Intelligent hybrid method to predict generated power of solar PV system

This paper presents a brand-new hybrid solar photovoltaic (PV) power forecasting model called empirical mode decomposition (EMD)-particle swarm optimisation (PSO)-adaptive neuro-fuzzy inference system (ANFIS). The model offers a solution to the challenge of accurately predicting generated solar PV power while considering the dynamic nature of environmental variables and solar radiation variability. As a solution, hybrid EMD-based PSO-ANFIS model is presented in this paper. Three different membership functions are compared to select the best input membership function, i.e., Gaussian. The input solar PV power data are broken down into intrinsic mode functions (IMFs) using EMD technique is fed into the PSO-ANFIS, along with important meteorological variables. The swarm optimisation is used to optimise the parameters of the ANFIS for enhanced accuracy. Utilizing empirical mode reconstruction of the ANFIS output, the predicted power of the solar PV system is computed. The suggested hybrid model's performance is evaluated and compared to alternative forecasting methods. It is discovered that the suggested model produces more accurate forecasts in terms of MAE, RMSE, and MSE. Additionally, the proposed model demonstrates robustness across various weather conditions, highlighting its applicability and effectiveness. Overall, this paper aims to explain the benefits of using a hybrid model instead of a standalone one, thereby enhancing the reliability and efficiency of solar PV power forecasting systems. Received on, 21 March 2025 Accepted on, 18 May 2025 Published on, 25 May 2025