Remaining Useful Life (RUL) estimation using OpenFAST and MLife

Wind & Water General
1

Hello Dr. Jason,

I hope you are doing great!

I need your advice and suggestions on the following questions:

  1. Is it possible to determine the Remaining Useful Life (RUL) using MLife and OpenFAST?
  2. Is there any new NREL tools available that can estimate the RUL of components?
  3. Any relevant papers published by NREL on the RUL estimation process

As always, your input is very valuable. Please suggest to me your opinion.

Regards
Ipsita Mishra

2

Dear @Ipsita.Mishra,

Regarding (1), MLife can calculate the short term lifetime damage and time until failure (TUF), which are related to remaining useful life (RUL), but RUL is not directly calculated by MLife.

Regarding (2), pCrunch (GitHub - NREL/pCrunch) is a tool under active development that combines many of the features of MCrunch, MLife, and MExtremes, but is Python-based (rather than MATLAB based). Again, RUL is not directly calculated by pCrunch, but short term and lifetime damage and time until failure are.

Regarding (3), I know NREL has done some work on RUL for wind turbine drivetrains, but I’m not familiar with the details.

Best regards,

3

Thank you Dr. Jason for your response. I would also explore pCrunch and look for the publication you mentioned.

However, I have a conceptual doubt about the time until failure (TUF) and the remaining useful life (RUL). May be RUL is a subset of TUF or maybe not.
Time until failure: This is the predicted duration from the present moment until equipment is expected to fail or become unusable.
Remaining useful life (RUL): Length of time equipment will operate before it requires repair or replacement.
So, using Mlife, how can I derive RUL from TUF/lifetime damage? Can you give some clues on this?

4

Dear @Ipsita.Mishra,

My understanding is that:

RUL = TUF - current operating time

That said, I normally using MLife (or pCrunch) to predict TUF in the design phase based on design load cases, before any actual operation has taken place. So, to use this equation to compute RUL based on TUF would require that you are computing TUF based on actual operational data rather than based on generic design load cases.

Best regards,

5

Thank you so much for the clarification.
Even I read many papers where the authors are relying on real-operating data to estimate the RUL. But as you know getting actual operating data like SCADA data is seldom possible for a research student. So, I was thinking, if I develop a digital twin model of an actual operating wind turbine and then calculate the TUF on design load cases and find out the RUL, will this be a correct way of doing? or any suggestion you have to calculate RUL if I don’t have real operating data.
I apologize if the question is too silly and looks weir to answer.

6

Dear @Ipsita.Mishra,

If you don’t have access to real operating data and you want to calculate RUL, I would guess you’d have to rely on design load case simulations to approximate the data and assume some sequence of operating conditions after the turbine has been commissioned to approximate what happens to the turbine over time.

Simulations are often a great way to validate digital twin models, but if don’t have access to real operating data, I don’t see the value of creating a digital twin.

Best regards,

7

Thank you Dr. Jason for all your valuable input.