Strong Convective Weather Forecasting Model Based On Bp Neural Network Research

Because severely convective weather is more of a sudden, short duration, intensity, the consequences are quiet serious, and it is significant to make timely and accurate forecasting. Due to the instability parameters can objectively reflect the degree of atmospheric instability conditions, and are closely related to the occurrence of severe convective weather development, it can perform approaching weather forecasting using convective instability energy indicators, and the critical value for these indicators of strong convective weather play an important role in the prediction instructions. It is limited that how to find the critical value of energy index of instability only though artificial accumulation of experience and it is of research value to find new predication method using existing computer technology.This thesis introduces the weather concept of severely convective weather, the meaning of researching severely convective weather, and summed up the past few decades'the main technical of meteorologists forecasting severely convective weather. In recent years, progresses has been made to forecast severely convective weather by calculating the energy instability index, and the prediction results are satisfactory, as a result, this thesis is based on the calculation of the instability indexes of severely convective weather and performs further study of severely convective weather forecasting's methods. Because the artificial neural network technology has a strong nonlinear mapping ability and parallelism, adaptability, fault tolerance and self-learning ability, it is especially appropriate for solving complex non-deterministic causal reasoning, judgments, prediction and classification problems. BP Neural Network (Back Propagation Neural Network) is a one-way transmission of multilayer forward neural network, and it is a error back propagation algorithm. BP algorithm is the most widely used neural network learning algorithm, it can solve convective weather prediction problem which is pattern recognition in nature, so this thesis used the improved BP neural network to build strong convective weather forecast model.Finally we built model of severely convective weather based on BP neural network, in which live data will be observed as a panelist on the BP neural network model training and testing, we do comparison test for the trained model, the forecast results are ideal. In this thesis, the model of severely convective weather forecast based on BP artificial neural network provides a simple, objective, practical prediction method.