newsletter - enginsoft · 7 - newsletter enginsoft year 12 n°3 case histories due to the black...

010100101110101001110101100101001101110010100101010100101110101001110101100101001

101110

01

0100101001010010

1010

10010111011001

011 Newsletter

Simulation Based Engineering & SciencesYear n°3 Autumn 201512

South Stream Project: the Flute System for laying of pipelines under the Black Sea

Seismic analysis of actuator for Ball Valves

The importance of quick responses and mechanical reliability: Structural analysis of a Spherical Valve

Structural analysis for an hoisting machine for lifts

Productivity and accuracy in piping modelling with ANSYS SpaceClaim Direct Modeler

A new 100,000-tonne press for forging turbine wheels for power generation applications

Axial flow fans design for automotive cooling

3 - Newsletter EnginSoft Year 12 n°3 Flash

LASHThe symbol of this year’s International CAE Conference is the Nautilus, a marine mollusc, considered a living fossil from the Triassic period. Its spiral shell is iconic and presents the finest natural example of a logarithmic spiral, also known as the growth spiral. To ancient Greek philosophers the logarithmic spiral personifies the ‘golden mean’ or the desirable mid-point between two extremes and was considered an attribute of beauty. With the modern capabilities of advanced simulation software to trade-off design objectives, the beauty of natural design can now be found in product design everywhere.

Engineering simulation today has developed tremendously since EnginSoft was founded, and I am delighted to share in this edition the continued advancements made through the implementation of engineering simulation technologies.

The ability to satisfy multiple constraints and design requirements is becoming more complex as market demand and pressures increase. On page 11, Cameron uses a simulation approach to investigate the durability of ball valves to meet the high technical standards required for reliability in the Oil & Gas sector.

In the fast evolving automotive industry there is high demand for improved design, at reduced cost, sooner to market. These objectives are driven by the consumer who demand quality, at an affordable price, today. Magneti Marelli Powertrain describes the use of the latest technology to develop an integrated air intake system for the internal combustion engine that will contribute to the improvement in the overall performance of the vehicle through reducing energy loss.

I look forward to seeing many of you at the International CAE Conference and hearing many more success stories.

Stefano Odorizzi, Editor in chief

FFor many industries today, engineering simulation is the most powerful tool used in product development. It is unparalleled in the speed in which designs can be understood and each design requirement met. This fact is no longer in doubt.

Slow and high cost developmental testing has in the large part been replaced by simulation. However, why stop there? Is there an even better method? What is the current ‘game changer’ in engineering? These questions are answered by a tool that allows the engineer to standing back from the problem, explore the whole design space through appropriate data mining tools, while an automated process considers all the relationships between each design variable and drives the design towards its optimum. This tool is modeFRONTIER.

A special edition of the Newsletter, dedicated to modeFRONTIER, highlights how it has impacted many industries in the way they do their design and development; indeed you will be amazed by how many familiar everyday products have benefitted from this approach and are featured in this edition – please enjoy and become part of the growing community of business leaders that are reaping the rewards in an increasingly competitive world market.

010100101110101001110101100101001101110010100101010100101110101001110101100101001

101110

01

0100101001010010

1010

10010111011001

011 Newsletter

Simulation Based Engineering & Sciences

Multi-Objective Optimization of Vehicle Handling & Comfort Performances

Structural Optimization of a Car-body High Speed Train:an innovative analysis

Robust Design Optimization of a Bumper System

Mechanical optimization of the injection system in acompression molding machine

Special Issue on

Pioneers in numerical optimization solutions

Parametric FEM model optimization for a pyrolitic oven

DOWNLOAD your free copy on:www.enginsoft.com/mf-newsletter

Simulation Based Engineering & Sciences

EnginSoft Newsletter

Special Issue on

5 - Newsletter EnginSoft Year 12 n°3 Contents

Contents

Newsletter EnginSoft Year 12 n°3 - 4

CASE HISTORIES6 South Stream Project - De Pretto Industrie & EnginSoft: the joined design of the Flute System to be installed under the S7000 JL-Tower11 Seismic analysis of actuator for Ball Valves14 The importance of quick responses and mechanical reliability: Structural analysis of a Spherical Valve16 Structural analysis for an hoisting machine for lifts18 Tolerances analysis on a nitrogen spring21 Productivity and accuracy in piping modelling with ANSYS SpaceClaim Direct Modeler24 Axial flow fans design for automotive cooling26 Forging Ahead - A new 100,000-tonne press for forging turbine wheels for power generation applications30 New Methodology: Intercooler Integration Space & Efficiency Optimization33 Numerical Optimization of Polymer Die Design for Multiple Objectives37 Blasting Simulation drives plants design for safer workplaces37 Multi-Objective Optimization of Space Frame Structures with Application to Motorcycle Chassis44 3D digitalization of your product to simulate its behaviour48 Calibration of Material Models for the Numerical Simulation of Aluminium Foams – MAT 154 for M-PORE Foams @ 3 Loads

RESEARCH53 Greenhouse Module for Space System57 RLW Navigator project A success story59 Disseminating the Music Project

SOFTWARE UPDATE60 ANSYS 16: Structures and Multiphysics focus62 What is System Level Thermo-Fluid Analysis?

HARDWARE UPDATE64 E4 Computer Engineering and AppliedMicro Demonstrate the Effectiveness of Power-efficient X-Gene®-based technology in Computational Fluid Dynamics Environments

COMUNICATE66 INDUSTRIO: A startup Renaissance for the Industry68 Ingredients and Recipes for Innovation: One Size doesn’t Fit All!

OUR ACKNOWLEDGEMENT AND THANKS TO ALL THE COMPANIES, UNIVERSITIES AND RESEARCH CENTRES THAT HAVE CONTRIBUTED TO THIS ISSUE OF OUR NEWSLETTER

Contents

Newsletter EnginSoftYear 12 n°3 - Autumn 2015To receive a free copy of the next EnginSoft Newsletters, please contact our Marketing office at: [email protected]

All pictures are protected by copyright. Any reproduction of these pictures in any media and by any means is forbidden unless written authorization by EnginSoft has been obtained beforehand. ©Copyright EnginSoft Newsletter.

EnginSoft S.p.A.24126 BERGAMO c/o Parco Scientifico TecnologicoKilometro Rosso - Edificio A1, Via Stezzano 87Tel. +39 035 368711 • Fax +39 0461 97921550127 FIRENZE Via Panciatichi, 40Tel. +39 055 4376113 • Fax +39 0461 97921635129 PADOVA Via Giambellino, 7Tel. +39 049 7705311 • Fax +39 0461 97921772023 MESAGNE (BRINDISI) Via A. Murri, 2 - Z.I.Tel. +39 0831 730194 • Fax +39 0461 97922438123 TRENTO fraz. Mattarello - Via della Stazione, 27Tel. +39 0461 915391 • Fax +39 0461 97920110133 TORINO Corso Marconi, 10Tel. +39 011 6525211 • Fax +39 0461 979218

www.enginsoft.it - www.enginsoft.come-mail: [email protected]

The EnginSoft Newsletter is a quarterly magazine published by EnginSoft SpA

COMPANY INTERESTSEnginSoft GmbH - GermanyEnginSoft UK - United KingdomEnginSoft France - FranceEnginSoft Nordic - SwedenEnginSoft Turkey - TurkeyVSA-TTC3 - Germanywww.enginsoft.com

CONSORZIO TCN www.consorziotcn.it • www.improve.itCascade Technologies www.cascadetechnologies.comReactive Search www.reactive-search.comSimNumerica www.simnumerica.itM3E Mathematical Methods and Models for Engineering www.m3eweb.it

ASSOCIATION INTERESTSNAFEMS International www.nafems.it • www.nafems.orgTechNet Alliance www.technet-alliance.com

ADVERTISEMENTFor advertising opportunities, please contact our

Marketing office at: [email protected]

RESPONSIBLE DIRECTORStefano Odorizzi - [email protected]

ART DIRECTORLuisa Cunico - [email protected]

PRINTING Grafiche Dal Piaz - Trento

Auto

rizza

zione

del

Trib

unal

e di

Tre

nto

n° 1

353

RS d

i dat

a 2/

4/20

08

The EnginSoft Newsletter editions contain references to the following products which are trademarks or registered trademarks of their respective owners: ANSYS, ANSYS Workbench, AUTODYN, CFX, FLUENT, FORTE’, SpaceClaim and any and all ANSYS, Inc. brand, product, service and feature names, logos and slogans are registered trademarks or trademarks of ANSYS, Inc. or its subsidiaries in the United States or other countries. [ICEM CFD is a trademark used by ANSYS, Inc. under license]. (www.ANSYS.com) - modeFRONTIER is a trademark of ESTECO Spa (www.esteco.com) - Flowmaster is a registered trademark of Mentor Graphics in the USA (www.flowmaster.com) - MAGMASOFT is a trademark of MAGMA GmbH (www.magmasoft.de) - FORGE, COLDFORM and FORGE Nxt are trademarks of Transvalor S.A. (www.transvalor.com) - LS-DYNA is a trademark of LSTC (www.lstc.com) - Cetol 6s is a trademark of Sigmetrix L.L.C. (www.sigmetrix.com) RecurDyn™ and MBD for ANSYS is a registered trademark of FunctionBay, Inc. (www.functionbay.org) - Kraken is a trademark of ESSS (www.esss.com.br)

Case Histories Newsletter EnginSoft Year 12 n°3 - 6

The South Stream Pipeline under the Black Sea had to be the offshore component of the South Stream PipelineSystem. The 931 km-long South Stream Offshore Pipeline was to connect the world’s largest natural gas reserves in Russia with consumers in the European Union.South Stream Transport B.V. was an international joint venture established for the planning, construction, and subsequent operation of the South Stream Offshore Pipeline. Throughout the pipeline’s design life of 50 years, the company had to provide transport services for natural gas from Russia across the Black Sea to the landfall site in Bulgaria. South Stream Transport brought together four leading energy companies: Gazprom from Russia with a stake of 50%, ENI from Italy with a share of 20%, EDF from France, and Wintershall from Germany –each with a stake of 15%.The pipeline had to start from the Russian Black Sea shore in the area of Anapa, to

cross the Turkish Exclusive Economic Zone in the Black Sea, and to land on the Bulgarian coast near Varna. Commercial operations were previously scheduled to start by year end 2015. The first line had to be ready by the end of 2015, the second and third lines by the end of 2016, and the fourth line by the end of 2017. When fully operational, the South Stream Offshore Pipeline will consist of four pipeline sections able to move about 63 billion cubic meter (bcm) per year, equivalent to the energy needs of some 38 million European households.In March 2014, Saipem was contracted for the construction of the first offshore pipeline, as well as the shallow-water parts, shore crossings and landfall facilities for all the four pipelines.

South Stream Project De Pretto Industrie & EnginSoft: the joineddesign of the Flute System to be installedunder the S7000 JL-Tower

Figure 1 - General view of Flute Structure with Roller Frames

Case Histories7 - Newsletter EnginSoft Year 12 n°3

Due to the Black Sea’s depth variation along the laying route, the offshore pipe laying had needs to be done using two different vessels.The first is Saipem’s Castoro Sei, an “S-Lay” vessel suitable for both shallow and deep waters, which can lay pipes up to a depth of 600 m. On board of the vessel, the individual pipes are welded together horizontally to create one continuous pipe to be lowered into the seabed in the shape of a large “S” by means of the stinger installed on the vessel.The second vessel is the Saipem 7000 (S7000), a “J-Lay” vessel suitable for ultra-deep water (that also laid the Blue Stream pipeline in the Black Sea) which can lay pipes from 600 m up to 2000 m (and more) water depth.During construction, batches of four pipes are welded together into ‘quad-joints’ of 48 metres long at the harbor. These should have then been shipped to the S7000 and placed vertically in a large J-Lay Tower (JLT) to be welded into the main string.All pipes should have been welded together with high precision using automated machines and each weld scanned with ultra-sound to verify they were free of any defects. The joints are then protected against corrosion with a special coating made out of a plastic fibre polypropylene before being laid on the seabed.For South Stream Project, the JLT system of the S7000 had to be equipped with a new End Roller System, named Flute System. This device had to be used during the laying operations in order to control the amount of pipe bending/ curvature. Moreover, the Flute had to operate/control the pipeline during accidental conditions (e.g. flooded pipe, Abandonment and Recovery – A&R).After De Pretto Industrie won the tender for the turnkey delivery of the whole Flute System (structural, mechanical, electrical parts and components), EnginSoft was asked to collaborate in the design and verifications (via code checks) of the Flute Main Structure and of the Interfaces with the Roller Frames which, in turn, have been entirely designed by De Pretto Industrie itself.The South Stream Flute System is, in fact, composed by six Roller Frames, placed at different levels and numbered from R1, FL0, FL1 to FL4, respectively starting from the top to the bottom, and by a Steel Rollers (named SR) located at the Flute Structure bottom level (this level was designed for the A&R condition and for the loads it experiences during this phase).Levels R1 and FL0 are included in the basket structural frame located just below the J-lay Tower, so that they were out of the scope of De Pretto Industrie supply, while levels from FL1 to FL4 and Steel Rollers level are included in the Flute Structural Frame.As anticipated, the main purpose of the Flute was to avoid any over-bending and kink effect at the JLT exit that could permanently damage the pipeline. Moreover, each Roller level needed to sustain the pipeline loads without damaging the PU pipe coating due to an excess of contact pressure. Finally, the Flute Structural Frame needed to use the existing structural connections on the J-Lay Tower.Hence the Flute System, which limits the J-Lay Tower inclination from vertical 90° to 110°, also had to guide, with the Rollers, both the pipeline and the A&R steel wire rope. The pipe-to-Rollers contact pressure must not damage the pipe coating during laying operations.Furthermore the Rollers had to guarantee a diameter opening in order to permit an easy passage for the A&R head system, for the Anodes, for the Buckle Arrestors (BA) and for the Hot Field Joint (HFJ).

Figure 2 - Flute structure: solid view of the frame finite element model

Figure 3 - Plan view with Roller Frames installed

Figure 4 - South Stream Offshore Pipeline Route


way, is). For each one of the above mentioned design conditions (Operational, Abandonment&Recovery, Survival, Transit) the following elementary load cases have been considered (and properly combined):

• Flute Main Structure dead load;• Frames dead load;• Buoyancy force;• Wave and current loads according to three incoming

directions (due to the symmetry of the Flute Structure);• Pipeline loads during laying;• Inertia loads due to accelerations imposed by the S7000

vessel to the Flute Structure.

Concerning the wave loads, the following table shows the significant wave height (Hs) and peak period (Tp) of the waves considered in calculating of wave forces. The table also details the velocity values of the assumed sea current for the Flute Structure design. Operational current velocity is given by the summation of the sea current velocity and of the vessel laying velocity.The table 1 reports the vessel Transit velocity as well. In fact, also in this condition, the Flute Structure experiences forces produced by relative current between vessel itself and sea water.

The wave forces have been computed according to the Morison equation which provides for the hydrodynamic drag forces (1st term) and the hydrodynamic mass force (2nd term):

F = (r v |v| D CD) / 2 + r a A CM

where

• F is the wave force per unit length acting normally to the invested member axis

• r is the water density• D and A are, respectively main dimension (Diameter for cylindrical

member, main dimension for other members) and Area of member invested by the wave

• CD and CM are respectively drag coefficient and mass coefficient• v is the wave water particle velocity relative to the member invested

by the wave (and current)• a is the wave water particle acceleration relative to the member

invested by the wave.

Figure 8 - FE Model of the lifting point - Von Mises stress distribution Figure 9 - View of upper region of Flute Structure with FL1 Frame installed

Figure 10 - Detail of connection between FL1 Frame and Flute Structure

Table 1 - Wave on sea current parameters


The spacing between different pipe Roller levels had been properly defined in order to open one level at once and to permit the passage for A&R head c/w buoy, HFJ, BA and Anode.

As previously said, the six Rollers levels (R1, FL0, FL1 to FL4) are coated with PU for sustaining and laying the pipe, where the Steel Rollers (SR) at bottom level are required only during A&R operation to avoid excessive deflections for the A&R cable. During normal lay, Steel Rollers have to be retracted in order to avoid contact with pipe. PU Rollers have been arranged with a diamond configuration in order to guarantee constant contact between the pipe (32” dia. and 39 mm thk) with two coplanar rollers, thus reducing the contact pressure between pipe and Rollers.

Considering that the capabilities of computational tools have reached levels extremely competent in reproducing reliable performance with reduced costs, the ‘simulation offers’ improved performances and increased reliability with a reduction in experimental testing. More importantly, the increasing sophistication of both tools and operators enable the development of more complex, dynamic and / or nonlinear analyses with respect to extreme behaviors and / or accidental events.Furthermore, it has to be stressed that trial-and-error procedures do not apply to this type of structure being unique of its kind and scope. This means that efficiency and risks had to be assessed upfront, during the entire design phase.Some shop tests were applicable at the end of the Flute Structure and Roller Frames fabrication, but they were only used to confirm the design. That shows that the simulation based methodology (which uses the Computer Aided Engineering approach to make the real accomplishment) was the only method which could efficiently determine the correct sizing of the structure and its components within the time constraints, while evaluating different what-if scenarios to ensure the required robustness is achieved in respect of the delivery schedule.

As consequence, the Flute Structure and the Roller Frames were designed via the virtual prototyping approach, taking into account all the targets reported above, exploring the Operational, A&R, Survival, Transit conditions experienced by the Flute itself during the pipeline laying phases. In addition, the Transportation, Lifting, Installation, Removal phases of the Structure have been investigated too. This is the reason why the Flute has been completed with proper bollards and padeyes with related lifting frames (just to facilitate the lifting, installation, removal phases). From CAD 3D model, specifically developed just for the design purposes (and progressively updated on the basis of the results of structural analyses and related verifications), an ad-hoc beam Finite Element model of the Flute Structure, completed with Roller Frames, was implemented by means of the tool suite ANSYS/ASAS, a FEM code developed for the extensive employing in the offshore environment for the design and code verifications of fixed offshore structures (like the Flute Structure, in some

Figure 5 - Displacement vector - Worst combination case LC16 (wave direction a=-135°)

Figure 6 - FE Model of Flute area where the pad-eye for vertical lifting is installed

Figure 7 - FE Model of the pad-eye - applied loads

11 - Newsletter EnginSoft Year 12 n°3 Case HistoriesCase Histories Newsletter EnginSoft Year 12 n°3 - 10

For each condition where wave loads govern over the force distribution (Operating, A&R and Survival) and for each heading, 8 wave crest-to-Flute positions have been considered. Crest position has been computed considering a different phase j in the formula simulating the crest profile [y = H cos (wt+j), where H is the wave height].For the pipeline loads acting at the Roller Frames level, the values detailed in table 2 have been considered (as defined by Saipem Design Specification; X is the axis along which the pipeline force acts, Y is the horizontal transversal axis, along which the friction force, related to the Fx force, acts, Z is the vertical axis along which the force due to BA acts).

Inertial loads acting on the Flute Structure and on the Frames due to acceleration imposed by the S7000 motion have been computed on the basis of the values provided by Saipem.Over the FE model of the Flute Structure, numerical analyses have been performed under the hypotheses of linear static behavior of the structure. Nonlinear changing status analyses have been anyway considered in order to take into account the contact between Mobile Frames (supporting the Rollers) and load cells, as load cells located at the interfaces between frames and Flute Main Structure act only in compression.On the basis of the analysis results for all the load conditions and combinations investigated, structural and mechanical checks have been performed.Code checks (carried out according to API RP2A WSD rule and AISC ASD Standard) have been referred to strength and buckling assessment of tubular members and to punching shear of tubular connections, while mechanical checks have been referred to the parts (members and components) where the interfaces with the Roller Frames were located.

Tubular members of the Flute Structure have been also checked to assure that no vortex shedding induced motion (in-line excitations and cross-flow excitations) occured within the existing environmental, operational and transit conditions. Vortex shedding checks have been performed according to DNV Classification Notes 30.5.With reference to the vertical lifting condition of the Flute Structure, dedicated FE models of lifting lugs and related frames transferring the Flute whole weight to the lifting slings have been implemented via dedicated shell elements.Detailed structural checks have, hence, been performed in order to guarantee proper safety for the operations which refer to the lifting of the Flute for the installation under the JLT system.Once the virtual prototyping and associated structural and mechanical checks were completed, the information gained in terms of dimensions and thickness of tubular members, beam sections and plates were transferred to the 3D CAD model for the final revision and the relevant issuing of the construction drawings, in such a way completing the design process of the S7000 Flute Structure for the South Stream Project (the stop of which belongs to another chapter of the story…).

For more information:Livio Furlan, EnginSoft

[email protected]

Figure 11 - Detail of connection between FL4 Frame and Flute Structure

Figure 12 - S7000 during pipeline laying

Table 2 - Forces on Rollers during pipeline laging

In recent years the equipment for oil & gas industry has extent more and more the range of operating conditions and the requested performances in terms of resistance respect to operational and accidental loads have been significantly increased. In the technical specifications provided by plants purchaser for large oil & gas project, for instance, the operability of mechanical components under severe exceptional conditions is considered as mandatory for safety reasons. Design for seismic scenario in particular is one of the critical aspect which has to be taken into account for the development of reliable and robust equipment such as motion control device.For actuators of high size ball valves, due to dimensions of the whole system and suspended masses of components in motion, seismic loads could significantly compromise the functionality and the mechanical resistance of the whole system. For this reasons accurate design respect to exceptional load must done by the actuator’s supplier. Therefore FE analyses can be a valid and reliable instrument to analyze complex geometries and actuator configurations by taking into account several load scenarios.Aim of this work is to present a methodology for the seismic design of hydraulic and pneumatic valve actuators which is based on response spectra analyses in compliance with well-known technical standard. Analysis approach and results on a actuator for a 48 inches API class 600 ball valves will be presented in following sections. The work has been developed in collaboration with Cameron Ledeen Facility. ASCE 7.10 and Eurocode standards have been used for the definition of seismic load and acceptance criteria respectively.Fig. 1 shows the geometry of the actuation system which includes the scotch yoke frame, valve connecting flange, cabinet support, spring support and actuator. A three dimensional model has been developed using Ansys rel. 16 FE code. A combination of solid, shell and beams elements has been used to model the actuator’s main parts. Concentrated masses have also been used for non-structural components of the actuation system. Developed mesh is shown in Fig.2. The FE model

consists of approx. 220000 nodes and 217000 elements.For actuator carbon steel has been considered.

Analysis workflowSeismic analysis has been carried out by means of a preliminary modal analysis followed by a series of response spectrum analyses (for each motion direction in compliance with ASCE 7.10 code). According to code’s prescription, spectra results have been than combined.The modal analysis performed using Block Lanczos eigenvalue extraction method combined with a sparse solver in order to extract

Seismic analysis of actuator for Ball Valves

Fig. 1 - Actuator applied on a ball valve

13 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 12 Case HistoriesCase Histories

According to EN 1993-1-8 technical standard following acceptance criteria has been assumed for welded material (Conservatively all welds have been considered as fillet welds).

Where a = 0.8, γM2 = 1.35 and fu is the UTS of welded material.

FE model and boundary conditionsAs stated in the introduction, all the actuator structural components has been directly modeled. Bolting between different parts, such as scotch yoke frame vs. connection flanges, actuator and cabinet support, have been modeled using beam joint elements. As regards boundary conditions, the actuation system has been considered fixed (all dof constrained) at the valve connection flange in order to simulate the coupling between the actuator and valve. This support acts as a fixed support for the modal analysis for the calculation of the natural frequencies and participation masses, while, for the response spectrum analyses, seismic ground motion accelerations has been applied.

ResultsAs reported in the description of analysis procedure, modal analysis is necessary to calculate the dynamic characteristics of the mechanical system. With first twenty natural frequencies about 96% of the total participation mass has been found. The result can be considered adequate and compliant with the request of at least 90% of the total mass. The first mode shapes in the load spectrum range (0 – 17 Hz) are shown in Fig. 4.

From the combination of response spectra analyses along main direction, results in terms of equivalent von Mises stress for different components have been collected and stress levels check has been carried out according the acceptance criteria. Fig. 5 shows the stress maps for scotch yoke frame, valve connecting flange and cabinet support. Valve connecting flange and cabinet support in particular proved to be the most stressed components, and considering the fact that their dimensions could vary to adapt the actuation system to the customer request, they could represent a critical components for the actuator resistance or efficiency under seismic conditions. Main welds of scotch yoke frame and cabinet support frame have been also checked. Generally in fact, welds are the weakest areas of a mechanical components and, especially for the system under investigation, failure of one of the joint welds of the cabinet support could significantly compromise the reliability of the system during operation. Example of weld stress extraction on cabinet support frame is shown in Fig. 6.

ConclusionsA simulation approach to investigate the durability of actuators for ball valves under seismic load has been proposed. The FE methodology proved to be reliable and allow the analysis of complex geometries under severe working conditions. As a results the prescriptions of technical standards, such as the extensively used ASCE 7.10 code, could be completely satisfied. The analysis carried out on a actuator for a 48 inches API class 600 ball valves pointed out that the design is

adequate to resist to seismic load typical for the actuator installation area. The results of the analyses show that stresses on components comply with the allowable stress limits. The equivalent stress induced by the combination of seismic load applied is globally under the allowable stress. Local intensifications areas, which a stress status higher than allowable limit but below the material yield, are present in correspondence on local notches or stress singularities and can be tolerated and do not compromise the structure global stability.

Described methodology can be effectively applied to assess the whole actuators production and offers a reliable instrument to verify the actuator design respect to the supplier requests and/or variation of installation site.

AcknowledgmentsThe work has been done in collaboration with Cameron LEDEEN Products & Systems

For more information:Stefano Cavalleri, [email protected]

Fig. 5 – Stress maps on actuator

Fig .6 – Detail of stress on welds

main modes and natural frequencies of the system and associated mass participation factors. Suitable number of modes has been considered in the modal analysis in order to properly describe the system dynamic behavior in compliance with the applied load spectrum and to obtain, as a results, a significant total participating mass (about 90% in X, Y and Z directions).The seismic ground motion has been estimated according to ASCE 7.10 – Chapter 11, Seismic Design Criteria . The spectrum analysis is based on a mode-superposition approach where the responses of the higher modes are neglected; therefore, part of the mass of the structure is missing in the dynamic analysis. This situation can be managed using a modal base with a total participated mass ≥ 70% of the total mass or using the “missing mass method”. The high frequency region of the spectrum is characterized by no amplification of the peak acceleration of the input time history. SRSS mode combination type has been used in the analysis. The acceleration mapped parameters (MCE) SS and S1 and site class have been assumed in compliance with technical specification provided by Cameron Ledeen Facility. Site coefficients Fa and Fv useful for the determination of spectral response parameters have been extracted respectively from tables 11.4-1 and 11.4-2 of ASCE 7.10 – Chapter 11. Resulting design response spectrum diagram for time period and frequency is shown in Fig.3 assuming a site class type E.According to ASCE 7.10 – Chapter 13, seismic load spectrum must be applied independently in at least two orthogonal directions. In addition, the mechanical system shall be designed for a concurrent vertical load equal to ± 0.2 of the horizontal one.Analysis has been carried out by applying the load spectrum (LS) calculated in previous section

separately for horizontal (x and y) and vertical (z) direction. Result of response spectrum analyses have been combined as following:

LC1 = LSx + LSy + 0.2 ∙ LSz

Acceptance criteriaTwo different acceptance criteria have been defined for the unwelded and welded material respectively. For unwelded material, stress status (equivalent von Mises stress) on the main mechanical components of actuation system due to load combination LC1 has been analyzed by assuming limits in compliance with Eurocode. Acceptance criteria has been considered satisfied if equivalent stress in the system components is below the allowable limit. Localized areas with stress higher than allowable limit could be allowed in correspondence of numerical singularities like sharp notches, constraints or contacts points.

Fig. 2 – Mesh developed for the FE model

Fig. 4 – Main obtained mode shapes in the range (0 – 17 Hz)

Fig. 3 – Diagram of applied seismic response spectrum


CompanyANDRITZ HYDRO is a global supplier of electro-mechanical systems and services (“water to wire“) for hydropower plants and one of the leaders in the world market for hydraulic power generation. Today ANDRITZ HYDRO is the legal successor of many former pioneers and leading companies in the hydropower sector. In fact, ANDRITZ HYDRO S.r.l. Unipersonale is one of these historical companies, founded in 1885 in Schio (Vicenza, Italy) from Eng. Silvio De Pretto which established a Foundry, initially repairing looms but soon after begun to produce Hydraulic Turbines and Paper turbines.

Nowadays ANDRITZ HYDRO S.r.l. Unipersonale specializes in Hydro electrical power plants installing all around the world Francis turbines, Pelton turbines and Kaplan turbines as well other components just like Spherical valves installed as high-pressure turbine inlet and pump valves.

Problem In a Hydro electrical plant, the Main Inlet Valve must guarantee functionality and reliability despite interruption in water flow caused by high upstream.pressureFor our company, the standard design of the spherical MIV by welding cast components was not competitive enough due to new competitors from low-cost countries. This type of valve has

a minimum head loss due to a completely free through-flow. They are usually installed as high-pressure turbine inlet and pump valves (where Flow is possible in both directions) and equipped with controlled service and maintenance sealing. The control system

is operated by oil-hydraulic pressure, or by a water pressure system, or both. The water for the valve control system can be taken from the upstream side and the Oil pressure from the hydraulic turbine governor system. Bypass, hydraulic, and electric control systems, as well as connecting pipes are usually part of the supply.

SolutionTherefore, to reduce manufacturing and quality inspection costs, forgings (“high structural quality” of the components) were used instead and assembled using bolts with controlled tightening (removal of the most of the welding’s). Hence, it was very important to conduct an in depth structural analysis of

The importance of quick responses and mechanical reliability: Structural analysis of a Spherical Valve

Figure 1- Nicola Pornaro, Safety design Engineer at Andritz Hydro

Thanks to a smart simulation process, ANDRITZ HYDRO developed a “family group” of valves with reduced design time and production costs. This complex mechanical product has seen an increase in sales, maintaining structural safety of the entire assembly with a focus on the solicitation on small important components


the entire valve, to check the effectiveness of each bolt’s coupling as well as optimize the thicknesses by parameterizing the main dimensions in the. The stress-state, deformations and pressure between sliding components are the main reasons why a deep inspection of the design concept via FEM analysis was completed using ANSYS Mechanical and ANSYS WB.

In this analysis, the sub modeling technique was fundamental. Performing the sub modeling analysis directly in ANSYS Workbench, without the need to use ANSYS Classic, allowed for a quicker setup process and obtained results to meet tight deadlines. This achieved detailed analysis on the sub models at an ”optimal point” between structural safety and design-to-cost.

The structural analysis started from the “Parent model” of the entire Valve with all components simplified, application of loads and boundary conditions coherent to the real working conditions (and some fictitious worst-case events). For the optimization of thicknesses of the main components, a system was developed of parameterizations between the different Workbench analysis with the target to accelerate and optimize the operations of verification. For the most complex and important zones of the Valve, some sub models were created with “boundary behavior” which derives from the “Parent model” applied to the coherent cut-boundary surfaces.

Business BenefitSince the development of the first valve, customers have shown great interest and the orders have grown. Continuing with the optimization of the integration between design and structural analysis we are continuously improving the knowledge and production of a “series of valve-size”, maintaining the structural-status control through precise FEM calculations in a limited time thanks to parameterization. We are therefore able to reduce time to market by decreasing design time. The set has enabled us to greatly increase the confidence on the design of the components and at the same time significantly reduce the overall costs of product development.

AwardThis project about FEM simulation has been evaluated by a panel of simulation expert and was selected as a best entry in the 2015 ANSYS Hall of fame Competition, ANSYS Inc., Canonsburg, PA.

Figure at the top of the article: Father model of the entire Valve, general stress-state distribution

Nicola Pornaro, Safety design Engineer at Andritz Hydro S.r.l. Unipersonale

Figure 4 - Submodel of the most complicated zone with boundary conditions: Trunnion

Figure 5 - Stress-state distribution on the Submodel of the Trunnion

Figure 6 - Submodel of the Trunnion, meshed model

Figure 2 - Stress-state distribution on the Submodel of the Main flanges coupling

Figure 3 - Image winner of the Ansys Hall of Fame competition 2015

Case Histories17 - Newsletter EnginSoft Year 12 n°3 Case Histories Newsletter EnginSoft Year 12 n°3 - 16

Sicor: quality, innovation, customer supportEstablished in 1981 in Rovereto (TN), Sicor is a hoisting machines manufacturer specifically designated for elevators. Since it was founded, the company has succeeded through its international approach, technological innovation and quality commitment toward its end customers.Throughout the years, an in depth research and development process has been established, leading the company to improve and grow its range of products in order to satisfy the market needs. Sicor today provides its customers with a wide range of products among traditional and gearless hoisting machines, capable of fulfilling even the most specific application requirements.In 2013 the new production plant was launched, with a total extension of 21.000 meters square and equipped with the latest and most efficient production solutions. Automation, optimized logistic and high precision processes assure uncompromised quality and shorter delivery time, matching the growing needs of speed and flexibility required by the market.Sicor’s mission has always been clear, that is to supply its customers with the best hoisting solution complying with specific application requirements, going through a precise and accurate assistance throughout all the sale process and post-sale service.

International vocation, with deep national rootsSicor’s gears are conceived, designed and assembled exclusively in Italy. This allows the company to offer each customer complete quality control. Sicor relies on a wide network of partners able to supply its raw materials, with over 200 Italian and foreign suppliers.

FEM analysis for products improvement and cost reductionToday the role of the FEM simulation has become a very common method to optimize the performance of the final result in many industrial areas.

In order to continuously improve the competitiveness of Sicor hoisting machine in a worldwide context, FEM solutions have proved to be the best way to help the product development, in order to reach the best results in terms of performance and cost reduction. This approach is also very useful for the optimization of existing products.

For this, EnginSoft has developed the structural modeling of the biggest hoisting machine in the Sicor product range called

Structural analysis for an hoisting machine for lifts

MR35. Starting from the 3D models provided by Sicor, the FEM structural simulation has analyzed the mechanical behavior of the components under the maximum static load on the machine in various possible directions. Every load configuration has then been analyzed, evaluating the most critical component in term of stress distribution. The result of the analysis has shown that the lateral load direction is the most critical for the machine, especially for the fixing bolts. According to this, the lateral component of the total static load has to be reduced in order to minimize the total stress on the bolts. Considering this situation, some improvement on the structural frame of the machine and of the fixing solution has been investigated and proposed in order to maximize the performance of the hoisting machine.

The projectIn order to evaluate the tensile state of the MR35 hoisting machine for lifts, several structural analyses have been carried out, applying the finite element method and taking advantage of the simulation software ANSYS 16.1. Particular attention has been paid to the main components of the hoisting machine: carter, supports, axis and bolts (Figure. 1).

Geometry and mesh After having completed a defeaturing operation on the hoisting machine geometry by means of the SpaceClaim software, the geometry has been discretized with hexahedral finite elements of quadratic form functions so to obtain a suitable accuracy of the calculated results (Figure. 2).

Boundary conditionsThe equivalent load acting on the pulley has been directly applied to the axis, according to four different directions with respect to the bearing plan of the hoisting machine: +60°, -60°, +90°and -90° (Figure. 3). Apart from the equivalent load of the pulley, also the clamping load of the bolts and the proper weight of all components have been taken into account within the FE model. The interaction between the hoisting machine components, as well as the

carter contact with the floor have been simulated by means of the mono-lateral constraints with friction (non-linear contacts).

Results Figure 4 shows the chats of the equivalent stresses of von Mises of the hoisting machine in the most critical load condition with the equivalent action of the pulley inclined of +60° with respect to the horizontal plan.

ConclusionsConsidering the carried out analyses, the fixing bolts of the carter on the floor have proved to be the most stressed components of the hoisting machine. On the contrary, the support axis of the pulley has turned out to be the less stressed one, in relation to the admittable load for the material constituting the component.Simulation has made possible the ability to identify the hoisting machine components that need to be modified in order to improve the performance (increase of the lifting load) and those that can be optimized in the view of a cost reduction.

For more information:Daniele Filugelli, EnginSoft - [email protected]

Fig. 1 – 3D Geometry and investigated elements of the FE model

Fig. 2 – Mesh

Fig. 3 – Load conditions Fig. 4 – Results

Case Histories19 - Newsletter EnginSoft Year 12 n°3 Case Histories Newsletter EnginSoft Year 12 n°3 - 18

Founded in 1978 to produce mechanical wire die springs, Special Springs has quickly became the first company worldwide to manufacture three different product lines complying to three different standards: ISO 10243, JIS B5012 and US Oval Wire. At the end of the 80ies, it also started the production of nitrogen gas cylinders to complete its offer. At the same time Special Springs introduced a series of hexagon socket head shoulder screws 12.9, still unreached for its geometric al mechanical qualities. In 2012 Special Springs gathered in a single plant the production of wire springs and nitrogen gas cylinders covering an area of 16.500 sqm. Raw materials and thermal treatment processes, as well as any other kind of analysis, are quickly and continuously controlled thanks to a complete internal metallographic laboratory, equipped with the most advanced tools and management software.Reliability tests are continuously performed in order to maintain and improve the performance of Special Springs products. The two test centres, specifically focused on wire springs and nitrogen gas cylinders, are always at customers’ disposal to perform duration tests in relation to the application specifications.The use of the most advanced software for 2/3D design allow Special Spring technicians to investigate, develop and design products according to the highest standards of reliability and safety.

The project Special Springs products catalogue includes several different kinds of nitrogen springs. Each type of spring is characterized by the force range that it is able to exert and by its size. As shown in Figure 1, the tolerance associated to the spring height is about ±0.25 mm, a value which is guaranteed by Special Springs to its customers. The width of the tolerance zone of 0.5 mm, applied to the vertical dimension of the springs, is an objective that Special Springs is able to reach with the productive processes and the related technologies currently available. This generates a very low number of scraps of “out of tolerance” products, about a 0.5 %. Something has changed in this very successful scenario, when an important customer made a challenging request to

the company: to guarantee a dimensional tolerance on the springs total height of ±0.10 mm, thus reducing the initial value of ±0.25 mm. At this stage it became necessary for Special Springs to understand the repercussions due to this reduction of the tolerance applied to the height on its own products, and in particular: how much would the “out of tolerance” scraps increase? Assuming that the percentage of scraps remains unchanged, which modifications should be required, thus generating the minimum possible costs?In order to find an answer to these two questions, Special Springs has instructed EnginSoft to investigate the tolerance analysis on a nitrogen spring. The tolerance analysis is based on a statistical approach in order to understand how the tolerances applied to the single nominal dimensions of the project affect, combining themselves, the final size, that may represent, as in this case, a very important product requirement.

Tolerances analysis on a nitrogen spring

Fig. 1 – Simplified CAD model of a nitrogen spring under investigation and definition of vertical dimension with related tolerance of ±0.25 mm

The development of the modelThe model of the tolerance analysis has been built using the Cetol 6s software, developed by the American company Sigmetrix. The development of the model consists of the following phases:

• Definition of the contacts between the single parts of the assembly. The software allows to define in a flexible way any kind of kinematic joint, managing the single degrees of freedom which are constrained. Cetol 6s has furthermore advanced functionalities that allow to automatize this phase, thus dramatically reducing the development time of the model. In particular, when the assembly constraints are set in the 3D CAD, Cetol 6s is able to read and to import them inside the model. The single parts of a nitrogen spring that are involved in the model are four: body, ring, bush and rod (Fig. 2). The rod rests on the bush in the area of the upper flange.; the bush is then constrained to the body by means of the ring that keeps it in the right place using the groove in which it is hosted.

• Definition of the scheme dimensioning of the single parts. This phase consists of the translation of the information included in the 2D drawings: the dimensions and the related tolerances (both dimensional and geometric) are included in the model and represent the quantitative information on the place variability and the relative orientation between plains, the axis, surfaces, etc. Furthermore, if datum notes (DRF – datum reference frame) are present in the 3D CAD model, dimensional and/or geometric tolerances, the software can one come identify and manage them automatically, Such information are imported in the model avoiding redundant action in the 3D CAD.

• Definition of the target size and analysis execution. Following the procedure described in the model, only some parts of the assembly are considered, that is those affecting the size to be calculated; the height, in this case. The method of constraints creation between the single parts of a set used by Cetol 6sallows

to take into account in an automatic way only the geometric characteristics that somehow influence the size. The advantage is that of neglecting the total geometric complexity of the single parts, since it is not required by the objective under investigation: practically speaking this means a great saving of time in the model set-up phase.

Achieved resultsA first calculation has been performed to validate the obtained model: at a nominal value of height of the nitrogen spring of 138.10 mm, a ±0.25 tolerance has been applied. The achieved result (Fig. 3) is in line with the experimental results obtained by Special Springs on the height size of the nitrogen springs. After the validation, a new customer’s requirement has been applied to the calculated distribution, that is the width of the tolerance zone of ±0.10 around the nominal value. This has allowed to estimate a scrap percentage of about 24.3% in the current production conditions and with the available technologies (Fig. 4)Summarizing, in the case Special Springs would satisfy the customer’s

requests, it would mean to forecast a 25% of scrap products.

This information can be used to estimate the costs and therefore for a more reliable quotation.The tolerance analysis does not only provide such information. It is possible to identify the tolerances that mostly affect the dispersion of the statistic distribution and to act so to meet the requirements with the minimum impact as far as costs are concerned.The tolerance with the highest impact on the dispersion of the height values around the average calculated value is corpo;alloggiamento_@anello;to_topPlane (see Fig. 5), with a nominal value at drawings of X mm (tolerance ISO 2768-mK: ±0.20 mm). Fixing the tolerance at this

Fig. 2 – Identification of the assembly parts that directly affect the dispersion of the height value around the average value

Fig. 4 – Statistic distribution of the nitrogen spring height requiring a tolerance of ±0.10 mm. The tolerance analysis foresees a percentage of scrap products of about 24.32%

Fig. 3 – Statistic distribution of the nitrogen spring height requiring a tolerance of ±0.25 mm. The tolerance analysis foresees a scrap value of about 0.35%


size at ±0.05 mm, the products satisfying the height requirement of ±0.10 go from about 75% to 92% roughly (Fig. 6).The second tolerance with the highest impact on the dispersion is corpo;topPlane;to_A with a nominal value at drawings of 85.00 mm (tolerance ±0.10 mm). Reducing this tolerance to ±0.05 mm, it is possible to reach an efficiency of about 96.5% (Fig. 7).

The average value of the calculated distribution is different from the nominal one. This is due to the presence of non-symmetrical tolerances in relation to the nominal values in the model. In particular the tolerance with the highest impact on the average value displacement is: boccola;bottomPlane;to_topPlane (23.5 +0.05 -0.00 mm). Keeping the width of this tolerance zone, but changing it into 23.50 +0.03 -0.02 mm, it is possible to achieve a centred process with an efficiency of the 97.6% (Fig. 8).

ConclusionBy means of the propagation of the tolerances inside the dimensional chain it has been possible to identify the dimensions and the tolerances of highest impact on the functional size and therefore it has been possible to act on the reduction of possible scrap products, avoiding to manage them in a general way, that is reducing tolerances in an indiscriminate way.

Such result can be achieved reducing just two tolerances, among the hundreds of them present in the project drawings and the displacement of the third tolerance, keeping its width value (no influence on the production costs, since it is not necessary any further precision machining).The activity carried out by EnginSoft has allowed Special Springs to identify the dimensions and the related tolerances to maintain and verify the quality to meet the product requirement requested by the final customer. The requirement is met assuming that all parts are produced respecting the tolerance they have been assigned by the project.

Enrico Boesso – EnginSoftDennis Lorenzi – Special Springs

Fig. 5 – Tolerances and related dimensions with the highest influence in terms of percentage on dispersion of the height value of the nitrogen spring in comparison with the average value. Percentage contribution on the standard deviation. In the “label” column, there are the drawing coordinates, where the dimensions are present

Fig. 8 – Definition of the dimension boccola;bottomPlane;to_topPlane and statistic distribution of the nitrogen spring height with tolerance requirement of ±0.10 mm as a consequence of the displacement of the tolerance zone from 23.5 +0.05 -0.00 mm to 23.50 +0.03 -0.02 mm

Fig. 7 – Definition of the dimension corpo;topPlane;to_A and statistic distribution of the nitrogen spring height with tolerance requirement of ±0.10 mm as a consequence of the tolerance reduction from 85.00 ±0.10 mm to 85.00 ±0.05 mm

Fig. 6 – Definition of the dimension corpo;alloggiamento_@anello; to_topPlane and statistic distribution of the nitrogen spring height with tolerance requirement of ±0.10 mm as a consequence of the tolerance reduction of X mm from ±0.20 mm to X ±0.05

About General ElectricGeneral Electric (GE) is an American multinational conglomerate corporation founded in 1892 in Schenectady, New York. The GE Oil & Gas business division provides plants and machinery for the petroleum industry and its TMS (Turbomachinery Technology Solutions) subdivision is mainly focused on compressors and gas turbines.

Case IntroductionGE aeroderivative gas turbines are used in electrical power generation. A complex piping system links a turbine to its auxiliary devices and needs to be carefully designed by taking account of operating loads (temperature, pressure) and transient loads to fulfill the specific ASME B31.3 regulations.

EnginSoft supported GE in the piping FEA-based design, providing engineering and analysis expertise during an iterative assessment-and-redesign process. In this situation, ANSYS SpaceClaim Direct Modeler (SCDC) was crucial in improving and simplifying the modelling activities, leading the design process to a validated solution in a much shorter time and thus meeting the strict project deadline.

Geometry Clean-UpThe very first modelling phase involved the raw geometry treatment. The huge and fully-detailed 3D CAD geometries provided by the GE design team needed a clean-up before it could be used in the analysis calculations. So the CAD was easily imported in the ANSYS SCDC environment and all the analysis-useless parts were

removed with a single-click operation. Also, every pipe support or clamp was deleted after its position had been imprinted on the related pipe branch. Instead, bodies that were entirely supported by the piping (i.e. valves, fittings) were kept and grouped in more manageable assemblies with an automatic procedure.

These first activities may seem trivial but mostly require long times when involving huge geometries: ANSYS SCDC sensibly sped up this clean-up phase.

Productivity and accuracy in piping modelling with the ANSYS SpaceClaim Direct Modeler

Figure 1 – An aeroderivative gas turbine

Newsletter EnginSoft Year 12 n°3 - 22 Case Histories23 - Newsletter EnginSoft Year 12 n°3 Case Histories

Every GE piping assessment is performed according to ASME B31.3 regulations. ANSYS provides a handy tool, the User Defined Result, that definitely makes post-processing calculations fast yet accurate. It automatically calculates any user expression and outputs the related contour plot, allowing the straightforward implementation of ASME B31.3 stress criteria for piping verification.

Then, once the simulation terminated, each pipe branch assessment was automatically performed and the results were straight graphically plotted for user validation.After a few assessment-and-redesign iterations, the whole piping system was validated and the final design was delivered to GE.

ConclusionsANSYS SpaceClaim Direct Modeler dramatically reduced the modelling time needed to process piping raw geometries and to make

them functional for an effective FEM representation with 1D-pipe elements and lumped masses. With ANSYS Mechanical, even a complex load scenario was simulated using simple functions available in the GUI and the analysis results post-processing was efficiently automated implementing ASME B31.3 regulations for an immediate piping verification.

Summing up, the whole design process was optimized and sped up by the ANSYS SCDC + ANSYS Mechanical duo, so leading the piping engineering to a validated solution in a very short time in order to meet the project deadline.

For more information:Riccardo Zoccarato, [email protected]

Figure 8 - ASME B31.3 stress criteria for piping verification and related User Define Result contour plots

Figure 7 - Operating static loads applied on the piping

ModellingThe main modelling operation for a piping FE analysis consists in converting the 3D-pipe geometry in a 3D model composed by 1D-pipe elements. In this case, the SpaceClaim BEAMS tool comes in handy in optimizing and speeding up the modelling process. Indeed, the Extract function activates a procedure that automatically recognizes the axis of the selected pipe bodies. Then the axis line is substituted by 1D-pipe elements which section is automatically detected and assigned.

In particular, all the straight pipes are converted to BEAM188 (linear pipe) or BEAM189 (quadratic pipe) ANSYS elements, while the curved pipes are converted to ELBOW290 (quadratic pipe with curvature stress intensification) ANSYS elements.By taking advantage of this powerful tool, the entire 1D-pipe model was generated in a few minutes.Proper piping boundary conditions were defined by applying remote points on the on the pipe section imprinted during the clean-up activities: couplings to non-pipe bodies were defined as rigid brackets (0 relative DOF) while supports and clamps were defined as elastic radial supports (4 DOF).The assemblies entirely supported by the piping were assumed to be much stiffer than the pipe branches, so each body group was converted to a MASS21 (point mass with also inertial properties) ANSYS element.The mass and the inertial properties of the MASS21 elements were easily characterized with a dedicated ANSYS SCDC procedure which analyzes a selected body group and automatically detects its mass, principal moments of inertia and related reference system position and orientation.

Piping Analysis and AssessmentThe piping FE calculations were carried out with an ANSYS Mechanical transient analysis involving both operating pressure and temperature (static loads) which were parametrically applied to the different pipe branches.Moreover, a set of dynamic loads completed the transient load case by applying a measured acceleration history to the whole model (inertial loads) and a set of relative displacements between the piping boundaries.

Figure 2 - Piping geometry after the clean-up process

Figure 3 - Piping 3D CAD geometry and related 1D-pipe elements created by the SpaceClaim BEAMS tool

Figure 4 - ANSYS element types implemented in the model

Figure 5 - Different boundary conditions applied to pipe sections

Figure 6 – Assemblies entirely supported by the piping and related point masses with inertial properties

ANSYS SpaceClaim Direct ModelerANSYS SpaceClaim Direct Modeler is the easiest and fastest solution to prepare geometry for simulation. SpaceClaim’s direct modeling solution speeds up time to analysis by removing the geometry bottleneck, and it shortens the time needed for analysis by allowing engineers to simplify models during pre-processing. SpaceClaim also provides a CAD neutral environment, freeing engineers to focus on the physics and modeling.By putting simulation engineers in control of geometry, CAE can drive product development by optimizing models before CAD and validating results post-CAD. With ANSYS SpaceClaim Direct Modeler, users can:

• Open data from any CAD system• Edit designs and prepare them for simulation no matter where the model was created• Simplify geometry by removing features, like rounds and holes, that lead to low quality mesh generations• Clean up and repair dirty geometry to create watertight solids• Create parameters on imported geometry to enable optimization of designs through analysis• Extract mid-surfaces/shells and beams from solid models to enable efficient meshing and solving• Extract volumes and create inner fluid domains and outer air enclosures for CFD analysis• Supply 3-D markups and compare models to document changes to design teams• Be up and running in a matter of days instead of months like big CAD

With SpaceClaim Direct Modeler’s STL Prep for 3-D Printing Module, users can also repair, and edit files for further FEA topological optimization and CFD analysis.

We approached then the mathematical modeling world looking for tools allowing the designer to better use information coming from 3D simulations.

5 . How does this affect your design process? We decided to improve the procedure by introducing different kinds of mathematical tools: Design of Experiments techniques, metamodeling and optimization algorithms support the designer knowledge in the most effective way, making possible to exploit 3D simulations’ information for the development of very effective tools, allowing to build 3D complex models.

6. Are you also thinking about applying mathematical modeling for new products and what expectations do you have?We are developing flexible techniques thinking about their possible application to other aspects (e.g. aeroacoustics) of the same product as well as to other products.Considering similarities existing among approaches followed within different design fields we can be confident about the possibility to enhance the design process level as we are doing for fans fluid-dynamics.

7. What’s the value that EnginSoft can give you?The valuable competence of the EnginSoft team and their ability to tailor their support to the customer level and needs has made possible an effective interdisciplinary collaboration allowing well-structured and fast progression of the design tools development. In particular, EnginSoft mathematical competences are employed together with best practices using open source frameworks.

8. Could you estimate the return on the investment related to these R&D activities?I can say, without going in details, that the level of the entire design procedure is being significantly increased. As a consequence, the designer can better control every characteristic of interest with a lower design timing.

The ongoing project with EnginSoft: design process innovationAxial flow fan design for automotive cooling is a complex task involving fluid-dynamic, aeroacoustic and structural aspects strictly interacting with aspects related to the design of the driving electric motor.Starting from customer’s technical requests (e.g. airflow rate, pressure rise) and constraints (e.g. available space, power budget, noise level) the fluid-dynamic designer has to define a suitable fan geometry whose main element is its blade.

To obtain the desired performance the designer uses different computational tools allowing to define, analyze and refine global as well as local blade fluid-dynamics.Relevant design parameters are the rotational speed, shape, stagger angle, length and radial stacking of aerodynamic sections used to define the blade surface.The traditional approach to this design problem, even involving advanced simulation tools, requires a certain amount of trial and error iterations which can take from several days to weeks, depending on the case complexities. Once defined, a fan geometry is prototyped and then tested.

Design tools at Johnson Electric Asti are in continuous progress; during last years the entire tools suite has been updated integrating self-developed, open source (Scilab) and commercial software (ANSYS, Siemens NX). This work involves several people and external suppliers; to support it with specific mathematical competences and to handle at best open source mathematical code a collaboration with EnginSoft Spa has started. The aim of the ongoing collaboration with EnginSoft, is to improve the fans design process, now structured in a new and non-standard way, namely coupling 3D CFD simulations with the study of physical phenomena, exploiting very flexible mathematical techniques.


Company profile Johnson Electric Johnson Electric is a global leader in motion products, control systems and flexible interconnects. It serves a broad range of industries including automotive, building automation and security, business machines, defense and aerospace, food and beverage, home technologies, HVAC, industrial equipment, medical devices, personal care, power equipment and power tools.

Established in 1959, Johnson Electric exports its products to more than 30 countries for use in hundreds of different product applications. Innovation and product design centers are located in Hong Kong, China, Switzerland, Germany, Italy, Israel, Japan, UK and USA. Globally, JE employs more than 35,000 employees and subcontract workers in over 23 countries.Within the Innovation Centers, design teams are organized into engineering centers of competence (ECC) based on specific technology. Johnson Electric Asti focuses on electric fan modules for automotive applications.

Interview with Gabriele Milanese, Johnson Electric Asti, Advanced engineering team 1. How long have you been using CFD simulation technologies and mathematical modeling in your company?CFD tools were introduced in the early 90s as a relevant part of our fans design procedure, while we recently started using mathematical models to better exploit simulation results.

In recent years, combining algorithms and methods of these two fields, we are significantly enhancing the fans design process.

2. What was the main reason for introducing these technologies? The aerodynamic design of fans for automotive applications has to deal with several flow complexities arising both from compact installations constraints and production process characteristics (e.g. plastic molding). The use of simplified models and of full 3D simulations supports the designer in the control of the flow characteristics of such fans type.

3. What kind of products are you using simulation for?Beside fans design, within the thermofluid-dynamic field, simulations are used to study and define cooling devices for electric motors and electronics. Simulation tools at JE Asti are widely used also within the structural field and for electric motors design.

4. Why did you decide to introduce mathematical modeling in the design process?In our design process the use of simplified physical models is a key element for designers to define preliminary geometries: 3D simulations provide valuable information but, being a verification tool in nature, it is not straightforward to use them as a design tool.

Axial flow fans design for automotive cooling


Fig. 3 - 3D CFD : postprocessing on a fan-stator configuration

Fig. 2 - 3D CFD : fan rotor mesh

Figure 1.Gabriele Milanese, PhD, Fluid-dynamic and aeroacoustic eng.

27 - Newsletter EnginSoft Year 12 n°3 Case HistoriesCase Histories Newsletter EnginSoft Year 12 n°3 - 26

Beside the well-known techniques and algorithms (e.g. DOE, optimization), graphics and innovative mathematical modeling tools have been developed, significantly enhancing the ‘full 3D’ design phase.

The initial results are encouraging. A design test has been carried out on a 460 mm diameter fan with a design speed of 2700 rpm and with multiple working point requirements. The design has been completed within a couple of days obtaining high efficiencies all over the specified working range.Further steps and related tests are ongoing aimed to build a fully integrated design environment where the designer can drive at the same time all the interacting aspects.

ConclusionsA complete and flexible suite of tools for axial flow fan design is under development, integrating open source and commercial software.Mathematical modeling is the key to allow the designer to effectively use information gained from 3D simulations.

Besides the ‘traditional’ use of optimization and metamodeling techniques, an effective approach has been developed to help the designer in the fan fluid-dynamic analysis and the subsequent design choices.

Gabriele Milanese, Johnson Electric Asti

Fig.4 - 3D fan design graphical interface

Fig. 5 - Fan design test : RBF model vs CFD vs Exp.

Forging Ahead

GIVA Corporation consists of several divisions capable of steel melting, forging, machining and a division for manufacturing of large industrial valves for oil and gas application. The forging divisions are unique in their wide range of available hot forming equipment that include ring rolling and open die press forging.

Open die press sizes range from 1500 tonnes to 100,000 tonnes. In which, the new 100,000-tonne press at GIVA’s Forgiatura A Vienna facility, is located in Rho (Milan), Italy. During 2013 and 2014 this press became operational and owns the title of the largest press in the world.To further explore the commercial applications of this equipment, GIVA established a new engineering group that consists of expertise in forging, Ni-base alloy metallurgy, and computer simulation modeling. This was an independent staff of engineers that could work cross-divisionally to access all of GIVA’s corporate resources. In the last two years, this team has invested a significant amount of time with training, investigating and education on the modeling technics and laboratory characterization of the Ni-base alloys. Their efforts have recently matured to a point where the first operational trials were completed by making large (one-metre OD) Ni-base alloy turbine discs.

Their process development method uses theoretical concepts of Process Mapping integrated with computer simulation modeling. These engineered process parameters were found to be easily implemented in the 100,000-tonne state-of-the-art press control system, allowing a reproduction of all forging parameters as given in the final computer model optimization. With this equipment and engineering tools they successfully controlled the Ni base alloy grain size and precipitate structure (delta phase) without the creation of tears or cracks during the final forging process.

Figure 1 - Final shaping

A new 100,000-tonne press for forging turbine wheels for power generation applications, located in Italy, is the world’s largest. Its engineers discuss the steps involved in working with it to produce a successful forging process


Their next steps are to integrate the engineering concepts and procedures for other Ni-base alloys and increase the size of the discs to two metres OD.

Critical materialsGIVA’s commercial interest was in turbine wheel development for power generation applications. Turbine wheels are part of the assembly of a turbine engine used to power the generator. In this case, the generator is 60 or 50 Hz running at 3600 RPM or 3000 RPM and connected to a main commercial/public power grid. As these parts run at such high speeds and temperatures, the materials necessary to perform in these environments are critical. Ni-base alloys are commonly used to overcome these issues. These alloys are designed to avoid distortion in high-temperature environments. However, this material property also makes them more difficult to hot form (forge), requiring higher-than-normal press power to achieve the final forge shape of the part. In addition, the alloys are more prone to adiabatic heat-up and tearing if not handled properly during the forging process.

These issues are the direct results of the first law of thermodynamics, which states that the more energy you put into the system (in this case press power), an equal amount will exhibit itself as heat, mechanical tearing, and/or structural changes. Where some subsequent changes are needed; because some are harmful to the final product.The concept of this energy distribution has been intensely studied within the last several decades. In which several tools have evolved to explain it, one of which GIVA found to be of considerable beneficial, technically named Process Mapping (PM).

Specifically, the energy distribution curves given in PM were excellent in explaining the hot forming process. The parameters used to explain the concepts were found to be the same (strain, strain-rate and temperature) as those used in computer simulation software. Therefore, a decision was made to incorporate/integrate the PM tool with the FEM code FORGE NxT to assist in the prediction of the wanted and unwanted effects of the forging process.Computer simulation of the manufacturing process (SIM) uses finite element analysis and CAD-CAM tools to project the process as a function of strain, strain rate, and temperature. Using SIM,

the entire forging procedure can be simulated using iso-curves to determine such physical characteristics as:

• Material flow;• Die life;• Strain-rate maps 2D and 3D;• Strain maps 2D and 3D;• Thermal maps 2D and 3D.

Figure 1 is an example of a SIM profile for a rotating disc application.This allows a prediction of the process prior to implementation or investment in raw material and manufacturing cost. For development programmes this tool has significant value in reducing the time and cost to bring a product to market. The cost of the manufacturing can be significant, especially when evaluating close-die or semi close-die forging processes when issues such as die filling, high strain regions or high potential laps areas are involved. Knowing this information prior to manufacturing avoids machining and re-machining of dies and retrials of the forging process.PM takes the finite element analysis one step further in developing hot forming procedures. It uses the strain rate, strain and temperature profiles to determine areas defined as instability, and determines the optimum process variables needed for manufacturing.Typical Process Maps are shown in Figure 2. On a two-dimensional scale, for given strain regions, power dissipation is shown as ISO lines. The power dissipation variable is a dimensionless parameter that is a function of strain-rate, strain and temperature. Power dissipation is a term used to express the efficiency of removing the mechanical and thermal energy put into the forging during

Figure 2 - Process Maps

Figure 3 - Final shapingFigure 4 - SIM-PM

the hot forming operation. As mentioned before, thermodynamic equilibriums must be met during the process. How the energy is distributed in the process is important to the success of the operation. If the press energy is distributed in a deleterious fashion, defects such as forging surface tears, internal cracking/tearing and unwanted structures are formed. The PM defines the strain-rates, strain and temperatures when these problems are most likely to occur. In Figure 3 the suspect regions have been darkened with hash marks. These regions are defined, in PM concepts, as instabilities and are recommended as regions to be avoided.

When profiling and modeling a forging procedure the critical processing variables (stress-strain, stress-strain rates and temperature) directly relate to the final structure in the forged part. Integrating the potential risk areas from the PM into the SIM modeling adds another dimension to predicting the outcome of the forging process. High-risk areas are where defects such as tears can be easily determined by means of color coding and three-dimensional imaging.Figure 3 is an example of this integration process as developed by the GIVA engineering group. The areas of dark red visually represent high-risk material flow zones where the energy distribution is unfavorable for material flow. In other words, GIVA engineering

has developed SIM software modifications that distinguish the unstable PM zones in a 3D color-coded strain-rate/temperature map. Internally, the high-risk areas are noted by superimposing the PM on the SIM, creating a new functional parameter which GIVA Engineering defined as:

SIM-PM = f (strain, strain rate, temperature, material instability, material efficiency)

The new process parameter was reported as a direct output of the results of the forging simulation process (SIM).

Saving cost and timeIt was during the two-year development program of the PG Ni alloy disc that the GIVA group developed and used this SIM-PM tool, specifically to determine if any modifications were needed to the forging process. As noted above, the goal was to save cost and time in determining the viability of this product line for GIVA upper management.The resultant SIM-PM integration is shown in Figure 5. This SIM-PM shows that the areas of high risk have been identified for further evaluation by Engineering. The locations of potential instability are shown in orange and red.Using this information, manufacturing procedures were selected to avoid processing variables that may aggravate the predicted unstable regions.These procedures would include the selection of the appropriate Ni-based alloy forging methods such as:

• Die temperature control;• Temperature;• Thermal cycle times;• Lubrication methods;• Etc.

The results - as shown in Figure 4 - were reviewed, and it was determined by GIVA engineering that the SIM-PM profile was a viable process operation, and therefore they could release the process with detailed instructions for controlling the manufacturing forging procedure.Communicating this complex information to manufacturing was not an easy task. The process of porting the SIM-PM had to be done carefully, with all involved personnel aware of the process procedures and how the exact SIM-PM parameters were to be implemented. This required detailed instructions that were reviewed and, in some cases, mock-ups were used to train the manufacturing personnel.In addition, the specific parameters had to be given to shop supervision and programmed into the operating systems of the 100 kt press. This was all done as part of the development process prior to the final forging procedure.It was in the latter part of 2014 that the final trial of these engineering concepts was used in processing the Ni alloy forging. The forging trial was a success.In the initial training, GIVA contributed SIM-PM engineering, equipment preparation and equipment SIM-PM parameter programming. A resulting disc is shown in Figures 5 and 6. This

Figure 5 - Final shaping

Figure 6 - SIM-PM


figure shows the final forging in the finished condition, directly after forging (Figure 6) and after the forging had cooled (Figure 7). A close observation of the forging in both conditions showed no defects that would be the result of forging instabilities defined by the SIM-PM process. The final result was an acceptable forging that was free of any defects that could be traceable to a PM instability (for example, surface tearing). Previous manufacturing tests had shown similar results.This successful result of the SIM-PM process was deemed by the GIVA engineering group to be a future tool and an engineering step to be used to predict the forging processes of Ni-based alloy products.The decision was taken to include a specific risk evaluation using SIM-PM in the determination of related PM instabilities. If the risks are deemed critical, modifications are made to reduce these risks at the SIM process stage. The SIM would be reinitialized and evaluated to determine the modification’s effectiveness. Once a

forging process was determined to be acceptable, the changes would then be formally approved by engineering and the details of the forge procedure would be released to manufacturing. The release procedure included necessary shop crew training and parameter implementation in the 100,000 tonne press control system.

Figure at the top of the article - The 100,000 tonne press at GIVA’s Forgiatura A Vienna facility

David Smith is Process Engineering Lead, Davide Colnago is Metallurgist/Process Engineer, and Daniele Brunelli is FEA Lead

Engineer at GIVA

Article Published in PEI - Power Engineering International Magazine - http://www.powerengineeringint.com/articles/print/

volume-23/issue-6/features/forging-ahead.html

Giva & Enginsoft: SW & HW success story

More than 10 years ago GIVA Group started to introduce Transvalor Forge to simulate their open-die forging processes. The pioneer company was OFAR with a standard workstation with only 4 cores, supported by ing. Marcello Gabrielli of Enginsoft from the early simulations, to give to the users some base training and after on-job specific training on specific topics. After some months of analysis made by ing. Daniele Brunelli on different parts, each one with useful results to improve the real produc-tion sequences, OFAR started to investigated non-standard materials and non-forging processes, like ingot heating and part re-heating, cooling and heat treatments. Enginsoft followed each request with continuous support, asking when needed to the software producer Transvalor improvements on the computation code. For non-standard materials, import from exter-nal database like J-Mat Pro and inverse analysis included in Forge have been used. For ingot heating phase, the optimization routine embedded in Forge has been used to tune the pause-in-temperature around 750°C and obtain a correct heating of the part. For quenching, new features have been introduced to follow the step of immersion and emersion from the quenching bath, responsable of part distortions, but also the some inverse analysis has been used to find the correct value of thermal exchange co-efficients of the quenching bath.

The needs to obtain even more advanced and precise results, togeth-er with an increasing request of simulation from the other Companies of GIVA Group, made it necessary to evaluate a complete re-thinking of the hardware and software architecture. After some meetings, needed to qualify who would have used the software and for what type of analysis, Enginsoft IT proposed a shared architecture based on an 60 core HPC cluster, used remotely by 3 sites for calculations, but with the possibility to share calculation tokens also on local machines. Enginsoft supplied the HPC hardware ready-to-use and configured the calculation network for the maximum efficiency. In terms of software, a customized proposal has been signed, with a progressive growth of available tokens based on a

continuous monitoring of software usage. Finally, but the most important point, a dedicated on-site training has been proposed, to increase the number of users from 2 to around 10 people and to treat in details topics specific for each site like ring-rolling, open-die complex sequences and the spray quenching process, where a completely new template has been finalized thanks to Transvalor Special Projects Team work directly linked to GIVA for validation purpose.

This “global approach” (SW+HW+training) is considered by GIVA Group a key factor to get the best out of these simulation tools: Giva consider Enginsoft as the right partner to be always updated on the last available options to improve the efficiency of this work and able to help them to investigate new needs on process simulations.Enginsoft is proud to see that this close collaboration with this customer and Transvalor continues to bring value to this company and offers this space on this number of the newsletter to thanks GIVA Group for this.

Topic Introduction (Issue & Solution), IndustryAutomotive industry is one of the fastest evolving industries when it comes to Passion, Expectation and Demand. Continuous addition of new techniques and technologies in the vehicle has shown exponential improvements in the past and this growth rate is increasing exponentially. With the introduction of electronics and its fine integration with mechanics there has been more evolution in the last 20 years than in a century. There is a highly progressive expectation of improvement in overall efficiency, a vision for global system’s size reduction along with the strictness towards the environmental impact. Optimized use of resources and precise crafting of technology is of utmost importance.

Magneti Marelli Powertrain is Magneti Marelli business line dedicated to engines and transmissions components production for cars, motorbikes and light vehicles. With an vision for the future of technology Nazario Bellato, Simulation Manager, Magneti Marelli Powertrain, Italy along with Vijay Raj Gupta, HOD, Magneti Marelli Powertrain, Mechanical, India are determined to work together and develop competencies in the Off-Shore development center at Manesar, India. An Indo – Italian Magneti Marelli team under their guidance is working to develop an integrated Air Intake System for Internal Combustion Engine (ICE) that will improve the performance at the vehicle level by reducing overall energy losses and will open the room to provide more technological alternatives for psychological comfort of the end users. Involvement of young engineers filled with passion for the technology in the team has given Magneti Marelli a vision to create and give hope for more challenging windows for the better future of technology. The team is working with the support of ANSYS to use special

New Methodology: Intercooler Integration Space & Efficiency Optimization

Figure 1 - CAE Team


3-D Computational Fluid Dynamics (CFD) features to simulate physically and numerically complex Thermo-Fluid Dynamic behavior to evaluate the overall system’s response and hence confidently analyze many virtual prototypes using DOE strategy in the time that would be required to build a single physical prototype.

Analysis Description & what we were trying to accomplishIntegration of the intercooler with Air Intake Manifold involved several benefits such as reduction of overall pressure loss, the possibility to use Air to Liquid Intercooler which has higher heat capacities compared to Air to Air Intercooler. In house manifold design flexibility gave us the opportunity to make air flow lines uniform, and hence optimize the heat exchange. The prime objective of this approach was to design the Air Intake Manifold and optimize the fluid dynamic behavior of the air inside it. For this, the first and probably the biggest challenge was to understand the fluid dynamics inside the component and find innovative ways to numerically calculate the thermal, static and dynamic parameters of the integrated system.

Developing the integrated system configuration through conventional process involving several iterations of design – analysis and testing was not the preferred choice because this project particularly involved large number of parameters in its Design of Experiment (DOE). An extensive use of numerical analysis in the system’s development cycle seemed to be the best compromise to optimize Quality – COST – TIME. For a CFD analysis, to adequately capture flow behavior on regions which will experience abrupt change in key variables such as pressure, velocity or temperature, it is necessary to have a refined mesh. This is done by generating inflation/boundary layer in those regions. This boundary layer preferably involves Hexa - Hedral prismatic elements as it is considered to be a better representation of the numerical domain for a CFD model. Quality of the boundary layer profile can be best understood in terms of y+ parameter which is a non-dimensional variable representing the distance from the wall to the first node away from it. It follows the following correlation:

Boundary layer mesh generation on one hand is basic criteria for an acceptable CFD analysis but for complex geometries (usualy the case in powertrain components) creation of a very high quality boundary layer mesh requires long time and large resources. From

the lesson learned in the past, here in Magneti Marelli we use a Hibrid meshing approach for CFD analysis as a trade off between QUALITY and TIME. The number of prismatic/hexa-hedral layers are decided uniquely for each analysis and component depending on the design and design change.

CFD analysis for this integrated system configuration was a mammoth task because of the presence of irregular fluid domain that comprised of micro channels of the intercooler. Modeling the air passage domain of the intercooler was a very complex task as they were too small to be meshed with a good quality hybrid mesh. Even the Hex-Dominant approach was not the preferred one because of the presence of flow hindrances that were designed to create turbulence and improve the heat transfer along the flow direction. Even after neglecting the presence of flow hindrances and meshing the micro channels as regular fluid passages the approach remained very complex as it led to very high element count, non-uniformity in the mesh density and hence long solution times.The numerical analysis team overcame these challenges by using special ANSYS features. Small sizes of the air passages in the intercooler domain were assumed to be the porosity in a porous

Figure 2 - ANSYS 2015 Italian User Conference: Massimiliano Di Paola

Figure 3 - The new approach, Intercooler integrated with Air intake manifold

Figure 4 - The boundary layer approach

medium. The directional loss model in ANSYS CFX that corresponds to the Darcy’s momentum loss equation for the fluids flowing in a porous medium was used for the analysis. The possibility to alter the linear and quadratic loss coefficients gave us the opportunity to study intercooler’s porosity levels based on its design and control the air pressure drop across the intercooler based on the experimental data.This special feature in ANSYS CFX proved to be a tool that substantially reduced the complexity of the numerical model. Not only a reduction of approximately 12 million elements and 24 working hours was observed for each simulation, we were also able to use the numerical and experimental results to study and develop an approximated correlation between design and porosity levels of the intercooler.

Now, as a next step towards further improvements in the system, our team is determined to study several other complex phenomena through numerical analysis such as the Condensation effect inside the air intake manifold, the water hammer effect in the intake system during the engine cranking. Studying and understanding these phenomena’s could help the team to suggest design improvements and reduce the present losses substantially. These physical phenomena’s are industry specific and hence it is quite uncommon to have the related codes and models in the regular ANSYS package. For this reason ANSYS has agreed to provide a direct support to our Magneti Marelli team which would involve a series of learning sessions and discussions one on one with ANSYS’s top researchers.

Outcome description & Helpful ANSYS featureANSYS tool once again proved to be a very useful platform for us to analyze some of the complex phenomena which once became a bottleneck in the component development process. Rational understanding of phenomena’s such as the flow uniformity,

intercooler permeability not just from the fluid dynamic point of view but also from point of design and structural stability of the system was studied. Further, with the use of ANSYS we were able to iteratively alter the system’s configuration and optimize the physical phenomena such as heat exchange through the intercooler and reduce the overall pressure loss.

Direct one on one support from the ANSYS team has given a lot of confidence and a vision to our engineers to solve highly complex physical problems numerically.

Massimiliano Di Paola - Senior CAE AnalystMagneti Marelli Powertrain S.p.A, Bologna, Italy

Nicola Mundo - Simulation EngineerMagneti Marelli Powertrain S.p.A, Bologna, Italy

Bhartendu Tavri - Simulation EngineerMagneti Marelli India Pvt. Ltd. Manesar, India

Figure 5 - Definitions for Darcy’s Law and its application for the Intercooler simulation

Figure 6 - Images Illustrates the Velocity profile of air for the Momentum Loss Model


Optimization can be beneficial for engineering design from the individual component level to the system level. Research engineers at Dow Chemical were interested in exploring the application of recent advances in engineering optimization using commercial Computer Aided Engineering (CAE) tools for their simulation based design phase. The design of a polymer die, also known as “coat hanger” die, was chosen as a test case for such an optimization study to improve more than one objective. The main objectives for the optimization were to maximize the flow uniformity of the polymer melt across the die exit while simultaneously minimizing the pressure drop. Polyflow, a commercial Computation Fluid Dynamics (CFD) code, and modeFRONTIER, optimization and process integration software, were used to optimize the design of one such coat hanger die.

The process started with the simple task of evaluating various optimization strategies to be used with Computational Fluid Dynamics (CFD), which subsequently led to more advanced, and efficient methods for Dow’s needs. The combined CFD-optimization methodologies were able to identify die geometries that predict more uniform flow profiles at the die exit than the base case geometry, with lower pressure drops. A comparison of the computational effort for the different methods was made. However, it was observed that some of the initial optimized results yielded geometries that were moving away from typical coat hanger die shapes and would be challenging to fabricate. This led to the addition of geometric constraints to push the optimization towards more feasible geometries. Therefore leading to a CFD-based optimization strategy, coupled with knowledge of die design technology and fabrication techniques, which can lead to the design of better performing polymer dies in a timely manner.

Introduction Coat hanger dies are widely accepted in the polymer processing industry because of their comparatively simple geometries

and proven ability to produce products with reasonably uniform thicknesses. However, the thickness uniformity is very dependent on the flowrate, the non-Newtonian polymer melt rheology, and the geometric shape of the die. Thus the design of new dies can be difficult and time consuming, often involving both numerical and experimental trial-and-error methods. The main geometric components and resulting flow patterns of a coat hanger die are depicted in Fig. 1. There are four major sections including the feed pipe, the distribution manifold, the land region and the die lips. The polymer melt flows along the relatively thick distribution manifold, with material continuously flowing into the thinner land region and finally entering the thin die lip region. The key to achieving uniform flow at the die exit is to create the flow

Numerical Optimization of Polymer Die Design for Multiple Objectives

Fig. 1 - Coat Hanger Die Features © Carl Hanser Verlag, Muenchen1

channel geometry in such a way that the same pressure drop is obtained along all flow paths. Over the last couple of decades, researchers have started to combine various numerical optimization methods with three-dimensional melt flow simulations to improve flow uniformity in film or sheet dies. In an early study, a coat hanger die model using a two-dimensional Hele-Shaw flow approximation was coupled with a Sequential Quadratic programming (SQP) algorithm for Newtonian and non-Newtonian fluids. More recent studies have used 3D finite element simulations results to create response surface models to use as the numerical engine for the optimization problem.

In this study, we evaluated the effectiveness of a number of optimization strategies by looking at both the optimized results and the computational resources required. We also varied the number of geometric parameters that could be modifies to reach an optimal solution. The primary objective was to maximize the flow uniformity, which was represented by a minimum to maximum flowrate ratio. This number varies between 0 and 1, with perfect uniformity having a value of 1. The secondary objective was to minimize the pressure drop within constraints.

Simulation Setup The coat-hanger geometry used for this study had a width of 25.9 cm and a fixed die lip gap of 0.152 cm. The initial CAD geometry for the baseline geometry was along the centerline to provide a half symmetry section. Additional cross sectional cuts were used to divide the geometry into multiple segments along the die width, as shown in Fig. 2a. The geometrical parameters for optimization were applied to the cross section geometries. This allowed the flow volume of new die geometries to be reconstructed, based on the

cross sectional cuts, using the skin-loft operation of the ANSYS Design Modeler tool. The parametric model of a cross-sectional die shape is shown here in Fig. 2b. In the present study, the die was split into either two segments (using one cross-sectional plane), three segments (using two cross-sectional planes), or four segments (using three cross-sectional planes).

Optimization Process, Methods and ResultsThe optimization started by automating the processes of geometry manipulation, meshing and running of the CFD analysis. This was accomplished by coupling the CAD, meshing and CFD software tools in modeFRONTIER, which serves as the automation manager as well as the numerical optimization package. In a previous study, a response surface model (RSM) was created based on a set of CFD results from the Design of Experiments (DOE) study. The RSM was then used as the computational engine to optimize the geometry to meet the two desired objectives. We wanted to investigate various parameterization and direct optimization strategies, looking at both the optimized results obtained and the computational resources and time requirements. In order to look at the effect of the degree of parameterization, the

Fig 2a - Segmented Coat Hanger Die Geometry based on two Cross-Sectional Cuts; 2b - Cross Section showing the Design Parameters. © Carl Hanser Verlag, Muenchen1

Figure 3 - modeFRONTIER Workflow

Fig 4 - Comparison of optimal solutions from cases with different Degrees of Freedom © Carl Hanser Verlag, Muenchen1

Understand the right solution for your needsWe understand that it is important for you to understand that the solution will meet your requirements and that value you can gain.

Request a demonstrationmodeFRONTIER is a very diverse tool and as demonstrated in this special newsletter issue, can be used in a wide range of applications from engineering design to business simulation model optimisation where our customers have reaped many benefits. Enquire about a demonstration and short presentation to your team to discover how modeFRONTIER has been applied in your sector.

modeFRONTIER WorkshopAttend a free introductory modeFRONTIER Workshop – the one day hands-on-training has proven to give engineers with the confidence to get started with modeFRONTIER. See the various dates over Europe http://www.enginsoft.com/events/

Pilot ProjectsUncover the benefits your company can receive and gain an insight into the potential ROI that can be achieved with modeFRONTIER. An effective and the most popular method for those seeking a optimization solution. Through the pilot projects we engage with your engineering team to understand your requirements and determine with you your criteria’s for success. The confidence you need to understand the right solution for your needs.

Further Training and SupportAll companies face different challenges, we strive to provide the necessary training and support you require to ensure the solution continues to deliver the maximum value and return. Discuss your requirements with us today insert appropriate contact details or contact your local EnginSoft Office:http://www.enginsoft.com/about-us/where-we-are.html

Uncover the possibilities with modeFRONTIER

Newsletter EnginSoft Year 12 n°3 - 36 Case Histories

problem was formulated in three different ways using an increasing number of adjustable geometric parameters (design variables) as listed below. The parameters used for each of the cases are indicated in Table 1.

1. One adjustable cross-section with 3 parameters,2. Two cross-sections with a total of 8 parameters, and3. Three cross-sections with a total of 12 parameters.

For each of these parameterization cases, various DOE and Optimization strategies were evaluated based on their applicability and adaptability for the number of design variables being investigated. Since CFD can be both time and resource intensive, a compromise may need to be made between numerical accuracy and the effort needed to reach an optimized solution.The process integration and optimization software, modeFRONTIER, was used to apply the various DOE/Optimization strategies, manage the automation of CFD runs, and analyze the data. One of the modeFRONTIER workflows for this project is shown in Figure 3.

Computation and Wall Clock Time for OptimizationFor CFD-based optimization (or optimizations involving other computational intensive compute engines), the total wall clock time required for optimization is primarily dictated by the time required for the CFD simulations. For this study, each CFD simulation took approximately 50 min on 8 processors. In comparison, the overhead time required for the optimization algorithm to run is a fraction of a second. The number of CFD simulations to approach a maximum in the flow ratio for the various cases is recorded near the bottom of Table 1, along with the estimated wall-clock times for each case. The three-parameter model was the first model that ran and after the optimization, a number of designs were identified that had a more uniform flow distribution and lower pressure drop than the baseline design. Eight-parameter and twelve-parameter models were then investigated to see if further optimization could be obtained. The eight-parameter case study was able to provide solutions with flow uniformity similar to the three-parameter case study but with improved pressure drop. The case study with twelve degrees of freedom was able to show improvements in both flow uniformity and pressure drop. The optimization progression with increasing

degrees of freedom is shown in Fig. 4. There is a computational resource penalty when considering more design parameters and using a more accurate CFD model as the compute engine. By using modeFRONTIER, the process can be fully automated so that analyst time is minimized and the computer resources are fully utilized, allowing for a higher level of accuracy in the analysis. For the eight-parameter case both the NSGA-II and FMOGA-II algorithms were used. The FMOGA hybrid strategy uses virtual optimization and exploration along with

validation and adaptive response surface training speeding up the research phase. The FMOGA-II

scheme was able to form the expanded Pareto front with just 250 CFD simulations, compared to the 660 simulations required by the NSGA-II scheme. Even though the FMOGA-II seems to be the most effective in minimizing the number of CFD cases required for the eight-parameter formulation, the NSGA-II scheme was able to identify a design with a higher flow ratio than the FMOGA-II scheme. This is shown Table 1.The third formulation of the die optimization problem involved twelve parameters. The previous analysis indicated that as the flow ratio increases, the pressure drop tends to decrease up to a certain value and beyond that any further reduction in pressure drop is possible only with lower flow ratios. It was decided to apply a direct single objective NSGA-II optimization strategy. Table 1 shows the resulting best design with a flow ratio of 0.96 and a pressure drop of 29.5 MPa. In order to push the optimization towards feasible solutions from a fabrication and designer standpoint, appropriate constraints on both the design and output parameters were imposed. This

Figure 5 - Die Shape and Flow Uniformity (a) Before Optimization, (b) After Optimization © Carl Hanser Verlag, Muenchen1

Table 1 - Baseline vs. Optimized Input & Output

was investigated using the twelve-parameter case by adding a constraint to require that the land length increased when moving from the edge towards the center of the die. The flow rate ratio and pressure drop for the constrained case was calculated to be 0.95 and 29.5 MPa vs. 0.96 and 29.5 MPa for the unconstrained case, so the constrained case has a slightly less optimal solution. The constrained showed a typical coat hanger geometry and could be altered to create a smooth curve. The flow distribution curves across the die width between unconstrained vs. constrained cases did not show any significant differences. Figure 5 shows the comparison between the baseline and optimal design in terms of both die shape and flow uniformity.

Conclusions This study has shown how CFD-based optimization can be used to improve the flow uniformity of a coat hanger die while minimizing the pressure drop. It was shown that more optimal solutions can be identified by increasing the number of geometric parameters that are allowed to change; however the computational cost also increases. Various optimization strategies were investigated, showing a trade-off between computational accuracy and the number of simulations required to reach an optimal solution. When dealing with a small number (< 5) of adjustable parameters and fairly linear responses, a simple DOE-RSM method may be adequate to identify an optimal solution in a short amount of time. The FMOGA method will be more accurate than the RSM method and shows promise when 10 or fewer adjustable parameters are involved. For a larger number of parameters, a direct optimization method, such as NSGA, may be the best option, although at a computational cost.

All the optimization methods investigated in this study were able to identify geometries yielding more uniform die exit flows than the base case. However, some of the initial geometries were not feasible from a fabrication standpoint. Additional constraints were placed on the allowable geometric parameters and solution space to push the optimization towards more feasible optimal solutions. modeFRONTIER has a rich library of optimization algorithms that can handle both single objective and multi-objective cases; thus it was found to be a powerful automation tool with its capability to generate a DOE scheme, and run an optimization study while integrating CFD into the process. This integration when compared to the typical trial-and-error method of finding an acceptable solution can result in significantly less calendar time and manpower to reach an optimal solution.

A more comprehensive review of this study is published in the following journal: Lee, P. C., Dietsche, L., Dooley, J., Parashar,

S., Improving Film Die Flow Uniformity Using Optimization Methods Coupled with Finite Element Computational Fluid

Dynamics (CFD) Analysis, International Polymer Processing: Vol. 30, No. 1, pp. 51. (2015).

Patrick C. Lee - Formerly The Dow Chemical Company; currently at the School of Engineering, University of Vermont

Laura Dietsche - The Dow Chemical Company, Midland, MIJoseph Dooley - The Dow Chemical Company, retired; currently

JDooley Consulting LLC.Sumeet Parashar - ESTECO North America


Blasting Simulation drives plants design for safer workplaces

IntroductionLS-DYNA (originated from the 3D FEA program DYNA3D), developed by Dr. John O. Hallquist at Lawrence Livermore National Laboratory (LLNL) in 1976. DYNA3D was created in order to simulate the impact of the Full Fusing Option (FUFO) or “Dial-a-yield” nuclear bomb for low altitude release (impact velocity of ~ 40 m/s). At the time, no 3D software was available for simulating impact, and 2D software was inadequate. Though the FUFO bomb was eventually canceled, development of DYNA3D continued. DYNA3D used explicit time integration to study nonlinear dynamic problems, with the original applications being mostly stress analysis of structures undergoing various types of impacts. In the recent decades, LS-DYNA is universally known as a reference solution to solve blast, vehicles (IED and mines), and home land security problems.In the last few years, EnginSoft has extended the use of the code to civil and industrial applications; blasting simulation is a key solution in

order to improve the safety of plants, workshops, off-shore platforms and in general, all buildings/rooms with explosion risks.New features are continuously implemented in LS-DYNA for faster and reliable blasting simulation. In this article an application example will be illustrated.

1. Blasting modelling methodsDifferent methods are implemented in LS-DYNA. These methods can be applied in different domains and coupled during the solving process. Turin Polytechnic and EnginSoft have validated all the approaches by comparing the numerical solutions with the analytical/experimental ones.

1.1. Load_Blast_Enhanced (LBE)The first method used for simulating a blast wave with LS-DYNA is called Load_Blast_Enhanced (LBE). It is based on the CONWEP function and allows the user to simulate bursts using an analytical formulation to include the distance from the center of the burst and the amount of the explosive used, through a parameter called “reduced distance”. This parameter is the ratio between the distance from the explosive charge and the TNT weight (in kg), the latter being raised to the power of 1/3. The algorithm is based on the equivalent TNT method; indeed, several kinds of explosive can be simulated by using an equivalent amount of TNT and appropriate conversion factors. This method allows the simulation of different kind of bursts:Fig 1 - LBE overpressure validation

Fig 2 - LBE pressure distribution on the target surface

• Hemispherical surface burst• Spherical air burst• Air burst with ground reflection• Air burst – moving non-spherical warhead

The development of the shock wave through the air is not explicitly simulated with finite elements using the LBE method, but it follows an analytical formulation. At a certain time, which means at a certain distance from the explosive charge, there will be a certain pressure level value.To get a tangible pressure value the shock wave should be applied over a target surface which has a defined distance from the center of the burst (figure 2): the computational time depends on that distance. The number of elements which composes the target structure also affect the computational time. The convenience of this method is its low computational cost, but it does not take into account the reflections due to the possible presence of objects situated between the explosive charge and the target. The presented numeric method has been utilized for the simulation of a free-air-spherical-burst: the results have been validated by comparing the experimental data taken from Kingery and Bulmash. (figure 1-3).

1.2 Multi Material ALE (MM ALE)The name of the method, Multi-Material, derives from the fact that different domains are involved in modeling the explosive and the shock wave propagation medium.It is based on the numerical solution ALE in which one mesh is a Lagrange-type-mesh and the other one is an Euler-type-mesh. One domain is created for representing the shock wave propagation phenomenon and is modeled according to the material of the

burst event environment; the second material, which is the Lagrange-type one, is used for simulating the explosive domain (figure 5). Both domains have an assigned material which allows the modeling of different kinds of explosive and those of the shock wave propagation space. Furthermore, a pressure propagation equation is used to characterize the shock wave propagation environment. This method is appropriate for simulating enclosed bursts, ground-level-bursts and free-air-spherical-bursts taking into account the interactions (the reflections) with secondary objects placed between the explosive charge and the simulation main object. The presented approach is a mesh-dependent method as it is based on a multi-

material domain modeling technique which is made by using solid finite elements ALE; this means that the convergence studies need to be made according to the type of simulation under examination. The computational cost is significantly higher if compared to the Load_Blast Method as the number of finite elements involved in the simulation model increases considerably when the explosive domain and the shock wave propagation environment are considered together.The same free-air-spherical-burst simulation has been conducted using this technique and the results have successfully validated the Load_Blast data-method. (figure 4-6).

1.3 Mixed Method LBE-ALE As a last validation test, a mixed method has been adopted. It has been developed by considering the two previously described simulation techniques: the LBE and MM_ALE methods. The idea is to simulate flat shock waves when at long distances from the detonation center. In this way, the computation is faster than the pure ALE method and requires a smaller number of finite elements. In the domain

Fig 3 - LBE trend of shock speed

Fig 4 - MM-ALE overpressure validation

Fig 5 - MM-ALE Time sequence of pressure contour

Fig 6 - MM-ALE trend of pressure in the time domain


which is not modeled by using finite elements, the shock wave is simulated with the LBE method, while for the modeled domain, the ALE_MULTI_MATERIAL method is employed.To better understand this technique, consider the configuration shown in the figure 7: the explosive is situated at a long distance from the target; furthermore the target is not placed within the direct field of view of the detonation point. The pressure pulses reach the target after reflecting on the surrounding structures. If the simulation were just based on the MM_ALE solver, the huge external environment, which also includes a quarter of the explosive domain, should have been taken into account within the Euler-type mesh modeling process. But, proceeding according to the presented mixed method, the two highlighted surfaces of the smaller Euler-type domain are processed by the LBE solver, while the reflected waves are processed by the MM_ALE solver.To simulate the air flow moved by the shock wave within the ALE domain, some receptor solid elements have been employed; they have the task of measuring the shock wave load which is computed by the LBE numeric method. At a later stage, they propagate within

the rest of the air domain which, in turn, is also modeled with solid elements. The receptor elements and the air domain make up the Multi-Material Domain.As mentioned before, EnginSoft is using numerical models extensively in investigating the blast effects for the assessment of industrial and civil structures.The numerical approach allows:

• to determine the blasting effects on the surrounding structures without experimental tests that are difficult or impossible to perform;

• to evaluate different plant/laboratory map from the safety point of view spending mainly CPU time.

• to evaluate different mitigation strategies including reinforcements, openings, collapsible windows, absorbers, etc etc

It is important to note that the physical phenomena is very complex and it is almost impossible to define an effective mitigation strategy without the contribution of the numerical simulation. For confidential reasons, a simplified model that represent the real

scenario

“by analogy” has been defined. The simulation methodology development and application involved EnginSoft, the Turin Polytechnic and other customers that want to remain anonymous.

2.1 Phenomena descriptionFrom the theoretical viewpoint, the explosion is an unexpected and aggressive emission of mechanical, chemical or nuclear energy, usually with production of high-temperature and high-pressure gases.

Such gases spread in the surrounding environment as a shock wave, which in the absence of obstacles, expands like a spherical surface centered in the explosion point. When a shock wave encounters an obstacle, the force is greater if more surface is exposed and if the distance from the explosion point is shorter. In fact, the pressure peak decreases when the distance from the explosion center increases. Considering the variation of pressure in time, fixed in a point in space, it changes with an exponential law achieving two load stage: the first one is positive due to overpressure

Fig 7 - Coupling scheme of the mixed method

Fig 8 - LBE_MM_ALE Time sequence of pressure contour

Fig 9 - Typical trend of the overpressure

Fig 10 - Pressure distribution in the closed room

while the second one is negative due to the depression caused by the explosion winds (figure 9). It is mandatory for the realization of a model to investigate the behavior of the structures under the effects of dynamics loads with high intensity and short duration, which are the effects generated by the explosions.

The problem is complex and can be interpreted in three basic prospective:

• Dynamics load evaluation due to the explosion and application on a structural element;

• Mechanical characterization of the structure behavior when it is exposed to high-intensity dynamics loads;

• Identification and implementation, through the use of the computer, of a numerical analysis method for the description and solution of the physical and mechanical problem.

The first investigates the scenario when an explosion occurs in a

restricted, closed environment (figure 10).This model represents several industrial incidents: explosion of a tank containing gas and/or chemical reagents, hydraulic system rupture under the effect of extreme pressure, storehouse explosion, etc etc.The simulation has been done with a spherical TNT charge explosion in a restricted surrounding, to study which methods can reduce the pressure wave intensity and preserve the boundary walls integrity.The numerical method used to perform this simulation has been: Multi-Material-Arbitrary-Lagrange-Eulerian.

The pictures show the pressure contour as function of the distance from detonation center. In this first run, the walls of the box are rigid.The walls strength (reinforced concrete) is inadequate to resist to the pressure peaks.

The first investigated strategy to decrease the pressure level has been the realization of some openings on one side of the room (figure 11).Subsequently, the pressure level with mitigation windows and steel safety walls have been compared. In a first time these walls are considered as rigid in order to verify the contribution and, in a second time, as deformable with steel rupture modelling (figure 12).

2.2 Results discussion and conclusionThe figure 13 shows the pressure peak on the room wall for the different configurations. As can be noticed, the mitigation strategies determine a lower wall stresses due to the wall pressure reduction:

B- Closed Box: This run corresponds to the higher measured pressure. The wall room that, in the real scenario are in reinforced concrete, are not able to contain the explosion and the structure collapses.A- Box with openings: This model correspond to the installation of collapsible windows on the wall and on the roof of the room. Obviously, it is possible to insert windows only on the building perimeter (the blast waves can escape in circumscribed external area). D-Rigid mitigation walls: this case has not a real correspondence, but is important to preliminarily understand the internal wall effect and investigate different architectures (number of walls, shapes, distances, heights, materials, etc etc).C- Deformable mitigation walls: Simulating different model, it has been learn that if the mitigation walls strength is too high, the blasting wave overload the room wall. In agreement with the customer, the final protective structure has been designed to absorb energy despite of its integrity. The priority is to preserve the external wall in order to protect workers in adjacent rooms. In this case, it doesn’t exists a protective structure, made of “traditional” materials ( concrete and/or steel) able to absorb energy without rupture and collapse risks; this means, from the safety rules point of view, that has to be forbidden to stay in the room if the system works.

E. Cestino, G. FrullaTurin Politecnico

G. Bolla, A. OrtaldaEnginSoft

Fig 11 - Pressure distribution in the room with mitigation windows

Fig 12 - Steel safety walls

Fig 13 - Comparison of the pressure peak for different configurations


A student from The University of Sheffield, UK presents a summary of his Final Year Project using modeFRONTIER

Space frame structures can be described as a collection of connected nodes. Parametric optimization of such structures is relatively simple as each node has 3 position parameters which can be varied by the optimization routine. As an initial investigation it was decided that two objectives would be used: One to minimize mass and the other to maximize the stiffness. It was noted that many other design considerations would be needed for a full optimization, but these were not considered in this project in order to keep the work to a manageable level for an undergraduate project. A 25% reduction in mass was achieved for a similarly stiff chassis compared to the baseline design.

Development of the Optimisation ProcessBefore the optimization could take place effectively several tasks were carried out:• Establish a two-way communication link between

modeFRONTIER and Z88 Aurora using Python scripts.• Define a feasible design region using a list of triangles and

associated nodes (Figure 1).• Create a routine to determine if nodes and elements are fully

inside a feasible design region (Point in Polyhedron test).

Initial tests on the optimization process were carried out with a simple box structure to reduce complexity (Figure 3). Later in the project a motorcycle type chassis was used to showcase a practical application of the techniques developed. It also demonstrates the use of a nested infeasible region, representing the engine block (Figure 2).Initial investigations show a tendency for nodes to group together for light weight designs so a script was written to combine these

‘almost coincident’ nodes. This script combined nodes at the mid-point and removed the resulting duplicate and zero-length elements. The results are shown in Figure 4 and Figure 5.

Multi-Objective Optimization of Space Frame Structures with Application to Motorcycle Chassis

Figure 2 - Baseline structure (green) and feasible region for the motorcycle chassis. Note the infeasible engine block region (dark red) inside the feasible region (red)

Figure 1 - Testing nodes with the feasible region. Green = valid nodal position, orange = invalid position

After subsequent optimizations it was noted that triangles of elements with very small angles were often formed which are undesirable in welded structures due to their inefficient use of materials and the difficulty of manufacture. These were removed with another script that looked for small angles in triangles of elements. The small angles were then fixed by either combining the nodes opposite the small angle node in the case of single small-angle triangles, or by removing the element bridging two small angle nodes in the case of dual small-angle triangles as shown in Figure 6.In all node combination cases, any applied loads were transferred to the new node. In the case of fixed boundary conditions, the non-fixed node was moved to the fixed one. Further scripts were written to correct the position of out-of-bounds nodes to lie just inside

the feasible region and calculate the proportion of wasted stock material for a given production run and stock tube length.

Motorcycle optimisation3 load cases were used to test the bending stiffness of the frame about the 3 principle axes. A single stiffness objective was used as the weighted sum of the maximum nodal displacements from these load-cases. The feasible region was defined as previously shown in Figure 2. The mass and combined weighted displacement values were both minimized and the material wastage was constrained to less than 5%. Hybrid algorithms were found to be more effective than MOGA-II and local optimization methods when comparing the final Pareto fronts after 2800 design iterations.The improvement over the baseline design and initial population can be seen in Figure 7 where the baseline design can be seen as a red circle. The rest of the initial randomized population are shown by black crosses and two pareto fronts can be seen in blue and green as the product of two different hybrid algorithms.

Design 2659 can be seen to achieve a 33% decrease in mass for a similarly stiff frame to the baseline design. However, as noted previously, several factors were not taken into account in this preliminary investigation. Indeed, upon inspection (Figure 8) it can be seen that this design is unsuitable due to high stress concentration at welds due to node removal. Additional optimization with the element section choice as a variable was carried out but the results were not as good as those with a single section choice due to the 3x increase in input variables Another optimization was carried out without node removal and

Figure 3 - initial box structure for simplified optimisation

Figure 4 - design from the simple optimisation with ‘almost coincident’ nodes on the loaded corners

Figure 5 - Before and after ‘almost coincident’ node combination

Figure 6 - small angle fix for single and dual small-angle element triangles

Figure 7 - The final Pareto fronts for two hybrid methods and the initial population

Sezione 45 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 44 Case Histories

resulted in structures with small weld angles that may be difficult to manufacture as shown in Figure 9.

RecommendationsTo add robustness to the optimization and ensure feasible designs, further load-cases should be added to model maximum loads and stresses. Further tests including fatigue, modal analysis, vehicle dynamics and the housing of additional components besides the engine could be incorporated into the modeFRONTIER workflow in future work. Additional investigation into multiple section choices could be carried out with pre-defined groups of elements sharing the same section to reduce the number of input variables.

Heavier but stiffer designs could be encouraged by introducing thicker sections to replace duplicate elements formed under the node collapse method and small angle removal methods.New algorithms such as JADE-II or modeFRONTIER’s pilOPT algorithm could help reduce the number of iterations needed for use with more computationally expensive simulations.

Thanks to EnginSoft IT, EnginSoft UK, The University of Sheffield

Barnaby LewisThe University of Sheffield

Figure 8 - design 2659 with a 33% mass reduction over the baseline design Figure 9 - Chassis produced without node removal script

Learn more at: www.prometechsoftware.com For queries: [email protected] EnginSoft is a non-exclusive distributor in Europe.

Robust simulation of free-surface flows at high resolutions. ✔

✔

✔

✔

Oil flow in HV transaxle Image courtesy of Toyota Motor Corporation

Excellent performance in the simulation of moving boundary problems. Capabilities for interaction between fluids and powders, or fluids and rigid bodies.

High-speed computing by multi-core CPU and GPU.

3D digitalization of your product to simulate its behaviour

Capturing the entire real object surface opens up a completely new dimension in the quality and detail of the information you get about your processes. Optical 3D metrology and in particular SLS systems (Structured Light System) give today the possibility to measure and virtualize accurately the real part in a very detailed way. In order to minimize the Time to Market and increase the quality of the production process, the 3D digitalization offers from one side, for your prototype, a full description of the part to better find out the adjustments needed, from the other side, for the final product, an accurate measurement of your critical dimension and keeps your process under control.This is the main field of application of Optical 3D metrology but not the only one. The complexity of numerical simulation is growing due to the need for real object geometry to be used as the input for the simulation task. In this case CFD, FEM-Mechanical or Process simulations (e.g. racing application, deformation analysis, injection moulding design, Casting process optimization) are someone of the most interesting new applications of 3D digitizing where the simulation of the real part, and not the theoretic one, allows us to reach the target in a precise and effective way.There are three fields of integration between 3D scanning technologies and the virtual simulation.

• In the design step, the theoretical CAD model represents the only one starting point to optimize the shape and performance, while the 3D scanning of the real model becomes

the only solution in the absence of CAD models or in case of modifications of the prototype.

• During the production step, the systematic measures of the products highlight the deviation, allowing the study of the impact of these “imperfections”, that are detected statically, in the behaviour by simulation with appropriate CAE tools, by establishing the acceptable tolerances on the real products.

• The third field of integration, 3D Scanning- CAE, is the comparison between the deformed measured and the simulated

Fig.1 - Evolution of the elementary geometries to complex form and shapes


at the end of the production process (residual deformations of the individual component or assembled) or during the usage of the product (deformation under load in operation).

Introduction in the 3D ScanningToday, design is the key strength of the products that surround us. Their functionality has become almost secondary, or taken for granted, as very often is the aesthetics to certify the success or failure. This constant research of style has transformed the products, composed of simple geometric shapes, into creations of depicting geometries and complex shapes.In this way the project manager and industrial designer has become modern artists from one side, from the other side this implicates the necessity to transform and adapt the production processes in order to realize that ideas.In this context, the optical 3D technology, and in particular the SLS systems (Structured Light System), that was for a long time a bleeding edge technology, has seen in recent years a huge success.Since 2002 QFP has introduced nationally and internationally new technologies in metrology and process control, integrating them in different production processes. In particular, the non-contact optical technology, commonly named 3D scanning, has seen an exponential growth in its use and in its progress in the last twenty years.This technology provides new, highly versatile possibilities. Capturing the entire object surface opens up a completely new dimension in the quality and detail of the information you get about your processes. Whether you work with classical nominal- actual comparisons, speed up your form finding process for the capture of design samples, or use reverse engineering to create a parameterized CAD file allowing the digital representation of “old” components for which no CAD data is available (anymore) – optical measurement technology moves your business forward in many ways.The first scanner 3D for industrial uses dated back about thirty years ago and was created with mechanical projection technology. The progress in electronic, optical and computer technology has allowed the transformation of these machines and today we have DLP projectors, optical calibrated and dedicated high-speed cameras to capture millions of points, with an accuracy level that was unimaginable thirty years ago.

A projection unit projects a fringe pattern, which is encoded using a special process onto the object to be measured. The topography of the object causes deformations and deflections in the fringe pattern. These make the object’s surface visible to the camera systems (one or more cameras) in 3D, which form the second key part of a white light fringe scanner. To determine the individual 3D point coordinates, a triangulation calculation (relationships in the right-angled triangle) is performed from each camera pixel based on beam intersection (projector – camera(s)) and encoding.Depending on the camera resolution and measurement volume that are used, a coarser or finer 3D point pattern (resolution) is placed over the object. The finer the pattern, the better intricate details (radii, edges, etc.) can be represented. This so-called lateral point spacing typically ranges between 20 μm and 250 μm, depending on the selected system.

In the automotive sector there was a need to accelerate and improve the process of reverse engineering in the field of style modelling with the scanning of the resin model or mock up. This has led today to the affirmation of CAD-CAM solutions that revolutionize the process. Since in the past there was only one time consuming solution through surface modelling in CAD, now it is possible to produce perfectly consistent surfaces for freeform area of the scanned model, parametric and regularized 3D geometry on the regular area or directly triangular mesh properly prepared to give place directly to the milling process, without having to go through surface modelling.However, it is not thanks to the reverse engineering that today these machines are advanced and widespread technology, but rather their growth and acceptance in metrology.Their use today covers a usage in different sectors and at different stages of the production process.

In automotive, aerospace, energy, industrial design today the first article inspection using this technology reduces the time to market and the process control allows to implement programs of quality management (Six Sigma), driven by growth in the rate of control.A ‘new’ and interesting field of application for the data generated from these systems is also represented by virtual experimentation or simulation.

Fig. 2 - Diagram of operation of SLS systems (Structured Light System)

Fig. 3 - 3D digitalization of the GP3 car from nosetip to rear edge of rear endplate

In fact, the 3D scanning enables in the first place to capture and virtualize the real geometry and therefore offers the opportunity to present in the simulation the real figure and not the theoretical one.The type of production process and the way in which this is implemented and parameterized determines, for the single component, the level of the deviation between theoretical data and actual data.

An object that is composed of n components presents in the final assembled the sum of the inaccuracies of the individual components. All this is well known and carefully assessed, as a rule, during design and product development. However, as analysed and evaluated, it will result in each case a difference between the theoretical and the real 3D.

CFD - AERO - FLUID DYNAMIC ANALYSIS OF BEHAVIOURIn some sectors, more than in others, to know this difference can mean winning or losing a challenge. In the world of racing, for example, it becomes a fundamental tool in order to analyse the real aerodynamic behaviour of the car. Here, where everything is taken to the limit, it is extremely important to know the real geometry, that goes wet from the air, because it means to know how to optimize the behaviour of the car in the race.Projects developed on cars for GP2, GP3, F3, Inshore boats for competitions as well as the optimization project of the handbike of Alex Zanardi, brought noticeable benefits to simulate the real object.

For example for the car GP3, it was possible to make a development of aerodynamic solutions through CFD analysis of the complete 3D survey of the car itself.The triangular surface mesh captured with scanning and allowed to introduce in the simulator CFDs the actual geometry of the vehicle.This starting point is a very important and a fundamental for the result, because it has been found that the differences between the theoretical model and the real were important.

Another fundamental aspect to obtain a CFD simulation, that is very adherent to the real behaviour of the car, which has been made possible using specific scanning equipment, was the possibility to dispose for the simulation of particular available only through the 3D relief. The Comet 5 11M, a scanning system with fringes projection by 11 megapixel, has in fact enabled, with its 11 millions of points for shooting, to capture ‘imperfections’ on the surface as for example the weave of the composites, the rivets, the coupling profiles (flush and gap).This information is very important and decisive to develop aerodynamic solutions of winning in the various conditions of loading and trim.

The simulation of the real figure represents a path to be made where the product does not meet expectations in terms of performance or behaviour.3D scanning of a hull of 45mt made it possible to have a dimensional control first, then the CFD simulation, identifying the reasons for which the boat could not reach the expectations of speed.Always through the 3D scanning of some ‘equal’ hulls of 18 feet, it was possible to identify the reasons for which these hulls showed different behaviours in terms of drivability.

TESTING In the automotive industry, testing laboratories, through crash tests or tests destructive impact, with rigorous methodologies established by guideline, occur and classify today’s levels of vehicle safety.3D scanning in this area is the tool adopted to validate the simulation model of deformation in comparison to the deformation undergone by the real component (and precisely detected by 3D scanning), as well as the tool used to measure the actual deformation of the vehicle in the tests d’impact for certification.

Fig. CM 1 – Definition of the virtual model from 3D scanning


SIMULATION IN THE PRODUCTION PROCESS Casting simulation in the field of foundry, die casting, plastic moulding is validated by the conformity of the simulation to the real object detected.In fact, it is necessary to set a loop where the 3D scan allows to measure through a dimensional control the effects of the simulation on the product actually made and to refine inversely the process parameter.In some cases, however, 3D scanning can be the starting point to study the production process with reverse engineering for the realisation of the equipment needed for production.The detailed scan of the actual object makes it possible to reveal the technology, the technique and the location of the equipment used for its production.In foundry, for example, with Comet L3D 5M it is possible to reconstruct the path that led to its production, starting from the core and optimizing the process based on the findings on the actual product. MAGMA is a worldwide leading developer and supplier of software for casting process simulation. As a powerful simulation tool, MAGMA C+M turns core production into a calculable and optimizable activity. Core shooting, core curing and the thermal control of core boxes can be analysed to find the best solution.MAGMA C+M supports core and mould manufacturers by offering process transparency, feasibility evaluation and rapid construction of core shooting tools.

The software is a design tool that allows to simulate the automatic moulding of foundry cores fully, from the shooting to the curing phase according to the various technologies available today. It offers an easy way to expand the knowledge of the process making it transparent. It was therefore possible to import directly the scanning of a core into the three-dimensional modeller MAGMA5 and define the configuration of shooting and curing directly on the imported geometry (Figure CM1)With version 5.3 the optimizer is integrated into the standard simulation environment of MAGMA5, so it is possible to vary the process parameters and evaluate the consequent effects on the final quality of the product, in an easy way.Trying some typical configuration for the chosen binder technology (Croning -Shell Moulding) the process parameters that allow to produce the core, both for shooting and subsequent curing, was traced automatically. It was therefore possible to understand the great importance of the shape of the shoot head from bottom on the filling of the cavity at the same shootting pressure (Figure CM2), as well as identify the causes of the burns observed on the scanned sample in function of the position and power of the heaters that control the temperature of the core box.

CONCLUSIONSVirtual experimentation today is an irreplaceable and sometimes essential tool for industry as it allows us to:• reduce the “time to market”• improve products through quick comparison with alternative

solutions• build a network with subcontractors, possible through the

simulation path, that allows you to experiment in a single project also just sequential stages, joining then at the end all the specific skills

• start the design according to the logic of “concurrent engineering”

• share the product with the customers and also the stages of their realization

• optimizing the production process, minimizing the risks of erroneous technical choices and avoiding neglecting aspects that could have significant impact in terms of cost and time

• enhance and develop their design know-how In these situations, the 3D scanning allows to validate the simulation model through the measure of reality, of the real object.

Andrea Pasqualetto, QFPFor more information:Nicola Gramegna, EnginSoft - [email protected]

Fig. CM 2 - Shooting tests with shoot head from bottom

Fig. CM 3 – Temperature distributions on the core box and core curing

Calibration of Material Models for the Numerical Simulation of Aluminium Foams – MAT 154 for M-PORE Foams @ 3 Loads1. IntroductionThe current activity is focused on calibration of the LS-DYNA MAT 154 material model with respect to M-PORE aluminium foams specimens submitted to three different loading conditions: uniaxial compression test, 3 points bending test and Charpy pendulum test. The final goal is to look for MAT 154 parameters values able to satisfy contemporaneously the experimental behavior of quasi static and dynamic tests. The calibration is performed both according to literature practices and according to an approach designed to highlight the relationships between the investigated phenomena without a priori constraints.The calibration procedure exploits a methodology based on LS-DYNA – modeFRONTIER coupling.

2. Calibration Procedure AssessmentThe assessment of the calibration procedure has been established according to the below key points:• exploitation of a set of experimental tests able to

characterize the static and dynamic behavior of aluminium foam samples. A step-by-step approach has been preferred by selecting a reduced number of experimental results with the intention to get confidence with material model calibration;

• set up of accurate and robust FE models for the selected experimental tests;

• building up of the tool dedicated to manage the numerical analyses (modeFRONTIER workflow);

• definition of the calibration strategy coupled with a multi-objective optimization.

2.1 Experimental TestsTo characterize the static and dynamic behavior of aluminium foam samples, a suitable set of experimental tests have been executed. In the current activity, the behavior of M-PORE aluminium foam samples with a 45 ppi density loaded along the foaming direction has been studied. The experimental tests are the following:

• uniaxial compression test vs. 40x40x10 mm specimen• 3 points bending test vs. 10x10x100 mm specimen• Charpy pendulum vs. 10x10x100 mm specimen

The Table 1 sketches the experimental tests with their basic informations and typical experimental curves.

Table 1 - Experimental tests


2.2 Numerical Set UpThe building of the finite element (FE) models has been done to get accurate results within reasonable computation times, especially in the perspective that a calibration process has to be carried out. Accordingly, an initial FEM robustness investigation has been performed with respect to three numerical models at the purpose to evaluate their behavior in terms of mesh element size, element formulation, loading rate, and further to realize how contacts policy affects the numerical performances. The development of the three FE models focuses on aluminium samples modeled after a first evaluation of the equipment numerical modeling, as explained below. All the models have been built using the [N, mm, tons, sec] unit system.

Assessment of Equipment Numerical ModelingThe models set up started from an assessment of the experimental equipment, that is a preliminary assessment of the test apparatus numerical modeling has been carried out by experimental tests accomplished on a simple and well-known material (i.e. homogeneous aluminium). Being aware of a standard material model parameters, it is possible to check if the numerical “casing” is accurate or induces an initial delta able to mislead the following results. The assessment has been performed for the 3 points bending and Charpy pendulum tests by using MAT_003 (PLASTIC_KINEMATIC). The numerical-experimental matching was quite good.

Material ModelThe MAT 154 (MAT_DESHPANDE_FLECK_FOAM) has been chosen because the material model is well suited for uniaxial compression loading but is still not well established for different loading conditions. Further, the above three experimental conditions can be supported by the constitutive equations embedded into the material model. Details of constitutive modeling are available in literature, where here is just recalled that the plastic flow of metal foams is related not only to the elastic shear energy but to the elastic volumetric energy as well. For this reason in MAT 154 the hydrostatic stress is embedded into the equivalent yield stress sas given by:

(1)where: σVM = Von Mises stress σm = mean stress (hydrostatic pressure) α = shape factor. It defines the shape of yield surface and

can be expressed in terms of the plastic Poisson ratio vp: (2) The values of α2 needs to belong to the range [0,4.5], otherwise

α is physical meaningless. The 0 value corresponds to the Von Mises criterion, while 4.5 means that lateral plastic deformation does not exist in uniaxial compression test. The last case vp = 0 is esteemed to be the usual condition for aluminium foams.

The yield stress function σy takes the evolution of yield surface Ф = σ - σy into account because it is the sum of the initial compressive yield stress and the strain hardening:

(3)where: σp = foam plateau stress (initial compressive yield stress) R(ε) = strain hardening ε = equivalent true strain εD = densification strain (true compaction strain). It is

theoretically the strain limit at which the foam density equals the density of base material. It can be expressed in function of the shape factor and loading case. For uniaxial compression loading:

(4) α2,γ,β = material parameters.The parameters of equation (3) can be calibrated from uniaxial compression tests. As referred in the section dedicated to the calibration strategy, the parameters σp,εD,α2,γ,β have been investigated according to different approaches.

Being required for a crushable foam model a fracture criterion at the goal to provide enough accurate results, especially in uniaxial tension, shear and flexural tests for which fracture do occur, in MAT 154 the fracture criterion assumes that elements are removed when the critical value of volumetric strain CFAIL is reached.

Uniaxial Compression FE ModelThe uniaxial compression test has been modeled according the following set up (main info): • aluminium foam sample by a 40x40x10 mm parallelepiped of

1024 solid elements with element formulation equal to 2 (fully integrated S/R solid).

• loading condition provided by a BOUNDARY_PRESCRIBED_MOTION_RIGID. A sinusoidal velocity curve has been exploited so that a zero initial acceleration was attained. By using a velocity of 0.01 mm/sec (quasi static condition), explicit FE model requires too much hours of CPU time to simulate all the compression test. Since the calibration phase needs to run a lot of analyses, CPU time has been dramatically reduced by augmenting the velocity according the scale factor SF.

The described FE model (referred here as Cal_Mod) represents the configuration applied during the calibration phase. At the only purpose to check the goodness of a reduced number of calibrated configurations, a more accurate FE model has been developed (Ch_Mod), whose corresponding CPU time is greater than 7 hours. Even if the Cal_Mod is less stable than Ch_Mod, the Force-Displacement curves are almost the same, and, above all, Cal_Mod is able to provide the true tendency of the compression behavior versus the examined free parameters.

3 Points Bending FE ModelThe 3 Points bending test has been modeled according the following set up (main info): • aluminium foam sample by a 10x10x100 mm bar of 5120

solid elements with element formulation equal to 1 (constant stress solid element)

• loading condition provided by a BOUNDARY_PRESCRIBED_MOTION_RIGID. As done for compression, the sinusoidal velocity curve has been scaled setting a suitable SF, so that a ca. 1 min 30 sec CPU time is required for a 0.06 sec simulation time.

• As for the compression case, the above calibration model Cal_Mod has been used for the massive computation phase, while a refined model Ch_Mod to check the goodness of a reduced number of configurations (Ch_Mod CPU time is greater than 10 hours).

Charpy Pendulum FE ModelThe Charpy Pendulum test has been modeled according the following set up: • aluminium foam sample by a 10x10x100

mm bar of 5120 solid elements with element formulation equal to 2 (fully integrated S/R solid).

• loading condition, embedded in PART_INERTIA, is given by an initial velocity=2.20 m/sec with a mass=3.44 kg (initial energy is equal to 8.333 J).

In the current case, a unique FE model (Cal_Mod) has been developed since it provides at the same time good accuracy of results and reasonable CPU time (ca. 1 min CPU time is required for a 0.01 sec simulation time). The numerical arrangement of the three models is reported in Table 2.

2.3 modeFRONTIER workflowThe material model calibration requires to perform a fitting between the numerical data and the experimental ones, basically the Force-Displacements curves for uniaxial compression and 3 point bending tests, and

the Force-Time curves for the Charpy test. Generally speaking that is a multi-objective problem since several objective functions can be defined for pursuing the best compromises amongst the different matchings. The multi-objective problem has been faced by using modeFRONTIER, by integrating all the three FE models and then performing the optimization process. The outcoming results have been evaluated by using the multivariate data analysis (i.e. data depending on multiple variables) tools available in modeFRONTIER. The workflow depicted in Figure 2 shows the process integration of the three FE models.

Once the MAT 154 free parameters are updated, their values are written into the LS-DYNA input models (.dyn files) and then the batch runs of the numerical models are executed. In the following also the post processing operations are performed in batch modality within the LS-PrePost environment. The curves numerical-experimental fitting has been implemented by splitting the curves into their main regions so as to get a local insight of the input-output correlations. Accordingly multiple objective functions have been defined for the three models, that is the 3 obj. functions mc_err1,2,3 for uniaxial compression test (UCT), as shown in Figure 1, and 2 obj. functions m3_err1,2 for 3 points and mch_err1,2 for Charpy tests.

2.4 Calibration and Optimization StrategyThe material model calibration has been performed both according to the literature practices, approach that is more correct from a theoretical point of view, and according to an independent methodology, aimed to assess the relationships between the investigated phenomena without a priori constraints. Basically the two approaches have been designed as in the following:

a) the “Standard Methodology” exploits the uniaxial compression test to calibrate the parameters embedded into equation (2), that is , but also to investigate the MAT 154 remaining parameters. Later, the calibrated material model is used for 3 points bending and Charpy pendulum simulations and discrepancies are evaluatedb) the “Experiments Driven Methodology” exploits contemporaneously for calibration all three experimental tests

Table 2 - Cal_Mod and Ch_Mod FE models

Figure 3 - Objective functions for UCT

Figure 2 - modeFRONTIER workflow


with the intention to identify the parameters values providing the best trade-offs between all configurations.

Even if different studies advise correlations amongst the “free” parameters (see Material Model description), the present study did not exploit them, letting on the contrary the calibration driven only by numerical/experimental fitting at the goal to explore the design space in an extensive way and looking for different values arrangements. Starting form literature data, suitable input dominions have been set.

The optimization strategy has been designed taking into account the requirement that all the analyses were multi-objective. The genetic algorithm MOGA-II (Multi Objective Genetic Algorithm) coupled with a SOBOL DOE (Design Of Experiment) has usually been exploited for the optimization process.

3. ResultsThe results here presented are focused on identification of suitable material parameters values for the calibration of the FE models for the three investigated loading conditions. Additionally, how the free parameters dominions affects the numerical/experimental fittings is shown.

3.1 Standard Methodology1st OptimizationThe analyses performed by using the FE model of uniaxial compression test can be straightaway evaluated by plotting into a 3D bubbles chart the three objective functions (OFs) mc_err1,2,3 (i.e. the least squares minimizations associated to the three typical regions produced by such test), where the best configurations should belong to left lower zone, see upper left part of Figure 3. At the purpose to get a deep insight of the correlations between input and output, the 3D bubbles chart has been combined with a vector plot chart where the numerical Force-Displacement curves are superimposed to the experimental one (black curve), and with a parallel coordinates chart where input and/or output values can be filtered at the goal to assess how they affect each other, see the whole Figure 3. Filtering out alternatively the greater values of mc_err1, mc_err2, mc_err3, as shown sequentially in Figure 4-Figure 6, the more suitable input dominions for the three OFs appear (framed in red). For the input labeled with “ys” (referred to yield stress function parameters), the charts are in agreement with the structure of the equation (3) and with literature data, that is mc_err1 is strictly dependent on σp, mc_err2 by γ, εD and β as well, while for mc_err3 some influence of α2 can be noted too.

For the remaining parameters, α is pushed to assume its maximum value (2.121), while the Poisson ratio, within the selected dominion, “jumps” between opposite values in respect of mc_err1 and mc_err2 minimizations, being on the contrary substantially indifferent for mc_err3 minimization. A sensitivity analysis relied on a factorial DOE,

Figure 3 - Integration of 3D bubbles, vector plot and parallel coordinates charts for uniaxial compression

Figure 4 - Uniaxial compression – filtering out mc_err1

Figure 5 Uniaxial compression – filtering out mc_err2

Figure 6 - Uniaxial compression – filtering out mc_err3

here not reported for sake of shortness, confirmed the previous results. Taking into account these conflicting trends between mc_err1 and mc_err2, the reason of their trade-off comes out.To be sure about the appropriate selection of the input dominions, the convergence values of the free parameters have been searched for, and the input dominions updated as in the next section.

2nd OptimizationThe second optimization relying on new dominions provides the results depicted in Figure 7 where only the Pareto designs (i.e. the best trade-off between all the objective functions) related to 1st and 2nd Optimization are given. In particular, a new subset of configurations, located in the left lower zone and framed in red, is being generated and the main parameter involving this performance is γ. A further observation concerns α and σp values: at the purpose to get Pareto configurations it is mandatory to set α=2.121 and σp= 2.5 MPa.

Looking just for the most fitting curves belonging to the red framed group, it comes out that β parameter is peaked around an average value of 7, while εD and α2 are nearly spread all over their dominion.Eventually, amongst the Pareto designs, the des 7148 has been selected. Even if its goodness for UCT is very high (both for Cal_Mod and Ch_Mod one), it is almost poor in terms of 3 points bending and Charpy pendulum tests. As reported in literature, possible reasons of discrepancy can be due to cell size effects and loadings not uniformly distributed, involving stress concentrations at specimen supports.

The further step was to look for a better balance between all 3 models and at this scope a multi-objective optimization involving all three experimental tests has been performed.

3.2 Experiments Driven MethodologyOnce the modeFRONTIER workflow has been updated in respect of free parameters dominion (modified taking into account the results coming from previous analyses), the optimization process has been addressed at the aim to minimize the three global OFs (labelled with tg_) representing the fitting of three experimental curves.Getting to the point, the new candidate solutions have been selected firstly looking for the best for every single FE model then searching for a suitable compromise among them, where the last ones, i.e. des 372 and 1387, are depicted in Figure 8. The main result coming from activity is that MAT 154 requires different inputs arrangements for the 3 experimental tests being not possible, from

physical point of view, to get univocal values for all cases. In particular the most demanding dominions are required for 3 point bending and, especially, Charpy tests. It comes out that MAT154 is characterized by some limits in terms of flexural behavior modelling.

4. ConclusionsA flexible and efficient methodology has been applied for MAT 154 material model calibration of M-PORE open cells aluminium foam at three different loading conditions.

Relying on efficient FE models of simple and cheap experimental tests, whose behavior has to be matched, the methodology has been implemented into modeFRONTIER, that allows

an immediate LS-DYNA integration and a very intuitive and powerful data mining of the outgoing multi-objective optimization analyses.

Material calibration has been performed firstly with respect to only uniaxial compression test and later taking into account at same time the 3 points bending and Charpy pendulum tests as well. Accordingly, the most suitable material parameters values have been assessed and investigated in terms of their influence on the different tests. Further, not only single calibration values have been provided but also the dominions in which the parameter values are “reasonable” for better numerical/experimental fitting. The accomplished results highlight that MAT 154 is not able, for the current M-PORE open cells aluminium foams, to support efficiently the flexural behavior given by simple experimental tests.

AcknowledgementsThe experimental tests have been carried out by the Department of Innovation Engineering, University of Salento, Lecce, Italy.

Vito Primavera, Marco PerilloEnginSoft

Figure 7 - 2nd vs. 1st Optimization - Pareto designs

Figure 8 - Best trade-off solutions

55 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 54 Research and Technological TransferResearch and Technological Transfer

To increase the possibilities for humans to explore space, the development of bio-regenerative life support systems is essential: these systems must fulfil all human needs to sustain a sufficient living condition. In this contest, a greenhouse module is the fundamental part of every concept for a stable and independent base in future space missions. Indeed a greenhouse can (re-)generate essential resources for humans by closing different loops within a habitat, like waste water recycling, CO2 reduction, food and O2 production.Following this aim this feasibility study has been carried out to investigate and develop the characteristics of a greenhouse module for a future lunar base. This project has been coordinated by the German Aerospace Center (DLR) and it is embedded in the MELiSSA framework of ESA research projects.

To investigate and define the technical concept of a greenhouse module on the Moon connected to an already existing habitat, different subsystems have been identified: Design/ Structure of the Greenhouse, Air Management System, Plant Health Monitoring & Quality Assurance, Nutrient Delivery System, Illumination System and Thermal/ Power Control System. Each one of these subsystem has been simultaneously developed and then combined together.During this project EnginSoft, thanks to its experience in the field of Heat Ventilation and Air conditioning (HVAC), has worked on the Air Management System and on the interface with the habitat infrastructure. In this article we will focus on these aspects.

Baseline design of the Greenhouse ModuleLooking at the global structure of the greenhouse, the design evolution during the study has led to a baseline design which is composed by a rigid core module with four inflatable petals, two connections to the lunar base structure and a solar collector to enable a hybrid natural and artificial illumination system. A sintered regolith cover around the structure was envisioned to provide micro-meteoroid and radiation protection, as well as facilitate the thermal management of the greenhouse (Figure 1).The inner core offers space for a three-level configuration (Figure 2-Left). The lower level is the main connection point to the

Greenhouse Module for Space System

Figure 1 - Baseline design of the Greenhouse Module

habitat, where two independent and pressurized corridors lead to the habitat infrastructure. The lower level of the rigid core houses the nutrient delivery system, the thermal management system, and the power subsystem with batteries for redundancy. The middle level of the central core houses the data handling and control system, and several working desks. Furthermore, the astronauts can access the petals from this level. The lower section of the petal is reserved for storage and placement of additional subsystem components, such as inflatable tanks and pumps, as well as interface panels for resource flows (e.g. water, cooling fluid) to and from the core module (Figure 2-Right). The upper section of each petal is dedicated to plant growth and contains the plant support structure, illumination system equipment, ducts and tubing of air and thermal management system, as well as piping of nutrient delivery system.

The Air Management SystemAll petals are independent from each other and operate as separate growth chambers. In this way each plant is cultivated in specific petals with a dedicated climate. In fact each crop requires a specific environment (relative humidity, temperature and air composition) to guarantee a proper growth. This functional requirement implies the generation of four air loops, one for each petal (Figure 3).

Inside each growth chamber (petal), the air management system has to satisfy these main tasks:• Control pressure, temperature and air composition

(concentration of O2, CO2 and RH).• Recover water through condensation of moisture.• Air cleaning from chemical and biological contaminants. Following these objectives, the air management system can create adequate climate conditions for the plant wellness and can restore the primary resources (water recovery and air revitalization). This can be done using the following list of components: fans, UV-sterilizers, particle filters, cooling coils, heaters, sensor packages,

trace gas filters, control valves, humidifiers and CO2/Air injection systems. Mass and power consumption of each component for the air management system are estimated, including a margin, from industrial datasheets (Figure 4-Left).

The characteristics of the components are selected considering the amount of cultivated area of each petal. In particular thermal aspects are taken into account due to the heat that is dissipated into the air. This heat especially comes from the illumination system which

is linked to the production of plants. A maximum transpiration rate (one petal) of 740[l/day] is calculated: this amount of water has to be removed by the cooling coil. Both these aspects have led to a specific cooling coil with a refrigerant capacity of 40.6[kW], which can be used to condensate the water and reduce the air temperature. To guarantee the complete moisture removal, the air mass flow should be at least 8200[m3/h]. Axial fans are selected to provide a mass flow between 2.0 [m3/s] to 4.22 [m3/s]. An UV sterilizer is introduced for the removal of biological contaminants,

Figure 2 - Left: Core design. Right: Petal internal layout

Figure 3 - Air loop for each petal

Figure 4 - Left: mass and total power for the air management system. Right: scheme of air loop


while trace gas filter is used for chemical contaminants. A heat exchanger is positioned to adjust the air temperature, while a humidifier has to increase relative humidity of recirculated air. Finally the CO2 level inside the petal is eventually increased by adding pure CO2, stored in tanks. Also extra air can be injected, if there are leakages in petal’s inflatable structure.All components are mainly placed inside the core where the air is recirculated in a loop (Figure 4-Right, Figure 5-Right). Pipping and plenum for air distribution/ collection are located inside the petal (Figure 5-Left). The connections between each petal and the air management system, located in the core, are established by two supply ducts (air form the core to the petal) and one suction duct (air from the petal to the core). With respect to the duct system two approaches have been foreseen: within the core, rigid aluminium-based duct systems are implemented, while flexible and semi-rigid ducts within the petals. The deployable ducts consist of flexible material reinforced by a wire cable built in a spiral configuration to maintain structural integrity.

To manage the air exchange between the greenhouse module and the habitants a direct gas connection is selected introducing the following operating modes (Figure 6):• NOMINAL MODE. Each petal

has a dedicated air loop with a continuous air recirculation. Temperature, pressure and air composition is monitored and regulated. In this mode there is no air exchange with the habitat.

• BREATHING MODE. Each petal is directly connected to the habitat infrastructure. In this mode O2-rich air can flow from the petal to the habitats and CO2-rich air can move in the opposite direction simultaneously. The breathing mode can be activated for a single petal at a time.

When nominal mode is active, the air inside the petal is completely recirculated. During the nominal mode each petal is a separate growth chamber with a dedicated air management system to control the climate. For each petal the sensor packages provide an instant feedback to the air management system in order to reach the desired temperature, pressure, humidity, and CO2 level. Due to

the complete recirculation in nominal mode, the O2 level can’t be continuously regulated, but it will be maintained in a predefined range, form 21% to 26.5%. When the maximum O2 level is reached, the nominal mode is modified into the breathing mode. In the breathing mode, through a set of regulation valves that are partially or completely open, the O2-rich air can flow from the petal to the habitat and the CO2-rich air can flow in the opposite direction (from habitat to petal). During the breathing mode the O2 level inside the petal decreases till the minimum value is reached and the nominal mode can start again. For example, to increase the level of O2 from 21% to 23% (nominal mode) inside the petal (with 2.85[kg/day] of O2 plant production rate) it takes approximately 90[h] (with lights on). So the breathing mode can be activated at least every 5-6 day to decrease the level of oxygen, moving it back to 21% of O2 level. The breathing mode itself lasts only several minutes, considering

Figure 5 - Left: air flow path inside the petal. Right: air ducts and distribution channels

Figure 6 - Left: nominal mode. Right: breathing mode

Figure 7 - Left: air flow path inside the petal. Right: CFD model of the petal

no extreme recirculation effects within the petal. With a mass flow of 8200[m³/h] and a petal volume of 618[m³] the total gas exchange would be around 4-5 minutes.

CFD analyses of a growth chamberCFD models of a growth chamber has been created to check internal climate of the petals: these simulations has allowed to calculate pressure, velocity, temperature, local concentration of O2/CO2 and humidity level.

When the nominal mode is active the air is treated in the core and then is blown down, through the supply ducts, into two plenum placed under the floor of the petal (Figure 7-Left). The air inside the (inflatable) plenum is pressurized by fans that are located just before the plenum entrance. Moreover the top surface of the plenum is perforated with circular holes (grid). The open area of the holes is the 40% of the total plenum area: this solution guarantees an uniform air distribution to the different parts of the floor. The air inside the petal is finally collected from a suction duct located at the top of the chamber.A geometrical model of the plants has been introduced to include the principal characteristics of the plants in the CFD analysis (Figure 7-Right). Plants are represented as boxes, taking into account the maximum volume that they can occupy respect to their growth. From a fluid-dynamic point of view, an isotropic pressure resistance is applied on boxes, considering that 85% of the total volume of boxes is occupied by plants.Plants are also sources/sinks of O2, CO2 and H2O according to the crop characteristic. In this case the cultivated crop is potato with O2 source =2.85[kg/day], CO2 sink =3.8[kg/day] and H2O source =740[l/day].

Plants are also sinks of energy: considering the quantity of evaporated water (740[l/day]) and a constant enthalpy of evaporation (2444[kJ/kg]), the total power absorbed is 35.8[W]. The heat due to lights (and other electrical equipments) is about 53[kW] and it is distributed uniformly in the part of petal volume that is illuminated by led panels.The flow is considered in a steady condition and in a turbulent state. The energy equation is solved, but no radiation is simulated. The reference pressure is set to the atmospheric pressure. The total mass flow at the two inlets is 8200[m3/h], the temperature is 20[°C] and the inlet air composition is fixed (O2= 21%, CO2= 0.12% and RH=70%). On the outlet a relative pressure of 0[Pa] is applied. Buoyancy effects are also taken into account using the gravity of the moon.The contour plots of the results show a velocity field in the range from 0[m/s] to 1[m/s] (Figure 8): no peak velocity values are present near plants. Higher velocity values are only present in the plenum for the supply air and around the suction duct (Figure 8). The total pressure loss between inlet and outlet is 130[Pa].The temperature inside the chamber remains in the range from 20[°C] to 30[°C], with a temperature of 26[°C] at the outlet duct (Figure 9). The highest

temperature values are present in the top part of the chamber where the heat is transported due to buoyancy effects. The change in the relative humidity is from 70% (Inlet temperature = 20[°C]) to 74% (Outlet temperature = 26[°C]). The temperature gradient remains below 10[°C], this should be compatible with the plant growth.Other CFD calculations have been performed with the same procedure considering bread wheat as cultivated crop: some critical aspects on thermal loads and temperature gradient have been pointed out and mitigation strategies have been proposed in this case.

ConclusionIn conclusion, during this project, a greenhouse module layout has been developed, defining all the necessary subsystems for the growth of plants. Relevant issues related to bio-regenerative systems are taken into account.In particular, the air management system has been designed with the selection and characterization of different components to control the climatic condition (velocity, temperature, air composition and relative humidity) inside each growth chamber.A direct interface with the lunar base has been defined for O2/CO2

exchange: a switch between two operating modes (nominal and breathing mode) is able to adjust the O2/CO2 levels inside the greenhouse.A CFD analysis has been finally performed to check the behaviour of the air management system from a fluid-dynamic point of view.

Lorenzo Bucchieri, Erik Mazzoleni - EnginSoft

Figure 8 - Velocity distribution on two section planes

Figure 9 - Temperature distribution on two section planes


RLW Navigator project A success storyDescriptionDuring the course of the past 3 and a half years, EnginSoft has had a team working on the RLW Navigator project funded by the European Commission under the ICT-Factories of the Future programme (http://www.rlwnavigator.eu/) along with many industrial and research partners from all over Europe and South Korea. The goal of the project was to develop an engineering platform for an emerging joining technology from the automotive industry, Remote Laser Welding (RLW), that could enable the exploitation of this technology and ultimately support other joining processes. The RLW Navigator accomplished this by integrating the universal simulation engine and experimental models to precisely model, configure, optimise and control process variation, production throughput and cost in multistage assembly processes. The RLW Navigator integrates manufacturing system design information (CAD/CAM) with statistical analysis/variation simulation thus helping to support the early introduction of new processes and fundamentally change these introductions from current trial-and-error to systematic math-based system and station configuration, optimisation and control.

EnginSoft RoleThe EnginSoft team within the RLW Navigator project consortium, had three distinctive activities, two of which were technical and focused on the study of the assembly layout as a whole, while the third had a managerial cut:• Developing a new stand-alone software for fast evaluation of

cost and feasibility of concept layouts, following feedback from STADCO, a project tier 1 industrial partner within the RLW Navigator consortium;

• Deploying optimization strategies and methods to be implemented together with Politecnico di Milano’s own analytic software for studying and evaluating assembly layout concepts;

• Planning, managing and deploying dissemination and exploitation activities.

All activities performed by our team proved over-successful, from both technical and managerial point of views.For example, the technology build from joining efforts with Politecnico di Milano and STADCO in developing a software suite able to guide the layout designer and cost evaluator to have at first a rough and fast idea of the performance of its designed layout, and to then offer him a more advanced optimization tool capable of highlighting bottlenecks of the design and to suggest further improvements, won this year’s “Best Application Paper Award” at the INCOM (IFAC Symposium onInformation Control in Manufacturing), in Ottawa, Canada.Furthermore, EnginSoft, in cooperation with the University of Warwick, laid out an innovative dissemination and exploitation strategy which had never before be implemented in EU co-financed projects, winning praises from the EU reviewers during the final review meeting of the project.

Layout Analysis SuiteAssembly Layout and Process EstimatorWithin the software suite developed, the first tool available to the user is the Assembly Layout and Process Estimator, which is a software developed by EnginSoft in partnership with one of the project industrial partner, STADCO. It is an integrated graphical environment where assembly line designers can perform an extremely fast evaluation of line concepts. The evaluation is made easy by monitoring the main design KPIs, which are computed taking into account user defined design constraints.

All aspects related to the preliminary line design are manageable within the same graphical environment. KPIs monitoring, Cost Drivers and Design Constraints definition, customizable and expandable database, together with fast and intuitive options for line layout population and task sequencing, offer a complete set of tools to shorten and ease everyday work.

Configuration OptimizationThe second tool available in the software suite is the Configuration Optimization: It capable of providing a unique framework for designing and optimizing hybrid assembly systems. Its core element is composed by two methods: an analytical performance evaluator and a Discrete Event System simulator. Both these elements provide a model of a generic hybrid assembly system, describing its dynamic behavior (machine failure, part flow, etc).These methods act as core evaluation modules for a multi-objective Figure 1 - Best Application Papaer Award won at INCOM 2015

optimization on costs and productivity, exploiting all modeFRONTIER capabilities. Data about reliability of machines and intermediate storages, come directly from the company’s databases, providing a direct integration between information and model. In addition, with Configuration Optimization it is possible to perform robustness and sensitivity analyses. The output of the optimization, is fully customizable within the context of modeFRONTIER.Configuration Optimization provides a quantitative system design procedure, enabling effective first-time-right designs. With this methodology, a company can provide the customer with a set of optimized solutions, tailored on customers’ needs: costs, productivity, floor occupation ecc. This method, provides valuable managerial insight for the organizations’ improvement, thanks to sensitivity and robustness analysis.

Dissemination & Exploitation Strategy DisseminationBoth Dissemination and exploitation of the project were very successful. Dissemination achieved and greatly exceeded the goals that were set at the beginning of the project. Over 50 peer reviewed publications have been published or submitted to date and prestigious journals accepted publications from the RLW Navigator Partners. Furthermore, the most influential conferences were targeted for dissemination and welcomed abstracts and presentations, often for several years in a row. A total of 8 prizes were won by RLW Navigator Partners, and among those 2 “best paper” were awarded to presentations of the project research. A detail of the prizes follows:1. SZTAKI 1st prize @ competition of the National

Scientific Students’ Committee at the Budapest University of Technology and Economics 2012 (Budapest, Hungary)

2. WARWICK 2012 WMG Department Prize for MSc dissertation (Warwick, UK)

3. WARWICK Ph.D. student Best Poster Award 2013 British Computer Society (BCS) University Challenge

4. SZTAKI 2nd prize @ conference of the National Scientific Students’ Committee 2013 (Budapest, Hungary)

5. WARWICK / POLIMI MSc Dissertation Award 2013 Italian Association of Manufacturing AITeM

6. WARWICK Best Paper Award at the 2014 International Joint Conference on Mechanical, Design Engineering and Advanced Manufacturing (Toulouse, France).

7. WARWICK / POLIMI MSc Dissertation Award UCIMU 2014 - Italian association for manufacturing, automation and ICT (Milan, Italy)

8. POLIMI / ENGINSOFT Best application paper award INCOM 2015 (Ottawa, Canada)

Presentations of project outcomes were also given at several relevant industrial fairs and exhibitions, and a number of symposia and workshops were organized and successfully carried out in Europe and South Korea.

ExploitationThe RLW Navigator consortium set ambitious objectives and benchmarks not only on the research, but most importantly on the exploitation of the results. An ambitious exploitation plan was set up and progressively refined throughout the three and half years project lifetime, in order to ensure that the benefits of European research will be consistently delivered to end users (Automotive OEMs and Tier One providers). The consortium was extremely effective in identifying, classifying and protecting the Intellectual Property generated from the project, at the same time devising an overall strategy and a detailed exploitation plan for the project results.

The overarching exploitation strategy proceeded through three main and coordinated steps:(1) Identification of Exploitable Results (ERs) Each Work Package (WP) has developed a number of results,

such as new methodologies, presented as technical deliverables. Some of these methodologies have been selected based on their novelty (IPR-related), relevance to industry, robustness in solving real industrial problems and other factors. We terms these methodologies as Exploitable Results (ER).

(2) Definition and selection of industry justified Exploitable Tools (ETs) The aforementioned ERs have been aggregated into industry justified

Exploitation Tools (ETs): this step allowed to carry out a preliminary market analysis by means of a series of industry workshops and other initiatives. This allowed to gather crucial information for example

Figure 2 - Assembly Layout and Process Estimator software developed by EnginSoft

Figure 3 - Configuration Optimization strategy developed by Politecnico di Milano

Figure 4 - Layout Analysis Suite result samples

61 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 60 Software UpdateResearch and Technological Transfer

for the further development of the Tools into commercial products (post-RLW navigator project). The ERs have been classified into the following categories:

(i) Process Feasibility and System Configuration

(ii) Workstation Design (based on ERs developed within WP2 and WP3);

(iii) Process Control (based on ERs developed within WP4);

(iv) Process Verification and Visualization (based on ERs developed within WP5 and WP6).

(3) Identification of suitable business strategies for exploitation and championing.

Leveraging the information received from the potential market, four different exploitation strategies have been identified: accordingly, the various Exploitable Tools have been aggregated into bundles that share the same strategy. For each different strategy, a Champion has been selected among Partners with the responsibility to progress exploitation, for example by coordinating the business planning process.

Francesco Micchetti, Giovanni BorziEnginSoft

Marcello Colledani, Andrea RattiPolitecnico di Milano

Disseminating the Music Project

The MUSIC project (MUlti-layers control&cognitive System to drive metal and plastic production line for Injected Components) has already completed its third year of activities and it is now time to intensify the dissemination of its achievements in terms of research and to present its technical results and demonstrators to a wide and expert audience, also taking advantage of three relevant events scheduled in the next months. The MUSIC consortium is committed to showing the Beta version of the MUSIC “Control&Cognitive System” (C&CS) in the Research Agorà of the International CAE Conference, on October 19th and 20th, also thanks to the support and sponsorship of Assomet Servizi srl. The C&CS, developed by EnginSoft S.p.A., will activate a quality control and cost efficiency loop in the high pressure die casting of light alloys (HDPC) and plastic injection molding (PIM) industry by introducing for the very first time an holistic approach to real time data monitoring, analysis and control of all the phases of the currently fragmented automated production lines. The two day event of the CAE Conference is the ideal context and venue to discuss the increasing relevance of ‘simulation based engineering and sciences’ and research and innovation as opportunities to be part of the future.http://music.eucoord.com/www.caeconference.com

On January 12th – 14th , MUSIC will be at the EUROGUSS International Trade Fair for Die Casting, to meet the wide public in the biggest trade fair for die casting in Europe, which has now reached its 11th edition.

The product spectrum shows innovative solutions for die casting processes like aluminium die casting, magnesium die casting or zinc die casting. The event also focuses on topics like rapid prototyping, die casting machinery, material testing and 3D printing. The MUSIC project in this context will count on the direct involvement of Oskar Frech GmbH + Co. KG, MOTUL Tech Baraldi Srl, IFAM Fraunhofer Institute, MAGMA GmbH, ELECTRONICS GmbH, REGLOPLAS AG so to create a sort of MUSIC tour, where to receive all project information, the “MUSIC Guide to key-parameters in High Pressure Die Casting” and get in touch with the project partners. The project is planning to participate in the parallel International Die Casting Conference with two different papers.https://www.euroguss.de/en/

The third great event will take place next June in Venice, that is the HTDC Conference 2016, a key-event for the international industrial and academic community involved in casting processes of Aluminium, Magnesium and other non-ferrous alloys. To sustain the development of the competitiveness in Die Casting production, HTDC 2016 will review several issues, strictly related also with the project, that will therefore sponsor some activities and at the same time present its achievement in terms of process control in one of the conference sessions. www.aimnet.it/htdc2016.htm

Figure 5 - RLW Navigator project toolkit schema

ANSYS R16.2: Structures and Multiphysics focus

IntroductionThe recent release of ANSYS 16.2 is the occasion to present the latest enhancements of the software and to point out some of the amazing new features implemented with the ANSYS 16.x family of products.Structures and Multiphysics enhancements will be focused on this document.

ANSYS MultiphysicThe new ANSYS AIM tool (Ansys Integrated Multiphysics) is the core of the new multiphysics capabilities available from 16.0 release.AIM is an integrated solution for 3D engineering simulation encompassing the breadth of ANSYS physics in a single, modern user environment. ANSYS AIM is built on the ANSYS Workbench Platform, and it is an evolution of the workbench strategy.AIM provides an immersive user environment that enables analysts to perform engineering simulations from start to finish in a single window user environment. Among the other features ANSYS AIM includes: geometry creation and preparation by means of ANSYS SpaceClaim, meshing for all physics, physics setup, solution and results processing for fluid flow, stress and vibration analysis, heat transfer, electric conduction, and fluid-structure interaction. AIM also provides design exploration enabling the evaluation of design alternatives and design optimization. ANSYS AIM provides a full suite of physics simulation in a unified, immersive user environment that is readily deployable across engineering organizations. In figure 1 the AIM interface is depicted.

All of the simulation capabilities of ANSYS AIM take advantage of parallel processing, from meshing generation to post processing.ANSYS AIM includes an integrated and context-sensitive help system that provides an overview of AIM concepts and workflows.The new fluids solver in ANSYS AIM is a cell-centered solver, which uses Fluent numeric as the foundation. However the new solver includes many improvements based on experience with both Fluent and CFX to enhance the robustness and convergence behavior of the solver. With regard to the structural applications the Mechanical APDL solver is the mechanical solver engine under ANSYS AIM.This product has benefited from the 16.2 releases in many fields. Among the new features is worthy to mention:

1. Conjugate Heat Transfer2. Compressible Flow3. Buoyancy4. Fluid Solution Robustness5. One-way Thermal FSI6. Nonlinear Contact7. Large Deflection8. Nonlinear Solution Controls9. Stress-Life Fatigue10. Solution Abort/Interrupt

As these new features clearly depict, from a purely linear solver, ANSYS AIM technology is rapidly turning to non-linear capabilities to capture the behavior of even sophisticated single physics solutions.

Ansys Structural enhancements: composite material analysisIn general composites are a great material alternative to drastically reduce weight in structural products. These materials are difficult Figure 1- ANSYS AIM interface: Multiple and multiphysics in a single window

Figure 2 – Composite beam model extraction

63 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 62 Software UpdateSoftware Update

to model due to non-homogeneous material properties and their exceptional dependence on the manufacturing process.ANSYS ACP (ANSYS composite PrePost) provides best in class technologies to model composite assemblies of even complex geometries.Some very useful enhancements have been implemented in ACP technology at ANSYS 16.2. Worthy to mention are:

1. Beam model extraction: In ANSYS 16.2 analyses of composite structures are speed up by Converting 3D Models into the Equivalent Beam Models. In Figure 2 this topic is depicted.

2. Composite Layer Definitions: in order to accelerate model setup a Flat Composite Layer can be projected onto a 3D CAD Geometry.

ANSYS 16 for fabricated structuresANSYS R16 brings to high efficiency the modeling of fabricated structures to the same level of 3D modeling. ANSYS R16 offers a new highly automated process that leverages faster transfer from geometry, new meshing algorithms, parallel meshing and efficient mesh connections. Sheet metal is a common “traditional” material used in many industries to minimize material and weight while delivering required performance. Structures made of slender plates are greatly used in industrial equipment (casings, crane, etc.) as well as in transportation applications, like trains or ships.These slender plates engage a different set of challenges for meshing than 3D structures. An effective workflow is required to manage a large number of bodies in an assembly and allow the user to create a meshed model in an automated way.With ANSYS 16 the user can automatically create a simulation model from a complex assembly in minutes.Typical workflow takes involves:

• Prepare the geometry, extracting midsurface or top/bottom surface.

• Imprint geometry where needed, especially at T-junction.• Add welds to join parts (optional).• Mesh all parts of the model Individually, leveraging parallel

meshing for high performance.• Connect parts after the initial individual meshes have been

created.• Check quality and manually edit the mesh if needed.• Apply loads and boundary conditions.• Solve the model.• Review the results.• Perform automated design variations by modifying geometry

dimensions, physical properties or loads or boundary conditions.

This process benefits from the new capability that allows user to connect mesh of different part directly in ANSYS Mechanical. This way the “nummrg” APDL ANSYS technology is exposed to ANSYS Workbench.As already mention mesh process benefits from the parallel meshing process that speed up the element and node model creation. The

parallel mesh generation is implemented in any ANSYS product and doesn’t need any special license.In ANSYS R16 element quality can be displayed on the mesh. After reviewing mesh quality plot, “Move node” option allows for local adaption of node location to manually improve mesh. In Figure 3 mesh connection and mesh quality are depicted.As fabricated structures can be very large – think for example of an offshore platform —sub-assemblies for a more efficient setup are needed. This way is possible to assemble all the various parts to create a model of the full structure.In Figure 4 a model assembly is depicted with the details of the mechanical interface.

Nonlinear Adaptive Meshing in ANSYS R16 Nonlinear adaptive meshing is a new feature in ANSYS R16. When activated, this capability refines the mesh automatically during the course of the nonlinear solution based on criteria set by the user. Adaptive remeshing is supported in ANSYS WB Mechanical interface for static structural analysis. Multiple non-linear adaptive regions allow user to define different criteria simultaneously.Three available criteria are available:

• Contact based: elements that come into contact with specified contact regions will be refined.

• Strain-energy-based: if the strain energy of an element exceeds a threshold its mesh is refined.

• Position-based: elements that come into a specified box are refined.

For more information:Roberto Gonella, [email protected]

Figure 3 – Fabricated model is depicted before and after mesh connection option (left). New tool displays mesh quality on elements (right).

Figure 4 – Model assembly in ANSYS WB. Three models are grouped together without losing the original features (connections, components, etc.)

The first part of this article, which was published in the previous issue of this Newsletter, we can see that there is a variety of fluid systems that can be studied using a system level approach. This class of problem is a broad one, ranging from large scale distribution systems down to distribution, lubrication and ventilation systems of all types and size. We also saw that real life components and fluids can be modelled in this approach by defining their characteristics.In what follows, we will examine in depth the topic of System Level Thermo-Fluid Analysis, speaking about pressure drops in a system and about time.

Pressure Drop: Paying Your DuesIf there is no pressure differential in a system, there is no flow. Analogously, a ball placed on a flat surface won’t roll anywhere without an external impetus, usually in the form of your booted foot. A pump or compressor performs much the same task in the world of fluids, though perhaps in a slightly different manner. The mechanism is more akin to tilting the surface the ball is resting on. The gradient of this slope is the pressure difference required to achieve a given flow rate.This pressure difference is effectively a toll that must be paid to the laws of thermodynamics. The fee is paid in energy lost variously to friction, heat and noise throughout the system. Therefore the trick is to design the system such that this fee is kept as low as possible.Pressure losses occur due to two broad effects: the first is frictional pressure drop. Moving a fluid through a component is conceptually no different from pushing a block of any material along a surface; the smoother the surfaces in contact, the faster the pace that can be achieved for a given energy input. Thus a smoother pipe interior makes for lower frictional pressure losses.The second element is a little more complex and while the details vary slightly depending upon the exact geometry in question, the principle is the same for all components. To illustrate the effect of secondary flows, let us consider a simple bend. The purpose of a

bend is to redirect the flow by a given angle. As the flow is forced on to its new course, the velocity distribution across the width of the pipe changes, becoming more asymmetrical with the flow toward the outside of the bend accelerating relative to that on the inside of the bend. This clearly isn’t the normal state of affairs for a straight length of pipe, which in turn means that the flow must expend energy reverting back to its equilibrium state after it exits the pipe (see Figure 1). This can take as much as 50-60 pipe diameters downstream of the bend to achieve, meaning that the influence of a simple bend exists well beyond the physical extent of the component itself.In conclusion, the pressure drop generated by a particular component is the sum of the frictional loss through that component and the energy expended by the flow to revert back to its equilibrium state downstream of it. If these two phenomena are accurately measured in the laboratory, the pressure drop associated with that component can be captured. This is significant as this value for pressure loss captures inherently three-dimensional effects. If the experiment is extended over a range of

What is System Level Thermo-Fluid Analysis?

Figure 1 - Flow through a 90° bend; note downstream secondary flows

Newsletter EnginSoft Year 12 n°3 - 64 Software Update

flow rates, it is possible to build up a model of that component in two dimensions: pressure drop vs. flow rate. This map can then be used to form the basis of a model of that component which can be used in system simulation. But the news gets better again: converting the ordinates of the map to non-dimensional values (Reynolds number vs. loss coefficient) makes the map applicable to any fluid within that Reynolds number range.The key, of course, is to ensure that the source used for the pressure drop measurements is a consistent and accurate one.However, the more engaged reader may have noted an issue: where components are closely spaced, the downstream pressure recovery described above may be affected. What happens in these cases? The short answer is that these component interaction effects (as they are generally known) must be accounted for either by further experimental measurements (often captured as interaction coefficients), or by three-dimensional CFD. This is in fact a good example of where 3D CFD can be used to effectively support system simulation; in this case by providing an understanding of the likely magnitude of component interaction effects for non-standard combinations.In summary, the overall loss of a system may not simply be the cumulative loss of each individual component, but it is entirely possible to account for the effect of component interaction in system level simulation.

Time: The Fourth DimensionHaving established that system simulation does in fact simulate a three-dimensional reality (it is hoped that it is by now clear that the oft used term 1D simulation isn’t an accurate description!) it’s time to consider the temporal dimension.Flow rates and pressures in systems can change for many reasons; it may be a basic design or operational requirement, or the result of planned equipment shutdowns or even unplanned failures at either a component or process control level. Whatever the root cause, it’s important to understand the response of the system for a number of reasons. At the most basic level, it’s important to know that the system is capable of operating across the required range. It may also be important to understand the time taken to achieve a given state, such as when filling or emptying tanks. These may even be safety critical issues. For example, imagine a situation where an aircraft fuel system is unable to move fuel between tanks and to the engines at a rate appropriate to the required demand.Rapid changes in flow rate, often caused by component failure or poorly designed shut down procedures, can even create severe surges in pressure. These surges can result in failures in the pipeline and so pose a significant risk to both personnel and equipment.Accounting for such surges often generically termed water hammer, requires that the system simulation tool be able to simulate and track pressure waves throughout the system. These high pressure peaks traverse through the system at the speed of sound for that fluid, reflecting off components such as pumps and closed or partially closed valves. A related phenomenon is that of the vapor cavity. Just as water hammer generates pressure peaks, it also creates pressure troughs. If the pressure drops below the vapor pressure of the fluid, a pocket of vapor can form (see figure 2 for an example). This can create two major hazards: if the vapor pressure of the fluid is less than the pressure on the outside of the pipe there is a real risk of pipes being crushed by this differential and imploding. This isn’t an unlikely consequence; the vapor pressure of water at 20°C is only 2% of that of the atmosphere. Conversely, even if the pipe walls can withstand the pressure differential, the

subsequent pressure peak generated as a result of the cavity collapsing can be extremely severe.In summary, thermo-fluid simulation that accounts for changes in system parameters with time can be essential for understanding both routine operation and safety critical events.

Who, What, Why, Where and When?Up to now it has been shown how a system level approach to system simulation captures three-dimensional loss data in a two-dimensional performance map. This method of working captures all the required loss information for a component, without the computational overhead associated with a three-dimensional geometry based approach. This in turn makes it feasible to simulate entire systems in both steady state and transient.The combination of low computational overhead and broad applicability, combined with the ubiquity of fluid systems, really does mean that thermo-fluid system simulation has a place across virtually all industries. System simulation has been used to assess and understand space vehicle propulsion, process plant operation, vehicle cooling, fire fighting water mains, gas pipelines, surge protection, lubrication of industrial and small scale transmission systems, aircraft cabin ventilation, submarine ballast tanks…it’s an almost endless list.Perhaps it’s more interesting to consider when and why it might be used. At the concept design stage, when available system information is by definition sparse, simulation can be used to indicate likely system performance. Operating in a virtual environment also means that evaluation of different or novel approaches can be achieved easily and quickly. As system details emerge and more information becomes available, the system model can be adapted to reflect this. The system model evolves with the project which translates to fewer design iterations and a reduction in the number of physical prototypes required.In the final analysis, the versatility of thermo-fluid system simulation means that its applicability extends to any point of the design and operation life cycle.

This article is a reprint of the original published by Mentor Graphics: http://bit.ly/YC5Qec

For more information:Alberto Deponti, EnginSoft - [email protected]

Perché rischiarepesanti sanzioni?

I sistemi anti-pirateria oggi consentono di geolocalizzare le

installazioni irregolari e forniscono tutti i dettagli

sul loro impiego.

Sei certo di voler rimanerenell’illegalità?

EnginSoft ti può aiutare a rientrare nella legalitàcon discrezione, attraverso percorsie strumenti dedicati.

www.enginsoft.it | [email protected] 2 - Pressure surge (Water Hammer) in a rising main due to a valve closure

Newsletter EnginSoft Year 12 n°3 - 66 67 - Newsletter EnginSoft Year 12 n°3 Hardware Update Hardware Update

for us to have tests carried out by Enginsoft as this firm operates very closely within manufacturing and automotive environments. The results show that we can really make an impact in terms of performance and TCO if we implement accelerated ARMv8 cluster with high speed networks.”

“The combination of a high performance, low power, cost-effective X-Gene ARM processor coupled with leading edge GPU compute enables excellent performance at compelling price points,” said John Williams, vice president of marketing at AppliedMicro. “AppliedMicro is excited to see continued validation of the value of the X-Gene processor in E4’s High Performance Computing platforms.”

The tests results were presented at Ter@tech forum, on June 23th, in Paris in the paper: Scalability & performances of general purpose fluid dynamics solver on low power cluster: new perspectives on combined CUDA-ARM architecture. EnginSoft is involved in H2020 Exascale projects for testing engineering codes on these new platforms and to improve efficiency on such kind of architecture.

About E4 Computer EngineeringEstablished in 2002, E4 Computer Engineering designs and manufactures high performance systems which aim to accomplish both industrial and scientific research requirements and to reach a variety of customers ranging from universities to computing centers. E4’s focus is on HPC although our expertise extends to all segments of IT. Thanks to the established experience and the outstanding quality of our solutions, E4 Computer Engineering is acknowledged and appreciated as a valuable technology vendor by prestigious organizations like C.E.R.N.in Geneva. E4 designs each system individually to deliver highly personalized, cost effective and power saving solutions. www.e4company.com

About AppliedMicroApplied Micro Circuits Corporation (NASDAQ: AMCC) is a global leader in computing and connectivity solutions for next-generation cloud infrastructure and data centers. AppliedMicro delivers silicon solutions that dramatically lower total cost of ownership. Corporate headquarters are located in Sunnyvale, California. www.apm.com

SANTA CLARA, Calif., Sept. 8, 2015 (GLOBE NEWSWIRE) – E4 Computer Engineering, the leading Italian manufacturer of High Performance Computing (HPC) and enterprise hardware solutions, and AppliedMicro (Nasdaq:AMCC), a global leader in computing and connectivity solutions, announced today the results of tests carried out independently on a GPU cluster provided to EnginSoft Italy, a premier global consultancy firm in the field of Simulation Based Engineering Science (SBES). The objective of these tests was to get an independent perspective from a large company working heavily on 3D simulations to determine if the performance and power-efficiency of the X-Gene-based ARM processor is suitable for Computational Fluid Dynamics (CFD) applications. Each GPU cluster provided by E4 incorporated nodes hosting one GPU, an 8-core, 2.4 GHz 64-bit X-Gene server SoC and a low latency network to enable comparative tests with standard x86-based clusters. The tests provided EnginSoft the opportunity to utilize their experience in CFD and HPC and examine the first production ARM64 processor architecture coupled with a GPU in a real cluster, connected via a FDR infiniband, and enabled E4 Computer Engineering and AppliedMicro to have independent tests on the new E4 system. To test the real capabilities of the cluster in an engineering environment, EnginSoft performed a complete porting of CFD code with a computational engine based on a GPU in order to prove scalability with low power, cost-effective CPUs. The CFD code used for tests was Sailfish, a free computational fluid dynamics solver based on the Lattice Boltzmann

method and optimized for modern multi-core systems including GPUs. After an initial software recompilation task, which was supported by E4’s R&D team, Enginsoft succeeded in having the full 3D CFD software solution working with approximately the same performance as competitor’s conventional high-end x86-based platforms and significantly better power efficiency.

“For EnginSoft it was an interesting opportunity to be the first company trying to successfully run CFD code on the low power ARM and GPU platform,” said Gino Perna, CTO EnginSoft. “The results were very promising in terms of scalability and performance, and there were no doubts that this outcome would have been impossible to achieve only a few months ago.”“E4’s mission is to deliver leading-edge high performing solutions to our wide range of customers,” said Piero Altoè, vice president of research & innovation, E4 Computer Engineering. “Our ARM-based X-Gene clusters coupled with GPUs are already implemented in universities and research centers, but it was a fantastic opportunity

E4 Computer Engineering and AppliedMicro Demonstrate the Effectiveness of Power-efficient X-Gene®- based technology in Computational Fluid Dynamics EnvironmentsLeading Hardware Manufacturer Provides a GPU cluster and 64-bit X-Gene® Server Class ARM SoC to an Independent Tester and the Results Prove Extremely Positive

Newsletter EnginSoft Year 12 n°3 - 68 69 - Newsletter EnginSoft Year 12 n°3 CommunicateCommunicate

Globalization challenges companies to constantly compete in order to gain and maintain a position in the marketplace. This competition occurs at different levels: companies need to be fast, precise, reliable, efficient and effective. Managerial and organizational weaknesses and slowness cause companies to exit the market or start-up initiatives to fail. There is no unique recipe for company success, even if some “must have” strategies and tactics have been identified as the basis for being competitive. Besides the efficiency of operational processes (i.e. production and logistic chain), the “condemnation of the innovator” is key: companies need to be convicted to innovate and continuously introduce new products and solutions to the marketplace that are more valuable than the ones offered by their competitors. At the same time, the ability to innovate should be efficiently organized with activities and processes properly coordinated and managed. A good design and engineering process management all over the organization – such as R&D, technical, manufacturing and marketing offices - determines a crucial path through which a constant flow of innovation and stable competitive position can be guaranteed.

The GeCo ObservatoryThis is the view of the GeCo Observatory (Management of Collaborative Product Development Processes, www.osservatorio-geco.it), that investigates how Italian companies manage their innovation processes.

GeCo Observatory is a non-profit research initiative of the School of Management in Politecnico di Milano, that collaborates with professors and researchers all over Italy (University of Bergamo, University of Florence, University of Rome “Tor Vergata”, University of Salento, Marche Polytechnic University); it is a super-partes independent initiative that aims to develop a better society through knowledge, awareness, and being meritocratic.

The GeCo Observatory was founded at the end of 2011, from a group of scholars interested in the research of methods and tools adopted by companies to foster R&D, design and engineering efficiency and effectiveness. The first research initiative launched by GeCo Observatory was run between March 2012 and April 2013. This exploratory analysis involved more than 100 companies in Italy and the results have been already discussed in this newsletter. This was followed by a second, deeper and vaster investigation that has taken place for the past 2 years and involved more than 400 companies in Italy and more than 700 designers, technicians, and R&D managers.This second stream of research, whose results have been published at the beginning of May 2015, has been partially financed by industrial partners (Dassault Systèmes, PLM Systems, PROFILE, PTC) and scientific partners (between them European Projects, namely Linked Design, ELICiT, Manutelligence and Diversity).

Ingredients and Recipes for Innovation: One Size doesn’t Fit All!Results from the GeCo Observatory of Politecnico di Milano

We are facing an industrial system that is changing and requires new models to continue to grow.In Italy, as in the rest of the world, the way business models are supporting industrial production methods and new product designs is drastically changing. This is mainly due to the change of the paradigm: industry-product-customer.We at Industrio are going to connect the industry with technology startups and talented entrepreneurs who can support this shift with new product and innovative services. Industrio is the First Hardware Startup Acceletator in Italy, established in 2013 and based in Rovereto at Polo Meccatronica. Industrio is focused on funding pre-revenue startups that are looking to grow and bring amazing Italian products to the market. Twice a year (in spring and autumn) we run a 5 months prototype-to-product program where we drive selected teams to achieve the most important things a hardware startup need to accomplish in the shortest time: business model, prototype, DFM, certification, go to market, fundraising.Our goal at Industrio is to build and scale startups through industrial growth rather than the traditional method of financial growth adopted in other countries.In Italy, we have the chance to work in one of the best industry and manufacturing platforms worldwide: this means that startups and young entrepreneurs can connect, build and grow their product inside a unique ecosystem made of companies and experienced people.Since 2014 Industrio has invested in #8 hardware startups in sector like: automotive, electronics, precision farming, mobility, robotics.Our investment and acceleration model has been made possible thanks to a network of international partners that are supporting our startups’ growth and product development.Through these partnerships with companies focused on product development, electronics manufacturing, microelectronics, mechanical engineering, prototyping, plastics, Industrio allows start-ups to gain access to a multi-million technology and knowledge platform, that wouldn’t be accessible for single entrepreneurs. Our Industry partners offer unique support, expertise and human-technology infrastructure availability for the different stages of development of a hardware startup.Thanks to our partnership with EnginSoft we can offer to our startups, Simulation-Driven Product Development and engineering simulation.Teams can work side-by-side with EnginSoft’s experienced engineers and technicians in order to validate, test and bring to market incredible products that are simulated first and then brought to reality to an amazingly high performance level.

Looking ahead, we think bigAt Industrio, we believe in a “Made GREAT in Italy” motto, an industrial model in which we can continue to building and bring to market well-made and performing products.Today we are living in an extraordinary moment in which there is the opportunity to support a large number of new business initiatives that can give birth to new industries that will create value and jobs for the next 50 years. With this vision, we at Industrio, supported by our incredible network of Partners, want to contribute to this new Industrial Renaissance, starting from Italy, where young entrepreneurs, startups and consolidated industry can work together to bring to market amazing products that can change people’s lives.

INDUSTRIO: A startup Renaissance for the IndustryHow lean methodology, talent, entrepreneurship and new business models can bring Italian industry back to the top

For more information: www.industrio.coJari Ognibeni - CEO Industrio

71 - Newsletter EnginSoft Year 12 n°3 Newsletter EnginSoft Year 12 n°3 - 70 CommunicateCommunicate

2014-2015 Research EditionIn the last two years, the researchers of the GeCo Observatory have been committed to an extensive investigation that firstly aims at understanding which design and product development practices are adopted within Italian companies, and finally aims to identify common behaviours that are able to describe organizational and managerial typologies (to be named archetypes). The methods used during this empirical analysis include case studies, use cases, semi-structured interview and surveys to gain statistical significance. A structured survey was sent to a selected audience of 700 companies of which 266 complete observations have been collected and analysed: 107 were small medium enterprises (SMEs) and 159 big enterprises, all distributed among four industrial sectors (98 mechanics, 56 electric and home appliances, 42 electronics and telecommunications, and 70 other sectors, such as fashion and chemical). More than 70% of the companies in the analysed sample follow the engineering to order (ETO) approach, meaning they design products when specific orders occur; mainly the companies in the sample design and produce products to be delivered globally.The cluster analysis performed on the collected data has identified four categories of recurring product development practices used within companies, and later called archetypes and described in the following session. The cluster analysis has been performed across 6 main clustering variables, determined through a statistical factor analysis.Each of the 6 factors include a specific set of product development practices (i.e. tools, methods, and methodologies) as follow:• Product Performance and Efficiency. This mix of design practices

is focused on managing and monitoring the progress of the product development process.

• Customer Focused Design. The methods in this factor aim at the customer satisfaction and the customization of the product offer.

• Multifunctional Skilled Team. These practices identify a set of organizational rules able to manage roles and competences of the actors involved in the development team.

• Sustainable Innovation. Mix of formal approaches focused on the delivery of highly innovative products through a holistic product lifecycle view.

• Collaborative Engineering. This factor involves practices able to support the concurrent management of development projects.

• Formalised Knowledge Management. These practices include lots of methods for knowledge representation, storage, usage and re-usage.

The “Fantastic Four” ModelThe identification of the 6 sets of product development practices, tools and methods is the result of the statistical analysis performed on the collected data. Actually such categories represent the “ingredients” a company can use to make its own “recipe of innovation”. The following

cluster analysis – run within a broader number of observations - has led to the identification of 4 recurring recipes, which describe 4 “archetypes of innovation”. The researchers of the GeCo Observatory get inspiration to name these archetypes from the Fantastic Four superhero team of the comic books published by Marvel, and namely:• Mister Fantastic. More than 50 companies (21 SMEs and 31

large, equally distributed among the four sectors) belong to this cluster. Companies following the Mister Fantastic recipe of innovation, strongly rely on multifunctional teams, where design knowledge is formally managed and shared in order to efficiently develop innovative products.

• The Thing. 76 companies (out of which 50 are large organizations) from all the four industrial sectors, appertain to this archetype.

• The followed innovation model is based on robust inclination towards customization and customer satisfaction (almost all the companies are ETO oriented), accomplished through the combined use of trivial practices, such as formal design rules, IT systems, knowledge formalization and so on.

• Invisible Woman. This recipe of innovation is the most followed in the analysed sample (77 companies, equally distributed between company size and industrial sectors). This cluster is based on structured and cross-functional collaboration of project team together with parallelization and synchronization of design activities along the development process. Not randomly the character symbolizing this kind of behaviour is a woman, naturally multitasking and careful to the needs of all actors involved in the process!

• Torchman. This is the most diffused archetype between the SMEs (around 50% out of the 61 companies belonging to this cluster) and apparently the most adequate to the less traditional industries (fashion, food, textile and pharmaceutical). This recipe of innovation is intensively addressed to product innovation, in the sense of creating new products, possibly in a green and sustainable way, with a good marketing strategy and a high attention to the customer, which mainly is the end user.

Figure 1 – Fantastic Four, the Radar Chart

Figure 1 depicts the 4 archetypes of innovation, resulting from the cluster analysis performed on the mentioned 6 variables, alias group of practices that companies adopt at different maturity levels (represented in the axis of the radar chart, with a increasing maturity from thecenter to the perimeter).

Competitive Factors and Performance of the “Fantastic Four”Together with the identification of the four archetypes of innovation, the researchers of theGeCo Observatory investigated the competitive factors and performance aimed and achieved by the interviewed companies. Two assessment models have been developed for this purpose, and submitted to 163 enterprises:• In academia, a “competitive factor” is

a strategic objective that a company aims to reach and on which it bases its long-term strategy and marketing positioning. The developed assessment model (named HEART) has been built on 9 recurring competitive factors –cost, time, quality, service, functional performance, flexibility, customization, segmentation, innovation- grouped into 4 classes and evaluated through 5 level of importance for each factor.

• A “competitive performance” is a specific result achieved in the market when implementing certain practices and operative choices aligned to specific strategic objectives. 14 different kinds of performance have been analysed within the sample, grouped into 5 classes – economic, time, innovation, cost, and quality performance - scored in 5 levels of accomplishment (SCORE model). The level of achievement of a specific performance has been estimated by comparing the individual company result with its direct competitors in order to obtain homogeneous, generalizable and comparable results.

Basically, competitive factors represent company’s desires and aspirations, while performances describe the real outcomes achieved in the marketplace.From the data analysis it is clear that different archetypes of innovation could achieve similar performances by aligning the used product development practices with the desired competitive factors.Practically, it can be stated that the same company could follow different “recipes of innovation” to obtain the same result. At the same time, certain archetypes better support the achievement of certain competitive factors. In other words, specific “recipes of innovation”, alias a superhero of the Fantastic Four, shall be preferred to others when pursuing determined goals. Figure 2 shows which are the archetypes that should be preferred, when aiming at reaching specific target performances according to a chosen factor the company would like to compete in the market.

The matrix in Figure 2 opens a broad range of discussions, such as for example:• For Italian companies, competitive factors such as quality,

service and high performance products are a must have condition to compete in the first place.

• Companies that aim to gain a market position based on offering segmentation and innovative products can gain economic, cost and time benefits when following the same managerial approach as The Thing.

• Enterprises whose purpose is to compete on cost and time in the marketplace, can effectively reach high time-to-market performance by adopting the typical design practices characterizing The Invisible Woman.

• Companies acting according to the Mister Fantastic character, who structure and organize their product development are able to achieve the higher performance in realizing innovative products; however if the main purpose of the company is to offer a high level of segmentation and innovation, then acting as Torchman is the proper example to follow.

ConclusionsThe research conducted by the GeCo Observatory examines different dimensions concerning the modality and practice used to manage product development processes. The main results have been used to define the Fantastic Four model which identifies four reference archetypes of innovation that a company can refer to when designing/redesigning their development processes.

Specifically, the matrix shown in Figure 2 is an intuitive and quick instrument for companies to reflect on their positioning, based on the experience of hundreds of companies and supported by statistical relevance. It is interesting to see how the archetypes of innovation, aliased by the Fantastic Four, are totally independent from a company’s size and sector, this debunks the myth that size and sector matter. In practice, SMEs and large enterprises can gain inspiration from each other, as well as companies from different industries can reciprocate how to innovate.

Department of Management, Economics and Industrial Engineering Monica Rossi, Sergio Terzi

Politecnico di Milano

Figure 2 – Matrix Competitive Factors/Performances across the Fantastic Four

Your opportunity to be part of the future!

19 - 20October 2015Pacengo del Garda,

Verona - Italy

www.caeconference.com

SPECIAL GUEST

Prof. Davide Cassi - University of ParmaCooking Hackers.The true story of molecular cuisine

During the last twenty years cooking art has evolved signi�cantly thanks to the contribution of scienti�c research.

Here, we retrace the milestones of this collaboration and look at the most important innovations that have been introduced.

newsletter - enginsoft · 7 - newsletter enginsoft year 12 n°3 case histories due to the black...

Documents