Graduate Theses and Dissertations Iowa State University Capstones, Theses and Dissertations

2019

Dynamic analysis of composite wind turbine blade Dynamic analysis of composite wind turbine blade

Divya Teja Pinnamaneni Iowa State University

Follow this and additional works at: https://lib.dr.iastate.edu/etd

Part of the Aerospace Engineering Commons, Oil, Gas, and Energy Commons, and the Statistics and

Probability Commons

Recommended Citation Recommended Citation Pinnamaneni, Divya Teja, "Dynamic analysis of composite wind turbine blade" (2019). Graduate Theses and Dissertations. 17542. https://lib.dr.iastate.edu/etd/17542

This Thesis is brought to you for free and open access by the Iowa State University Capstones, Theses and Dissertations at Iowa State University Digital Repository. It has been accepted for inclusion in Graduate Theses and Dissertations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact digirep@iastate.edu.

Dynamic analysis of composite wind turbine blade by

Divya Teja Pinnamaneni

A thesis submitted to the graduate faculty in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Major: Aerospace Engineering

Program of Study Committee Sheidaei Azadeh, Major Professor

Pouya Shahram Farzad Sabzikar

The student author, whose presentation of the scholarship herein was approved by the program of study committee, is solely responsible for the content of this thesis. The

Graduate College will ensure this thesis is globally accessible and will not permit alterations after a degree is conferred.

Iowa State University

Ames, Iowa

2019

ii

TABLE OF CONTENTS

LIST OF FIGURES iii

LIST OF TABLES v

ACKNOWLEDGMENTS vi

ABSTRACT vii

CHAPTER 1. INTRODUCTION 1

CHAPTER 2. BLADE MODELING 5

CHAPTER 3. AERODYNAMIC LOADS 12

Blade Element Momentum Theory 12

CHAPTER 4. BLADE STRUCTURAL ANALYSIS 18

Static Structural Analysis 18

Blade Fatigue Analysis 23

Yearly Fatigue Analysis 26

Monthly Fatigue Analysis 32

CHAPTER 5. CONCLUSION 38

REFERENCES 39

iii

LIST OF FIGURES Figure 1: Wind turbine failure type distribution incidents recorded between 1980

and 2016 [6] 2 Figure 2: Blade analysis procedure 4

Figure 3: Blade planform 6

Figure 4: Blade specifications 6

Figure 5: The station wise variation of chord length [5] 7

Figure 6: The station wise variation of thickness [5] 7

Figure 7: The airfoil details of the blade [5] 8

Figure 8: The S830 airfoil section model [5] 8

Figure 9: The S831 airfoil section model [5] 9

Figure 10: A sample blade cross section as defined by Perkins and Cromack [7]. 10

Figure 11: Blade model with imaginary planes showing the distance (in mm) of sections from the root 11

Figure 12: Blade model with Spar 12

Figure 13: The relations between angle of attack (α), the inflow angle (ϕ), twist angle (ϴ), the velocities and forces acting on a wind turbine blade element [8] 14

Figure 14: Flowchart for calculating blade forces and moments [8] 17

Figure 16: Material properties of the blade 18

Figure 17: FE Model after applying load case 1 19

Figure 18: FE Model after applying load case 2 19

Figure 19: Deformation plot of yearly analysis of load case 1 20

Figure 20: Von-Mises stress plot for yearly analysis of load case1 21

iv

Figure 21: Deformation plot of yearly analysis of load case 2 22

Figure 22: Von-Mises stress plot for yearly analysis of load case 2 22

Figure 23: Fatigue analysis procedure using Ansys 23

Figure 24: Goodman mean stress correction 24

Figure 25: S-N curve for skin material 25

Figure 26: S-N curve for spar and stiffener material 26

Figure 27: Thrust loads for load case 1 28

Figure 28: Thrust loads for load case 2 28

Figure 29: Sectional moments for load case 1 29

Figure 30: Sectional moments for load case 2 29

Figure 31: Safety life of the blade for yearly analysis of load case 1 30

Figure 32: Damage of the blade for yearly analysis of load case 1 31

Figure 33: Safety life of the blade for yearly analysis of load case 2 32

Figure 34: Damage of the blade for yearly analysis of load case 2 32

Figure 35: Thrust loads application in Ansys for load case 1 33

Figure 36: Moments application in Ansys for load case 1 34

Figure 37: Safety life of the blade for monthly analysis of load case 1 35

Figure 38: Damage of the blade for monthly analysis of load case 1 35

Figure 39: Safety life of the blade for Monthly analysis of load case 2 36

Figure 40: Damage of the blade for monthly analysis of load case 2 37

v

LIST OF TABLES Table 1: Monthly RMS values for both the load cases 27

Table 2: Comparison of yearly versus monthly fatigue results for load case 1 37

Table 3: Comparison of yearly versus monthly fatigue results for load case 2 37

vi

ACKNOWLEDGMENTS

The Iowa Atmospheric Observatory towers and associated instrumentation were funded by

an NSF/EPSCoR grant to the state of Iowa (Grant #1101284) and a follow-on NSF/AGS grant

#1701278.

vii

ABSTRACT

The purpose of this paper is to establish a basis for determining the accuracy level

required in load prediction models by comparing fatigue results of a composite material

wind turbine blade for a set of experimental wind loads. Wind data for two successive

years is considered separately as two load cases to calculate aerodynamic loads. These

loads are used in monthly and yearly cycles to perform fatigue analysis and the difference

in safety life and damage for successive years is calculated, to check, if faithful prediction of

hourly or daily wind speeds for a wind forecast model is required. Thus, the dynamics of

wind forecasting can be improved for safe, reliable and economical operation of the wind

turbines. This report details blade modeling procedure, calculation of aerodynamic loads

from the collected wind data and fatigue analysis of the wind turbine blade.

1

CHAPTER 1. INTRODUCTION

Numerous researchers have studied that irregular wind patterns have caused

abrupt aerodynamic effects and wind turbulence on the blade which act as fatigue loads on

the wind turbine blade [1] [2] [3]. Precise wind speed prediction will be very effective for

safe and reliable working of wind turbine as featured in [4]. The conventional fatigue life

projection of wind turbine blades relies on a set of wind load distributions that does not

fully capture wind load uncertainty. This could lead to early blade fatigue failure and

ultimately increase wind turbine maintenance costs. In the production of renewable energy

around the globe, wind turbines are becoming progressively common and significant. In

the last 20 years, there has been 1305.4% increase in wind energy generation capacity in

United States of America and looks promising for further growth. The future looks very

ambitious and to reach these targets, more efficient wind farms are required. Accurate

prediction of the wind is one parameter that helps in reducing operating costs and makes

power generation systems more reliable and efficient [5]. The evolution in wind prediction

models has eventually helped building stable and reliable wind turbines but a further scope

of improvement is observed [6]. A chart showing the failure type of wind turbines is shown

in Figure 1. Blade failure, structure failure and environmental damage account for the

majority share of wind turbine collapses. An advancement in wind prediction models can

help to get this percentage down. Therefore, considerable effort goes into wind pattern

analysis and forecasting to establish blade loads, blade design optimization, fatigue analysis

and manufacturing techniques for safety and reliability, which is why the accurate

2

prediction of wind patterns and safety life of wind turbines plays a vital role in economical

running of wind turbines.

Figure 1: Wind turbine failure type distribution incidents recorded between 1980 and 2016

[6]

While blade efficiency is susceptible to the unsteady nature of the wind and its

intensity of turbulence, blades’ fatigue life may not be as sensitive to the unsteadiness

nature of the wind. In other words, it may not be essential to faithfully predict hourly or

daily wind speeds for a wind forecast model used to predict blade fatigue. The objective of

this work is to test this hypothesis in two successive years using accessible measured wind

speed data. First, averaged monthly wind speed data is used to predict fatigue life of a

wind turbine blade and then process is repeated for a yearly average without considering

the seasonality of the data. Wind data from September 2016 to August 2018 are gathered

from “The Iowa Atmospheric Observatory towers”. September 2016 to August 2017 is

considered as load case 1 and September 2017 to August 2018 is considered as load Case 2.

Blade Failure18%

Human Health4%

Human Injury7%

Fatal Accidents

6%

other20%

Environmental damage

10%

Transport9%

Ice Throw2%

Structure Failure

9%

Fire15%

3

The analysis was performed on both the load cases to check if the results observed in the

first case would be valid in the second case to make strong conclusions. Aerodynamic loads

are calculated for these two load cases using Blade Element Theory. Hence, four fatigue

analysis are performed in Ansys software and the results are compared to determine what

level of accuracy is required in wind speed data to predict blade fatigue life once the

forecast model is developed.

The blade analysis procedure is described in 4 steps as shown in Figure 2. First, a

three-dimensional computer-aided (CAD) model of a wind turbine blade is designed in

Solidworks commercial software based on the design inputs from a SCANDIA Blade Systems

Design Studies report by Derek S. Berry. This is a horizontal axis wind turbine blade which is

the focus of our study. The CAD model is imported in Ansys software for Finite Element

Analysis. Material properties and boundary conditions are applied to the model and

meshing is performed. Static structural analysis is performed to calculate deformation and

stresses. In the third step, details about wind data gathering are presented. Finally, a

Matlab soft- ware program based on BEM theory. This program is used to extract thrust

forces and bending moments at each section of the blade by providing wind data and blade

specifications as the input. This dynamic load cycles are used in Ansys to calculate fatigue

damage and safety life of the wind turbine blade and critical locations on the wind turbine.

For this analysis, two sets of wind data are gathered, each consisting of one-year wind

speed, wind direction, temperature and humidity. Analysis was performed separately for

monthly and yearly cycles for both the years to compare the difference in damage and

study the most effective way.

4

Figure 2: Blade analysis procedure

Modelling

•Geometric Parameters Gathering

•Designing 3D CAD Model in Solidworks

Load Cases

•Wind Data Collection

•Aerodynamic Loads Calculation

FE Model

•Material Properties

•Boundary Conditions

•Meshing in Ansys

•Calculation of Stresses

Fatigue Analysis

•Defining fatigue properties

•Calculation of safety life and damage

5

CHAPTER 2. BLADE MODELING

In this section, blade design parameters and procedure followed to design three-

dimension model of the wind turbine blade are detailed. The CAD model of the blade in

this study is referred from a SANDIA report on blade system design [5]. This blade is

particularly selected as we wanted to conduct the analysis on a horizontal axis wind turbine

blade. Furthermore, it’s convenient to design a CAD model as all the blade specifications

and properties are clearly detailed, and the model can be used to perform fatigue analysis.

The wind turbine blade is divided into 10 sections along the span-wise direction as shown

in the Figure 3. The loads are applied on these sections while performing structural

analysis. The blade specifications, station wise variation of chord length and thickness are

shown in Figure 4 to Figure 6. The airfoil types used in the blade model are shown in Figure

7 to Figure 9.

6

Figure 3: Blade planform

Figure 4: Blade specifications

7

Figure 5: The station wise variation of chord length [5]

Figure 6: The station wise variation of thickness [5]

8

Figure 7: The airfoil details of the blade [5]

Figure 8: The S830 airfoil section model [5]

9

Figure 9: The S831 airfoil section model [5]

The blade model consists of skin, spar and blade stock. The skin is an outer layer on

the spar, it’s a hollow tapering section and runs from the blade stock to the end. Skin is

composite structure made up of fiberglass epoxy matrix with low bending modulus. Spar

runs through the blade at the leading edge and acts as a stiffener. Spar is used to provide

stability and increases the strength of the blade. A high modulus fiberglass epoxy

composite material is used for spar to attain the required strength to the blade. The blade

stock is the section attaching the blade to the wind turbine tower. The blade stock is made

up of fiberglass epoxy with steel sleeve to make the root section very durable as the entire

blade is joined to the tower through the blade stock. A sample blade section is shown

Figure 10.

10

Figure 10: A sample blade cross section as defined by Perkins and Cromack [7].

Referring to the Scandia report following steps were followed to design the CAD model

in Solidworks commercial software:

• Airfoil types shown in Figure 3 were chosen for the different spanwise locations of

the blade and calculated the spanwise station locations.

• Blade stations are modelled using planes feature and the airfoil shapes are

modelled using splines feature in Solidworks

• The airfoil sections are scaled to match the chord details shown in Figure 3 and

Figure 5. Each section is then applied with twist angle given in Figure 3.

• A three-dimensional shell modal was built on the associated airfoils using the lofted

surface feature. The blade model is shown in Figure 11.

11

Figure 11: Blade model with imaginary planes showing the distance (in mm) of sections

from the root

12

Figure 12: Blade model with Spar

CHAPTER 3. AERODYNAMIC LOADS

Wind data is collected from an agricultural location 30 miles from Ames, Iowa.

More information about the data resources can be found in https://mesonet.agron.

iastate.edu/projects/iao/ . Two-year 1 hertz data is collected from one of the two towers

located in above mentioned location. The monthly average of the data is calculated, and

RMS is determined. These loads are used to calculate the aerodynamic loads (thrust and

moment forces) on the blade. BEM theory is used to determine loads at each section of the

blade.

Blade Element Momentum Theory

In the BEM theory, the flow is assumed to take place in independent streamlines

and the loading is estimated from two-dimensional sectional airfoil characteristics. BEM

13

method is used to calculate the aerodynamic loads (thrust and moment forces) that are

generated at the sections of the blade due to interaction with wind stream. To determine

the forces and moments, we should know the local angle of attack (α) and flow velocity

relative to the blade (Urel), along with chord length and blade twist angle [8]. As we know

the design parameters of the blade i.e., is twist angle and chord length, we should

determine angle of attack and flow velocity.

14

Axial Induction Factor

The axial induction factor a is defined as the loss in axial speed due to presence of

the blades and is given by the formula [8]:

𝑎 = 𝑈∞ − 𝑈𝑡

𝑈∞

Angular Induction Factor

At the rotor plane, relative wind speed’s angular velocity component can be defined

in terms of blade’s rotational speed and induced rotational speed. A fraction of the blade

rotational speed (ωra’) is the angular component of the blade, where a’ is defined as the

angular induction factor. Hence, the tangential velocity component can be defined as

Wt = ωr (1 + a’) [8].

i.e., 𝑎′ =𝑊𝑡

𝜔𝑟

Figure 13: The relations between angle of attack (α), the inflow angle (ϕ), twist angle (ϴ), the velocities and forces acting on a wind turbine blade element [8]

15

Using axial induction factor (a) and angular induction factor (a’), relative inflow

velocity (Urel) and inflow angle (ϕ) can be determined. Relative inflow velocity (Urel) and

inflow angle (ϕ) in terms of a and a’ can be defined as:

𝑈𝑟𝑒𝑙 = √[𝑈∞(1 − 𝑎)]2 + [𝜔𝑟(1 + 𝑎′)]2

tan 𝜙 =𝑈∞(1 − 𝑎)

𝜔𝑟(1 + 𝑎′)

Angle of attack (α) can be determined by subtracting twist angle (ϴ) from inflow

angle (ϕ). Aerodynamic forces can now be calculated using relative inflow velocity and

angle of attack. Using trigonometric relations, the calculated lift and drag forces can be

converted into normal and tangential forces. The forces acting in flow direction are defined

by normal force coefficient (Cn), which is thrust force. The forces acting in tangential

direction are defined by tangential force coefficient (Ct), which is moment acting on the

blade section. The normal force coefficient and tangential force coefficient are given by the

formula shown below [8]:

𝐶𝑛 = 𝐶𝐿 cos 𝜙 + 𝐶𝐷 sin 𝜙

𝐶𝑇 = 𝐶𝐿 sin 𝜙 − 𝐶𝐷 cos 𝜙

By integrating normal forces (thrust) and tangential forces (moment) along all the

sections of the blade, total thrust and moment can be determined.

16

For a finite length blade, the circulation created by a rotating blade tends to

exponentially to zero close to the tip as proved by German engineer Prandtl. On this basis,

we can add Prandtl correction factor to the BEM equations. The approximate formula given

by Glauert for Prandtl correction factor is shown below [9].

𝐹 = 2

𝜋 𝑐𝑜𝑠−1 [𝑒

−𝐵(𝑅−𝑟)2𝑟 sin 𝜙]

Where B is number of blades and r is local radius. Prandtl correction factor is very

efficient and is proved to give good results for wind turbine [10].

A Matlab code is developed following the flowchart shown in Figure 14 to calculate forces

and moments on the blade section.

17

Necessary Inputs

𝛩 𝐵 𝑐 𝑟

Local pitch of blade Number of blades

Chord length Radius

𝛥𝑟 𝛺 𝑉1

𝜌

Length of element Rotational Speed

Wind Speed Wind density

Table of lift and Drag Values

Guess induction factors

𝑎 = 0 𝑎’ = 0

Find flow angle

𝜑 = tan−1 [(1 − 𝑎)𝑉1

(1 − 𝑎′)𝜔𝑟]

Find angle of attack

𝛼 = 𝜑 − 𝜃

Find force coefficients

𝐶𝑛 = 𝐶𝐿 cos 𝜑 + 𝐶𝐷 sin 𝜑

𝐶𝑇 = 𝐶𝐿 sin 𝜑 + 𝐶𝐷 cos 𝜑

Find force solidity

𝜎(𝑟) =𝑐(𝑟)𝐵

2𝜋𝑟

Find tip loss factor

𝐹 = 2

𝜋cos−1 [𝑒

−(𝐵(𝑅−𝑟

2𝑟 sin 𝜑)]

Calculate new values for a and a’

𝑎 = 1

𝐹4𝑠𝑖𝑛2𝜑𝜎𝐶𝑛

+ 1

𝑎′ = 1

𝐹4 sin 𝜑 cos 𝜑𝜎𝐶𝑇

+ 1

Find tip loss factor

𝑇 = 𝐹𝐶𝑛

1

2𝜌𝑉1

2(1 − 𝛼)2𝑐𝐵∆𝑟

𝑠𝑖𝑛2𝜑

𝑀 = 𝐹𝐶𝑇

1

2𝜌𝑉1(1 − 𝛼)𝑐𝐵

∆𝑟

sin 𝜑 cos 𝜑

Figure 14: Flowchart for calculating blade forces and moments [8]

Retrieve lift and drag values from the table CL CD

18

CHAPTER 4. BLADE STRUCTURAL ANALYSIS

The Finite Element Analysis is performed using ANSYS commercial software. CAD

model is imported to Ansys and material properties are assigned as defined in the Figure

15. The blade sits on the turbine hub and is considered as a cantilever body. To attain this

boundary condition, the edge of the blade is arrested for all degrees of freedom.

Figure 15: Material properties of the blade

Static Structural Analysis

The analysis was performed with the maximum forces for monthly data. The gravity

force (m/s^2), thrust force (N) and sectional bending moments are applied at each section

at center of the section. The Finite Element Model after applying the loads for September

2016 to August 2017 is shown in Figure 16 and September 2017 to August 2018 is shown in

Figure 17.

19

Figure 16: FE Model after applying load case 1

Figure 17: FE Model after applying load case 2

20

The total deformation of 73.162 mm and von-mises stress of 85.32 MPa are

observed on blade under the application of load case one. The deformation and stress

plots are shown in Figure 18 and Figure 19 respectively.

Figure 18: Deformation plot of yearly analysis of load case 1

21

Figure 19: Von-Mises stress plot for yearly analysis of load case1

The total deformation of 123.1 cm and Von-Mises stress of 156.07 MPa are

observed on the blade under the application of load case two. The deformation and stress

plots are shown in Figure 20 and Figure 21 respectively.

22

Figure 20: Deformation plot of yearly analysis of load case 2

Figure 21: Von-Mises stress plot for yearly analysis of load case 2

23

Blade Fatigue Analysis

Fatigue analysis procedure is majorly divided into 5 steps in Ansys as described in

[16] and shown in Figure 22.

Figure 22: Fatigue analysis procedure using Ansys

Fatigue Analysis Type: Stress life approach is used to perform this analysis as it’s easier

than strain life approach and the results obtained will be enough to calculate the fatigue

parameters. In this approach, empirical S-N curves are used along with a variety of factors.

Loading: “Non-Constant Amplitude, Proportional loading” technique is used to apply loads

on the blade sections due to dynamic and cyclic loads induced in the blades by the wind.

This loading technique allows the load to vary over time to calculate a cyclic load. Using this

method, critical fatigue location can be determined with only one set of FE results but the

loads that cause critical damage cannot be easily seen. To calculate total damage and loads

causing it, cumulative calculations like cycle counting using Rainflow algorithm and damage

summation using Miner’s rule must be performed.

24

Mean Stress Corrections: Mean stress corrections factors are usually used to calculate a

mean stress correction, this will be helpful in determining the effective alternating stress.

This stress will be used with an S-N curve to obtain fatigue results. Goodman’s criteria are

used here as it’s the best for brittle materials as compared to Soderberg’s equation or

Gerber’s equation. Using Goodman’s equation with graph in fig, effective alternating stress

(σeff) can determined.

𝜎𝑒𝑓𝑓 = 𝜎𝑎 [ 𝜎𝑢

𝜎𝑢 − 𝜎𝑚 ]

Figure 23: Goodman mean stress correction

Multiaxial Stress Correction Factor: The aerodynamic loads are usually uniaxial whereas the

FE results are usually multiaxial. Multiaxial stress state should be converted to uniaxial, so

that Von-Mises stress can be compared with the uniaxial stress value.

Results: There is a wide variety of options to choose for calculating results. In this analysis

fatigue life and damage are calculated. Fatigue life contour plot shows the available life of

the given component under certain loading. Fatigue damage is given as designed life

25

divided by available life. A value greater than 1 indicates that the component will fail

before the design life.

The thrust and moment forces acting on blade for the two load cases are calculated

in this section using the procedure described in section Aerodynamic loads and stresses

calculated in previous section are used for the stress life fatigue analysis. The material

properties used for fatigue analysis are shown in Figure 15. These material properties used

in this analysis are from SNL/MSU/DOE Composite Material Fatigue Database [13] for

equivalent materials. The stress amplitude versus number of cycles (S-N) curves used in this

analysis are determined using Constant Life Diagrams [14]. The S-N curves for blade

components are shown in Figure 24 and Figure 25 respectively. Following the procedure

described above the fatigue analysis was performed for two load cases to calculate safety

life and damage. These results are compared for yearly and monthly analysis for the two

load cases to study how it impacts the wind blade.

Figure 24: S-N curve for skin material

26

Figure 25: S-N curve for spar and stiffener material

Yearly Fatigue Analysis

The Root Mean Square value is calculated for the collected monthly data. The RMS

monthly data is shown in Table 1. The calculated RMS is the velocity input for the Matlab

code shown in Figure 14. The RMS along with blade dimensions generate Thrust force (N)

and Sectional Moments (N-m) at each section. 12 load points depicting 12 months are

generated at each section. For yearly analysis, cyclic load consisting of all the maximum

loading values at each section is considered to calculate fatigue. Thrust force and section

moments calculated using Matlab code that incorporates BEM method are shown in Figure

26 to Figure 29. The loads are applied on each section of the blade and fatigue analysis was

performed.

𝑅𝑀𝑆 = √∑ 𝜈2𝑛

𝑖=1

𝑛

27

Where,

𝜈 = wind velocity

n = Number of days

Table 1: Monthly RMS values for both the load cases

Month RMS Month RMS

Sep-2016 5.42 Sep-2017 4.78

Oct-2016 6.27 Oct-2017 10.00

Nov-2016 8.52 Nov-2017 8.44

Dec-2016 10.22 Dec-2017 9.85

Jan-2017 7.55 Jan-2018 9.54

Feb-2017 9.10 Feb-2018 6.56

Mar-2017 11.66 Mar-2018 8.81

Apr-2017 9.75 Apr-2018 11.45

May-2017 8.84 May-2018 7.73

Jun-2017 7.19 Jun-2018 6.40

July-2017 3.22 Jul-2018 3.76

Aug-2017 3.33 Aug-2018 3.85

28

Figure 26: Thrust loads for load case 1

Figure 27: Thrust loads for load case 2

29

Figure 28: Sectional moments for load case 1

Figure 29: Sectional moments for load case 2

30

Yearly analysis for load case 1

Load case 1 is generated from the acquired wind data for September 2016 to

August 2017 period. The thrust and moments used in this analysis are shown in Figure 26

and Figure 28 respectively. The Life of the blade is observed as 706 months i.e., 58 which is

in line with the industry standard of 20 to 40 years [15]. Damage is observed to be 0.0957

which implies that the blade will not fail before the design life. The safety life and damage

simulations under this loading condition are shown in Figure 30 and Figure 31 respectively.

Figure 30: Safety life of the blade for yearly analysis of load case 1

31

Figure 31: Damage of the blade for yearly analysis of load case 1

Yearly analysis of load case 2

Load case 2 is generated from the acquired wind data for September 2017 to August

2018 period. The thrust and moments used in this analysis are shown in Figure 27 and

Figure 29 respectively. The Life of the blade is observed as 958 months i.e., approximately

78 years which is way above the industry standard. This is because of the lower average of

considered loads. Damage is observed to be 0.0125 which is less than 1 and hence, blade

will not fail before the design life. The safety life and damage simulations under this

loading condition are shown in in Figure 32 and Figure 33 respectively.

32

Figure 32: Safety life of the blade for yearly analysis of load case 2

Figure 33: Damage of the blade for yearly analysis of load case 2

Monthly Fatigue Analysis

The Root Mean Square value is calculated for the collected monthly data. The RMS

monthly data is shown in Table 1. The calculated RMS is the velocity input for the Matlab

33

code shown in Figure 14. The RMS along with blade dimensions generate Thrust force (N)

and Sectional Moments (N-m) at each section. 12 load points depicting 12 months are

generated at each section. For Monthly analysis, a cyclic load consisting of all the 12 load

points at each section is considered for the fatigue analysis. The thrust loads and sections

moments applied in Ansys are shown in Figure 34 and Figure 35 respectively.

Figure 34: Thrust loads application in Ansys for load case 1

34

Figure 35: Moments application in Ansys for load case 1

Monthly fatigue analysis for load case 1

The Life of the blade is observed as 45 years which is in line with the industry

standards. Damage is observed to be 0.022 which implies that the blade will not fail before

the design life. The safety life and damage simulations under this loading condition are

shown in Figure 36 and Figure 37 respectively.

35

Figure 36: Safety life of the blade for monthly analysis of load case 1

Figure 37: Damage of the blade for monthly analysis of load case 1

36

Monthly fatigue analysis for load case 2

The Life of the blade is observed as 41 which is in line with the design life standards.

Damage is observed to be 0.0243 which implies that the blade will not fail before the

design life. The safety life and damage simulations under this loading condition are shown

in Figure 38 and Figure 39 respectively.

Figure 38: Safety life of the blade for Monthly analysis of load case 2

37

Figure 39: Damage of the blade for monthly analysis of load case 2

Table 2: Comparison of yearly versus monthly fatigue results for load case 1

Sept 2016 to Aug 2017 Yearly Monthly

Safety Life (Years) 58 45

Damage 0.096 0.022

Table 3: Comparison of yearly versus monthly fatigue results for load case 2

Sept 2017 to Aug 2018 Yearly Monthly

Safety Life (Years) 79 41

Damage 0.0125 0.0243

38

CHAPTER 5. CONCLUSION

The fatigue results for monthly and yearly analysis are summarized in Table 2 and

Table 3. Based on the analysis results for two years’ data, it appears that both monthly and

yearly averaged data predicts a similar fatigue damage. However, a significant difference is

noted in the safety life. The difference can be because of the dynamic nature of the load

cycles. The change in nature of the wind loads plays a very vital in the fatigue analysis. The

average design life of a wind turbine is considered to be approximately 20-40 years [15].

The monthly analysis safety life is observed to be in line with this approximate design life

while for yearly analysis it was observed to be very high than the industry standards.

Difference of 28% and 92% was observed in safety life of yearly and monthly analysis for

the two load cases respectively. The further study considering wind data for long term

period can be helpful to safely conclude that hourly or daily prediction of wind speeds is

not required to predict wind forecast models. As a scope for the future work, we can

perform fatigue analysis for a daily load case with 365 load points to study the variation of

fatigue results.

39

REFERENCES

1. Mo W, li D, Wang X, et al. Aeroelastic coupling analysis of the flexible blade of a wind turbine. Energy 2015;89: 1001-1009

2. Liu X, Lu C, Liang S, et al. Influence of the vibration of large-scale wind turbine blade

on the aerodynamic load. Energy procedia 2015; 75: 873-879

3. Liu X, Lu C, Liang S, et al. Vibration-induced aerodynamic loads on large horizontal axis wind turbine blades. Appl Energ 2017; 185: 1109–1119

4. Kaneko T, Uehara A, Senjyu T, et al. An integrated control method for a wind farm to

reduce frequency deviations in a small power system. Appl Energ 2011; 88: 1049–1058

5. Derek S. Berry, Blade System Design Studies Phase II: Final Project Report, Sandia

Report, SAND2008-4648, July 2008

6. Ma Y, Martinez-Vazquez P and Baniotopoulos C, Wind Turbine Tower Collapse Cases: A Historical Overview. Proceedings of the Institute of Civil Engineers – Structures and Buildings, https://doi.org/10.1680/jstbu.17.00167

7. Perkins FW and Cromack DE. Wind turbine blade stress analysis and natural

frequencies. Wind Energy Center Reports, Paper-11, 1978, http://scholarworks.umass.edu/cgi/viewcontent.cgi?article=1010&context=windenergy_report

8. John Amund Karlsen, Performance Calculations for a Model Turbine, June 2009

9. W Z Shen, R Mikkelsen, J N Sørensen, and C Bak. Tip Loss Corrections for Wind Turbine

Computations. Wind Energy, 8:457 – 475, 2005

10. M O L Hansen. Aerodynamics of Wind Turbines, 2nd ed. Earthscan, 2008

11. Pravin A Kulkarni, Weifei Hu, Ashwinkumar S Dhoble and Pramod M Padole, Statistical Wind Prediction and Fatigue Analysis for Horizontal-Axis Wind Turbine Composite Material Blade under Dynamic Loads, Advances in Mechanical Engineering, 2017, Vol. 9(9) 1–26

12. J.F. Manwell, J.G. McGowan and A.L. Rogers, Wind Energy Explained - Theory, Design

and Application, Second Edition, 2009

40

13. Mandell JF and Samborsky DD. SNL/MSU/DOE composite material fatigue database.

Albuquerque, NM: Sandia National Laboratories, 2014 14. Mandell JF, Ashwill TD, Wilson TJ, et al. Analysis of SNL/MSU/DOE fatigue database

trends for wind turbine blade materials. Sandia report, SAND2010-7052

15. R. H. Crawford, Life-Cycle Energy Analysis of Wind Turbines—An Assessment of the Effect of Size on Energy Yield, doi:10.2495/ESUS070161

16. ANSYS, C. (2011). "User Manual Release 12.1." ANSYS Inc