Use of On-line Sensor Technology for Oil Machinery Condition Monitoring - Case studies on Real World applications, and their use to predict Machinery Failure and extend Oil Change Intervals

摘要

This paper will show how well the sensors have performed, and prove that real time assessments of machinery condition can be made from this data. Ever increasing demands to cut costs and manpower means an inevitable reduction in time consuming laboratory based oil analysis. An alternative is therefore required. A range of on-line condition monitoring sensors have been designed, tested and placed in real world situations. These sensors detect parameters such as metallic wear debris, emulsified and dissolved water, viscosity and oil condition. Oil and machinery condition data has been collected from slow/medium speed diesel engines, industrial and wind turbine gearboxes. This will show how oil change intervals can be extended and maintenance can be scheduled in advance. Tests have shown for the first time the rate at which gear wear debris is produced as load stages increase, as well as the profile of wear during each load stage. Data has been collected on wear particle evolution in bearing failure conditions. This will be used to illustrate prediction of the onset of machinery failure by providing detailed analysis of particle evolution from a bearing surface. The ability to detect particle size will be demonstrated and how this can be used to provide very early detection of failure, well before any vibration is evident. A novel oil condition sensor has been developed for predicting oil change intervals and contamination events. The sensors has been tested on operating machinery and has been shown to reliably detect a variety of changes to the oil such as water content, oxidisation, degradation. The sensor development project described in this paper has now been completed and resulted in full commercialization of a suite of on-line oil conditions sensors.