Analysis Of Spatiotemporal Characteristics And Diagnostic Criteria Of Distribution Network Fault

In our country, influenced by the old idea "attach importance to generation of electricity, undervalue power distribution", the developments of generation and transmission of electricity are advanced, but the development of distribution network is in the initial stage. In recent years, our country has paid much more attention on distribution network, and lots of new technology, new devices and advanced managements have been applied to distribution network. All of these new measures have improved the degree of automation of distribution network a lot. However, some problems appear with the automation of distribution network. Firstly, a great deal of data will be recorded with the application of automation equipments; Secondly, the operation of distribution power system automation system relies on the monitor and control units which are installed in the whole distribution network. When these monitor and control units can’t work well, or communication system is faulty, how should we guarantee the reliability of power supply?To solve these problems, this paper presents a theory that study the spatial and temporal characteristics according to the faulty data of distribution network which is recorded in the process of fault treatment. To analyze the spatial and temporal characteristics of distribution network fault, we design a software. This software can analyze the data recorded in the distribution automation systems such as GPMS and SCADA, and it can analyze the fault characteristics from many different aspects such as time, space, cause, relay protection and so on. The software can also analyze the characteristics among different fault properties. The analysis results can be output in many different ways. The main output modes include chart and statement, and these output modes make the analysis results more visual. This software is easy to master, and it can be used in many different areas.In addition to the study of spatial and temporal characteristics, this paper also studies the diagnostic criteria of distribution network fault with the method of data mining. We can find lots of useful information contained in the fault data with the method of data mining, and find the incidence relations among different fault properties. The process of data mining can be finished off-line, because the fault data is historical data. The more data is contained, the more precise and useful the analysis results will be. We can build a database of association rules with distribution network fault data, and diagnose distribution network fault with these rules. In this way, we can diagnose distribution network fault without real time monitoring information supplied by distribution automation system. The reliability of power supply will be improved a lot.We can get spatial and temporal characteristics of distribution network fault through analyzing fault data. Then study the incidence relations among different distribution network fault properties. These findings can diagnose distribution network fault in the absence of distribution automation system, and the analysis results can provide theoretical basis and technical support for other fault diagnosis methods.