多模态数据融合驱动的电网强降水灾害风险预测模型分析

打开文本图片集

中图分类号：TM711 文献标识码：A文章编号:1672-3791(2026)02-0182-03

Analysisof Heavy Rainfall DisasterRiskPrediction Model for PowerGridbyDriven Multi-Modal Data Fusion

GAOZhen HAN ChenhuiMENG Jie DUPeide NIUYun

Shanxi Meteorological Service Center, Taiyuan, Shanxi Province, O3ooO2 China

Abstract: The terrain of Shanxi Province is complex,and extreme heavy rainfals posea serious threat to the safe operation of the power grid.This paper constructs a multi-modal data fusion-based risk prediction model for power grid disasters induced by heavy rainfall It integrates multi-source data such as meteorology,power grid operations,and geographical environment.A spatiotemporal data fusion methodis employed tooptimize feature representation,and combines deep learning frameworks including Long Short-Term Memory(LSTM)and Transformer to enhace the accuracyof heavy rainfalldisaster prediction.Experiments based on historical disaster datasets from Shanxi Province hasverified the applicabilityof different models in mountainous areas,basins,and river vallys, quantify prediction errors,and establisha high-precision power grid risk asessment system to improve disaster early warning capabilities.

Keywords: Multi-modal data fusion; Power grid; Disaster risk prediction; Heavy rainfall; Deep learning山西省地形以山地和高原为主，地势起伏，其强降水极端天气具有突发性和局部集中特征，对电网基础设施的稳定运行构成严重威胁。（剩余5046字）