Fault detection for manufacturing home air conditioners using wavelet transform

摘要

This paper presents a wavelet transform based method of fault detection for manufacturing home air conditioners in a noisy environment. Wavelet transforms have proven to be very sensitive to impulse signals and many signals from faults in air conditioners have impulse signal like characteristics. Normal operational sound levels for air conditioners are substantially lower than a typical manufacturing environment which requires methods that are robust to noise. We have developed several features that have good sensitivity to impulse signals. The effectiveness of the developed method was demonstrated on sound measurements taken from a real manufacturing site.