A multi-layer perceptron based deep learning model to quantify the energy potentials of a thin film a-Si PV system

Abstract

The advantageous thin-film PV modules like a-Si suffer from degradation over a span of time resulting in less efficiency. A deep learning model based on Multi-Layer Perceptron (MLP) is proposed for forecasting year ahead solar radiation. After the forecasting, corresponding degradation-influenced energy of the a-Si PV system is evaluated. Clear sky Global Horizontal Irradiance (GHI) data and historical data are used as input for the prediction. The paper aims to show the comparison of the deep learning and traditional machine learning techniques in solving the aforementioned problem. The proposed deep learning method for solar radiation forecasting and degradation-influenced energy estimation outperformed the machine learning method, showing the efficiency of the proposed method.

Keywords

• Deep learning
• Machine learning
• Multi-layered perceptron
• Solar radiation forecasting
• PV energy estimation
• Degradation rate