面向高速铁路道岔的隐患分析与故障定位改进

Hidden-security-risk analysis and fault location improvement for high-speed railway switch

以高铁常用S700K型交流转辙机为例,通过测试道岔动作过程中关键继电器的时序周期,探究造成交流道岔中途停转的深层次原因。提出基于动量BP(Back Propagation)算法的神经网络故障定位的改进方法,并在以往微机监测系统测试的电气参数基础上补充故障样本。对神经网络进行训练并测试,由误差比对结果可知,该方法使得高速铁路道岔智能故障定位更加精确。

Through testing the time sequence of the key relay in the course of switch action,the reason why the AC switch-machine such as S700K switch stopping on the way was explored.Then,an improved method to locate switch-fault based on the neural network algorithm of Momentum BP(Back Propagation)was proposed,and the fault samples were supplemented on the basis of the electrical parameters measured by the Maintenance and Monitoring system.Through training and testing this kind of neural network,the error comparison shows that the intelligent fault location of high-speed railway switches can be more accurate by this way.