The Research On Intelligent Fault Diagnosis Method Of Power Grid Based On Multi-source Information Fusion

As a kind of special clean and green energy,electric energy is widely used in national production.The progress of society,economy and many other aspects make the power system more and more complex.As the direct carrier of power transmissing,power grid will inevitably fail due to its long-term exposure to the natural environment.Because of the interconnection between power grids that becomes more and more close,the impact of power grid fault on power system is also increasing,which requires the dispatcher to quickly and accurately judge the fault in the massive abnormal data and take corresponding measures.Therefore,with the help of intelligent technology to quickly identify the fault components,eliminate the fault timely and restore effectively the normal operation of the fault area,it is particularly important.The traditional fault diagnosis methods of power grid mostly focus on the analysis of protection and circuit breaker action information.However,in the actual power grid operation,the correctness of fault diagnosis results is seriously affected by the occurrence of protection,circuit breaker misoperation and refusal to operate,information loss and temporal error.Electrical quantity has a good advantage in information accuracy and fault tolerance.Therefore,this paper introduces the information of electrical quantity to carry out comprehensive fault diagnosis of electrical quantity and switch value,and proposes an intelligent fault diagnosis method based on multi-source information fusion.(1)A multifractal fault diagnosis method based on electrical quantity is proposed.Firstly,Hurst index is used to measure the self similarity of the current signal.Then,the fault features of current waveform before and after fault are extracted by using variational mode decomposition and multifractal theory,and the fault degree is calculated by correlation coefficient and normalization.Finally,the effectiveness of the proposed method is verified by the fault analysis of the New England 10 machine 39 bus system.(2)A Petri net fault diagnosis method based on switch value is proposed.Firstly,the general temporal fuzzy Petri net model of components is established.Secondly,the time sequence constraint relationship of alarm information is introduced,and the time sequence constraint checking method is proposed.The time sequence constraint checking method is used to check the suspicious fault components respectively,which can automatically eliminate the wrong alarm information.Then,the initial confidence of the event is evaluated by using the time sequence correlation characteristics,and the fault probability representation of the components is obtained by reasoning in the form of matrix.Finally,through the test of several groups of simulation cases,it shows that the method can accurately identify fault components in simple and complex fault scenarios.(3)A fault diagnosis method based on information fusion is proposed.Firstly,aiming at the influence of abnormal evidence on fusion results,an improved D-S evidence theory method based on Cityblock distance function is proposed.Then,the improved D-S evidence theory is used to fuse the information of decision-making level according to the fault degree obtained from the analysis of electrical quantity and switch value,and the final result of fault diagnosis is obtained.The effectiveness of the proposed method is verified by a number of simulation cases,which shows that the fault probability of the actual fault components can be effectively improved and the identification ability of the fault components can be improved.