Design And Implementation Of Wind Farm Power Forecasting System Based On BP Neural Network

With the increasing installed capacity of wind power in China,wind power has become a part of the total electricity consumption of the society.However,due to the volatility and intermittent of wind power,it is difficult to integrate wind power into social power.Therefore,accurate prediction of wind power is of great significance.This thesis establish the wind power prediction system of Sichuan Dechang Wind Farm in accordance with the needs of China Institute of Water Resources and Hydropower Research.The system can predict the wind power of Sichuan Dechang Wind Farm in the next 24 hours.At the same time,the user can determine the corresponding power dispatching and power management scheme based on historical meteorological data and system analysis results,and improve the immediacy and accuracy of power regulation measures.This thesis follows the software engineering specification,on the basis of the system feasibility analysis,the requirement analysis of the wind power prediction system of Sichuan Dechang Wind Farm was carried out,including functional requirements and non-functional requirements;the software architecture design,problem domain design,persistence design,system interface design,and BP(Back Propagation)neural network model weight and threshold optimization design are introduced;according to the wind power data push module,wind power information display module,wind power power prediction module and wind power information statistics module of the system designed system test cases and test them requirement;according to the preliminary requirement,design and test,the wind power prediction system of the entire Sichuan Dechang Wind Farm is realized.In order to satisfy the requirement of the system and the performance of the system in Sichuan Dechang Wind Farm,by comparing and analyzing the wind power prediction models of wind farms at home and abroad,this thesis finally used BP neural network model to predict the wind power of wind farms.Aiming at the problem that the weight and threshold have large randomness in the actual process,the improved fruit fly optimization algorithm(IFOA)is used to optimize the weight threshold of BP neural network model to establish IFOA-BP model implements automatic optimization of weights and thresholds.The experimental results show that the IFOA-BP model has better predictive ability than the BP neural network model and the improved BP neural network(FOA-BP)model based on the traditional fruit fly optimization algorithm;Compared with the improved BP neural network(ABC-BP)model based on artificial bee colony algorithm,and IFOA-BP model is better than ABC-BP,under the System performance.At present,the system operation is normal,and the wind power prediction effect of Sichuan Dechang Wind Farm is good.It can provide digital decision support for wind power prediction of Sichuan Dechang Wind Farm.