当前位置: X-MOL 学术Eng. Geol. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Landslide displacement prediction based on multi-source data fusion and sensitivity states
Engineering Geology ( IF 6.9 ) Pub Date : 2020-06-01 , DOI: 10.1016/j.enggeo.2020.105608
Yong Liu , Chang Xu , Biao Huang , Xingwei Ren , Chuanqi Liu , Baodan Hu , Zhe Chen

Abstract Owing to the complexity of the coupling relationship between multiple external triggering factors and the internal sensitivity state of landslides, it is difficult to accurately predict the displacement response of landslides to triggering factors, using existing methods. To overcome this setback, two new concepts, namely the trend sequence and sensitivity states, were introduced to quantificationally characterize landslide displacement caused by external factors and the landslide internal states, respectively. The support vector regression method was used to predict the trend sequence, while the long short-term memory neural network was employed to predict the sensitivity state. Thereafter, by fusing the predicted trend sequence and the sensitivity state, a non-linear model for landslide displacement prediction was proposed; moreover, with the Baishuihe landslide, located in the Three Gorges Reservoir area, as a case study, the proposed model was evaluated and validated using a large quantity of rainfall, reservoir water level, and displacement monitoring data, spanning a period of over 11 years. Based on the results obtained, the performance of the proposed model with respect to landslide displacement prediction was satisfactory. Furthermore, compared with three existing traditional prediction models of landslide displacement, the proposed model achieved a higher accuracy. Therefore, this study is helpful because it provides new insights that can be used to develop deep data-mining approaches for landslide displacement prediction.

中文翻译:

基于多源数据融合和敏感状态的滑坡位移预测

摘要 由于多种外部触发因素与滑坡内部敏感状态耦合关系的复杂性,现有方法难以准确预测滑坡对触发因素的位移响应。为了克服这一挫折,引入了趋势序列和敏感性状态两个新概念,分别量化表征由外部因素引起的滑坡位移和滑坡内部状态。采用支持向量回归方法预测趋势序列,采用长短期记忆神经网络预测敏感性状态。此后,通过融合预测的趋势序列和敏感状态,提出了滑坡位移预测的非线性模型;而且,以位于三峡水库区的白水河滑坡为例,利用跨越11年以上的大量降雨量、水库水位和位移监测数据,对该模型进行了评估和验证。根据获得的结果,所提出的模型在滑坡位移预测方面的性能令人满意。此外,与现有的三种传统滑坡位移预测模型相比,该模型具有更高的精度。因此,这项研究很有帮助,因为它提供了新的见解,可用于开发用于滑坡位移预测的深度数据挖掘方法。使用跨越 11 年以上的大量降雨量、水库水位和位移监测数据对提议的模型进行了评估和验证。根据获得的结果,所提出的模型在滑坡位移预测方面的性能令人满意。此外,与现有的三种传统滑坡位移预测模型相比,该模型具有更高的精度。因此,这项研究很有帮助,因为它提供了新的见解,可用于开发用于滑坡位移预测的深度数据挖掘方法。使用跨越 11 年以上的大量降雨量、水库水位和位移监测数据对提议的模型进行了评估和验证。根据获得的结果,所提出的模型在滑坡位移预测方面的性能令人满意。此外,与现有的三种传统滑坡位移预测模型相比,该模型具有更高的精度。因此,这项研究很有帮助,因为它提供了新的见解,可用于开发用于滑坡位移预测的深度数据挖掘方法。所提出的模型在滑坡位移预测方面的表现令人满意。此外,与现有的三种传统滑坡位移预测模型相比,该模型具有更高的精度。因此,这项研究很有帮助,因为它提供了新的见解,可用于开发用于滑坡位移预测的深度数据挖掘方法。所提出的模型在滑坡位移预测方面的表现令人满意。此外,与现有的三种传统滑坡位移预测模型相比,该模型具有更高的精度。因此,这项研究很有帮助,因为它提供了新的见解,可用于开发滑坡位移预测的深层数据挖掘方法。
更新日期:2020-06-01
down
wechat
bug