Study of vibration signatures of bearings for fault diagnosis and condition monitoring

Abstract

Rolling element bearings are used in a wide variety of applications. Failure of such bearings needs to be diagnosed at an early stage to avoid machine failures. Literature survey has indicated that about 40 % of all machinery failure are due to failure of bearings. From the detailed literature survey, seven different tasks are taken for specific study. The studies are carried out experimentally using the experimental facility created or using the existing machines in case of case studies. In study-1: Experimental study to identify a particular type of defect has been carried out. Specific defects are created artificially. Vibration signatures are obtained experimentally and are compared with theoretically computed fault frequencies. In study-2: Effect of rub on rotating shaft is attempted experimentally. Two rub materials; Aluminium and mild steel; are tested against the mild steel shaft. In study-3: Effect of shaft misalignment and method of correcting the same is attempted both experimentally and numerically, using vibration signatures. In study-4: A method of deducting the unbalance and correcting the unbalance in a CNC machine spindle system is attempted experimentally using vibration signatures. In study-5: Effect of unbalance with mass addition or mass subtraction on a multi-rotor system on vibration and power consumption has been studied experimentally. In study-6: Use of Artificial Neural Network for defect identification and for knowing the severity of defect using vibration signatures as input is attempted. In study-7: Effect of solid contaminants on a grease lubrication system is attempted systematically with a view to use the vibration signatures for condition monitoring. The vibration amplitude of the bearing is found to increase with increase in concentration of solid contaminants, particle size of contaminants and speed of the system.