Forecasting of Photovoltaic Output in Plateau Area Based on RBMO-LSTM

This paper provides a summary and comparison of various methods for predicting photovoltaic (PV) power. It focuses on summarizing the methods for predicting PV power using artificial intelligence and selects the long short-term memory (LSTM) algorithm as a means of predicting PV power. Following an examination of the fundamental principles of LSTM, this paper proceeds to analyse the function of the hyperparameters associated with LSTM, before proposing the RBMO(Red-billed blue magpie optimizer)-LSTM algorithm. The central concept of the algorithm is to utilise the RBMO optimisation algorithm to optimise the hyperparameters of LSTM over numerous iterations, with the objective of achieving the most optimal LSTM parameter configuration. In this study, the meteorological and photovoltaic data from Tibet were selected as inputs to assess the efficacy of the optimization algorithm proposed in this paper.