Resumen
Hydropower is one of the most important renewable energy sources. However, the safe construction of hydropower stations is seriously affected by disasters like rockburst, which, in turn, restricts the sustainable development of hydropower energy. In this paper, a method for rockburst prediction in the deep tunnels of hydropower stations based on the use of real-time microseismic (MS) monitoring information and an optimized probabilistic neural network (PNN) model is proposed. The model consists of the mean impact value algorithm (MIVA), the modified firefly algorithm (MFA), and PNN (MIVA-MFA-PNN model). The MIVA is used to reduce the interference from redundant information in the multiple MS parameters in the input layer of the PNN. The MFA is used to optimize the parameter smoothing factor in the PNN and reduce the error caused by artificial determination. Three improvements are made in the MFA compared to the standard firefly algorithm. The proposed rockburst prediction method is tested by 93 rockburst cases with different intensities that occurred in parts of the deep diversion and drainage tunnels of the Jinping II hydropower station, China (with a maximum depth of 2525 m). The results show that the rates of correct rockburst prediction of the test samples and learning samples are 100% and 86.75%, respectively. However, when a common PNN model combined with monitored microseismicity is used, the related rates are only 80.0% and 61.45%, respectively. The proposed method can provide a reference for rockburst prediction in MS monitored deep tunnels of hydropower projects.