Why Online Oil Quality Monitoring is a Best Practice for Reliability Programs
Fluid monitoring programs have long been a challenge within the reliability and operations community. For example, oil samples provide a significant amount of value in diagnosing root causes, the presence of contaminants and the measurement of oil quality. Yet, for these programs to be implemented properly, the samples must be taken at intervals more frequently than assets or personnel are available. This leads to a reliability program lacking the information required to meet reliability targets. Best in class reliability programs have found success addressing this challenge by implementing online fluid or oil monitoring programs.
What Is Online Oil Quality Monitoring?
Online oil quality monitoring provides real-time monitoring of oil condition. Online oil monitoring systems have come a long way in recent years. Early systems relied on dielectric sensors and chip detectors to provide oil health and wear debris monitoring capabilities. Viscometers, water in oil, optical particle counters, inductive coil wear debris monitors and other sensor technologies were later introduced, proving some insights into general oil quality. However, they were not sensitive enough to many of the oil health properties that reliability programs were looking for.
These sensors helped establish the value of online oil quality monitoring, but more recent sensor technologies have improved sensitivity, range of detection and correlation with oil lab samples currently relied upon for decision-making. Modern inductive coil sensors are now twice as sensitive as the previous generation’s sensors, allowing for a wear debris detection size roughly twice as small for the same bore diameter. The latest oil quality sensors, using an impedance spectroscopy technique, are now capable of correlating or directly measuring multiple critical aspects of the oil, such as overall health, oxidation, total base number (TBN), contaminants (e.g., water, soot), etc. These sensors are often bundled together as single offerings and/or packaged with other technologies, such as vibration, smart Industrial Internet of Things (IIoT) devices, data science, etc.
Challenges with Off-Line Oil Sampling
Oil quality and wear debris are typically not static, consistent, or slow changing within assets. Changes typically revolve around events, especially when transitioning from healthy to non-healthy states. Wear debris is lost in bursts over a period of time and is quickly filtered out by the filtration system. Whether or not the fault is detected, depends on when a sample is taken (see Figure 1). Contamination events or changes in oil properties, such as water contamination or additive dropouts, also occur in time periods much shorter than standard oil sampling frequencies. This often results in undetected fault conditions, causing further damage that, in many cases, cannot be reversed and may lead to catastrophic failure.
Figure 3: Wear debris timeline of a bearing raceway axial crack fault within an industrial gearbox
Figure 4: Water contamination event
Where Is the Wear?
Wear debris is one of the best indicators of asset health. A study commissioned on a fleet of 137 industrial gearboxes examined the previous two to four years of oil samples, along with fault history. Figure 2 summarizes the findings with iron concentrations color coded by known healthy (in blue) and faulty (in red) gearboxes. Using traditional approaches, one would expect to find higher iron concentrations within faulty gearboxes. However, the study showed a lack of utility from these samples for determining asset health.
Online debris monitoring has proven itself as a perfect solution. During a 14-month timeline, as shown in Figure 3, periodic oil samples taken for analysis indicated no significant findings. In fact, the reported oil cleanliness actually exhibited improvement, based on International Organization for Standardization (ISO) cleanliness codes, as concentration levels reported by the online debris monitor increased. It is important to note that although the peak debris concentrations were increasing, the momentary concentrations remained highly variable.
Online monitoring of metallic wear debris enables these events to be observed and tracked in real time, allowing for adjustments in operation to prevent catastrophic failure and extend operating life until a repair or replacement can occur. Online wear debris monitors provide the additional benefit of allowing the analyst to correlate wear debris data with operational data to pinpoint the cause of fault progression, as identified by debris concentration spikes. In the Figure 3 use case, the asset was derated until maintenance could be performed.
Figure 5: Overall oil quality
Figure 6: Oil life extension example
Most of the value from oil samples is centered on oil health itself. Key properties and contaminant levels, such as oxidation, TBN, total acid number (TAN), additive packages, viscosity, water, fuel, soot, etc., are all used to judge the remaining useful life (RUL) and identify preventative and life extending actions. Just as with wear debris, many of these are event driven. Events, such as water and fuel contamination, happen within minutes and can be missed hours later, as shown in Figure 4.
Decision-making between oil changes, bleed and feed, top offs, etc., is often difficult because trending of the oil’s critical properties and RUL estimation are not available. With online oil quality monitoring, many aspects of the oil can be trended and RUL estimated. In Figure 5, online oil monitoring allows for oil change optimization based on actual oil quality measurements, as compared to traditional time-based changes. With traditional time-based oil changes, oil is almost always exchanged too early or too late. Real-time monitoring allows for optimized oil change intervals for each asset.
Utilizing this data, RUL can be estimated and life extension actions can be optimized based on operational needs. In Figure 6, the asset was derated as a way to extend oil life by almost 100 operating hours. Actions, such as a bleed and feed, can be measured and overall oil life extended through many tools available to reliability teams.
Periodic off-line oil analysis provides significant value within reliability programs, but often it is not a sufficient tool on its own for meeting the program’s reliability goals. Online oil quality monitoring systems have proven to be critical, cost saving tools, providing the data necessary to make optimal maintenance decisions.
Many industries, such as energy, mining, rail, marine, etc., have all started adopting an online oil monitoring program and it is expected such programs will become standard practice over the next few years. With new sensor technologies, the payback period for investment is usually less than one year, making them one of the best investments for any reliability team looking to adopt the latest best practices.
Stephen Steen is VP of Industrial IoT for Poseidon Systems, LLC. He has diverse experience in reliability technologies, such as CM, PHM, ML, AI, and fluid analysis spanning automotive, ag, heavy equipment, and energy. Stephen’s current role is focused on working with customers with adopting state of the art IoT fluid solutions.