Fault Diagnosis Of Smart Grid Based On Intuitionistic Uncertainty Rough Sets

Recently, Smart Grid has been paid much attention because of the ever increasing of the energy consumption, more pressure on the environment and rapid development on new energy technology. Presently the researches on Power Grid are targeting on traditional grid but barely on Smart Grid. One of the main reasons is that Smart Grid is being developed without being constructed or utilized. Another reason is that Smart Grid’s quite different with traditional grid, some fault diagnosis methods of which will not be applied to Smart Grid and new methods have to be brought forward to adjust to this change. With a large number of distributed generators brought into Smart Grid being the main difference, the participance of the distributed generation (DG) makes the original distribution systems which are traditionally designed as radial systems into the systems with multiple power supplies. This thus makes the power flow appear as multiple directions and decreases the detection sensitivity of the main power supply fault current. When fault happens in power systems, a great deal of useful information may be lost due to improper measurement and transmission, which causes the traditional fault diagnosis methods based on relay protection devices not applicable and creates lots of uncertain factors and fault conditions which are hard to judge. All the issues above bring enormous difficulties to the fault diagnosis of Smart Grid. So fault diagnosis methods which can deal with the uncertain information are of magnificent significance.This thesis firstly applies the rough sets in fault diagnosis systems. After the simulation in the specific fault diagnosis system, the rough sets have been proved realistic value and application value. Also, the application of fuzzy theory and rough sets theory in the fault diagnosis of traditional power grid has been studied. With the simulation in the power grid, it has been validated that the two theories are effective in the fault diagnosis of traditional power grid. Combining the definition and characteristic of Smart Grid, this thesis shows a possible structure of Smart Grid. And then, the uncertainty rough sets have been brought forward which contain rough sets theory and fuzzy theory to conduct fault diagnosis on Smart Grid. However, owing to multiple power flows and some uncertain information, this method has been proved not good enough for fault diagnosis on Smart Grid.With the impact of random factors added, considering the features of Smart Grid and especially the denying information judging factors, this thesis puts forward the amelioration of uncertainty rough sets--intuitionistic uncertainty rough sets and a way to reduct which realizes the designing of reduction methods of original rules on fault diagnosis on Smart Grid. At the same time, every important conditional property has been rationally converted into the form of intuitionistic uncertainty rough sets. Also, the nodepend-degree and norely-degree contained in this method effectively solve the problem above and this method has been proved effective in the fault diagnosis on Smart Grid.