Fault trending of an automatic double threaded trampoline webbing machine

This paper applies signal processing methodology on a specially designed webbing machine to evaluate and improve the trampoline manufacturing process. Digital Signal Processing (DSP) methods such as Fast Fourier Transform (FFT), and Statistical Process Control (SPC) have been applied to find singular frequencies, but the results are not satisfactory. Instead, a Continuous Wavelet Transform (CWT) based methodology is used to predict the fault before it occurred. Unlike the traditional wavelet scalogram method, the collected data in the research is separated into small segments, and the divided segments are used to generate scalograms. By observing and counting the pulses in different frequency ranges in scalograms, the singular frequency ranges can highlight the trending to fault occurrence. Additionally, a timer is used in the experiment to record the time when the fault occurred. This information validates accuracy of the method and demonstrates that this novel fault trending detection method has a promising usage in recognizing the trending of faults.