基于遥感解译的盐湖地区输电线路杆塔地面沉降易发性评价  

Susceptibility Assessment of Land Subsidence of Transmission Line Towers in the Salt Lake Area Based on Remote Sensing Interpretation

作　　者：金必晶 殷坤龙[1] 桂蕾[1] 赵斌滨[1,2] 郭宝瑞 曾韬睿 Jin Bijing;Yin Kunlong;Gui Lei;Zhao Binbin;Guo Baorui;Zeng Taorui(Faculty of Engineering,China University of Geosciences,Wuhan 430074,China;China Electric Power Research Institute,State Grid Corporation of China,Beijing 100192,China)

机构地区：[1]中国地质大学工程学院,湖北武汉430074 [2]中国电力科学研究院有限公司,北京100080

出　　处：《地球科学》2024年第2期538-549,共12页Earth Science

基　　金：国家电网公司总部管理科技项目(No.52280721000A),合同编号:SGQHDKY0SBJS2100034。

摘　　要：跨越察尔汗盐湖地区的750 kV柴鱼输电线路是国家西部能源运输通道上重要的一环,受盐湖地区特殊的地质环境与人类活动影响,使得部分杆塔塔基发生不均匀沉降,严重威胁到输电线路的安全运行.针对盐湖地区目前存在的杆塔地基变形破坏问题,利用小基线集合成孔径雷达干涉测量(SBAS-InSAR)技术对杆塔基础变形失稳前2018年的Sentinel-1A数据开展遥感解译,获取了盐湖地区地面沉降分布情况.基于频率比法,筛选出与地面沉降相关性较强的8种评价因子构建盐湖地区地面沉降易发性评价指标体系,采用多层感知器神经网络(MLPNN)、逻辑回归(LR)、贝叶斯网络(BN),对比分析了盐湖地区地面沉降的易发性评价效果和精度.评价结果表明,MLPNN、LR、BN的评价精度均较高,分别为0.85、0.84、0.82.这表明,通过遥感解译获得地面沉降样本数据与机器学习相结合的方法是盐湖地区输电线路杆塔地面沉降易发性评价的有效手段;同时,评价结果可为输电线路杆塔监测、运行管理及新塔选址提供参考.The 750 kV Chaiyu transmission line across the Qarhan Salt Lake area is an important part of the energy transportation channel in the western part of the country.Affected by the special geological environment and human activities in the salt lake area,some tower foundations have uneven settlement,which seriously threatens the safe operation of the transmission line.Aiming at the problem of deformation and failure of tower foundation in salt lake area,small baseline integrated aperture radar interferometry(SBAS-InSAR)technology was used to carry out remote sensing interpretation of Sentinel 1A data in 2018 before the deformation and instability of tower foundation,and the distribution of ground subsidence in salt lake area was obtained.Based on the frequency ratio method,eight evaluation factors with strong correlation with land subsidence were selected to construct the evaluation index system of land subsidence susceptibility in the salt lake region.The multi-layer perceptron neural network(MLPNN),logical regression(LR)and Bayesian network(BN)were used to compare and analyze the evaluation effect and accuracy of land subsidence susceptibility in the salt lake region.The evaluation results show that the evaluation accuracy of MLPNN,LR and BN is high,which are 0.85,0.84 and 0.82,respectively.This shows that the method of combining the sample data of land subsidence obtained by remote sensing interpretation with machine learning is an effective means for evaluating the susceptibility of ground subsidence of transmission line towers in the salt lake region.At the same time,the evaluation results can provide reference for transmission line tower monitoring,operation management and new tower location.

关 键 词：盐湖地区 杆塔地基变形破坏 遥感 频率比 机器学习 

分 类 号：P627[天文地球—地质矿产勘探]