Producing perfect PV fleet forecasts – at what cost?


In our last blog, we talked about the performance of SolarAnywhere® v4 PV fleet forecasts, and discussed how accuracy was a function of the PV fleet’s geographic footprint. For instance, SolarAnywhere hour-ahead forecasts in the Western U.S. were shown to have a capacity-normalized mean absolute error (MAE) of 3.5% for a…

Continue reading