Need help
Hi everyone,
I'm a fourth-year undergraduate student pursuing a Bachelor's in Artificial Intelligence and Machine Learning, and I'm planning to do my final-year project in the wind energy sector.
The idea I'm currently exploring is building a hybrid digital twin for wind turbine predictive maintenance, where a physics-based simulator is combined with AI/ML models to detect anomalies, estimate component health, and eventually predict failures (e.g., gearbox, bearings, blades, etc.).
So far, I've started looking into OpenFAST as the physics simulation engine and have been reading about SCADA data, condition monitoring, and predictive maintenance. My tentative architecture is to use OpenFAST for the physics-based model and build AI models on top for anomaly detection and Remaining Useful Life (RUL) estimation.
However, I'm still in the early research phase and would really appreciate guidance from people working in this domain.
Some questions I have are:
- Is OpenFAST the right place to start for this kind of project, or are there other frameworks/tools I should explore?
- What are the standard software stacks used for building industrial digital twins in the wind energy sector?
- Are there any publicly available SCADA or condition monitoring datasets that are commonly used for research?
- What papers, books, courses, or GitHub repositories would you recommend for someone entering this field?
- Are there any common pitfalls or misconceptions that beginners should be aware of when building a digital twin for predictive maintenance?
My goal is to build something that's as close to industry practices as possible while remaining feasible for an undergraduate project.
I'd really appreciate any advice, resources, or suggestions. Thanks in advance!