fatigue calculation

[Junyi.Wu]

Hello everyone,

I am now working on the theory of MLife and feeling confused with some problems.

1. I find that in the MLife theory manual, it uses load range rather than stress range to calculate fatigue. Does the ‘load range’ equal to ‘stress range’?
2. Is there any method to decide how many time-series to be divided in MLife?
   Could someone can help me please?

Kind Regards
Junyi

[Jason.Jonkman]

Dear Junyi,

Loads and stresses are not equivalent. FAST calculates loads (lumped forces/moments at a cross section), not stresses. Of course, the loads can be used to calculate stresses either analytically for simple stress state (e.g. stress = force/area for a simple rod undergoing axial loading) or through a more general cross-sectional analysis (e.g. 3D FEA) for a more complicated multi-component stress state (axial, shear, bending, etc.), especially for a nonaxisymmetric cross sections composed of an anisotropic material. Often, FAST output is used directly in MLife, in which case MLife is processing loads, not stresses. MLife could also be used to process stresses, if the stresses are calculated from the loads output from FAST before post-processing with MLife.

I’m not sure I understand your second question, but the wind turbine design standards (e.g. IEC 61400-1) specify loads-analysis requirements, including requirements on the number and length of the simulations needed per load case.

Best regards,

[Junyi.Wu]

Hello Jason,

Thanks you a lot. I mean we process all the input data files in MLife and it will read a time-series. Does it mean if you have 5 input data files, so it has 5 time-series? Or for each input data file, it is separated by many time-series?

Regards
Junyi

[Jason.Jonkman]

Dear Junyi,

I’m still not sure I understand your question, but a single MLife input file can be used to process many FAST time series data files.

Best regards,

[Junyi.Wu]

Dear Jason,

I am really sorry about my English! The problem that I wonder is in rainflow counting calculation, how to define the length of each time-series. Another one is that MLife is based on S-N curve approach, and in the manual I can not find anything about stress range. I just wonder how to get stress range in MLife. Thanks a lot.

Regards
Junyi

[Jason.Jonkman]

Dear Junyi,

Again, the wind turbine design standards (e.g. IEC 61400-1) specify loads-analysis requirements, including requirements on the number and length of the simulations needed per load case. For example, the load-case 1.2 for fatigue during normal operation involves running at least six 10-minute simulations per 2-m/s wind speed bin from cut-in to cut-out wind speed.

I’m not sure I understand your question about stress range.

Best regards,

[Junyi.Wu]

Dear Jason,

I just read your past paper and found the answer. Thanks a lot!

Regards
Junyi

[Junyi.Wu]

Hello Jason,

I am now running some cases by using MLife. However, it seems not going well. The Matlab said:

Attempted to access Fatigue.TimePP(828); index out of bounds because numel(Fatigue.TimePP)=784.

Error in compute_windspeed_bins (line 56)
Fatigue.TimePP(WSbin) = Fatigue.TimePP(WSbin) + Fatigue.ElapTime(iFile);

Error in mlife (line 565)
Fatigue = compute_windspeed_bins(Fatigue, windChannelMeans,FileInfo.DLCs(1).NumFiles, FileInfo.DLCs(2).NumFiles );

Error in er3 (line 1)
mlife(‘Test For TLP.mlif’, '.\Data', '.\Results\Testtlp');

I don’t know what’s wrong with my code. Could you help me please? Another question is how to define WSMaxBinSize?
Here is my Mlife file.

Regards
Junyi

[Junyi.Wu]

@Junyi.Wu

Hello Jason,

I am now running some cases by using MLife. However, it seems not going well. The Matlab said:

Attempted to access Fatigue.TimePP(828); index out of bounds because numel(Fatigue.TimePP)=784.

Error in compute_windspeed_bins (line 56)
Fatigue.TimePP(WSbin) = Fatigue.TimePP(WSbin) + Fatigue.ElapTime(iFile);

Error in mlife (line 565)
Fatigue = compute_windspeed_bins(Fatigue, windChannelMeans,FileInfo.DLCs(1).NumFiles, FileInfo.DLCs(2).NumFiles );

Error in er3 (line 1)
mlife(‘Test For TLP.mlif’, '.\Data', '.\Results\Testtlp');

I don’t know what’s wrong with my code. Could you help me please? Another question is how to define WSMaxBinSize?
Here is my Mlife file.

Regards
Junyi

MLife.txt (8.22 KB)

[Srinivasa.Sudharsan]

hi bonnie sir;
I am working with the MLIFE postprocessor while using i get the following error which prevents from creating the shortDEL file kindly help me to resolve this error:
Running MLife (v1.01.00a-gjh, 30-Oct-2015)

Processing file “F:\FAST8.10\MLIFE\mlife1\CertTest\Data\DLC2.3_1.out”.
Reading “F:\FAST8.10\MLIFE\mlife1\CertTest\Data\DLC2.3_1.out” (1.409390 MB).
Rows=1201, Cols=135
Done reading
Generating statistics for “F:\FAST8.10\MLIFE\mlife1\CertTest\Data\DLC2.3_1.out”.

Writing summary statistics to: F:\FAST8.10\MLIFE\mlife1\CertTest\RootFxc1_Statistics.txt
Writing summary statistics to: F:\FAST8.10\MLIFE\mlife1\CertTest\RootFyc1_Statistics.txt
Writing summary statistics to: F:\FAST8.10\MLIFE\mlife1\CertTest\RootFzc1_Statistics.txt
Writing summary statistics to:
  F:\FAST8.10\MLIFE\mlife1\CertTest\Test01_Summary_Statistics.xlsx

Undefined function ‘wblcdf’ for input arguments of type ‘double’.

Error in compute_windspeed_bins/compute_bins (line 74)
bins(iBin) = wblcdf(WShi, c, k) - wblcdf(WSlo, c, k) ;

Error in compute_windspeed_bins (line 32)
[nBins1, binWidth1, binProbabilites1] = compute_bins(0, Fatigue.WSin,
Fatigue.WSMaxBinSize);

Error in mlife (line 565)
Fatigue = compute_windspeed_bins(Fatigue,
windChannelMeans,FileInfo.DLCs(1).NumFiles, FileInfo.DLCs(2).NumFiles );

[Srinivasa.Sudharsan]

Hi sir;
I am working with Mlife but i have a doubt. I am working on controller to reduce the tower deflection and blade root bending moment. while using Mlife can we give the deflection as a fatigue input or only the force should be given as fatigue input. And also i need to know whether the Life time damage means the accumulated value of damage over 20 years of life time and also what does the time until failure? kindly help me with this aspect.

[Jason.Jonkman]

Dear Srinivasa,

I’m hoping that Greg Hayman can comment on your earlier questions, but I’ll respond to the questions from your most recent post.

The fatigue calculation is important for loads; I’m not sure there is much value in processing deflections for fatigue as there is nothing similar to e.g. an S-N curve for deflection.

The lifetime damage in MLife is based on the user-specified design life. We typically set DesignLife in the MLife input file to 630720000 s, which is equivalent to 20 years, but the user may specify this differently.

The time until failure computed by MLife is in seconds, so, hopefully this is more than the specified DesignLife.

Best regards,

[Greg.Hayman]

Dear Srinivasa,

wblcdf() is a MatLab function found in the statistics toolbox. You will need to have a license for this toolbox, or replace the wblcdf() call with your own code to calculate the weibull distribution.

Greg

[Jesse.Harrison]

@"[url=http://www.lawsuitsettlementfunding.co/

This[/url]"]

@Junyi.Wu

Hello Jason,

I am going to run a few cases by using MLife. However, it seems not going well. The Matlab said:

Attempted to access Fatigue.TimePP(828); index out of bounds because numel(Fatigue.TimePP)=784.

Error in compute_windspeed_bins (line 56)
Fatigue.TimePP(WSbin) = Fatigue.TimePP(WSbin) + Fatigue.ElapTime(iFile);

Error in mlife (line 565)
Fatigue = compute_windspeed_bins(Fatigue, windChannelMeans,FileInfo.DLCs(1).NumFiles, FileInfo.DLCs(2).NumFiles );

Error in er3 (line 1)
mlife(‘Test For TLP.mlif’, '.\Data', '.\Results\Testtlp');

I don’t know what’s wrong with my code. Could you help me please? Another question is how to define WSMaxBinSize?
Here is my Mlife file.

Regards
Junyi

Hello,
I am getting this error also. I don’t know how to solve it, I don’t understand. sorry for my bad English. Edit: actually, got it, thank you.

[Koen.Lane]

Dear Jason,

I am also interested in the question raised by O.P., what are the basic assumptions to use loads (e.g. blade root bending moments, tower base monents) instead of stresses for the evaluation of DELs?

Best regards,
Koen

[Jason.Jonkman]

Dear Koen,

I’m not sure I really understand your question, but as discussed in my post dated Mar 24, 2016 above, one can convert from a load to a stress via a cross-sectional analysis.

Best regards,

[Jiantao.Liu]

Dear @Jason.Jonkman

I am currently calculating the fatigue damage of the onshore NREL 5MW wind turbine blade by referring to the documents “Definition of a 5MW/61.5m Wind Turbine Blade Reference Model” and “The Sandia 100-meter All-glass Baseline Wind Turbine Blade: SNL100-00”. I have encountered a problem and hope to receive your guidance.
The “Definition of a 5MW/61.5m” document calculates the fatigue damage of the blade material in the flapwise and edgewise directions at eight airfoil sections located 0m, 1.3667m, 4.10m, 6.8333m, 10.25m, 14.35m, 18.45m, and 22.55m from the blade root for the NREL 5MW blade under DLC 1.2 conditions. The calculation results are shown in the figure below:

As shown in the figure above, the relationship between material fatigue damage and blade airfoil section position is as follows: the degree of blade fatigue damage first increases with the distance of the airfoil section from the blade root, reaches its maximum value at the 10.25m airfoil section, and then decreases as the distance of the section from the blade root increases. My current calculation results differ significantly from those in this document. My OPENFAST calculation divides the blade equally into 16 segments and outputs the flapwise and edgewise bending moments at the blade root and at the midpoints of these 16 blade segments (located at distances of 1.92m, 5.77m, 9.61m, 13.45m, 17.29m, 21.14m, 24.98m, 28.83m, 32.67m, 36.52m, 36.52m, 40.36m, 44.2m, 48.04m, 51.89m, 55.73m from the blade root). Following the calculation process of the “Definition of a 5MW/61.5m” document and the “The Sandia 100-meter” document, I calculated the fatigue damage for the main spar and trailing edge materials at each airfoil section. The calculation result shows that the maximum damage for the E-LT-5500 material in the edgewise direction occurs at the 11th section (36.52m), located in the mid-to-tip region of the blade. My main calculation process and intermediate results are as follows:

For DLC 1.2 conditions, wind speed conditions ranging from 3-25m/s need to be generated at 2m/s intervals. Since the material damage trends under different wind speeds are similar, the 11m/s wind speed condition is used as an example. The main calculation process is as follows:

1. Change the BldFlDmp value in the blade structural parameter definition file NRELOffshrBsline5MW_Blade.dat within OPENFAST to 1.5. Use OPENFAST to calculate the flapwise bending moment SpnkMLyb1(k=1-9) and the edgewise bending moment SpnkMLxb1 at each section.

2. Open the blade structural parameter definition file BldFile(1), extract the pitch axis position (PitchAxis), flapwise stiffness (FlpStff), and edgewise stiffness (EdgStff) at different blade sections. Use linear interpolation to calculate the pitch axis position and flapwise/edgewise stiffnesses at the midpoints of the 16 blade segments.

3. Open the blade aerodynamic-structural parameter definition file ADBlFile(1), extract the chord length (BlChord) and section type (BlAFID) at different blade sections. Use linear interpolation to calculate the chord length and thickness at the midpoints of the 16 blade segments.

4. Use the following formula to convert the flapwise and edgewise bending moment data output by OPENFAST into stress data for each material:
   

   
   
   In the formula, c for the flapwise direction is half the thickness of the airfoil section, while c for the edgewise direction is the blade chord length multiplied by (1 minus the pitch axis position).
   The bending moment data for sections 4 and 11 under the 11m/s wind condition for E-LT-5500 are shown in the figure below.
   
   
   
   The stress data for sections 4 and 11 of E-LT-5500 are shown in the figure below. Continue translating as per the above requirements.
   
   

5. The calculation method for damage in the “The Sandia 100-meter” document and the definition of the S-N curve are shown in the figure below:
   

   
   
   The definition of the S-N curve and the parameters related to each material in the “Definition of a 5MW/61.5m” document are shown in the figure below.
   
   
   
   
   
   
   The definition of the S-N curve and the calculation method for damage in MLife are shown in the figure below.
   
   
   
   Based on the above information, it can be concluded that C in Equation (2) of the “The Sandia 100-meter” document corresponds to Lult in MLife, and b corresponds to m. Therefore, the input content for the MLife file is as follows:
   
   
   
   
   
   
   
   

6. The fatigue damage for the E-LT-5500 material in the edgewise direction at each section, as output by MLife, is shown in the figure below.
   

   
   
   As shown in the figure above, the current results exhibit two issues: first, the damage to the material over a 20-year design life is too small; second, the location of maximum material damage is abnormal, as mentioned earlier. What might be wrong in my calculation process that could be causing these results?

Best regards,

[Jason.Jonkman]

Dear @Jiantao.Liu,

I’m not an expert on the SNL reports you are referencing and am not familiar with their details, so, I can’t really comment on how your analysis compares. Just a few comments on your approach that you describe:

• Is the approach you are using (stress = M * c / EI) the same as what is applied in the SNL reports. Seems like this approach is quite simplistic for a composite blade cross section and would important effects of the multi-component stress state, composite materials, and airfoil shape.
• Your time series samples look odd to me. I see some oscillation, but why is there a linearly growing trend as a function of time? I would think if you were simulating in turbulence for a given mean wind speed, the time series would be stochastically distributed about a given mean.
• When you calculate lifetime fatigue damage from DLC 1.2, you’d have to post-process time series across a variety of wind conditions, from cut-in to cut-out wind speed. Are you doing that? If so, how many conditions are you simulating?

Best regards,

[Jiantao.Liu]

Dear @Jason.Jonkman ,

Highly appreciate your advice and response. After inspection, I identified that the oscillation in the time series was caused by the excessive blade pitch angle setting (90°) during my OPENFAST simulation. Currently, after setting the pitch angle to 20°, the bending moment data for sections 4 and 11 under the 11m/s wind condition for E-LT-5500 are shown in the figure below.

The stress data for sections 4 and 11 of E-LT-5500 are shown in the figure below. Continue translating as per the above requirements.

In the DLC1.2 condition (calculated over the entire cut-in to cut-out range), the fatigue damage of the E-LT-5500 material at each airfoil section is as shown in the figure below.

From the above figure, it can be seen that the location of maximum damage for the current E-LT-5500 material is basically consistent with the ‘Definition of a 5MW/61.5m Wind Turbine Blade Reference Model’ document, but the issue of the damage value being too small still exists.I plan to adjust the damage to around 0.5 to 1 by increasing the safety factor. Highly appreciate your reply.

Best regards,