System of Automatic Detection of PV Mounting Configuration

A complete set of reliable system configuration parameters is a prerequisite for the holistic analysis of photovoltaic (PV) power production systems. However, in practice, it is common to have an incomplete or incorrect configuration on record. We propose an automatic system for determining basic mounting parameters using PV yield time series and location data. The system consists of quality control and inference modules, which we describe in this paper. The core inference module, based on solving an optimisation problem, is illustrated. Lastly, we demonstrate the utility of the proposed system with an evaluation exercise in the context of PV yield forecasting over a real asset portfolio.