Research On Signal Processing And Imaging For Testing Signals Of Locomotive’s Rims

Digital processing and imaging for testing signals of locomotive’s rims are important part of thelocomotive wheels online automatic testing device, so they are the key of extracting the defects echoeffectively and detecting the machine wheel rim automatically. This article focus on solving the keytechnology and difficult point of the signal processing, such as digital filtering process, digitaldemodulation and the ultrasonic A-Scan.Ultrasonic flaw signal is often mixed with the complex interference noise, which is collected inthe industrial field. If this signal is directly used for testing, the result may be an error. In order tosolve the noise problem, the paper introduces the FIR digital filter to reduce the noise firstly. Whiledue to the FIR digital filter is based on the single window function of noise reduction processing, aswell as nonstationarity of noise signals and other reasons, the FIR digital filter can’t extract the rimtesting signal effectively. For this reason, combined with the analysis of the signal-to-noise ratio andthe real-time systems, wavelet transform is adopted to denoise firstly and then digital demodulation isbased on cubic spline interpolation. We can extract the envelope curve of the signal effectively withguaranteeing data integrity and detect the position of the peak point of the defect echo. So we are ableto obtain the distance of the defect to the rim tread on the PC interface and the signal can be seen bythe A-Scan. As the key technology of the signal processing is done in MATLAB environment, it can’tbe used as a industrial control and imaging interface. In this paper, the control and imaging interface isdone by C#and it can automatically call the MATLAB packaged dynamic link library of signalprocessing. Therefore, this paper realized commands sending, retrieving data, signal processing,signal feedback, imaging display, defect positioning and other functions.Finally, the paper completes experiment with the signal which is collected by the locomotivewheels online automatic testing device. We get the result is that the detection method described in thispaper has noise reduction obviously, the detection curve clearly, defective positioning accurately, fastresponse and able to meet the performance indicators of industrial design.