Construction of a Soil and Water Loss Probability Model and Landslide Risk Prediction for Hybrid Pumped Storage Engineering Based on an Online Monitoring Database

Lei Lei1, Xiaolong Wei1, Liang Wang1, Qingyun Chen1, Lv Wang1, Duozhi Kang2
1State Grid (Xi’an) Environmental Protection Technology Center Co., Ltd., Xi’an, Shaanxi, 710100, China
2Electric Power Research Institute of State Grid Shaanxi Electric Power Co., Ltd., Xi’an, Shaanxi, 710100, China

Abstract

In the context of building a new type of energy system, pumped storage projects have been widely adopted as a form of energy storage with the most mature technology and the most economical investment. In this paper, a hybrid pumped storage project online monitoring system based on multisensors is proposed, and an online monitoring database is designed and constructed. Based on the data in the online monitoring database, the soil erosion of the hybrid pumped storage project is calculated and analyzed by combining the CSLE model. Then the attention mechanism is combined with BiLSTM model to construct the landslide risk prediction model of hybrid pumped storage project. The soil erosion during the construction of the hybrid pumped storage project is mainly distributed on the construction land, garden land, grassland and cropland, among which the construction land has the largest area of soil erosion (132.19 km²), followed by the area of soil erosion of cropland (29.24 km²). The MAPE is between 0.002% and 0.005% when predicting landslide risk deformation of hybrid pumped storage project using CNN-BiLSTM-ATT model. And using the model in this paper can minimize the error of rainfall on the prediction of landslide risk deformation and realize the safe and stable construction and operation of hybrid pumped storage projects.

Keywords: online monitoring database, CSLE model, attention mechanism, hybrid pumped storage project, soil erosion, landslide risk prediction