The increasing proportion of bifacial photovoltaic modules (Bi-PVM) in new projects makes the operation of photovoltaic system (PVS) more complicated, and it is difficult to accurately predict the power of the PVS. To solve this problem, this paper proposes a new power prediction method for PVS based on Bi-PVM. Firstly, we construct the equal proportion digital twin model of the example project, analyze the factors affecting the power generation performance of Bi-PVM by using the superposition principle, and construct the feature engineering according to the analysis results. Secondly, we reduce the parameter error caused by Bi-PVM to the prediction model by introducing the bifacial correction coefficient. On this basis, we establish a power prediction machine learning model based on bi-directional gated recurrent unit (Bi-GRU) network. Finally, we carry out a simulation experiment on Tensorflow machine learning platform. We use the actual operation data of a PV power station in Jiuquan, China to conduct simulation analysis under four weather types, namely sunny day, rainy day, snowy day and complex and changeable day, respectively, and demonstrate the correctness and excellence of the proposed method.

No abstract