关键词: 光伏系统, 发电量预测, 混沌自适应粒子群优化算法, 反馈型神经网络

Abstract: Considering the characteristics of photovoltaic （PV） systems and various factors which will affect the power generation of PV systems, a short-term generation forecasting model of PV systems was proposed based on recurrent neural network （RNN） and chaos adaptive particle swarm optimization （CAPSO） algorithm. In the model, the weights and thresholds of RNN were optimized by the global optimization ability of CAPSO algorithm in order to avoid the disadvantages of the traditional RNN, such as slow convergence and prone to local minimum. The fuzzy membership function was used to process the data in the temperature evaluation, which can improve generation forecasting precision. Forecasting results show the high accuracy of the proposed model.

Key words: photovoltaic system, generation forecasting, chaos adaptive particle swarm optimization algorithm, recurrent neural network