DATA TRENDING AND DIAGNOSTICS

An effective oil analysis system requires that data is efficiently stored and is readily accessible to all personnel involved with equipment monitoring, maintenance and reliability analysis. A fully functional oil analysis system serving a large fleet generates vast quantities of data for processing and interpretation. This can be most effectively done automatically using computer techniques, freeing up personnel for other duties.

The prime objective of an oil analysis program is the early detection of oil degradation, contamination and wear problems that lead to equipment failure. From this early detection it must be possible to predict the rate at which deterioration is occurring in order to pull the unit from service before catastrophic failure happens. This can only be achieved by looking at trends and not by limits alone, and doing this manually can be very labor intensive and time consuming. Expert systems provide an ideal environment to capture and validate human knowledge and they also provide an excellent means to utilize knowledge for automatic problem solving. It is important to remember that expert systems have better memories and are more consistent than human evaluators.

The most effective predictive maintenance programs trend the data looking for signs of early failure, allowing the equipment to be repaired at minimal cost and down time. In order to best utilize trend analysis, data must be available on a regular basis. Obviously the more frequently the sampling is performed the more accurate the analysis becomes. However, weekly sampling has been found to be the most cost effective. Diagnostic reports from the expert system on the condition of the machinery assist maintenance personnel in making critical decisions regarding equipment health conditions.