Hyperbolic Hopfield Neural Network for Enhanced Solar Power Generation Forecasting
Raj Kumar Gupta, K. Naveen, S. Praveen Kumar, C. Yamini, Raghapriya NR, Vijay Kumar Dwivedi · 2024
Solar power is growing more essential for renewable energy, but measuring its output is difficult due to irradiance, temperature, and weather fluctuations. Planning and controlling energy distribution requires accurate short-term estimates. An innovative Hyperbolic Hopfield Neural Network (HHNN) model for short-term solar power prediction is presented in this paper. The procedure includes preprocessing, feature extraction, and model training. Remove linearly dependent characteristics and filter nighttime data during preprocessing. The HHNN model is trained after feature selection is optimised using PCA and wavelet decomposition. The new HHNN model surpassed the HNN and DNN models in solar power forecasting with an average accuracy of 92.41%.HHNN is a solid method for calculating solar power output. This model has improved solar power management systems by increasing projected accuracy.