The renewable energy industry has only recently started to rely on data-driven models on applications that have traditionally required complex physical solutions.
In this talk, we would like to show how we leverage from Spark, Keras and (in our case, on-prem) high performance computing (HPC) infrastructure to tackle a common and interesting problem in the wind-related industry (saving hours of CPU-consuming simulations).
The whole process is semi-automated at the moment of writing.
We are working on submitting a patent application for this work, and so we will be happy to provide more details regarding the use case and evaluation results very soon if required.
- Inspiration on how to use deep learning for this and similar use cases.
- Lessons learnt (e.g. limitations we have seen of using Hive, possibilities of data representation,…).
- Tailored-made infrastructure to learn hyper parameters of deep neural networks.
- Evaluation results.