Short-term PV power forecasting based on sky-conditions using intelligent modelling techniques

Abstract

The work in this paper involves intelligent modelling techniques i.e. fuzzy logic, Artificial Neural Network (ANN) and Adaptive Neural Fuzzy Inference System (ANFIS) methodologies for estimating the power in a solar photovoltaic (SPV) system. Since, the generation of power is subjective to environmental factors such as ambient temperature, variation in sky- conditions and solar insolation, therefore, an intelligent modelling techniques have been proposed for forecasting the power of a solar photovoltaic system employing 210 W Heterojunction with Intrinsic Thin layer (HIT) photovoltaic modules for different sky-conditions such as clear sky, hazy sky, partially foggy/cloudy sky and fully foggy/cloudy sky conditions respectively for composite climate zone and performance has been evaluated using statistical indicators.

Keywords: ANFIS (adaptive neural-fuzzy inference system); ANN (artificial neural network); forecasting; fuzzy logic; sky- condition; SPV system.