Islanding detection method of distribution generation system based on logistic regression

Abstract

Islanding operation will harm the safety and stability of the power system, in the occurrence of island operation which must be in the specified time to monitor it. Here, the logical regression algorithm is introduced into the islanding operation monitoring. First, an accurate distributed generation system model is established in Matlab environment. The model considers various possible island operating states, minimum monitorable area, multiple distributed power supplies, various operating conditions, and different network topologies, and then uses. The data feature quantity under various island operating conditions is trained in the TensorFlow environment. Finally, the trained model is used to monitor the islanding status. The results show the feasibility of the method.

Keywords

• regression analysis
• power grids
• distributed power generation
• power distribution faults
• islanding detection method
• distribution generation system
• logistic regression
• safety
• stability
• power system
• island operation
• specified time
• logical regression algorithm
• islanding operation monitoring
• accurate distributed generation system model
• possible island operating states
• minimum monitorable area
• power supplies
• island operating conditions
• islanding status