ML as a tool is being used in general solar energy research. e.g. one recent application is in digital twin implementation of photovoltaic power plants using ML models to detect faults, optimize maintenance, scenario simulation etc. DTs involve heavy use of ML and is a multi-faceted domain from data acquisition, analysis, visualization, and using the data to train an ML model and update it with real-time measurable parameters like, Temperature, windspeed, Voc, Isc etc.

The DT field can still be considered in its infancy.

Similarly, for power electronics, I recently read a paper that implemented a DT of a Boost converter ( "DC-DC Boost Converters Parameters Estimation

Based on Digital Twin" ). There are other works as well that are using ML based DTs for different power electronic circuits.