A Case Study on the Impact of Working Conditions for Machinery Fault Diagnosis

Variation in working conditions imposes great challenges for machinery fault diagnosis, especially for data-driven fault diagnosis and prognosis. This paper presents an experiment which employs a rotary torsional fatigue test bench to study the impact of working conditions on the dynamic behavior of a mechanical system. Vibration signals were acquired under a normal condition and a compound bearing fault condition with a working-condition matrix composed of 6 working loads and 5 rotating speeds. The impact of working load and rotating speed is qualitatively revealed by the frequency spectrum and time-domain features of the acquired vibration signals.