RESEARCH ON THE APPLICATION OF A BILSTM-INFORMER HYBRID MODEL IN REAL-TIME ENHANCEMENT PREDICTION OF PHOTOVOLTAIC POWER GENERATION OUTPUT

Engineering notes

Engineering notes will be added by the Power for Solar editorial team.

Chinese explanation / 中文解读

中文解读待补充：本站会优先为光伏效率、钙钛矿太阳能电池、储能技术、太阳能热利用、BIPV、并网技术等高价值论文补充中文说明。

Original abstract

Accurate prediction of photovoltaic power generation is crucial for ensuring the safe and stable operation of the power grid.To address the shortcomings of traditional Long Short-Term Memory (LSTM) networks in capturing long-term temporal correlations in power sequences, this paper designs a hybrid architecture (BiLSTM (Bidirectional Long Short-Term Memory)-Informer) that integrates a bidirectional LSM network and an Informer to improve the real-time prediction performance of power generation.The innovation of this method is reflected in three aspects: First, by integrating the bidirectional context-aware capability of BiLSTM and the long-range dependency capture mechanism of Informer, the expressive power of time series features is significantly improved; second, a dynamic data augmentation strategy is introduced, utilizing real-time noise injection and sequence reconstruction techniques to optimize the robustness and generalization ability of the training process; finally, an adaptive loss function and parameter tuning scheme are designed to address the inherent non-stationary fluctuation characteristics of photovoltaic data.Experimental verification based on actual photovoltaic power plant operation data shows that the hybrid model constructed in this paper performs excellently in key performance indicators, with a mean absolute error (MAE) of 0.52 kW, a root means square error (RMSE) of 0.68 kW, and a mean absolute percentage error (MAPE) of 0.71%, all outperforming all compared benchmark models.The results fully demonstrate the significant advantages of the proposed method in terms of prediction stability and real-time performance, providing reliable technical support for the engineering practice of intelligent energy control systems.

Links and sources

• PDF from original source

Comments