Dear Jason,
Thanks a lot for your quick reply!
I have some more questions:
1- DEL about fixed mean or zero mean
Like in the previous question we are unsure about when it’s appropriate to use one or the other. If my understanding about fatigue theory is correct the DEL about a fixed mean is the same as a sinusoidal wave with a certain mean stress (L_MF) and an alternate stress (L_MF ± DEL/2) while the diference with the DEL about a zero mean is that there is no mean stress (0 ± DEL/2). I believe that a load with null mean stress causes less damage than another with it. So, is one more conservative than the other?
2- Ultimate Strength and S/N Curve Slopes
According to the article “Model Development and Loads Analysis of an Offshore Wind Turbine on a Tension Leg Platform, with a Comparison to Other Floating Turbine Concepts” in Section 4.3 it is said that no data on the component ultimate strength (LUlt) of the components of NREL’s 5-MW baseline turbine is available. Is it correct to assume that the component’s ultimate strength can be obtained from the extreme loads of the land-based turbine appearing in “Dynamics Modeling and Loads Analysis of an Offshore Floating Wind Turbine” – Appendix F?
The other uncertain factor in the fatigue analysis is the S/N (stress/number of cycles to failure) curve slope m, a material-dependent factor. Commonly, values of m = 3 for steel and m = 10 for composite materials are chosen. However, I’ve read that you can also conduct a parametric study were m varies from m = 3, 4, 5 for the steel components of the tower and LSS and m = 8, 10, 12 for the composite blades. So, which option would you recommend?
3- Aggregate short-term damage-rate
If I want to compute the fatigue damage suffered by the turbine in last year, how should I proceed? I have the wind speed at hub height in hour resolution from the entire 2022. Is the aggregate short-term damage-rate useful in this situation?
I hope I’ve been clear enough.
Best regards,
Ignacio
Dear Jason
Thanks a lot for your answers.
Please let ask you some more questions:
1-I’ve accomplished my parametric study with different values of m (Whöler exponent). The results show that the bigger m is the bigger the DEL become. I think this is the logical result to expect, but I want to make sure I’m correctly interpreting what the Whöler exponent means. In Spanish we use a different nomenclature and I’m not sure I’m correlating well. The Whöler exponent is a material dependent parameter, and its value indicates the slope of the S/N curve diagram. So a material with m=10 like fiberglass means that it can support more fatigue damage than a material with m=3 like steel. Am I right?
2- What I meant with my last question is that I’d like to perform a similar study like the one described in “Extrapolation of Extreme and Fatigue Loads Using Probabilistic Methods”, last chapter, “One-year Simulation”. My limits are that I only have access to data like wind speed at hub height with an hour-long resolution. This study is part of some fatigue preliminary calculations and that’s why I assume stationary conditions (mean wind speed changes each hour but turbulence is not considered during that hour) with a Normal Wind Profile (NWP from IEC 61400-1) considering the change of the shear exponent during each hour (it can be calculated because I also have data of mean wind speed at 10 m).
According to the article I mentioned before, it runs 10 min simulations with a total of 52,596. I don’t know how long the simulations should be in my case and how many should I perform.
Best regards,
Ignacio
Dear Jason,
Thanks a lot for your answers.
There is something I don’t quite understand. I’ve finished my parametric study where I analyzed through MLife the effect of the wind shear on short-term DEL. I’ve conducted a series of simulations with different wind shears for the same wind speeds. I’ve plotted the results and I observe a similar behavior in all plots, here is an example:
For m=3 the short-term DEL increase as the wind shear do, just like for the other values of m. When I analyze the Short-term Damage Rate for m=3, 4 or 5 it occurs that it increase with larger values of the short-term DEL just like it is supposed to do. But what I don’t understand is why does the short-term DEL increase with m but it produce less short-term damage rates. Here is the raw data of MLife:
I would expect higher short-term damage rates when the short-term DEL increase but somehow different values of m affect it in a way I don’t understand.
Hope you can help me.
Best regards.
Dear @Ignacio.Lopez,
The damage-equivalent load is rising with increasing value of m because the higher-amplitude cycles are more influential for higher values of m than for lower values of m.
The damage rate is dropping with increasing value of m because it takes more cycles to cause fatigue failure for higher values of m than for lower values of m.
Best regards,
Hi Jason.Jonkman,
Is that possible to calculate DEL for a bi-slope steel connection (e.g. m=3 and m=5) with MLife or PyFAST? if I want to implement that to the pyfast fatigue script code, do you have any literature for how to that, like formulas?
Thank you.
Dear @Milad.Moghtadaei,
Unfortunately, bi-slope SN curves have not been implemented in MLife or pCrunch at the moment. I agree that this would be a good feature to add, but I have not planned out the required formulas myself.
Best regards,
1 Like
Dear Jason sir,
I am using Mlife to calculate and compare the lifetime fatigue damage of fixed and floating openfast models, I tried calculating the Lult using the method of multiplier to time series maximum value upon the Mextremes calculated value . I am having this doubt," is the LUlt value calculated for YawBrMyn for fixed turbine applicable for the floating turbine as well". I want to calculate the time until failure for these two turbines for my controller.
Best Regards
Dear @Nitin.Sivakumar,
I would generally expect the loads at the yaw bearing to be a bit different between a fixed-bottom and floating offshore wind system (with the loads from the floating offshore wind turbine being a bit higher). So, if you are calculating LUlt in MLife via scaling of the highest loads, I would not use the loads from the fixed-bottom system for the floating system.
Best regards,
Dear Jason Sir,
Then do you think that if we calculate Lult seperately and calculate the fatigue loads and time until failure for fixed and floating, they are comparable?
I asked so because, currently I adjusted the Lult value and thereby calculated the time until failure for both floating and fixed, so do you recommend same scaling factor be used ,in order for them to comparable. Dont mind if my doubt is trivial. Thanks in advance
Best Regards
Also since I found that the changing of the Lult value can help change the fatigue loads and time until failure. I am unable to figure out what should be done to compare different cases. In one of the older posts, the author said about , setting one of the time until failure to 20yrs by adjusting Lult as a reference and use that Lult value so as to get other time until faiure . Is that recommended?
Thanks in advance.
Best Regards
Dear @Nitin.Sivakumar,
I would generally recommend to compute LUlt based on the cross-sectional properties of the member, e.g., as discussed in my post dated March 25, 2020 in the following forum topic: Mlife - User Defined Distribution. If that is not possible, scaling LUlt based on the highest load experienced across all load cases is an option.
Best regards,
