Study On Settlement Prediction Of Concrete Face Rockfill Dam Based On Multiple-monitoring Point Model

Dam safety monitoring is a necessary project for the whole life cycle safety management of concrete face rockfill dam(CFED).Through the systematic monitoring and data prediction of the dam structural state,it is helpful for the management to accurately evaluate the dam operation state,and provide abnormal warning and implement maintenance measures.With the development of automatic monitoring system,the number of dam deformation monitoring sensors has been greatly increased,and the method of dam structure health monitoring has gradually changed from the single monitoring point statistical model to the combination of multiple monitoring point(MMP)model and machine learning model.Therefore,it is of great scientific significance to carry out the research on the spatial-temporal distribution MMP model for analyzing the overall deformation trend of CFRD.The main contents of this paper are as follows:(1)The MMP model integrates spatial and temporal information to make predictions according to the corresponding relationship among the coordinate position,environment values,and settlement.Combined with the strong nonlinear mapping ability of PSO-SVM,the accuracy of MMP model fitting and prediction was improved.In order to define the selection range of monitoring points and impact factors in MMP model,five test schemes were designed,and the influence of monitoring points selection on the prediction performance was verified by example analysis.The results show that the prediction ability of the MMP model is greatly affected by the correlation degree of the monitoring points,and it is particularly important for the deformation monitoring model to reasonably select the points with high similarity as the training samples.(2)The MMP model is established to predict the long-term missing data in malfunctioning settlement sensors by using the method of time-space series data analysis to make up for the lack of collected data.In view of the mixed data in the training set of the traditional MMP model,panel data clustering analysis is used as the measurement method to determine the settlement similarity to screen the appropriate data.Taking the settlement monitoring data of Jishixia CFRD as an example,the effects of different prediction models,different clustering groups and different number of monitoring data on the prediction results of the model are compared,and the effectiveness and reliability of the model are verified and evaluated.The result demonstrate that the MMP model is suitable for the long-term data prediction of failures in rockfill dam settlement monitoring.After the spatiotemporal panel data clustering analysis,the model prediction accuracy is significantly improved.This model provides a new method for dam settlement prediction and analysis.(3)The MMP prediction model based on the physical cause analysis of CFRD settlement and the the expansion of spatial components.The influence of water level load transfer,rockfill rheology and soil properties on settlement during the operation period of impoundment is comprehensively analyzed,and the space-time distribution model of CFRD during the operation period under the action of multiple factors is established.The XGBoost model was used for fitting prediction and eigenvalue contribution analysis,and the model was evaluated by various performance indicators.The results show that the spatial parameters such as the upper filling height,rockfill thickness,panel-point distance and soil material accord with the deformation characteristics of rockfill dam,and the new influence factor form coupled with the spatial parameters has higher gain contribution to the model prediction.Taking the settlement monitoring data of Liyuan CFRD as an example,the new MMP model under the action of multiple factors can predict the settlement of full section points with higher accuracy,which has certain application and popularization value for related projects.