Research And Application Of Swarm Intelligence Algorithm In Short Term Power Load Forecasting

In the rapid development of economic and social situation,with the continuous progress of the construction of electric power enterprises,power industry scale expands day by day,and electric energy is difficult to long-term mass storage,Power short-term load forecasting has become an important part of the power management system.Accurate load forecasting,can make the power generation keep up with the change of the load of power system at any time and balance the power and the load of the line,and then help the power department to formulate the reasonable distribution plan,ensure the safe and stable operation of power system,reduce energy waste,improve social and economic benefits.Firstly,this paper outlines the research background of short term load forecasting of power system,development of load forecasting and research status at home and abroad,and introduces the classification,characteristics,influencing factors and error analysis and other basic theories of power load forecasting.Secondly,this paper focuses on the research of BP neural network(BPNN)in artificial neural network,including the neuron model,the network model and the training process of the network;The design of hidden layer number and number of neurons in the input layer,output layer and hidden layer and the choice of transfer function.Then,aiming at the shortcomings of BPNN method,such as slow convergence speed,trapped in a local minimum point easily,poor robustness,and so on,several main swarm intelligence algorithms,like genetic algorithm(GA),particle swarm optimization(PSO),and bat algorithm(BA)are proposed to optimize the weights and thresholds of BP neural network.After introducing the basic principles and characteristics of these algorithms,through the comparison of experimental performance,a novel swarm intelligence algorithm based on Simulation of echo location in bats,namely bat algorithm,has better performance.Because there are some problems in the basic bat algorithm,such as premature convergence,poor global searching ability,slow convergence speed in the later period and so on,the adaptive inertia weight coefficient is introduced to correct the velocity and position equation of the algorithm,and the selection of relevant parameters and the final velocity structure are determined by convergenceanalysis,so as to improve the convergence performance of the algorithm and realize the balance between the global and local search of the bat population.The improved BP neural network is used to optimize the structure and parameters of the neural network.This method can avoid the blindness of selecting the network parameters in order to achieve the efficiency and accuracy of short-term power load forecasting.Finally,in order to verify the performance of the IBA-BP load forecasting model established by the BPNN optimized by improved bat algorithm,not only compares it with BA-BP and classical BP neural network load forecasting model in the prediction results and the relative error,but also with other load forecasting models,like PSO-BP and GA-BP load forecasting model in terms of performance.,are respectively established by other swarm intelligence algorithms,such as particle swarm optimization algorithm and genetic algorithm for BPNN optimization.The power load data provided by the EUNITE contest is used as the experiment sample to carry out the simulation experiment,and predict the load value of the whole point at 24 hours of a day.The results show that the IBA-BP neural network load forecasting model has faster convergence speed and higher prediction accuracy,can be very good to the short-term power load forecasting,is widely applied in the power system of high accuracy requirements.