A methodology for determining rolling element bearing limits applied to inductive wear debris sensors

摘要

Machinery condition monitoring (MCM) is the process of measuring, trending and assessing key parameters that indicate the health of a machine. Wear debris analysis is a common element of MCM programs and can involve the assessment of size, quantity, morphology and composition of debris shed from dynamic components such as gears and bearings as they deteriorate. Traditionally, wear debris has been detected by the laboratory analysis of routine oil samples, visual inspection of magnetic chip detectors or other off-line techniques which all have significant technical or logistic limitations. The Inductive Wear Debris Sensor (IWDS) is a relatively recent development in this field and commercial examples have demonstrated some significant advantages over traditional techniques. These sensors are installed in the return line of a lubrication system so that wear debris liberated from deteriorating components can be detected in real time. Applications for these sensors currently include aviation machinery, wind-turbine generators, marine propulsion systems and some aspects of general industry. One aspect that has received virtually no attention in the literature is a validated methodology for determining a suitable limit (or suite of limits) for a particular application. Limits are a critically important aspect of MCM programs that can be used to trigger further detailed interrogation of the machine or maintenance action. This research has developed and validated a new methodology for determining rolling element bearing deterioration limits that have physical meaning and can be applied to a system containing an IWDS. The application of the general methodology to other dynamic components and complex machines has also been explored. In addition, a metric consisting of three subordinate condition indices has been developed that can be used with data obtained from an IWDS. This metric addresses some of the idiosyncrasies of IWDS data that have not previously been addressed.