摘要
为有效掌握古滑坡复活体发育特征及其变形发展规律,在古滑坡现场调查、无人机影像解译基础上,先开展复活体发育特征分析;其次,结合地表变形监测成果,通过FA-GRU-CT开展滑坡变形预测,并以预测结果开展其后续变化趋势等级划分。分析结果表明:古滑坡具显著的复活特征,按照地形、变形差异,主要复活体可分为6个区,且结合各区发育特征,其总体呈由前向后的牵引变形特征;同时,通过FA-GRU-CT的变形预测,得到预测结果的相对误差均值介于2.08%~2.17%,训练时间介于180.46~200.58 ms,具较高的预测精度和较快的收敛速度,并由预测结果计算得到滑坡变形趋势评价指标R的范围为0.789~1.081,变化趋势等级为Ⅱ级~Ⅳ级,说明此滑坡后续变形仍具较大的增加趋势,潜在失稳风险较大,建议尽快开展其防治处理。研究为此类古滑坡防治提供了一定的理论基础,具有推广应用价值。
In order to effectively learn the development characteristics and deformation development laws of ancient landslide resurrection bodies,this article first conducts analysis of the development characteristics of resurrection bodies based on on-site investigation of ancient landslides and interpretation of drone images.Secondly,based on the results of surface deformation monitoring,landslide deformation prediction is carried out through FA-GRU-CT,and its subsequent trend classification is carried out based on the predicted results.The analysis results show that ancient landslides have significant resurrection characteristics.According to differences in terrain and deformation,the main resurrection bodies can be divided into 6 zones,and combined with the development characteristics of each zone,the overall characteristics are characterized by traction deformation from front to back.At the same time,through the deformation prediction of FA-GRU-CT,the average relative error of the prediction results was between 2.08% and 2.17%,and the training time was between 180.46 ms and 200.58 ms,with high prediction accuracy and fast convergence speed.Based on the prediction results,the evaluation index R for landslide deformation trend was calculated to range from 0.789 to 1.081,with a change trend level of II to IV,indicating that the subsequent deformation of this landslide still has a significant increasing trend and a high potential risk of instability.It is recommended to carry out its prevention and treatment as soon as possible.Through the study of this article,a certain theoretical basis has been provided for the prevention and control of ancient landslides,which is worthy of promotion and application research.
作者
张莹
张宁晓
ZHANG Ying;ZHANG Ningxiao(The Fourth Geological Exploration Institute Co.,Ltd of Henan,Zhengzhou 450000,China)
出处
《甘肃科学学报》
2024年第4期26-32,40,共8页
Journal of Gansu Sciences
关键词
古滑坡
复活体
发育特征分析
神经网络
变形预测
Ancient landslides
Resurrected body
Analysis of developmental characteristics
Neural network
Deformation prediction
作者简介
张莹(1986-),女,河南新郑人,工程师,研究方向为水文地质、工程地质、环境地质及生态修复。E-mail:xiaohema19901210@163.com。
