Research On Micro-grid Load Forecasting Based On Extreme Learning Machine

Making a reasonable and reliable generation plan is of great significance to the economic dispatch of microgrid system.With the popularity of smart grid,the demand for power is increasing,and the sharp increase of power load will have a huge impact on power system maintenance and capacity allocation.At the same time,the wide application of microgrid and renewable energy makes the load more vulnerable to the interference of meteorological and environmental factors,which increases the complexity of load forecasting and affects the forecasting efficiency to a certain extent.Therefore,accurate short-term load forecasting is of great significance for the stable operation of microgrid system and the reasonable distribution of power consumption on the supply and demand side.In this thesis,the short-term load forecasting of microgrid is taken as the research content,and the forecasting accuracy is deeply studied.This main research contents are as follows:Firstly,the load data obtained by the data acquisition device is affected by the factors such as power failure,system failure,etc.,which will lead to the problems of missing data,abnormal data and so on.Therefore,it is necessary to identify and process the abnormal data and missing data in the load data.At the same time,in order to speed up the training speed of the prediction model,the load data and environmental factor data are normalized.Secondly,because the stability of extreme learning machine(ELM)is related to hidden layer parameters,improper selection of hidden layer parameters often leads to large prediction error.Therefore,based on ELM,kernel extreme learning machine(KELM)is established by introducing kernel function mapping instead of random mapping of hidden layer parameters and improve the generalization performance.At the same time,based on the offline ELM model,an online sequential extreme learning machine(OSELM)model is introduced.When new samples are input,the online learning algorithm is used to update the offline model automatically,which improves the adaptability of the prediction model in the dynamic environment.Thirdly,prediction accuracy of short-term load is critical to the normal operation of the microgrid due to the strong randomness of load.KELM prediction model based on complementary ensemble empirical mode decomposition(CEEMD)and regional-division self-adapting variation particle swarm optimization(RSVPSO)is proposed.The load sequence is decomposed into several smooth subsequences by using complementary ensemble empirical mode decomposition to reduce the mutual influences among different local information.Aiming at the problem that particle swarm optimization is easy to fall into local optimization and is slow in converge,a inertial weight and learning factor based on regional-division are utilized to improve the global search ability and search efficiency.Furthermore,combining with the adaptive mutation operation,the prediction accuracy of the kernel extreme learning machine is enhanced,yet the problem falling into the local optimum is avoided.Finally,in order to improve the adaptability and forecasting efficiency of the load forecasting model under complex environmental factors,a microgrid load online forecasting model based on online sequence extreme learning machine optimized by principal component analysis(PCA)and particle swarm optimization algorithm(PSO)is established.Through an example,it is proved that principal component analysis can effectively reduce the input dimension of the forecasting model and eliminate the correlation between environmental factors.The optimized online sequence extreme learning machine algorithm has better prediction efficiency by dynamically updating the model training sample.