Brush wear detection by continuous wavelet transform

This paper is concerned with the singularity detection of a motor current signal. These singularities are related to the brushes condition, which are one of the sensible parts of the motor to wear out. As the mechanical contact between brushes and commutation segments increases, the rising parts of the current commutation waves become wider and this is a signal of brush wear. We successfully implement an algorithm to localise the singular points of the commutation waves based on the wavelet maxima lines. The width of the rising part of the commutation waves is then calculated from these maxima lines.