Research On Power System State Estimation With Non-Gaussian Error And Multiple Data

With the development of social economy,the requirement of safe and reliable operation of power system is higher and higher.The performance of state estimation directly or indirectly affects the establishment of real-time database,the accuracy of computing software and the reliability of the results.It is an one of the core functions of the energy management system,Load forecasting is an important part of power system economic dispatch,the accuracy of short-term load forecasting will directly affect the safe and economical operation of the system and the quality of power supply.Therefore,it is of great theoretical significance and practical value to research on load forecasting,state estimation and organic combination of them.On the basis of a large number of domestic and foreign literatures,this paper mainly carries out the research work from the following two aspects:(1)The problem of early-maturing can not be completely solved by using chaotic particle swarm optimization algorithm,which leads to the decrease of precision,when using least square support vector machine to do the short-term load forecasting of power system.In this paper,based on the logistic function,a particle inertia weight model is proposed to improve the weight adjustment of the existing chaotic particle swarm optimization algorithm.Then,the proposed inertial weight model,fitness variance and mean particle moment are introduced into the chaotic particle swarm optimization algorithm to enhance its optimization ability.The improved chaotic particle swarm optimization algorithm is used to predict the short-term load of power system based on least squares support vector machine.The feasibility and validity of the method are verified by the historical load data and meteorological data of a real power grid in East China.(2)In view of the fact that the current power system state estimation methods,the measurement error of non-Gaussian distribution is not considered thoroughly,and that there is no effective use of existing data such as load forecasting and historical data,this paper presents a two-layer robust state estimation.In order to deal with the problem of uncontrollable non-gaussian error in power system,a Power System Expectation Estimation method is proposed to further utilize load forecasting data,its own parameters and historical load data.Then improved the projection statistics proposed adaptive project statistics to fine-tune the projection statistics in the face of Jacobian matrices of different dimensions,and then designed a generalized maximum likelihood estimate of APS-LH type considering the non-gaussian error correction.So that,the proposed method can deal with the non-gaussian error in state estimation.Finally,the effectiveness and superiority of the proposed method are demonstrated by IEEE14,IEEE30 and IEEE57 cases.