Font Size: a A A

Research On Short Term Load Forecasting Based On Extreme Learning Machine

Posted on:2020-08-09Degree:MasterType:Thesis
Country:ChinaCandidate:W W WangFull Text:PDF
GTID:2392330620454946Subject:Electrical engineering
Abstract/Summary:PDF Full Text Request
As an important part of power system,the importance of power load is self-evident.However,because of its time-varying and instability,it is easy to cause fluctuation of power load curve,which is not conducive to the smooth operation of power system.Therefore,only by better,faster and more accurate forecasting of power load,can we escort the stable and reasonable operation of the entire power system,and thus better improve the safe and high-quality power supply for the national economy.Over the past decade,researchers at home and abroad have done a lot of research on load forecasting.Various load forecasting methods have emerged in large numbers.The representative methods are neural network method,support vector machine method and so on.However,the neural network method needs a large number of iteration operations as the basis,the overall structure is complex,involving more parameters,and the network learning speed is slow.Therefore,this paper considers the use of extreme learning machine model based on Particle Swarm Optimization on the basis of neural network,combined with actual power load data for short-term power load forecasting.Firstly,this paper gives a brief introduction to the artificial neural network,briefly describes its theory and derivation,and finally leads to the basic principle of the extreme learning machine used in this paper,and introduces its characteristics.Secondly,the real historical load data are preprocessed,the outliers of data samples are processed,and the data features are extracted by regression coefficient method and thermal chart,and the processed data are used for load forecasting later.Then,we try to achieve the combination of extreme learning machine algorithm and particle swarm optimization algorithm,and use the optimization ability of particle swarm optimization algorithm to optimize the random parameters of the existence of extreme learning machine to get a better extreme learning machine,that is,to complete a better prediction model.By combining the advantages of particle swarm optimization and extreme learning machine,it has the characteristics of simple parameter adjustment,global range optimization and strong generalization ability.Finally,the implementation process of the optimization of the extreme learning machine model based on the improved particle swarm optimization algorithm is described in detail,and the feasibility of the prediction model is verified by an example analysis using specific data.
Keywords/Search Tags:Short term load forecasting, Neural network, Extreme learning machine, Particle swarm optimization
PDF Full Text Request
Related items