Research On The Key Technologies Of Fault Diagnosis For Transmission And Distribution Network Oriented To Dispatching Emergency Disposal

Over the years,great progress has been made in power system dispatching automation and safety and stability control technologies,the online decision of the operators can not be replaced.When the power system encounters such big disturbance as complex fault,it is very important for the dispatching department to carry out emergency disposal work in time and effectively,thus to ensure the operation safety of the power system and avoid large-scale blackout.Particularly,the fault diagnosis technology is the most critical supporting technology for dispatching emergency disposal,it is of great value to improve the timeliness,accuracy and online adaptability of fault diagnosis.However,faced with the massive power system operation monitoring information and complex fault scenarios,it is a serious challenge for operators to accurately analyze the fault scenario and its safety level,and formulate emergency disposal strategies in a short period of time.Under this condition,it is of great theoretical and practical significance to study the targeted dispatching emergency disposal support system and its key technologies of fault diagnosis.By screening and processing all kinds of monitoring signals,the useful key information can be condensed and transferred to the operators.The processed information can provide decision support for dispatching operators in mastering and controlling the system status and taking emergency measures.Therefore,the pertinence and effectiveness of the emergency emergency disposal work can be effectively improved.For this reason,the research of this dissertation is focus on the key technologies of fault diagnosis for transmission and distribution network oriented to dispatching emergency disposal.Considering the demand of two-level dispatching of transmission and distribution system,the functional framework of dispatching emergency disposal support system is studied,and the key fault diagnosis technologies for both transmission network and distribution network are studied in this dissertation.Since the protective relay is the first defense line of system security,and the emergency disposal work can be meaningful only if the protection removes the fault.Therefore,in the first time of dispatching emergency disposal,the fault diagnosis only considers action and reported information of the protective devices to infer the real fault scenarios.As for the fault diagnosis of transmission network,the further research on performance improvement is carried out for both the analytic model based method and the Petri net graphical modeling based method,which are widely used at present,and the diagnosis performance can be greatly improved through the cooperation of the two kind of methods.As for the fault diagnosis of distribution network,the fault diagnosis method considering both the timeliness and fault tolerance is studied for the conventional radial distribution network,based on which the improvement scheme is proposed to deal with the characteristics of active distribution network with multiple power sources in parallel.The detailed researches of this dissertation are presented as the following aspects.First of all,the demand target of the dispatching emergency disposal support system is analyzed from the dispatching emergency disposal focus of the two-level dispatching of transmission and distribution system.Based on the three modules of data access,security early warning control and emergency control recovery,the functional framework of power system dispatching emergency disposal support system is carried out,which considers the coordination of transmission and distribution dispatching coordination.Besides,the key technologies involved in the support system are summarized and analyzed,and the technical objectives and overall implement scheme of the fault diagnosis in the dispatching emergency disposal support system is discussed.The basic goal of transmission network fault diagnosis is to estimate the actual fault scenario and its protection removal process,which is the important premise and basis for operators to deal with the fault urgently.Considering that the existing analytic model for transmission network fault diagnosis needs to expand the dimension of optimization variables,it is difficult and time-consuming to solve the model.Therefore,a staged analytical model for transmission network fault diagnosis is proposed in this paper.In the first stage of the model,the influence of different supposed fault elements on the objective function is analyzed,and the protection correlated index and the circuit breaker correlated index are combined to establish the fault measurement index,which can quickly screen the suspicious elements without iteration.In the second stage of the model,the actual state of the protection and the circuit breaker is introduced into the fault hypothesis,and the diagnosis objective function reflecting the action logic error and communication fault of the protective system is established,which can be solved by intelligent optimization algorithm,and the fault elements as well as the actual state of protection and circuit breaker can be obtained.Through the improvement of staged analytical modeling,the dimension of the optimization model can be effectively reduced,and the timeliness of fault diagnosis can be improved.Since the existing Petri nets graphic modeling based fault diagnosis methods for transmission network are mainly contructed based on single element offline,which requires high hardware of storage and is difficult to adapt to the changes of network topology.Therefore,a power grid fault diagnosis model based on the information fusion of topology graph elements is proposed.Based on the topological structure of the power grid,the topological correlation matrixes of the sections,protections and circuit breakers are formed.Taking the topological correlation matrixes as the graphic unit,the information fusion based diagnosis model is then constructed.Besides,according to the logic and export of the protective system,the topological mapping rules of the secondary backup protection and the complete reasoning flow are defined.The proposed mehtod can make full use of the network topology information to realize the online automatic modeling of the diagnosis model,and determine the fault elements quickly without traversal reasoning.Besides,the network topology description can be updated in the process of fault diagnosis,therefore it can adapt to the changes of network topology and is suitable for cascading faults diagnosis.Since the existing Petri nets based fault diagnosis methods are difficult to achieve efficient temporal reasoning,a fuzzy time Petri net(FTPN)based fault diagnosis method of transmission network is proposed in this paper.With the application of SOE in some substations,the alarm signal of the protection and breaker can be labeled with a unified time stamp,which can be fully used to improve the accuracy of fault diagnosis.Therefore,the timing attribute is introduced into the places and transitions of the Petri net model,thus to describe the timing constraint relationship of the alarms.Besides,the correlated reasoning operation of the confidence probability and the timing constraint is defined,and the hierarchical reasoning process of the FTPN model is established based on the model structure.The proposed algorithm does not need to check the temporal information of each alarm by the complicated forward and backward temporal reasoning,and can realize the efficient temporal reasoning process based on the matrix description of Petri net,and also it can make full use of the temporal reasoning results to improve the accuracy of fault diagnosis.In addition,the comprehensive application method of transmission network fault diagnosis technology in the support system is studied,and the cooperation mode between the analytical model method and the Petri nets graphical modeling method is proposed,which can improve the comprehensive performance of transmission network fault diagnosis..Different from the transmission network,the basic goal of distribution network fault diagnosis is to locate the fault section,based on which the operators can accurately isolate the fault and restore the normal area,so as to reduce the load loss and improve the reliability of power supply.Considering that it is difficult to reconcile both fault tolerance and timeliness in the existing diagnosis technology for transmission network,a fault diagnosis method for conventional radial distribution network is proposed in this paper,which combines both matrix algorithm and optimization algorithm.Firstly,the matrix description of distribution network is established based on the causal relationship,and a novel matrix criterion is constructed for fault daignosis.The realization process of the criterion is very simple and the internal meaning is clear,and the fault section can be located quickly and accurately with normal alarms.Besides,if there are distorted alarms within the fault scenario,the suspicious sections can be screened out according to the matrix criterion,and on this basis,the optimization model can be established using the matrix description to locate the fault sections with strong fault tolerance.The proposed method can make full use of the complementary advantages of matrix algorithm and optimization algorithm in timeliness and fault tolerance,so as to improve the performance of section location.In view of the problem that the current of active distribution network with multi sources in parallel can flows in two directions and the fault diagnosis method of conventional distribution network is difficult to apply,a fault diagnosis method suitable for the active distribution network with multi sources in parallel is proposed.With the development of renewable power generation technology,distributed generators and energy storage equipment are connected to some distribution networks,which changes the conventional distribution network into the active distribution network with multi sources in parallel.Therefore,based on the fault diagnosis method for conventional distribution network,the referenced positive direction of the switch current is defined for the active distribution network considering the directional characteristics of switch over-current alarm,and the extended matrix description of the active distribution network is established based on the causality.On this basis,the matrix algorithm criterion and fault-tolerant optimization model proposed above are respectively improved in this paper,and the particularity of multiple faults in active distribution network is alao analyzed,which can not only reconcile both fault tolerance and timeliness,but also be applied to the active distribution network with multi-source operated in parallel.Finally,the main work of the dissertation and the distinctive results are summarized.The functional modules of the relevant dispatching emergency disposal support system have been applied in a regional power grid in China.On this basis,the work prospect of the next step is given.