Early Warning Research For Geohazards Based On The Dynamic Data-driven Technology

Monitoring and early warning of geological disasters plays an important part in the process of disaster prevention and mitigation.With numerous human activities to mountain region.Geohazards,more and more than before,causing great threat to the local people.Especialy in the Southwest China,the geological condition is complex,and earthquakes caused lots of geological disasters and Cracted Mountains,which can be easy to become debris flow under rainstorm.Under such background,this paper present a real-time monitoring and early warning system based on the factors of geologic disasters and formation mechanism,combined with the new dynamic data-driven technology.The system can deal with monitoring data quickly and dynamicaly,select warning model and evaluate the risk level to promote geological disaster prevention and mitigation by information technology.Main reseuls are:(1)A data-driven methodology for real-time monitoring and early warning has been presented.Data-driven methodology is a mathmetica model for the known system,which can detect and control the system faults.Thus,many advantages can be found by using data-driven technology in geohazards early warning.(1)Dynamicaly data-driven application system combines the simulation system with the real system.Simulation system can receive data and output response during operation process,and simulation result also can control the process in real system,which can make the caculation result more reliable.(2)Based on the dynamic data-driven application system,geohazards early warning system include four parts as follows: Data collection? History data analysis ? Build primary model(monitoring service flow,warning model service flow)? Real-time monitoring and feedback correction.(2)Summary the monitoring technology in geohazards,mainly analyze the methods for graghs.The monitoring types for slope include displacement monitoring,underground water monitoring and triggering factors monitoring.The debris flow include rainfall monitoring,material source monitoring,and movement feature monitoring.Common methods for monitoring data analysis as follows:(1)Comparative method,graphing method,statistical characteristics and influence factor method are the conventional methods for geohazards data analysis.(2)Common monitoring graphs include single index analysis and mutil-index combination analysis.The single graph includes rainfall,soil water content,pore pressure and surface displacement.The mutil-index combination analysis includes rainfall and soil water content,pore pressure and displacement for the interaction relationship understanding.(3)The intergration trend forcasting is important for geohazards displacement analysis,and the methods include grey model,multivariable correlation analysis,kalman filter method and time series analysis.(3)The processing of monitoring data by data-driven technology is to relize the fault diagnosis,error elimination and curve generation in a real-time and automatic way.Service engine has been used to establish the flows of monitoring data service for the key information extraction.(4)Present three levels early warning model for geohazards,including threshold warning,process tracking and temporal-spatial comprehensive analysis.Based on these model,the geohazards early warning model database has been established in order to realize the dynamic model calling,and optimal correction for higher accrate and reliable in a long monitoring time.(1)Threshold warning: Based on the understanding of formation mechanism,ensure the mainly control factors,and establish the warning threshold by statistics and numerical simulation.The methods mainly contain rainfall threshold,displacement threshold and velocity threshold.(2)Process tracking warning: Based on the real-time monitoring data,judge the displacement state of geohazards by mechanism and data analysis.The methods mainly contain grazing angle of S-T graph and T-t graph.(3)Temporal-spatial comprehensive warning: Based on the threshold warning and process thracking,the surface cracks has been added in the comprehensive warning.Because of the developing cracks indicate that the changes of geohazards,like the order,location and size of cracks.(4)Early warning system has been developed,which has already relized the data query,monitoring data analysis,graphs and automatic warning.Compararing to the traditional system,the system use the dynamic data-drvien technology to modify the system for a better use,higher efficient process and data service.(5)Heifangtai landslide and Wenjia gully have been selected as the application study areas in order to test the monitoring data service and warning model database.The system has successfully warned the rotational slide in Heifangtai region and debris flow in Wenjia gully,which showing its applicability to a certain extent.The results also testifies and supports,theoretically and practically,the presented methodology in this study.