基于分层免疫机制的无绝缘轨道电路补偿电容故障诊断系统

Fault Diagnosis System for Compensation Capacitors of Jointless Track Circuit Based on Layered Immune Mechanism

本文基于轨道电路及其读取器间的信息传输模型,借鉴生物免疫机制,提出基于分层免疫的轨道电路补偿电容故障诊断方法。在特征提取层,通过分段线性拟合,对常见故障下的仿真数据和实际数据进行特征提取,以形成"抗体"和"抗原",并构建相应的"抗体库";在细胞免疫层,基于"抗体库"对"抗原"进行"自己/非己"的识别,实现对常见故障的诊断和监测;在体液免疫层,针对非常见的组合故障,利用克隆选择算法获得与"抗原"具有高亲和力的"抗体",以此实现相应故障识别。实验表明:本算法可实现对补偿电容的状态检测、常见故障与组合故障诊断等多层面操作,并可提高对复杂故障的二次反应速度,具有较高的适应性、灵活性和准确率。

Basing on the transmission model between a track circuit and itrs reader and learning from the biologi- cal immune mechanism, the fault diagnosis method for compensation capacitors of track circuits using the hier- archic immune algorithm was proposed. In the feature extraction layer, piecewise linear approximation was used for feature extraction of simulation data and actual data in common fault modes to obtain ＂Antibodies＂ and ＂Antigen＂ and build an ＂Antibody library＂ accordingly. In the cellular immune layer, ＂self or non-self＂ recog- nition of the ＂Antigen＂ based on the ＂Antibody library＂ was conducted to realize diagnosis and monitoring of common faults. In the humoral immune layer, for the uncommon combined faults, the clonal selection algo- rithm was used to obtain the ＂Antibody＂ that had the highest affinity with the ＂Antigen＂ to achieve the corre- sponding fault identification. Experiments show that the algorithm introduced in the paper has high adaptabili- ty, flexibility and excellent accuracy, while it can realize multi-level operations such as state monitoring of compensation capacitors, common and combined faults diagnosis and so on, and it can also improve the sec- ondary reaction speed for complex faulty conditions.