Distributed Fault Analysis System for SS4B Locomotives Based on Sensor Networks

In response to the fault diagnosis requirements of the SS4B heavy-duty freight locomotive, this paper proposes a distributed fault analysis system based on a sensor network. The system assists traction rectifier fault diagnosis by analyzing the harmonic characteristics of grid-side currents. An onboard intelligent diagnostic terminal is employed, which integrates the LZW lossless compression algorithm to compress collected data, while the Flower Pollination Algorithm (FPA) is used to optimize Support Vector Machine (SVM) parameters, thereby enhancing diagnostic accuracy. In addition, compressed sensing technology is applied to denoise signals contaminated with noise. Experimental results demonstrate that after compressed sensing denoising, the diagnostic success rate of rectifier faults using the FPA-SVM method exceeds 98%. Finally, the paper presents the circuit design and PCB layout of the onboard signal acquisition node, and establishes a ground-based simulation diagnostic system in the laboratory. The results validate the coordinated operation of data acquisition, transmission, and fault diagnosis.