Study On Power System Fault Diagnosis And The Optimization Of Protective Relay Setting

Power system fault diagnosis is an aid of dispatcher while processing grid accident, it can shorten the processing time of the accident, so it’s the premise of quickly resume power supply after faults, it’s essence is use the action information of protective relay and circuit breaker to infer the possible fault location, identify faulty components, and analysis the behavior of the protection system. The algorithm is the heart of the power system fault diagnosis; optimization model-based fault diagnosis and expert system-based fault diagnosis are two major theories which are more mature currently. PSO algorithm and QPSO algorithm are important branch of swarm intelligence algorithm, they have the advantages of few parameters, high accuracy and fast convergence rate, this paper describes the basic principles and the process steps of these two algorithms.Through the analysis of the operation of the protective relay and circuit breaker rules, establish0-1integer programming mathematical model of fault diagnosis. Use PSO and QPSO algorithm to solve the problem of fault diagnosis, and discuss the impact of the different rounding ways. The optimization algorithm can improve the fault tolerance margin of information in the process of diagnosis; it also improves the reliability of the identification of the faulty components.In addition, the paper focuses on the problem of the optimization of inverse-time over current protective relay setting, consider the basic requirements and operating characteristics of protective relay, construct the mathematical model of the optimization of protective relay setting, which abstract "the four properties of protective relay" as to optimize target, and set the characteristic parameter values of relays in the system as the decision variables, establish a multi-objective optimization model combined with the distribution network example, this model reflects the basic requirement for protective relay of power system. Then, the paper select the QPSO algorithm to solve the problem of optimize protective relay setting.In view of the defect of the existing model which lack of basis while assign each target item weight, in order to assess the importance of the items in the objective function, introduce the concept of the grid connection degree indicators where protective relay installed. Use two separate populations to optimize time dial settings and current setting multiplier of the protective relay. The standard QPSO is possibly falls into a local optimum, so several improvements are made in order to capture the global optimum point stably. The standard test function and protection setting example show that those three modified QPSO algorithm’s performance of optimization is better than that of standard QPSO algorithm.Finally, have a brief summary of the research in this paper, and point out the issues that need to be further discussed.