Application of multi-fuzzy system for condition monitoring of liquid filling machines

In this paper a novel approach is implemented for investigation of failures in Stork bottle filling machines. Fuzzy based system is used to detect the abnormalities present in machine by using time and frequency domain statistical features. Statistical analysis of vibration data determined the gearbox failure which correlated with engineer´s findings. The method used has shown promising results to predict the failure in this case of low speed rotary machines. It has been concluded that statistical based analysis of vibration signal is a suitable for predicting machine faults with low rotating speeds. This paper presents a system, implemented on the industrial process machine, which has successfully predicted the faults in the gearbox before the catastrophic failure.