A mutated hybrid Cuckoo Search Artificial neural network for Grid-Connected Photovoltaic system output prediction / Norfarizani Nordin

This thesis presents a hybrid technique for predicting the AC power output from a Grid-Connected Photovoltaic (GCPV) system. Initially, the prediction was conducted using six classical Multi-Layer Feedforward Neural Network (MLFNN) models. These models were developed based on different sets of inputs. A key feature for developing these models is the inclusion of time-series inputs. The inclusion of time-series inputs to the network is important as the solar irradiance, ambient temperature and module temperature have different time-constant; i.e. they have different rate of change as the climate changes.