Risk Assessment And Damage Prediction Model Of Typhoon Disaster

Typhoon disaster is the most frequent and serious natural disasters in the world.Typhoon accompanied by strong winds and typhoon precipitation often cause secondary disasters such as mountain torrents,landslides,debris flows,causing casualties,house damage,farmland disaster,seriously threatening the safety of life and property of human society,restricting the sustainable development of social economy.Guangdong Province is located in the southernmost part of the mainland of China,and tropical cyclones are extremely frequent.Therefore,based on the historical disaster data,natural geographical data,socio-economic data and remote sensing geographic data of Guangdong Province,this paper evaluates the risk of typhoon disasters in Guangdong Province by combining the evaluation model and the relevant technical methods of GIS.And the neural network algorithm was used to predict the losses caused by typhoon disaster,so as to provide scientific planning decision-making basis for the formulation of disaster reduction policy of Guangdong Province.The research work of this paper mainly includes:(1)I have analyzed and studied the theories and methods of natural disaster risk assessment and relevant technologies of geographic information system(GIS).In view of some problems,for example,the weight determination method in the existing risk assessment methods cannot give full consideration to both subjective and objective aspects,and the degree of refinement of risk model is not enough,this paper carried out research and improvement.Firstly,data were collected,including disaster and socio-economic statistics related to typhoon disaster,remote sensing map data and other geographical data in Guangdong province.I chooses disaster risk,environmental sensitivity of disaster pregnancies,vulnerability of disaster-bearing bodies and regional disaster prevention and mitigation capabilities as evaluation indicators.And chooses socio-economic and natural ecological factors related to the above evaluation indicators as their relevant factors.Then,in view of the shortcomings of the existing weight determination algorithm that it is difficult to take both subjective and objective aspects into consideration.I propose FN-AHP-EW subjective and objective weighting method to obtain the weight of the relevant factors of each evaluation index.Next we combine disaster index to determine the weight of the above four evaluation indicators.Finally,based on GIS,refined data processing and model establishment are carried out to obtain the Guangdong Typhoon Disaster Risk Assessment Model,which evaluates the regional risk of Guangdong Province.(2)The basic principles and characteristics of ANFIS neural network algorithm,PSO algorithm and chaos theory are deeply studied,and the existing defects of ANFIS and PSO algorithm are summarized.A new method named ICPSO-ANFIS algorithm based on Improved Particle Swarm Optimization(ICPSO)and adaptive neuro-fuzzy inference system(ANFIS)is proposed.The chaotic improved particle swarm optimization(ICPSO)algorithm is used to optimize the parameters of ANFIS to avoid falling into local extremum and improve the performance of the model.ICPSO-ANFIS algorithm is used to predict typhoon disaster losses.The experimental results show that ICPSO-ANFIS algorithm improves the prediction accuracy,provides a new method for the prediction of regional typhoon disaster losses.