Research On Power Grid Fault Diagnosis Technology For Complex Data Environment

Power grid fault diagnosis is one of the basic subjects about making intelligent decision for power grid dispatching.It can find the failure cause and the fault component through analyzing and judging the failure data collected by dispatching automation platform,which has a great significance to shorten the interrupt time and improve the efficiency of fault handling.With the development of smart grid,the diagnosis system is in the more complex data environment,which makes the uncertainty factor increasing during the process of fault diagnosis and makes the modeling and maintenance of the fault diagnosis system more onerous.In this context,the accurate solution and modeling of power grid fault diagnosis have been put forward more urgent practical demands.This paper summaries various methods of power grid fault diagnosis.And then aiming at improving tolerance,operational efficiency and maintainability of the fault diagnosis system,the design of the overall structure of power grid fault diagnosis model,the corresponding diagnostic modules and their solution methods,and modeling and testing of the diagnosis system are studied.The main work and achievements are summarized as follows:A hierarchical immune model for power grid fault diagnosis is proposed drawing lessons from the multiple important mechanism of immune system and its layered defense mechanism.This hierarchical immune model uses a two-layer structure and transmits diagnostic parameters,signals and results through the inter-hierarchical interaction.The first layer is the inherent diagnostic layer including the immune learning mechanism and the establishment and maintenance of the fault diagnostic knowledge base,which can diagnose rapidly known faults with model matching method.The second layer is the adaptive diagnostic layer,which can adaptively identify the new faults that are not recognized by the first layer.And it includes the identification of unknown faults using multi-objective optimization method,and the identification of multiple solutions using electrical quantities to judge multiple solutions.The hierarchical immune model also adopts the hierarchical diagnosis idea,which uses all kinds of fault information related to implement fault diagnosis of the corresponding immune layer.This idea makes reasonable use of the information resources,and improves the accuracy and efficiency of fault diagnosis effectively.In the inherent diagnostic layer,a power grid fault diagnosis based on humoral immune principle is proposed by using the principle of humoral immunity.The method firstly sets up the relationship between the process of humoral immune response and of power grid fault diagnosis;and then the method establishes a power grid fault diagnosis model based on the humoral immune response mechanism via simulating the mechanism and structure of humoral immunity resisting the invasion of antigens.This proposed model has a higher fault-tolerance ability,which can not only detect known faults according to the prior knowledge,but also judge unknown faults through the continuous learning function of the fault diagnosis system.The continuous improvement and supplement of fault knowledge can overcome the incompleteness of fault knowledge,which increases the maintainability of the diagnostic system.A fault diagnosis approach based on immune clonal constrained multi-objective optimization method is presented in the identification of unknown faults in the adaptive diagnostic layer.This method can effectively solve the problem that weight values in existing mathematical models for analytic model-based methods are remarkably affected by artificial subjective factors.Firstly,this paper analyses objective functions of original models,then employs multi-objective method to transform the fault diagnosis problem into a multi-objective optimization problem.Pareto method is used to solve the problem,avoiding errors caused by artificial setting of weights in procedure of fault diagnosis.Secondly,considering shortcomings of existing mathematical models applying penalty function method to deal with constraints,this paper transforms the constraints into an objective function,converting the constrained multi-objective optimization problem to an unconstrained one.And then an immune clonal constrained multi-objective optimization algorithm is put forward to achieve optimal solution.The procedure of the optimization algorithm for power grid fault diagnosis is formulated.Finally,fuzzy set theory is used to evaluate Pareto optimal solutions.The design method of electrical criteria for identifying suspicious fault components is studied in the identification of fault set in the adaptive diagnostic layer.For the transmission line,a new pseudo root identification algorithm is presented,and it can be constructed as the electrical criterion to identify the fault branch and the non-fault branch.For the busbar,the electrical criterion of sum of currents flowing into busbar is applied to judge.For the transformer,the electrical criterion of phase difference of fault components of positive sequence currents is used to identify.Combining the constructed electrical criteria with the developed analytic model-based method,a graded power grid fault diagnosis approach is proposed.This method employs a graded diagnostic model which can gradually call switch and electrical quantities in procedure of fault diagnosis.Firstly,switch quantities are implemented the first level fault diagnosis based on the developed analytic model-based method.Then electrical criteria are adopted to perform the second level diagnosis for possible multiple solutions.The graded diagnostic model enhances the pertinence of fault diagnosis and reduces the demand for electric quantities via gradually calling fault information,so as to effectively improve the efficiency of fault diagnosis.The hierarchical immune fault diagnostic model and its solution approach studied in this paper are simulated and tested by utilizing the simulation module of D5000,and the results verify the effectiveness and correctness of the proposed model and method.