Our reliability and maintenance group use a mental model for an assets’ life cycle called the D-I-P-F curve. In this illustration, you can see how design, installation, probability of failure, and failure are related over the life cycle of an asset. There are critical points in the life cycle. When explaining these, we will often use humanistic terms to personify assets. This helps operations people relate to the concept of assets and the care we can provide.
"Predictive technologies are a balance between the cost of data, the ability to address issues found, and the cost of failure"
In particular, we are interested in condition-based monitoring point on the curve. Condition-based monitoring is much like a routine physical with your physician. Based on your physical characteristics, age, and basic vitals, doctors will prescribe tests to understand a person’s state of health. If the tests are out of acceptable criteria, the doctor can confirm with additional proofs or begin a treatment plan. The treatment plan can vary from medications to lifestyle changes. We all know early detection saves lives. It’s the same for machines, early detection allows for a controlled, planned repair or replacement at a cost-effective point in time, often extending the life cycle of an asset.
For an asset like a pump, the essential element of condition-based monitoring is an operational check. Consider these to be the pumps vitals, pressure and flow rate. How does it sound? How is it performing? An odd sound or performance degradation may prompt additional tests. In some instances, performance can be monitored by a computer triggered by the reliability limits of operation. As flow rate deteriorates, we can predict they wear out for pump rebuild or replacement versus a calendar or run to failure.
On more critical assets, we may also choose to perform vibration analysis to check the bearing condition. Indications from the vibrational analysis can get ahead of issues months to years before functional failure. Even before there is a vibration due to bearing wear, there is a metal loss. The metal loss can be detected as particles in the oil. Oil analysis can be a corroborating test or a leading indicator before vibrations are present. Oil analysis can also keep the grease in good shape to mitigate premature wear.
A common pitfall is to over monitor or misapplies predictive technologies. Predictive technologies are a balance between the cost of data, the ability to address issues found, and the cost of failure. Application of predictive techniques should come from a reliability strategy and make sense. The reliability strategy should define the technology and acceptance criteria. Unfortunately, many times, a predictive technology or service is brought to a site in an improvement effort. The team feverishly looks for all assets to cookie cutter into the program to leverage the technology. This generalization process can dilute the activities, causing a cursory look at the data, leading to missed or misdiagnosis. A more targeted approach is a better solution to make gains. The focused approach starts with a criticality analysis to identify the most critical assets. Once the list is created, reliability strategies can be developed. The reliability strategies then drive predictive technologies with acceptance criteria. Once there is an indication, the correct response can be implemented in line with the finding and consequence of failure.