Study Of Fault Diagnosis In Power Grid On Information Fusion

Diagnosing the fault component accurately and real-time can reduce the power interruption time and enhance the reliability of power supply,which has great significance to the stable operation of power grid. When the power grid is at fault, large alarms and fault information swarm into dispatching centre in a short time,that are beyond the dispatching personnals’handling ability,even in the complex fault,it raises the difficulty to the fault diagnosis.The paper is based on this consideration,it brings the information fusion theory into fault diagnosis field to solve complex fault diagnosis.This paper goes on with new exploration and research on fault diagnosis, aiming at the currently existing problems,studys a method based on fuzzy integral information fusion of fault diagnosis in power grid,which combines with methods of flow fingerprint and RBF neural network,then gets the flow fingerprint matching degree and fault reliability separately,finally achieves decision-level information fusion on the above two fault using the fuzzy integral and getting the branch fault degree, at last realizes the power grid fault diagnosis.The major work of this paper I have done is as follows:1).It introduces the topic’s research background and the developing trend of both at home and abroad.2).From the perspective of biological and information theory.this paper demonstrates the information fusion technology’s feasibility and validity in fault diagnosis field,then introduces information fusion algorithm that often used.Combining them,choose fuzzy integral method for information fusion algorithm of this paper. It establishes the general framework of information fusion fault diagnosis based on the characteristics of the fault diagnosis.3).According to thought of fingerprint fault diagnosis,the paper researches the power grid fault diagnosis model based on the flow distribution,which generates the fault flow fingerprint database by the ways of DC flow method,then gets the flow distribution operating through the PMU real-time measurement information.next quickly dispatches with fault samples,finaliy obtions the flow fingerprint matching degree.By means of simulation results,proving the feasibility and insufficiency of the method. 4).This paper also introduces the artificial neural network and the advantages of the RBF neural network,it points out that the conventional neural network models have some diagnose problems.The RBF neural network model uses the switch information of PMU point for fault diagnosis.and gets the fault reliability by simulation calculation. By means of simulation results,proving the feasibility and insufficiency of the method.5).Using the fault diagnosis result as primary, which are flow fingerprint matching degree on electrical measurements and RBF fault reliability on switch information measurements,and then combining the power network topology to comprehensive diagnosis using the information fusion method,this is the method of information fusion fault diagnosisin power grid,whose principle and the concrete implementation steps are introduced. By means of simulation results,proving the feasibility and superiority of the method.