Using MLife - Input Files

Hello everyone,

I’m a new user to OpenFAST and MLife so I’m currently how these software work.
I was wondering what should be my input files in MLife to calculate the lifetime damage of a turbine. I know I need my time-series data of loading but do I also need to put the wind files associated with this loading ? Should I add an another input file ?

Thank you already for the answers,
Sarah

Dear @Sarah.Gilles,

I’m not sure I really understand your question, but typically the hub-height wind speed is included in the OpenFAST output file, so, MLife can access the wind data from that.

MLife needs to access the set of OpenFAST output files that you wish to post-process for fatigue.

I would review the MLife examples in the CertTest and the user and theory documentation of MLife for further guidance.

Best regards,

@Jason.Jonkman thank you for your quick answer. I
have some difficulties to understand the operating principle of the software.

If I got it right from the guide, I have to specify to MLife which load time series in the OpenFAST outputs files to take into account for the calculation of the lifetime damage. Based on those, Mlife applies the rainflow cycle counting algorithm to determine the cycles in the data. The part which is a bit blurry for me is how the software extrapolates the cycle counts over the design lifetime. MLife first matches the different load cycles from the time-series with the corresponding wind speed (included in OpenFAST output file). Then it uses wind distributions (such as Weibull distribution) to extrapolate the short-damage to the design lifetime. This extrapolation is also done for the power production time serie if I understood well ? It is also binned according to the wind speed.

And Mlife has a direct access to the wind data as well to the power data, there is no need to specify it in the input file ? The extrapolation process is automatically done ?

Best regards,
Sarah Gilles

Dear @Sarah.Gilles,

I generally agree with your description of how MLife operates, except that I would say that MLife is targeted at computing fatigue quantities like damage equivalent load and time until failure, which is post-processing generally applied to structural loads, not power production.

Best regards,

Dear @Jason.Jonkman, thank you for your answer.
I have an other question concerning the output files from Mlife. I was currently passing through the different tests available in CertTest and I was just wondering if I correctly understand the output data as I’m new in this domain.

For example, let’s take the test18. The Test18_Lifetime excel file shows two tables : Lifetime Damage and Time Until Failure.

  1. For the first table (Lifetime Damage), is the second line showing the total lifetime damage for different channels over all the time-series and the total lifetime period ? Also, is it expressed in N ?

  2. Concerning the second table, the second line shows the time until failure for the different channels but why is it bigger than the lifetime period ? Does that mean that the conditions given in the inputs files do not occur much fatigue damage ?

  3. Now if I take a look to the second sheet of the Test18_Lifetime excel file, the lifetime DELs, what does the n_ref represent ? And again, does the last line show the DEL for each channel and is it expressed also in N ? What does the number of equivalent cycles represent ?

I apologize if my question seems quite simple and straigthforward but I just want to make sure to understand well.

Kind regards,
Sarah Gilles

Dear @Sarah.Gilles,

Here are my responses.

  1. The lifetime damage is not variable “N” in the MLife Theory Manual; rather, damage is denoted by variable “D”. It is unitless where a value of D equals unity corresponds to failure.
  2. The time until failure is expressed in seconds (s). The time until failure would equal the use-specified DesignLife if D^Life = 1. In general, time until failure = DesignLife / D^Life.
  3. n_ref is the number of damage-equivalent cycles over the DesignLife. In this simple example, n_ref is very low because the mean wind speed of the Weibull distribution is very low (last worksheet) and only data for one bin is specified. Yes, the last row shows the DEL for each channel; the unit of the DEL will correspond to the unit of the load being post-processed, which for the blade root forces output from FAST is kN (not N).

Best regards,

1 Like

Dear @Jason.Jonkman

Can MLife be used to compute lifetime fatigue from the output of other time domain simulation programs? HAWC2 for instance?

Kind regards
Abhinav

Dear @Abhinav.Abeendranath,

Yes, MLife is a general purpose tool, not specifically tied to OpenFAST. That said, MLife requires the time series data to be in a format that MLife can read, so, that format is the only constraint. See the MLife documentation for more information.

Best regards,

Dear @Jason.Jonkman

Thanks for the reply. Good to know that MLife would potentially work with HAWC2 output as well. I will read the MLife documentation and see what can be done.

Sincerely
Abhinav