Study On Ultra Short-Term Wind Power Prediction Based On Relevance Vector Machine

Wind power is the most large-scale development of new energy power generation and most commercial development way of generating,but wind power to a large depends on the weather conditions.With the increasing proportion of wind power in power grid,wind power grid will cause influence to the stable operation of power system.The accurately grasping of the trend of wind power changes for a very short period of time can reduce the need for wind power access to system reserve capacity,and reduce the neative influence of wind power grid in power system and improve the economic performance of power system operation.So it is necessary to research on a certain wind power ultra short-term wind power prediction.This article analyzed the influence of wind farms to the grid of the active power related factors,and a wind farm in northwestern China measured history data was used as the training sample.The BP neural network,SVM,RVM algorithm were set up the algorithm of short-term wind power prediction model and expounded the implementation steps in this paper.By analyzing the compare and contrast the simulation results of three kinds of algorithms and the advantages and disadvantages,this paper proposed a relevance vector machine algorithm based on mixed kernel function.Through the comparative analysis of simulation results,the simulation results verified the relevance vector machine based on the mixed kernel function is the effectiveness of the proposed wind power prediction algorithm,this algorithm calculation model error is small,high precision,the method is easy to realize.Finally,the wind power prediction software is designed,data can be flexible selected in the software for the wind power prediction and analysis of error online.Verifies the algorithm which helps more convenient and simplify test steps,more intuitive observation test results.