Grid-Type Fuzzy Models for Performance Evaluation and Condition Monitoring of Photovoltaic Systems

Abstract

This chapter describes a model-based approach to performance evaluation and condition monitoring of the photovoltaic systems as typical examples of multi-mode industrial processes. The main idea is to construct special grid-type fuzzy models with a partial (incomplete) fuzzy rule base, where the number of the fuzzy rules and their locations depend on the amount of the available data and their distribution in the input space. In the chapter a special iterative algorithm for learning the singletons of the fuzzy rules of these partial grid-type fuzzy models is presented and analyzed. Then each typical operating condition (mode) of the photovoltaic system can be represented by respective partial model and all these models create a Model Base. In the performance evaluation procedure, each new unknown operating condition is checked against all partial grid-type fuzzy models in the Model Base and they produce their own estimations. The model with the best estimation refers to the operation that is most similar to the new unknown operation. This concept is also used for creating a real-time condition monitoring system with a Model Base, consisting of four typical seasonal models and its performance is estimated by using real data. The main novelty in the chapter is focused on the concept and the learning algorithm for the partial grid-type fuzzy models as well as on the calculation steps for condition monitoring.