I’m just starting my PhD research at the Cork Institute of Technology, Ireland.
My research is based on predictive maintenance of offshore windfarms, I’d like to build a machine learning and scheduling system that uses data from the farm or turbine (scada etc.) and combines this with a model of historical and real weather and ocean conditions.
I have plenty of weather, wind and ocean data, but I am finding it really very difficult to find any turbine data e.g. accelerator or sensor data that is already used in predictive monitoring systems.
Is there any freely available datasets that have this kind of turbine failure data? Some of the usual sources of data e.g. UCI or OpenEI have nothing suitable.
I have also contacted academics (whose papers I’ve read) and manufacturers and operators.
All are reluctant or cannot give me their data.
I don’t even mind if its very old data…
Any help would be greatly appreciated.
I suspect that maintenance data are kept secret in general, but I’ll ask our analysis team if they know anything.
My coworker, Ben Maples, said:
“Refer them to ECN. That’s where we got our data, and I think they got it from the Reliawind studies among others.”
He also said the our Gearbox Reliability Collaborative team may have something. I can contact them if it interests you.
You mean this company ecn.nl/energy-innovation/wind-energy/ ?
I’ll get in touch, thanks!
That would be really great. I’d appreciate it!
Seems that way, we (as a machine learning team) have just about given up on finding a reliable source of data and we were going to settle for a simulated model.
Maybe in a year two someone, local to us, may release something.
In the meantime, I’ll follow up with your pointers, thanks!
Your ECN link is correct. I have also sent email to two others here at NREL to see if they have anything to add to this discussion, so stay tuned.
Shuangwen Sheng suggested you try this web site: http://en.openei.org/datasets/node/970
Thanks on all accounts.
I am currently involved in a research project which has similar objectives. Can I still get access to the available experimental data?
You are resurrecting a quite old forum topic and Marshall Buhl has since retired from NREL. I suggest reaching out to others mentioned in this topic for help.
Thanks for your reply. I will try to reach out to the project members.