Isolation Detection Using Decision Tree Algorithm in Micro-Grids with Variety of Distributed Production Resources

Isolation detection methods are divided into three active, passive and telecommunication categories. In most cases, passive methods are used more than other methods due to low cost, simplicity, high speed and no harmful effect on the power quality of the power system. In these methods, a parameter of the system is considered as an indicator and the system is analyzed by setting threshold values for this parameter. An island system is detected if the desired parameter exceeds the threshold values for it. The major problem with these methods is their large detection area. In this paper, a method for identifying islandization is proposed based on the decision tree algorithm. This method determines islandization by using one or more indicators used in other methods of inactive detection in terms of network complexity. This method has been tested under various possible incidents in grid-grids and various loading systems and distributed generation sources. The results indicate the fast and accurate performance of the proposed method with its inadequate detection area compared with other inactive methods of island identification.

Isolation Detection Using Decision Tree Algorithm in Micro-Grids with Variety of Distributed Production Resources Keywords:

Data mining , Distributed generation , Decision tree algorithm , Inactive detection method , Non-detection area , Grid. , داده کاوی , تولید پراکنده , الگوریتم درخت تصمیم , روش تشخیص غیرفعال , ناحیه عدم تشخیص , ریزشبکه.