Research On Parameter Identification Method Based On Gross Error Reduction Index And MIQP Model

Along with rapid development of power system,the structure and the operation mode is becoming more and more complicated in recent years.With scale of power system becoming enormous,automation level of dispatching center is increasingly improved.Many advanced appliances rely greatly on the accuracy of the network model and state estimation in Energy Management System(EMS).Network parameter errors are the main factors influencing the accuracy of the network model and state estimation.Therefore,it is very important to identify and correct the possibly erroneous parameter with relatively large errors.This dissertation makes a study of parameter identification methods based on the gross error reduction index and the Mixed Integer Quadratic Programming(MIQP)model.Specifically,the main focuses can be summarized as follows:1)A new parameter identification method based on the gross error reduction index is proposed.It studies and analyses this method from three aspects,which are theoretical derivation,case simulation and actual system simulation.Compared with the normalized Lagrange multiplier method,it verifies the priority on the aspects of identification efficiency and accuracy when this method identifies single and multiple erroneous parameters and bad data simultaneously.2)Joint parameter identification model on multiple measurement scans is proposed based on the parameter identification method on the gross error reduction index.It studies and analyses this model from three aspects,which are theoretical derivation,case simulation and actual system simulation.In addition,it verifies the priority on the aspects of identification accuracy when this model identifies single and multiple erroneous parameters and bad data.3)A new parameter identification method based on MIQP model is proposed.It studies and analyses this method from two aspects,which are theoretical derivation and case simulation comparison.In addition,it verifies the priority on the aspects of over-identification avoidance and identification efficiency on this method.