| Intelligent monitoring and early warning of landslide disasters can reduce or avoid human and property losses to a large extent.Due to the complexity of the slope structure,the limitations of the sensors or observers,and the imperfection of the information acquisition technology,the changes of the slope monitoring parameters usually show random,imprecise,fuzzy and other uncertainties.However,the traditional landslide disaster prediction and early warning technologies and methods usually do not take the uncertainty of monitoring parameters into consideration enough,which makes it difficult to give accurate prediction and early warning results based on these monitoring parameters.Aiming at the uncertainty and timeliness of monitoring parameters,This thesis studies the method and application of landslide disaster early warning based on evidence reasoning,and the main work is as follows:(1)The prediction method of disaster-causing factors based on evidence reasoning.Considering the uncertainty of changes in disaster-causing factors(key monitoring parameters),an evidence reasoning(ER)model is built to describe the uncertain and nonlinear mapping relationship between historical,current and future disaster-causing factors;ER rule is used to fuse the evidence activated by the input quantity,and the prediction values of the disaster-causing factors in the future are accurately calculated according to the fusion result;The online updating method of ER model is given,and the model parameters are updated online according to the real-time change rule of disaster-causing factors.(2)Landslide early warning method based on evidence fusion of multiple disaster-causing factors.On the basis of prediction for the future change trend of disaster-causing factors in(1),the reliability distribution(stability evidence)of the predicted values of disaster-causing factors supporting the slope stability grade is extracted;ER rule is used to fuse the stability evidence provided by each disaster causing factor,and the landslide early warning decision is made according to the confidence maximization rule.The comprehensive early warning is carried out by integrating the prediction information of multiple disaster-causing factors,which effectively increases the comprehensiveness and reliability of early warning decision-making.(3)Landslide disaster monitoring and early warning platform.The wireless sensor network is constructed to conduct all-round dynamic real-time monitoring of disaster-causing factors through multi-source disaster factor information collection and transmission module;Combined Wireless Local Area Networks with Internet of Things technology,the collected data will be transmitted to the data server of the control center in real time,and the change trend prediction module will be constructed using the prediction method of disaster-causing factors proposed in(1)to predict the changes of disaster-causing factors in real time;Use the landslide early warning method proposed in(2)to build the landslide early warning module,and timely capture the precursor information of landslide disasters for early warning.By using the monitoring data collected on typical slope structures(engineering slopes and mine slopes),the above monitoring and early warning platforms and core intelligent fusion prediction and early warning algorithms are tested and applied,which shows the effectiveness of the platform and methods studied. |
