Hetero-Junction with Intrinsic Thin-Layer Photovoltaic Module Temperature Prediction and Performance Evaluation using ANN and Hybrid ANFIS

Abstract Radiation intensity and panel ambient temperature affect solar power generation. Predicting solar panel temperature will enhance the power generation forecasts. This research work involves various data prediction models for estimate the temperature of photovoltaic (PV) modules, depending on ambient conditions like ambient temperature, wind speed, and radiation intensity. A total of 18,778 data samples were used to train and validate the Artificial Neural Network (ANN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) models for predicting the temperature of Hetero-Junction with Intrinsic Thin-layer (HIT) photovoltaic modules. Before implementing those algorithms, data pre-processing is performed to prepare the datasets, and the selection of input features is undertaken based on a correlation matrix. Results show the comparison of the coefficient of determination (R2) score, mean square error, mean absolute error, and mean percentage error. Overall, the results show that ANFIS nonlinear models predict better than ANN models. As R2 scores of ANN and ANFIS 0.88 and 0.94, respectively. These two models are the best-fitting models on actual data. Compared to model, the ANFIS model has a minimum Root Mean Squared Error of 3.75, and the model also achieved minimal deviations in efficiency (0.02 %) and power (0.33 %), performing better than ANN and other comparative models. The research reported in this study is valuable for predicting the temperature and power of Photovoltaic modules. It will also be useful for energy management.