Construction And Verification Of Point Machines Data Acquisition System Oriented To PHM Technology Application

With rapid development of urban railway transportation, it provides great convenience forcitizens in daily work and life in metropolis. Meanwhile, the turnout safety and reliabilityreceived much concern, since traffic delays and safety problems caused by the turnout arisesfrequently. In particularly, as the actuator of turnout control system, point machines failure rateperforms a remarkable rising tendency. However, the conventional preventive maintenance andbreakdown maintenance cannot satisfy the requirements of equipment maintenance because oftheir resource-consuming and low efficiency. Therefore, it is urgent that some effective methodsabout point machine state monitoring, fault diagnosis and fault prediction should be proposed toimprove the operational efficiency and reliability of railway transportation.To track with the above problems, this paper proposes the overall architecture ofprognostics and health management technology (PHM) for point machines, and presents thepower data acquisition system construction and its verification, and then performs the faultdiagnosis on site data using support vector machine(SVM) algorithm. It is expected that the dataacquisition system and its related PHM techniques will provide some supports in point machinemaintenance decision. Specifically, the main contents are developed in the following three folds.1. The power data acquisition principle for the S700K electric point machines is obtained.The fault mechanism and performance parameters of S700K electric point machines is firstlyanalyzed, and the data, i.e., the three-phase operating voltage and current signals of motor, arecollected, then the fitting algorithm is applied to get the power signal curve, this power signalcan also be viewed as sensitive parameter signal in S700K point machines fault diagnosis andprediction.2. The power data acquisition system is constructed and verified based on virtual instrument.According to the system design requirements, the overall framework of data acquisition systemis proposed based on LabVIEW and PXI，and the hardware construction and software programsdesign are discussed in detail. In laboratory, the debugging and simulation are performed, and theerror modification is conducted according to the comparison of simulation data and theory value.The power data acquisition system is tested on-site by injecting faults, and the S700K pointmachine voltage, current and speed signals are collected, displayed and storage online underinitial state and fault state, respectively. 3. Data mining methods are applied to S700K point machines fault diagnosis. The powercurve is divided into four sub-intervals according to the motor speed, and the data characteristicsin each sub-interval are extracted to perform fault diagnosis in a more precise way. The principalcomponent analysis (PCA) and SVM algorithm is introduced in features extraction and faultdiagnosis, respectively, and the fault diagnosis ratio is calculated. The analysis result on theon-site experimental data demonstrates the feasibility and effectiveness of the data acquisitionsystem.In a word, the PHM sample machine for point machine based on PHM technology ispresented in this paper, which aims to perform state-centered maintenance. This PHM samplemachine consists of fault mechanism analysis, sensitive parameters analysis, data acquisition,feature extraction, fault diagnosis, and so on, and the construction and verification of power dataacquisition system is discussed in detail. It is expected that this research results will providegood supports for point machine PHM application in engineering.