Oil analysis provides a huge payback when deployed through a proper strategy. While an extremely valuable tool in today’s reliability programs, it is sometimes applied in an ad-hoc manner. This is a dangerous approach, as the program can quickly become quite costly due to overtesting or even show little value due to inadequate testing. Let’s take a look at both situations.

Overtesting

A recently visited paper mill had a rather robust oil analysis program. This program was further optimized by the corporate reliability manager. The maintenance manager had a positive feeling about the benefits of predictive technologies and was supportive of the oil analysis program. While this was all seemingly positive data, the drawback was that the manager decided he wanted all equipment to be incorporated in the oil analysis program, including small centrifugal pumps containing less than even a quart of oil.

Taking this approach would have meant that the mill would run hundreds of oil samples on at least a quarterly basis. Adding to this, when following proper sampling procedures, we understand that the sampling hardware must first be flushed. When sampling small reservoirs, such as those in small centrifugal pumps, following the flush portion and then sampling, a complete oil change would have occurred on every pump each quarter. Considering the increased lubricant consumption coupled with the additional cost of testing the oil samples, you can see how the overall costs would add up quickly.

Although the maintenance manager should be commended for his aggressive drive toward equipment reliability, moving forward with the initially desired approach would have been costly, significantly reducing the program’s overall return on investment (ROI).

Inadequate Testing

During a recent oil analysis program benchmarking exercise, it was asked how machines were selected for inclusion in the testing program. The initial response was, “We use criticality.” When the process used for criticality assessment was investigated, it was revealed that there was no real process. The machines were selected based on what we like to call “perceived criticality.” This resulted in a very small group of components initially being tested, although the program was growing in a methodical manner. When a machine component failed that was not part of the analysis program, the replacement component was then put on the program. So there was no real methodology at all.

This plant was experiencing a significant number of failures that could have been avoided had the program been put together properly in the first place. By taking this approach, the total cost of program development and optimization was incredibly high once the costs of missed opportunities were included into the equation.

Moving Forward

Oil analysis comes in three basic forms:

  1. Commercial Lab Testing— Samples are collected and sent to a third-party laboratory for testing and analysis. This can take place on a routine basis or to confirm screening data from select on-site testing. 
  2. On-site Testing— Samples are collected and tested at the plant site using a number of potential on-site test equipment. Many advances have occurred in on-site test equipment that will be explored in a later issue of Machinery Lubrication. 
  3. Online Testing— Specialty meters (usually particle counters), moisture meters and dielectric testers are installed in a circulating system in order to capture “live” lubricant conditions. As with on-site testing equipment, this technology has grown significantly over the past 5 years. 

Each of the basic types of oil analysis has an intended function and can offer significant benefit to the end user if deployed properly. For companies with a large number of lubricated components included in the oil analysis program, it is vital to incorporate some level of each of these categories for a well-rounded program.

Utilizing the criticality of machines that has been assigned through a documented method provides the best starting point in the decision-making process regarding which form, or combination of forms, is best for each component.

A plant with a well-developed criticality system already has the foundation for establishing an equally well-developed oil analysis program. Some of the primary decisions related to oil analysis that criticality can assist with include:

  • Machine selection 
  • Reliability objectives 
  • Test slate selection 
  • Sample frequency 

The days of the common test slate and frequency are over. The largest ROI will be achieved by using criticality to fine-tune an existing program and to get a new program off to an optimized starting point. The plant that does not have an established criticality assigned to machines should consider this foundational element. Without it, the entire predictive program is at risk of supplying less than the desired effect on overall reliability and ROI.

Understanding Operational Criticality

The oil analyst should know a machine’s operational criticality. This can be broken down into two basic elements. The first is mission criticality, which considers the consequences of failure (production losses, safety, etc.) in relation to the machine’s intended mission. The second is functional restoration, which basically asks in the event of failure, what would it cost to replace, repair and rebuild the broken machine.

These two elements of operational criticality don’t always go hand-in-hand. Because of redundancy and standby equipment in some processes, an expensive repair may not always result in costly downtime. Likewise, in other cases, huge production losses may be triggered by small throw-away machine components.

Operational criticality is best defined by the asset owner, not by outside oil analysts or other non-stakeholders. For instance, consider using a scale from one to five for both mission criticality and functional restoration. A rating of one might mean failure is inconsequential, while a rating of five alerts that failure could have devastating consequences. The cost, frequency and quality of oil analysis will likely vary in accordance to how the machine is rated for operational criticality.