Research On Flood Disaster Risk Assessment Method

Zhejiang Province is one of my country’s economically developed coastal regions,which is more severely affected by natural disasters,especially typhoons.The heavy rainfall caused by typhoons induces floods and causes huge losses to people’s lives and social economy.Therefore,it is of great significance to carry out flood disaster risk assessment and prediction.In order to explore effective methods for predicting and assessing the risk of “typhoon rainstorm” flood disaster in Zhejiang Province,in this paper we construct a “typhoon rainstorm” flood disaster risk dynamic prediction and assessment model from a data-driven perspective and a knowledge-driven perspective.The main work of this paper is as follows:(1)Aiming at the task of identifying the main controlling factors of cumulative rainfall in different periods of the “typhoon rainstorm” flood disaster gray system,this paper uses the gray correlation analytical method combined with comparative experiments to determine the main controlling factors.(2)From data-driven perspective,a method based on the artificial neural network generated by one typhoon event to forecast another typhoon event is proposed,and the risk prediction model of mountain flood disaster is constructed,This model aims to rely on real-time rainfall forecast data to improve the accuracy of disaster prediction.The reliability of the model is verified through 2 cases,and the AUC value obtained by using ROC curve to evaluate the performance and results of ANN model are all above0.8,the model is verified to be reliable;(3)Yongjia County is selected as the research area,and the hourly rainfall forecast data is based on RUC,the artificial neural network generated by the mountain flood scene training under Typhoon Meranti conducts dynamic prediction and evaluation of the flood disaster risk in different rainstorm scenarios of Typhoon Lekima,The dynamic results of mountain flood risk prediction are generated.The results show that as the prediction time goes on,the prediction results are also dynamically changing and are close to the time and location of the actual disaster.In summary,the dynamic model of mountain flood disaster prediction can effectively and dynamically predict the impact of“typhoon-heavy rain” flood disaster.(4)From knowledge-driven perspective,we use Analytic Hierarchy Process(AHP)to construct a flood disaster risk dynamic assessment model in which an objective evaluation index system factor weights dynamically updated with the changes of“typhoon-heavy rain” scenarios.The results show that,compared with the model where the weights are not updated dynamically,the evaluation results obtained by updating the weights of the factors adapted to different rainstorm scenes have a higher accuracy.The above methods have certain application value and reference value for disaster risk assessment applications.