Syncretic Application of IBAS-BP Algorithm for Monitoring Equipment Online in Power System

With the integration and development of sensor technology and control technology, the construction of smart grid is in the ascendant. But existing technology is still insufficient in the field of equipment monitoring. Therefore, it is difficult to accurately determine if SF
₆
breaker or other equipment have faults in its operational mechanisms which can cause false action and thus impact the safety of power grids. In this paper, the IBAS (Improved Beetle Antennae Search) algorithm and the BP neural network are combined and used in the monitoring system for the first time. To improve the beetle search algorithm, a single beetle is improved into a population in the iterative process of algorithm. The measured opening (closing) current data is used to verify the accuracy of different algorithms. The results show that compared with PSO-BP (particle swarm optimization, PSO) model, GA-BP (genetic algorithm, GA) model and BAS-BP model, the IBAS-BP model not only effectively avoids the possibility of local minimums but also has higher prediction accuracy and better robustness. The number of iterations in the IBAS-BP algorithm is only 38, and the average error is only 0.1%. According to this, it is possible to make real-time diagnosis of faulty states in SF
₆
breaker including iron core jams, low operating voltages, poor contact of auxiliary switches, jammed operating mechanisms, and barometer failures in SF
₆
. At the same time, the IBAS-BP algorithm can be applied to wind turbine power prediction and other occasions since model regression determination coefficient R
²
of the training set is up to 0.9753 and the relative average error is only 0.25%. The results prove that the IBAS-BP algorithm has obvious advantages and fairly good universality. It can be further promoted and applied to power systems to provide reference for optimizing the online monitoring of power equipment.