基于动作电流波形树表达的道岔转辙机状态检测方法

State Recognition Method for Point Machines Based on Representation of Shift Current Waveforms by Relational Trees

给出了一种基于波形树的随机波形表达算法，即根据波形的特征点递归求出其波段组成的波峰集合。针对信号微机监测设备测试的道岔动作电流曲线，分析了ＺＤ６型电动转辙机电流曲线的波动特征，通过计算电流曲线的波形表达识别出道岔转换的各个阶段，检测出波形中蕴涵的故障征兆。最后给出一个故障检测的示例，并指出该方法的适用范围。

Shift current is an important parameter in testing and monitoring point machines,but it is difficult for an actual method to detect fault.This paper first proposes in detail an algrithm for representation of random waveforms by relational trees,which can find out recursively the peak set in curve segmentations making up the whole current curve in term of the characteristic points.Here the shift current data sampled by the dedicated equipment for testing and monitoring the railway signal appliance are used,then the features of shift current waveforms of ZD6 point machines are analyzed.By computing the representation of shift current wavefoms by relational trees, each stage in switch shift process is located and recognized,therefore the fault signal is discovered.A diagnosis example is also presented.Finally the applicable range for this method is given.