Photovoltaic power prediction based on sliced bidirectional long short term memory and attention mechanism

• DOI: 10.3389/fenrg.2023.1123558
• Publication Date: 2023-03-16T04:36:31Z
• Authors: Wengang Chen; Hongying He; Jianguo Liu; Jinbiao Yang; Ke Zhang; Diansheng Luo
• Keywords: photovoltaic power generation system; A; sliced recurrent network; attention mechanism; 7. Clean energy; power prediction; bidirectional long short term memory; General Works

Solar photovoltaic power generation has the characteristics of intermittence and randomness, which makes it a challenge to accurately predict solar power, is difficult achieve desired effect. Therefore, by fully considering relationship between data climate factors, new prediction method proposed based on sliced bidirectional long short term memory attention mechanism. The results show that presented model higher accuracy than common models multi-layer perceptron, convolution neural network, memory. cyclic network high low root mean square error absolute 1.999 1.159 respectively. time cost only 24.32% 13.76% network.