饱和铁心型超导限流器故障电流快速模式识别

Rapid Pattern Recognition of Fault Current for HTS Three-Phase Saturated Iron Core Fault Current Limiter

分析了开断直流励磁以及断开速度对该类超导限流器限流效果的影响,指出短路故障电流需要快速识别,设计和制造了三相小样机和相应的基于Labview和NI板卡的实时检测控制系统。提取了短路故障电流的三个重要特征:电流实时幅值、电流变化率、故障时高电流变化率的持续时间,根据特征进行神经网络感知机模式分类,离线在Matlab环境下训练神经网络找出线性分类面,在NI板卡内部的FPGA上实现分类神经元,并综合采用阈值模糊化的方法,对故障进行复合决策,以提高辨识准确性和速度。对阈值法和复合决策法进行的比较实验,表明了该方法的有效性。

The current-limiting effect by turning off DC excitation and the speed of turning off for superconductor fault current limiter （SCFCL） are discussed. The necessity of rapid recognition of fault current is presented. Proposed SCFCL and its real-time data acquisition and control systems based on Labview and board of NI are designed and fabricated. Three important characteristics such as the value of current, changing rate of current and the lasting time of keeping high changing rate of current are extracted. According to the characteristics, off-line perceptron training in Matlab is done to find the classification surface. Also the neural ceils and the method of fuzzification of the threshold current value are synthesized in FPGA of NI board with the trained weights and biases to increase the speed and veracity of pattern recognition. The experimental comparative results between complex decision method and threshold method indicates the former is more timely and veracious.