基于Relief特征量优化及SOM网络的断路器操作机构状态评估

Condition Assessment of the Circuit Breaker Operating Mechanism Based on Relief Feature Vector Optimization and Som Network

为提高断路器操作机构状态评估的效率及准确性,文中提出了采用Relief优化方法对分合闸线圈电流特征向量进行优化及SOM网络对线圈电流样本进行分类的高压断路器操作机构状态评估方法。仿真研究结果表明:基于Relief的分合闸线圈电流特征向量优化方法和SOM网络分类方法的操作机构状态评估能够实现操作机构状态的准确评估,同时采用优化后的电流特征向量数据进行状态评估减小了运算量和缩短了运算时间,提高了断路器机构状态评估的效率。

To improve the condition assessment efficiency of a circuit breaker operating mechanism, Relief is employed to optimize the feature vector of a high voltage circuit breaker and SOM neural network is adopted to classify the samples of switching coil current. Thus, an operating mechanism condition assessment method for high voltage circuit breakers is proposed. Simulation results show that this Relief and SOM based operating mechanism condition assessment method can recognize the operating mechanism condition accurately, and application of the optimized current feature vector can reduce the computing amount and time and improve the assessment efficiency.