Vibration analysis for condition monitoring of an automatic press machine for thermoplastic polymers

Vibration analysis is recognized as a powerful tool for fault diagnosis and quality control of industrial machineries. Accelerometer signals can provide a measure of the majority of causes of inaccuracy in manufacturing processes and of components' faults, which can in turn affect the product quality. This paper provides a fault identification method based on time and frequency domain analysis. This method is applied to an automatic press machine for soft thermoplastic polymers. Pilot experiments conducted in laboratory conditions allowed to select the proper number and the position of the accelerometers, as well as the signal features needed for a preliminary threshold-based maintenance strategy.