Study On Visualization Of Landslide Dam Formation Process And Rapid Assessment Method Of Dam Breach

Landslide dams are a major natural disaster that often occurs in mountainous area.Once a catastrophic dam-break occurs,it will cause great damage and seriously affect the economic and social development and ecological environment protection of the downstream areas.In order to reduce the damage caused by landslide dams,it is necessary to conduct a rapid assessment of the formation of dams,the stability of potential dams and breaching parameters before the occurrence of geological disasters,and to develop prevention and control measures in advance to minimize losses.Although scholars at home and abroad have made many achievements in the research of the above issues,the particularity of emergency rescue work of landslide dam requires a more rapid and accurate assessment of the hazard before the failure occurs.Therefore,further research is conducted on the problems of dam formation,stability assessment,and estimation of breaching parameters in this paper.A visualization simulation tool based on Arc GIS was developed for the formation process of landslide dam.For the dam breach assessment problem,a new assessment model of dam stability is established,and the existing estimation model of breaching parameters is improved.Eventually,a rapid assessment method was established for the process from dam formation to breach.The main research work and results are as follows:(1)A visualization simulation tool for the formation process of landslide dam was developed based on Arc GIS.Combining the calculating function of Mass Mov2D model with the visualization capabilities of GIS software,the formation process of landslide dam was successfully calculated and displayed on Arc GIS,and the applicability of the tool was verified by the Baige landslide.(2)The machine learning algorithm was introduced into the study of landslide dam stability assessment.Based on the dam case database,the decision tree algorithm of ensemble learning was improved by using the sparrow search optimization algorithm,and a rapid assessment model of dam stability based on the SSA-XGBoost was established.Compared to the common models,the new model was proved to have higher accuracy.The feature importance analysis revealed that the dam height has the greatest influence on stability among all parameters.(3)For unstable landslide dams,the existing empirical model for estimating the breaching parameters was improved by considering the riverbed gradient and the material composition of the dam.Comparing the estimation results of different models,it was verified that the estimation accuracy of the model in this paper is significantly improved.And a simplified model was established through the importance analysis of influencing factors.(4)The developed simulation tool for dam formation,stability assessment model and breaching parameters estimation model were combined into a rapid assessment method for dam breach,which is applied to the Caijiaba landslide.The results showed that three large deformation areas,namely Kuashan,Luojialiang and Xuetangyuanzi,may form landslide dams after sliding,and all three dams are unstable.In case of a breach,the maximum flood peak discharge is 719.7m~3/s.