Medium And Long-term Power Load Forecasting Based On Fuzzy Genetic Algorithm And BP Neural Network

Power load forecasting is the premise of economic dispatching,real-time control,operation planning and development planning of power system.With the refinement of various work of power system,more and higher requirements are put forward for load forecasting.Improving the accuracy of load forecasting plays an important role in ensuring the safety,stability and operation of power grid.Especially the medium and long term load forecasting is critical to the future development of power grid,so it is extremely important to establish an accurate and reliable medium and long term load forecasting model.Based on the historical load data of Chuanhui District,Zhoukou City,this thesis predicts the load in January 2021.According to the characteristics of the historical load data in Chuanhui District of Zhoukou City,the BP neural network medium and long term load prediction model(BP model)was first established,and the load output results were predicted by the MATLAB simulation software.Then according to the BP model load forecasting root mean square error and average absolute percentage error and the actual value of the load results for comparative analysis and summary,the sum of the root mean square error and the average of the average absolute percentage error are 4.30% and 2.11%,respectively.The results show that the BP model has long step length,large error,low accuracy,slow convergence speed,and local vibration and local minimum problems in the training process.It shows that BP model prediction is easy to fall into the local minimum in the training process,and the global optimization ability of the model is poor.Secondly,the BP neural network is easy to fall into the local minimum problem,the global optimization ability is poor and the convergence speed is slow in the training,and the genetic algorithm is proposed to optimize the BP neural network.The medium and long term load prediction model(GA-BP model)based on genetic algorithm optimization BP neural network was established,and the load prediction results of genetic algorithm optimization BP neural network model were predicted by using Matlab simulation software.The results showed that the predicted load results of medium and long term load prediction of GA-BP model were higher than the actual load.According to the root mean square error and the average absolute percentage error of the GA-BP model and the BP model,the prediction accuracy is compared.The sum of the root mean square error and the average of the average absolute percentage error are 4.02% and 2.07%,respectively.The mean square error of the GA-BP prediction model is 0.04% lower than that of the BP prediction model,and the average absolute percentage error of the GA-BP model is 0.04%lower than that of the BP model.The mean square hex error of GA-BP prediction model is 0.28% smaller than that of BP prediction model,while the mean absolute percentage error of GA-BP prediction model is 1.37% smaller than that of BP prediction model.The results show that the prediction of GA-BP model is better than that of BP model,and has better prediction accuracy and convergence speed.However,local oscillations still occur in the training process,which is caused by the premature convergence of the genetic algorithm in the training process.Finally,for the problem of premature convergence of genetic algorithm,a medium and long term power load prediction model(FGA-BP model)optimized by fuzzy genetic algorithm(FGA)was proposed,and the load prediction results of optimized BP neural network model by fuzzy genetic algorithm(FGA)were predicted using Matlab simulation software.The prediction results show that the FGA-BP model has a high degree of fitting with the actual load,which is better than that of GA-BP model.By comparing the load,root mean square error and mean absolute percentage error predicted by FGA-BP model,GA-BP model and BP model,the mean of the sum of root mean square error and mean absolute percentage error are 3.70% and 1.80%,respectively.The RMS error of FGA-BP model is 0.32% lower than that of GA-BP model and 0.60% lower than that of BP model.The average absolute percentage error of FGA-BP model is 0.27% lower than that of GA-BP model and 0.31% lower than that of BP model.FGA-BP load prediction model is the optimal load prediction model.