defined in the FEA model file by an existing group, then there would be

no need to create a new one

12.5 Define the Welds



Geometry menu. A new dialog box will be displayed:

Figure 12-5



Issue: 12.0 Date: 08.09.2016


use the Delete All… button to clear the tree. Click Yes when prompted to delete.

To create a weld, click the Add… button and select Line Welds (From Weld Fillet Elements) as the type

of the weld to be defined. This will add a new weld definition to the tree with the name New Line

Welds:

Note: If the New Line Welds options do not appear as shown in Figure 12-6, then expand the tree view

to show more details, and click the checkbox to show all configuration options.

Figure 12-6





Issue: 12.0 Date: 08.09.2016

Fillet Group

Double click on the blank Fillet Group value and then select the previously loaded both_fillets group.

Click OK.

Figure 12-7

Failure Modes

By default Toe failures are automatically enabled, but for this tutorial we will analyse both Toe and

Fusion failure modes, so double click the Auto-Detect Fusion Failures value fields, and select On from

the drop-down menu box which then appears.

Ensure that the Shell Root Failures and Throat failures are set to off. If not, double click the fields and

change the value.

The analysis definition can also be obtained by using the Open button in the Verity definition dialog and

loading <DataDir>\weld_fatigue\t-tube_sld_solid_weld_toe-fusion.wdf (you may

need to change the file type filter from .def to .wdf).



Issue: 12.0 Date: 08.09.2016

The updated weld definition dialog box should look like Figure 12-8 below.

Figure 12-8

Evaluate the structural stresses

The weld definition for both fillets is now complete – fe-safe will extract the weld lines and domains

automatically. Note that nodes will be added to the weld domain either until a specified plate thickness

is reached, or a surface is reached, or any optionally specified domain is exhausted. In this case we

simply want to go all the way through the (hollow) tubes, so it is not necessary to specify any plate

thickness. Click Save and Analyse.



Issue: 12.0 Date: 08.09.2016

The weld toe element and node domains will be automatically defined along with the associated

through thickness nodes (when necessary) for each weld line node. Based on the extracted domains

and line geometry, Verity will evaluate the structural stresses. After the structural stresses are

calculated, information similar to that shown below will be displayed in the Message Log window. This

information is also written to the file <ProjectDir>\Model\verity-diagnostics.log.










…



Classification of line 17 is Fillet_Edge

Weld Element Domains:

Domain: 1

Name: both_fillets

Elements:

1667-1711,3378-3422,5089-5133,6800-6844

Output Parameters:

Output Directory: C:\installations\user_directories\fe-

safe.version.2016_15943\projects\project_07\jobs\job_01\fe-results

Output Printed Structural Stresses to file verity-ss.txt

Processing weld New Line Welds_Toe_1.




Processing weld New Line Welds_Fusion_1.





Processing forces for weld New Line Welds_Toe_1.






Issue: 12.0 Date: 08.09.2016

Processing forces for weld New Line Welds_Fusion_1.







In total, 8 weld lines have been analysed by Verity consisting of two Toe and two Fusion failure modes

on both fillets. The structural stresses evaluated by Verity have been merged in with the stresses read

by fe-safe and two new groups named WELD[Fusion] and WELD[Toe] have been created. A summary

of the Verity analysis run can be found in the verity-diagnostics.log file (Weld Preparation >>

ViewStructural Stress Diagnostics Log menu) and a table of the structural stresses calculated can be

found in the verity-ss.txt file (Weld Preparation>> View Structural Stress File menu).

The automatically determined Toe domains are illustrated in Figures 12.10 and 12.11.

Two Weld groups have been added to the analysis panel (for Toe and Fusion failures). The analysis

panel will look as follows:

Figure 12-9



Issue: 12.0 Date: 08.09.2016

Figure 12-10



Issue: 12.0 Date: 08.09.2016

Figure 12-11


After evaluating the structural stresses using Verity we should have the two groups named

WELD[Fusion] and WELD[Toe] in the groups parameters area of the Fatigue from FEA window. The

material will have defaulted to the 50% failure equivalent structural stress S-N curve, and the algorithm

to SSM-Normal.

In this tutorial the mean life curve will be used for the weld. The remainder of the model will be

configured for no analysis.


The loading will be a fully reversed load with R=-1. We will cycle between the stresses in the dataset

multiplied by 1 and multiplied by –1. This is the normal default loading in fe-safe, so some of the

following steps may not be necessary.

In the Current FE Models window, select the stress dataset corresponding to step 1 increment 1 of the

loaded FE model: Dataset 1: (1.1) S : Stress.



Issue: 12.0 Date: 08.09.2016


mouse click on the loadings tree and select Clear all loadings to remove any existing loading. Click Yes

when prompted to Delete Loadings.

Click the Add... button and then select the A user-defined LOAD * dataset. Click OK when prompted

with the Automatic Block Creation dialog box.

In the Dataset Embedded Load History dialog box, specify 1 and –1 in the Loading Scale box:

dialog box

Figure 12-12

Click OK. This will add an item to the newly created Elastic Block as shown:

Figure 12-13




Issue: 12.0 Date: 08.09.2016

12.6.2 Limit analysis to the Weld groups

Select the non-Weld groups in the Algorithm column,right-click the Algorithm column and select Edit….

In the displayed dialog box select Do Not Analyse and click OK.

Figure 12-14

Click OK.


adjustments are required;


the Fatigue from FEA dialog box should now look similar to the following:

Figure 12-15



Issue: 12.0 Date: 08.09.2016

The results filename will have defaulted to a new .rst file.




Summary

=======


at Element 2809.3


The analysis log shows that the worst-case life for the eight weld domains is:


A copy of the original .rst file excluding the original steps was created, onto which a new step

containing the fatigue results was written. The exported fatigue results are written to the Ansys variable

SX and contain the log10(life) contour of fatigue lives, which should look similar to Figure 12-16 and

Figure 12-17. Both sides of the welding are visible in Figure 12-17, but the model loading was applied

from the left end, so the left side (seen in Figure 12-16) has the lowest life elements (seen in red). Note

that the values appearing on the surface of the fillet are from the Fusion failure analysis.



Issue: 12.0 Date: 08.09.2016

Figure 12-16



Issue: 12.0 Date: 08.09.2016

Figure 12-17



Issue: 12.0 Date: 08.09.2016

13 Tutorial H: Automatic Weld Finder for Shell Welds (Using Nastran)

13.1 Introduction

This tutorial outlines how to analyse welded structures in a Nastran .op2 file using the fe-safe shell weld

finder and Verity®. See also Tutorial C, in which the legacy manual line definitions are used to perform

a similar analysis..

Figure 13 -1 shows a linear shell model which is based on models created for the SAE Fatigue

Challenge in 2003 and 2004.

Figure 13-1

The weld line definitions will be generated from a group of weld fillet elements (yellow material above).

Whereas the legacy line definition required a separate complex line definition for each weld line and

failure mode, the weld finder only needs a group of all fillet elements. The weld finder then locates weld

lines and determines all other information needed. The located domains in this example will be

analysed for weld fatigue at toes. The remainder of the model (except the weld material shown in

yellow) will be analysed using a local strain approach and the material Manten steel.

The tutorial uses a Nastran OP2 model. However, the same techniques can be applied to all FE formats



Abaqus ODB files

Nastran OP2 files



Issue: 12.0 Date: 08.09.2016

13.2 Preparation










Start a new project, if an existing project is loaded select [File >> Project >> New Project…] and choose

a new location for a project.

13.3 Open the Model

13.3.1 Select the datasets


file <DataDir>\weld_fatigue\t-tube_2k4_pn2.op2 and click Open. This will display the Select

Datasets to Read pre-scan dialogue:

Figure 13-2



Issue: 12.0 Date: 08.09.2016






Figure 13-3




Issue: 12.0 Date: 08.09.2016

A dialog box will appear prompting to edit element groups loaded from the model, click Yes.

Figure 13-4

Use the button to move the group associated with Property 3 (property definition 3 in the

Nastran input file) which represents the elements in the weld fillet, and the group associated with

Matierial 1 which represents all elements in the whole model with material definition 1. The Default

group will be used to limit analysis in the weld elements which are not part of the remaining Property 1

and Property 2 groups. Click OK.

If the Current FE Models Properties dialog box dialog box does not appear automatically, then it can be

displayed by right-clicking on the icon in the Current FE Models window and selecting Properties.

If the Edit Group List prompt does not appear, the Select Groups to Analyse dialog box can be opened

using FEA Fatigue >> Manage Groups… menu selection.



Issue: 12.0 Date: 08.09.2016

13.3.2 Define the welds


structural stresses. This can be done using the Weld Preparation>> Define and Analyse Weld

Geometry menu. A new dialog box will be displayed. Initially this may contain the weld definitions that

were used in a previous Verity run (e.g. a previous Tutorial):

Figure 13-5



For this tutorial, click the Add… button, and then select the Line Welds (From Weld Fillet Elements)

option

A new weld finder definition will appear called New Line Welds. Double click on the name and change it

to TV_to_LG_tube. Click on the arrow icon to expand the property tree so that you can set the fillet

group.



Issue: 12.0 Date: 08.09.2016

Figure 13-6

Double-click on the Fillet Group field to obtain the group selection dialog, and then select Property 3 as

the group (see Figure 13-1 for groups).

Figure 13-7



Issue: 12.0 Date: 08.09.2016

This completes the weld definition, as shown below in Figure 13-8. Select the Merge all groups into

‘WELD’ group checkbox, deselect the Output Extensive Diagnostics checkbox, then click Save and

Analyse , which will generate the weld line definitions and associated element domains, and then

invoke Verity to calculate their associated structural stresses. A group will be created called

WELD[Toe]. Information similar to that shown below will appear in the Message Log, and in the

structural stress diagnostics log (file <ProjectDir>\Model\verity-diagnostics.log).

Verity Diagnostic Log

Deleting Verity warning groups:

Weld Element Domains:

Domain: 1

Name: Property 3

Elements:

1145-1204,4589-4648

Output Parameters:

Output Directory: E:\testprojects\Verity\shelltut\jobs\job_01\fe-results

Output Printed Structural Stresses to file verity-ss.txt

Verity structural stress file backed up to

E:\testprojects\Verity\shelltut\model\verity-ss.arc.

Processing weld TV_to_LG_tube_1[Toe].




Processing forces dataset 1.1.

Processing forces for weld TV_to_LG_tube_1[Toe].








After evaluating the structural stresses using Verity there should be a group called WELD[Toe] in the

Fatigue from FEA window. The algorithm will have defaulted to SSM-Normal (SSM means Structural

Stress Method), and the material will have defaulted to the 50% failure equivalent structural stress S-N

curve. The remainder of the base materials will be configured for analysis with Manten steel.



Issue: 12.0 Date: 08.09.2016

13.4.1 Define the loading







box:

Figure 13-6


Figure 13-7




Issue: 12.0 Date: 08.09.2016

13.4.2 Define the algorithm




Double-click the cell defining the Default group in the Algorithm column and select Do not analyse and

click OK.



material and in the Group Parameters table select the Property 1 and Property 2 groups, then right click

and select Edit. You will be asked if you wish to set the material, click Yes.



Figure 13-8




displayed in the Message Log window, ended with the followng summary.

Summary

=======


at Element 1262.4:1




Issue: 12.0 Date: 08.09.2016

This shows the worst model life or 294075.8 repeats occurred at element 1262 node 4, bottom layer. A

copy of the original .op2 file excluding the original steps was created, onto which a new step containing

the fatigue results was written.

13.5 Specific line analysis

The next section shows how the weld finder can be configured to create more specific line groups, so

that different master S-N curves can be assigned. Suppose that it is desired to analyse the weld toes on

the TV tube with the 50% Steel-Weld mean material, whilst the toes on the LG tube require the more

conservative 2.2% failure equivalent structural stress S-N curve. This requires us to generate individual

line groups that allow this distinction. Unchecking the Merge all groups into ‘WELD’ group checkbox on

the Weld Definition dialog would generate four separate line groups, but the group naming in terms of

line index counts does not make it immediately apparent which lines are from the TV tube, and which

from the LG. For this it is necessary to provide two separate weld finder definitions. Each definition uses

the same fillet material group, but the toe domains are specifically defined as coming from the TV tube

or LG Tube respectively. Note that the domain groups provided to the weld finder will be the complete

set of all elements in the TV or LG tube (Property 2 or Property 1 groups respectively). It is not

necessary to restrict the domain group to only the actual weld domain, as this will still be determined by

the weld finder; but the domain definition restricts the weld finder selection to those elements contained

in the defined domain. This contrasts with the Legacy Line Definition, which would require the specific

weld domain.



Issue: 12.0 Date: 08.09.2016

13.5.1 Define the line welds

Reopen the Weld Definition dialog (Weld Preparation > Define and Analyse Weld Geometry…), and

press the Delete All button to remove the previous definition. Click Add.. then select Line Welds (From

Weld Fillet Elements) to create a new weld definition. Double click on the default name of New Line

Welds and change it to TV_to_LG_tube_in_TV_tube to create a definition for the TV tube. Repeat the

process to create a definition for the LG tube, modifying the weld definition name to

TV_to_LG_tube_in_LG_tube_open. Expand the tree for each definition in turn and select the Fillet

Group to be group Property 3 as before to obtain the definitions below.

Figure 13-1

Next deselect the Merge all groups into WELD group checkbox so that individual line groups will be

created.

In order to associate each definition with its respective tube, it is necessary to set the Toe Domain for

each. Click the Show all configuration options checkbox which will extend the options displayed. In the

first TV_to_LG_tube_in_TV_tube, double click on the Toe Domain value field, and select Property 2 (the

TV tube) from the disaplyed list of element groups.



Issue: 12.0 Date: 08.09.2016

This results in the definition below:

Figure 13-12

Similarly set the Toe Domain Group for the LG Tube in TV_to_LG_tube_in_LG_tube_open to be

Property 1.

This weld definition configuration may also be obtained by opening (click Open…) weld definition file

<DataDir>\weld_fatigue\t-tube_2k4_pn2_auto.wdf.

Click Save and analyse to invoke the weld finder’s generation of the line definitions, and their

consequent processing by Verity to generate structural stresses and associated fe-safe analysis

groups.



Issue: 12.0 Date: 08.09.2016

13.5.2 Analyse the Model

After running the weld finder and evaluating the structural stresses using Verity we should have four

groups starting with WELD in the Fatigue from FEA window, one for each individual line. The algorithm

will have defaulted to SSM-Normal, and the material will have defaulted to the 50% failure equivalent

structural stress S-N curve for both groups. The analysis panel should look as below.

Figure 13-13

This default weld material will suffice for the toe of the TV tube, whilst the toe on the LG tube will need

to be represented by the 2.2% failure equivalent structural stress S-N curve.

The remainder of the base materials will be configured for analysis with Manten steel.

Figure 13-9



material and in the Group Parameters table double-click the column titled Material. You will be asked if

you wish to set the material, click Yes.

Figure 8-12 shows the current list of S-N curves in the Verity database. This tutorial uses the default

material, Steel Weld (50%) for the two_TV_to_LG_tube_in_TV_tube derived line element groups and

Steel Weld -2SD (2.2%) for the two_TV_to_LG_tube_in_LG_tube_open line element groups.



Issue: 12.0 Date: 08.09.2016

Figure 13-10

From the Material Databases window in the verity.dbase file select the Steel Weld -2SD (2.2%)

material and in the Group Parameters table double-click on the cell defining the

WELD_TV_to_LG_tube_in_LG_tube_open _3[Toe] group in the column titled Material. Click Yes. Then

repeat similarly for the other TV_to_LG_tube_in_LG_tube_open derived weld group.



Figure 13-4



Issue: 12.0 Date: 08.09.2016





Summary

=======


at Element 1262.4:1




94530 repeats of the loading, at element 1262, node 1, bottom layer.

A copy of the original .op2 file excluding the original steps was created, onto which a new step

containing the fatigue results was written. Note that the life is now lower than in the previous analysis

because a more conservative material has been used for the toe domain in the LG tube.



Issue: 12.0 Date: 08.09.2016

References


Issue: 12.0 Date: 08.09.2016

14 References

1. P Dong & JK Hong (2006). A Robust Structural Stress Parameter for Evaluation of Multiaxial

Fatigue of Weldments, in Journal of ASTM International 3(7):1-17, 2006.

2. JK Hong & TP Forte (2014), Fatigue Evaluation Procedures for Multiaxial Loading in Welded

Structures using Battelle Structural Stress Approach, in Proceedings of the ASME 2014 33rd

International Conference on Ocean, Offshore and Arctic Engineering.