Improved Extreme Learning Machine Power Load Forecasting Based On Artificial Intelligence Algorithms

Power load forecasting is important for power system security,stability and ensuring efficient operation,related to the normal functioning of all walks of life.At the state advocates energy conservation and environmental protection to save energy consumption of existing situation,the power load forecast accuracy is related to the economic,efficient operation and safe operation of power grids throughout the entire plant,that the current situation for electric power load forecasting accuracy put forward higher requirements.In this paper,extreme machine learning and artificial intelligence algorithms limit the establishment of a new electricity load forecasting model,designed to improve the accuracy of load forecasting.The main contents are as follows:(1)Extreme Learning Machine(ELM)is a single hidden layer feed forward neural network learning algorithm,is a more accurate nonlinear fitting method,and has good learning ability and generalization ability,so this article use ELM for power load forecasting.ELM is the same as the neural network,however,is based on the experience of minimization principle,so easily lead to excessive fitting,and extreme learning machine input weighting matrix and hidden layer offset for the model of random assignment,making the whole model to lack of specific learning the sample data,thus affecting its generalization ability.In order to improve the ELM learning ability and generalization ability,improving the power load forecasting accuracy,this paper first introduces Glowworm Swarm Optimization(GSO)extreme extreme learning machine(GSO-ELM),powerful ability of global optimization using GSO find the hours of training error which ELM model input weighting matrix and hidden layer offset matrix,a certain region in China for a period of time of power load simulation experiment,proves the validity and superiority of the model.(2)The GSO-ELM of power load forecasting model than the limit of the simple ELM model has achieved good result,however,prediction accuracy is still not very high,this is due to the defects of GSO.In view of the defects of GSO,the Artificial Fish Swarm Algorithm(AFSA)was introduced to the extreme learning machine,which composed of AFSA-ELM model,based on the same area at the same time the power load simulation experiment,not only verify the AFSA can enhance the learning ability and generalization ability of ELM,and verified the model of AFSA-ELM for power load prediction effect is better than that of GSO-ELM model.