Research And Application Of Short-term Load Forecasting Method In Power Plant

Load forecasting,especially short-term load forecasting,plays an important role in the operation and load distribution optimization of power plant generator sets.However,the increasingly complicated fluctuation law of power load and the increasing requirements of energy conservation and emission reduction bring great challenges to accurate power load forecasting.Aiming at the problems of large generalization error and local optimum of traditional forecasting model,based on intelligence theory,a short-term load forecasting method of power plant is studied.A short-term load forecasting model based on Improved Beetle Antennae Search(IBAS)to optimize BP neural network is proposed.The prediction results are applied to the load distribution of power plant generator sets,and good results of load prediction and distribution optimization are obtained.The main work completed is as follows:(1)Research on short-term load foundation forecasting model of power plant based on BP neural network.Based on the analysis of the annual load law,periodicity and holiday characteristics of a power plant,as well as considering the influence of temperature and date factors,the main factors on forecast are selected through principal component analysis.Based on BP neural network,short-term load prediction model is built.Also,BP network structure and input of the prediction model are established.By using practical data,the prediction performance of BP neural network is evaluated so as to summarize the defects and deficiencies of the BP prediction model.(2)Short-term load prediction of IBAS-BP power plant based on Improved Beetle Antennae Search algorithm(IBAS).In order to solve the problems of high randomness of power load and poor prediction accuracy of BP,the population elite strategy and nonlinear step decreasing strategy are introduced to improve the Beetle Antennae Search algorithm,and a short-term load prediction method is proposed to optimize the BP network by using the Improved Beetle Antennae Search algorithm.The experimental results and practical calculation verifies the effectiveness of the IBAS-BP model.(3)Optimal load distribution of power plant units based on IBAS-BP forecast results.IBAS-BP prediction results are applied to load optimization distribution of power plant units,and a distribution model is established based on unit characteristics.Particle Swarm Optimization is used to allocate unit loads.The experimental results show that the load distribution based on the IBAS-BP prediction results can improve the fuel efficiency and the operating efficiency of the unit.