结合MIDAS仿真和BP神经网络的输电杆塔基础滑坡风险评估

近年来，暴雨及洪涝等异常灾害使得滑坡现象频发，导致野外输电杆塔发生倾斜、位移及沉降，严重威胁电网安全运行。为解决现有输电杆塔基础滑坡风险评估方法存在的普适性差、主观性强等问题，本文构建了一种结合仿真计算和数据驱动的杆塔滑坡风险评估模型。首先，将杆塔基础荷载纳入杆塔基础滑坡影响因子，结合MIDAS软件仿真得到坡体安全系数。其次，基于卫星、气象等多源信息融合，提取8 个杆塔基础滑坡影响因子，搭建按误差逆传播算法训练的多层前馈（back prop-agation,BP）神经网络模型。最后，对比测试样本集安全系数的仿真结果与模型预测值，误差范围为0.01%~0.24%，并结合合成孔径雷达干涉技术 (interferometric synthetic aperture radar,INSAR)验证了模型有效性和可行性。研究结果表明利用该神经网络模型可实现对任意输电杆塔基础所在坡体安全系数的精准、快速预测，同时具有普适性强的特点，研究成果可为输电线路防灾减灾提供新思路。

Landslide Risk Assessment of Transmission Tower Foundation Based on MIDAS Simulation and BP Neural Network

In recent years, abnormal disasters such as rainstorm and flood caused frequent landslides, which led to the inclination, displacement and settlement of transmission towers in the field, seriously threatening the safe operation of power grids. In order to solve the problems of poor universality and subjectivity of the existing landslide risk assessment methods for transmission tower foundation, this paper constructs a tower landslide risk assessment model combining simulation calculation and data drive. First of all, the tower foundation load is included in the tower foundation landslide influence factor, and the slope safety factor is obtained by combining MIDAS soft-ware simulation. Secondly, based on the fusion of satellite, meteorological and other multi-source information, 8 tower foundation land-slide impact factors are extracted and BP neural network model is built. Finally, the simulation results of the safety factor of the test sample set are compared with the predicted values of the model, and the error range is 0.01%~0.24%. Combined with INSAR, the va-lidity and feasibility of the model are verified. The research results show that the neural network model can be used to accurately and quickly predict the safety factor of the slope where any transmission tower foundation is located, and has the characteristics of strong universality. The research results can provide a new idea for transmission line disaster prevention and mitigation.