Ball bearings are the most critical components of rotating machinery in oil and gas companies. Typical research has focused on bearing failure detection based on bearing failure frequencies derived from the velocity spectrum. However, most bearing failures are caused by improper or insufficient lubrication. The current research utilizes a case study demonstrating when ball bearings must be replaced or relubricated due to poor lubrication conditions. Poor lubrication is the cause of natural fre- quency excitation in bearings, where rapid bearing damage is typically induced by poor lubrication film. According to experimental data in this study, the bearing failed due to natural frequency excitation. In addition, when analyzing a signal with the velocity spectrum, high frequencies are displayed. Bearing failure is detected without bearing failure frequencies using the natural frequencies of the bearing in the velocity spectrum signal. Moreover, an experimental investigation of the bearing failure of a liquid ring compressor was conducted utilizing a VIBXPERT II vibration analyzer and the Omni trend software. The velocity spectrum is derived based on a fast Fourier transform from a time signal. After lubricating natural frequencies must be disappeared from the velocity spec- trum otherwise, the bearing is failed and must be changed.
