Propose of unsealed deep groove ball bearing condition monitoring using sound analysis and fuzzy logic

Monitoring of bearing condition is important for machine damage warning. Generally, vibration and acoustic emission analysis are used for bearing condition monitoring. However, those methods need costly apparatus and experienced operators. Therefore, economical devices and effective method for monitoring of bearing condition is necessity. This paper aims to propose unsealed deep groove ball bearing condition monitoring by integrating of sound analysis and fuzzy logic. The condenser microphone was used to detect sound signal and sent sound wave to a computer through a sound card. The sound wave was analyzed by FFT method; and numerical integration was applied to compute the spectral density. The normalization technique was used to transform the spectrum intensity into percentage. Two input parameters, consisting of percentage of load and spectrum intensity were established as input membership functions. Each input parameter had three levels, including maximum, middle and maximum. The nine fuzzy rules were created for classification of bearing condition. The defuzzification of output membership functions, including damaged, fair and normal were performed by using centroid method. In the experiments, three kinds of bearing condition were examined by the proposed method. The results revealed that the proposed method is an effective method for monitoring of bearing condition with high accuracy.