| The fast and accurate diagnosis of power grid faults plays an important role in maintaining the reliability of power supply.Aiming at the problems that the current power grid fault diagnosis algorithm has poor fault tolerance and slow calculation speed,the fault type identification algorithm is susceptible to various noises and the lack of analysis of the fault cause,this paper designs a set of power grid fault diagnosis algorithms that integrate fault alarm information,fault current waveform and weather warning information.By adding fault diagnosis criteria and improving optimization algorithms,the current power grid fault diagnosis methods are optimized to realize the rapid and accurate identification of faulty components of the power grid,identification of fault types and quantitative analysis of fault causes.The main work of this paper includes the following aspects:Firstly,aiming at the problem of poor fault tolerance of the traditional power grid fault component identification algorithm for fault alarm information,the research is carried out to improve the fault tolerance of the algorithm by taking the time sequence of the action alarm information of protection and circuit breaker received by the power grid system into consideration.The alarm value and expected value of the fault alarm information sequence criterion are defined,and the difference between the two is used as the matching item of the alarm information sequence criterion and added to the original objective function,which can effectively reduce the probability of incorrect fault hypothesis being selected in the case of error or omission of alarm information and improve the fault tolerance of the algorithm for wrong alarm information.Secondly,aiming at the problem of insufficient speed and accuracy of the traditional optimization algorithm to solve the power grid fault diagnosis analytical model,the research on the power grid fault component identification algorithm based on the improved genetic algorithm is carried out.Explore the speed and accuracy of traditional optimization algorithms in seeking mathematical analysis models,and improve the basic genetic algorithm by designing crossover probabilities related to evolutionary algebra and adaptive mutation probabilities related to individual fitness which improves its search ability and convergence speed,so as to improve the speed and accuracy of the fault component identification algorithm.Thirdly,based on the identification of the faulty components,a method of identifying the fault type in the power grid based on the morphological-wavelet algorithm is designed to reduce the interference of signal noise on the identification of the fault type.Aiming at the various noises contained in the fault current,a multi-scale and multi-structural element compound morphological filter is constructed using semi-circular and rectangular structural elements to filter them.Then,use Haar wavelet and db4 wavelet to perform wavelet transformation on the filtered fault current,and use the transformed wavelet parameters to identify the fault phase and grounding type.The algorithm combined the morphological filter with wavelet transform,which has effectively improved the anti-interference ability and sensitivity of the algorithm.Finally,in view of the lack of analysis of fault causes in current fault diagnosis algorithms,a quantitative analysis algorithm for fault causes of line faults caused by meteorological factors has been proposed.Introducing meteorological information as the criterion for the analysis of the external causes of the fault,and then read the alarm information of the external meteorological factors associated with the faulty equipment to form the external meteorological factor failure risk assessment vector,and finally use TOPSIS method to comprehensively evaluate it to realize the quantitative analysis of the cause of the grid failure,which provides effective auxiliary information for the grid operators. |
PDF Full Text Request