Output power Prediction of a PV Power Plant by using MLP and MLP-ABC algorithms

Prediction of the output energy generated by a PV power plant in different times and various places have been of great importance due to fluctuations of the PV system performance in different geographic and climatic conditions. In this paper, the output power of a 3.2 kW PV power plant is predicted by using MLP (Multi-layer perceptron) and MLP-ABC (Multi-layer perceptron - Artificial Bee Colony) algorithms. In this regard we used the data gathered from photovoltaic power plant of Tehran University during four months between September 22nd, 2012 and January 14th, 2013. In the analysis, 10665 data which are equivalent to 35 days are used after validation. The output power of PV was predicted in different time periods of a day by constructing three models by using MLP-ABC and MLP algorithms (three models for each algorithm), which resulted in better precision by MLP-ABC with about 91.5% and 2.1 in terms of R2 (Co-relation Co-efficient) and MBE (Mean bias error) respectively. The accuracy gained by our proposed model for dividing the day into three durations is also increased by about 1 percentage in comparison with the model which is covering the whole day.

Output power Prediction of a PV Power Plant by using MLP and MLP-ABC algorithms Keywords:

Photovoltaic , Power Plant , Prediction , MLP-ABC Algorithm