Research On Landslide Monitoring And Early Warning Technology Based On Internet Of Things

Landslide is the third largest geological disaster ranked after earthquake and volcano,and it is also one of the geological disaster most directly affected by human activities.The intensity,scale,economic losses and casualties of landslide disasters in China are among the highest in the world.It is necessary to carry out landslide disaster prevention and control to reduce disaster losses.As an important means of landslide disaster prevention and control,landslide monitoring and early warning,which can achieve the maximum benefits at the least cost,has gradually become a research hotspot in the field of landslide prevention and control at home and abroad in recent years.With the development of technology,landslide monitoring and early warning has entered the era of Internet of things.Landslide monitoring and early warning system based on Internet of things is widely used,and has achieved good results in practice.However,there are still three problems: First,the robustness of landslide monitoring and early warning system is poor,which cannot meet the needs of working in the wild;Second,the communication protocol is incomplete,resulting in "information island" and lack of "four remote functions" which stands for "telemetry "," remote signaling "," remote adjustment " and " remote control " which is not conducive to information sharing and system management;Third,the application scope of landslide early warning based on threshold method is limited.To solve the above problems,the landslide monitoring networking technology,landslide monitoring communication protocol and landslide online early warning technology based on time series observations are studied in this thesis.Firstly,by introducing adaptive wireless networking technology,a low-power adaptive wireless networking landslide monitoring network is developed to improve the robustness of landslide monitoring system.Secondly,a completely landslide monitoring communication protocol is designed based on the investigation of the existing landslide monitoring system and the protocol stack is also developed to solve the problems of "information island" and lack of "four remote functions".Finally,the fast deployed landslide monitoring system based on the adaptive networking and landslide monitoring communication protocol is developed,and the landslide online prediction model based on displacement time series observations is studied.The achievements and conclusions are as follows:(1)A landslide monitoring system based on low-power adaptive wireless networking technology is constructed.The system adopts the network design of the combination of cluster wireless mesh network and mesh gateway backbone network,realizes multichannel redundant communication and ensures the reliability of communication.The landslide monitoring system based on adaptive wireless network includes three working modes: normal communication,fault communication and central communication.The three working modes are automatically switched according to different working environments,which effectively solves the problems of poor robustness and poor continuous working ability existing in the conventional landslide monitoring network.(2)The design principle and reference protocol model of landslide monitoring communication protocol are proposed,the landslide monitoring communication protocol meeting the "four remote functions" is formulated,and the communication protocol stack is developed.The protocol has been adopted by Yunnan Geological Environment Monitoring Institute as the geological disaster monitoring standard of Yunnan Province,and has been applied by nearly 1000 equipment at hundreds of geological disaster monitoring points in Yunnan Province,which effectively solves the problems of "information island" and lack of "four remote functions" and promotes the unified and efficient management of the landslide monitoring and early warning system.The protocol has been promulgated as an industry standard by China Geological Disaster Prevention and Control Engineering Association in 2019,and has been popularized and implemented throughout the country.(3)KM-FFT-SVM(Kalman filtering-fast Fourier transform-support vector machine)landslide disaster early warning model is proposed.Through the analysis of landslide surface displacement data,the model extracts the frequency domain landslide precursor information from the acceleration data of displacement to realize real-time landslide early warning.The KM-FFT-SVM model is trained and verified by using the landslide displacement data measured in the study area.The results show that the landslide real-time alarm technology based on KM-FFT-SVM model can realize highprecision imminent landslide early warning for landslides without three-stage change characteristics or emergency monitoring.Compared with the threshold early warning method,it has a wider application range.(4)A simple and fast deploying monitoring device is designed,and the fast deploying monitoring and early warning system based on the above research is built and applied in the research area.The system works stably and monitors the second sliding of Baige landslide in Jinsha River,which provides necessary early warning information for landslide emergency relief.