The Controlling Factor Of Geological Disaster Of Typhoon Rainfall Spatial Distribution And Dynamic Prediction In Wenzhou Based On BP Neural Networks

Wenzhou is located in the southeast coast of China, mainly mountainous and hilly terrain, geological disasters are mainly caused by rainfall. The mainly factor which caused disaster are typhoon activity and the heavy rainfall associated with typhoons in the region. Typhoon rainfall seriously threatens to people's life and property. Traditional weather forecast typhoon rainfall areas in the accuracy is not high, the need for new means of Wenzhou typhoon rainfall forecasting.This paper studied typhoon rainfall forecast in Wenzhou, Zhejiang Province. First, filter and summarize typhoon rainfall data, analyze the characteristics of typical typhoon and rainfall in Wenzhou. Then based on the artificial neural network theory, analyze the BP neural network principle and characteristics, according to the requirements initially set the network structure and training parameters, and use MATLAB to prepare building, training BP neural network program.Then, select representative rainfall data as a sample of typhoon rainfall spatial distribution prediction and dynamic changes prediction. The sample will be divided into training and test samples. Using the BP neural network model of typhoon rainfall prediction in Wenzhou, adjust the parameter so as to achieve the best effect. Finally, use the samples to predict, and analyze the result. And to judge the accuracy of BP neural network method based on the available date to determine the applicability of the established model.This study shows that, based on BP neural network method, according to the existing typhoon information and rainfall data in Wenzhou, it's more accuracy in case of typhoon rainfall spatial distribution prediction and dynamic changes prediction to predict whether a typhoon in the rain. But in single station rainfall forecast, the error is larger and the prediction reference value is low.