The objective of a root cause analysis (RCA) is to identify what happened, why it happened, and what can be done to prevent it from happening again. RCA entails examining the problem and considering evidence from different perspectives. Oil analyses using as many technologies as possible (Table 1) can play a crucial role in the RCA of the failure of lubricated machines. Vibration analysis, thermography, ultrasonic analysis, metallurgical analysis, equipment/component inspections, and operating histories are among the most valuable sources of this kind of data.
Fluid degradation is responsible for many kinds of lubricated equipment failures. Lubricants are subjected to a wide range of conditions—extreme heat, air, incompatible gases, moisture, contamination by dirt and wear particles, process constituents, radiation, and inadvertent mixing with a different oil—that can degrade their base oil and additives. Changes in viscosity and the development of insoluble particulates are among the first oil degradation problems to affect equipment performance. Therefore, it is vital that appropriate diagnostic analyses be performed to detect these conditions in critical and sensitive lubrication systems.
Italian job
The Italian economist Vilfredo Pareto observed that the unequal distribution of wealth in 19th-century Europe could be stated as follows: 20% of the people own 80% of the wealth. Pareto's rule of thumb has since been applied to countless situations—including equipment reliability, where 20% of equipment failures indeed account for 80% of total losses. A good manager understands that having the right knowledge and tools to eliminate that critical 20% represents a big opportunity for reliability improvement.
RCA has been bandied about in lubrication circles for several years. However, in the authors' experience, it is a tool that few people use correctly. RCA should be performed on chronic problems or failures that recur and often become more serious over time. Yet, most lubrication-related RCA is based on a rudimentary analysis of the oil and some poor assumptions.
The failure to correctly identify a root cause is typically not due to a lack of knowledge on the part of the individual or team tasked with the RCA. In fact, in many cases too much expertise is the problem, because it may lead the investigation in a predetermined direction. To be sure, a person or team chosen to conduct a RCA should have expertise in multiple disciplines and training and experience in RCA. But they should also be persistent and have an engineer's objectivity. False assumptions and predetermined conclusions are the primary reasons why root cause analyses end without soundly documented results or correct findings.
An RCA must begin with an understanding of the problem and a list of all of its possible causes, however remote (Figure 1). Care should be taken not to assign blame for the failure at this point in the investigation. For each potential root cause, experiments and tests should be devised to test its potential existence. Keep in mind that using data to prove a theory false (the devil's advocate approach), is as viable as taking the normal, causative approach—because that data can eliminate incorrect assumptions.
Slippery slope
Analyzing lubricants for degradation is a cornerstone of many machinery condition-monitoring programs, and RCA is a powerful tool in the analysis toolbox. Measuring chemistry changes and detecting insoluble contaminants in the lubricant is the first step in identifying fluid degradation, which in turn is an important part of a lube oil failure RCA.
Oil analysis professionals should already be familiar with the priority list of fluid-testing techniques and procedures presented in Table 1. For those less conversant with fluid-testing options, the available tests are categorized as primary, secondary, and "other useful." Cross-check these tests against some of the common fluid degradation mechanisms listed in Table 2 and shown in Figure 2 to help in the selection process. Consult your fluids professional to determine which tests are right for you.
Usually, it's impossible to pick a solution to a problem before it is identified. Oil analysis is an exception: A well-designed program can select the optimal testing regimen (whether performed in-house or outsourced) for the specific needs of your plant.
When attempting to solve a nagging or catastrophic lube oil–related problem, it's tempting to jump quickly to a solution based on conjecture or a rush to get the plant back on-line. The RCA process is designed to help you resist the temptation to skip steps or make unsubstantiated assumptions. O&M professionals often find, inexplicably, that although there "wasn't time" to fix a problem once and for all, there was time to fix it again.
Following are two case studies that illustrate the points outlined above.
Email
Comments
Print
Reuse
