Could digital checklists be the way to go in your operations?
By Jane Alexander, Managing Editor
Whenever terms associated with reliability such as P-F (potential-to-failure) interval, Mean Time Between Failure (MTBF), Weibull, Markov, and the like are discussed in industrial operations, all heads turn to reliability engineers for explanation. But, despite being the people in a plant who understand these concepts, these individuals may not feel confident in the numbers shown on reliability dashboards. That’s where the power of maintenance checklists can shine.
As an example, asset management experts with Meridium, from GE Digital (Roanoke, VA, meridium.com) point to this hypothetical, but all-too-common, plant scenario: When a value is found to be increasing in pump A, a condition-based maintenance (CBM) engineer informs operations of the situation. Operations hands the asset over to maintenance to change the bearing. Maintenance personnel determine the mechanical seal is good, but that the oil seal is damaged. Since they are already changing the bearing, a decision is made to also change the mechanical and oil seals. Unfortunately, while pump A is being aligned, the mechanical seal on the operational standby unit (pump B) fails due to a process upset.
To avoid production loss, pump A is returned to operations, despite not having been completely aligned. The standby pump B, in turn, undergoes maintenance, and the seal is changed. The unit is then boxed up, aligned, and handed back to operations. The maintenance supervisor fills in the equipment asset management (EAM) system’s long text with observations and closes the notification for both pumps.
One month later, vibration in pump A starts to increase 3 to 4 mm/sec., probably because of misalignment. Upon opening the EAM history and reading long text that states “Pump bearing replaced,” “Vibration came down from 8 to 3 mm/sec.,” and “Mechanical seal replaced,” the CBM engineer decides to trend the vibration and not generate any recommendation to correct the alignment. After all, the unit had been maintained just one month before, and the vibration was still within limits.
Fifteen days later, the mechanical seal fails due to poor alignment. (Although vibration was to have been checked within 20 days, instead of 30, the seal failed before it could be done.)
Three months later, a reliability engineer is assigned to reduce the number of mechanical-seal failures in the plant by 30%.
This individual starts the study by reviewing the history of all pumps at the facility and learns that pumps A and B had three seal replacements within a month. The engineer, however, still had questions:
Regarding Pump A maintenance due to high vibration:
• Was the seal found in failed state? What is the reason of failure?
• What was the reason of bearing damage? Did seal failure result in bearing damage or bearing damage result in seal failure?
• Was the seal leaking when the maintenance was planned?
• Did personnel create a root-cause analysis for seal and bearing? (Probably not, as it was a corrective action initiated by a CBM recommendation.)
• Can I find the as-is condition of the spare anywhere, in photo or observation log or checklist? (Probably not, as we don’t have a policy to fill out a maintenance checklist for CBM recommendations.)
If maintenance had replaced the seal:
• Why did it fail within 15 days?
• Was the installed spare faulty?
• Was the maintenance technician new or lacking the proper training?
• With three failures in one month, is operations doing something wrong and hiding the truth?
Regarding Pump B breakdown maintenance:
• Before removing pump A from service, was the performance of pump B checked and confirmed?
• Was the seal leaking when the pump was taken off line?
• What was the “process upset” that damaged the seal and is it now being monitored?
According to Meridium, reliability-engineering personnel deal with similar questions daily. To complete their work, they end up performing an undocumented root-cause analysis, meeting with people, spending time just trying to understand what might have happened, and subsequently making assumptions, censoring some data, and formulating conclusions based on individual perspective rather than actual data points.
If your reliability teams have ever faced this type of dilemma, it might be time to revisit the effectiveness of your maintenance checklists.
A checklist’s lifecycle
The lifecycle of a maintenance checklist typically begins when a breakdown occurs and a technician notes important readings on a piece of paper. The piece of paper moves from one pocket to another as the shift changes. When the machine is handed back to operations, the supervisor is handed a piece of paper with numbers and checkboxes only. By the time the EAM history is updated, the count of such papers with the supervisor is in double digits. This person, in turn, closes the notification with minimum details and considers the work done. The piece of paper is then stacked somewhere for audit purposes.
In a traditional setup, the most valuable insight regarding an asset’s reliability is often not recorded and almost never trended. Items/information including and related to as-is, as-left, failed-part pictures and issues, first observations, immediate actions, component(s) replaced or repaired, and basis of maintaining a component should, ideally, be kept in a checklist and available for CBM and reliability engineers to consider while performing their respective studies—without wasting time to collect the information.
Keep in mind that a paper checklist is not just something with checkboxes and blanks for filling in information. It’s a valuable tool to track what was first seen, an asset’s condition, and what else that wasn’t scheduled was observed. It is not only for overhauls or breakdowns, but also for every opportunity when a maintenance technician goes to the machine with a permit.
The as-is information gives insight to the CBM engineer to back his or her analysis. There are peaks that go unnoticed because the as-is information may not have been available for the engineer to complete the analysis. Every CBM recommendation, if accompanied with as-is pictures and maintenance findings, will provide a basis to define other pre-dominant peaks in the vibration signals, therefore, bringing confidence to the existing CBM program. A reliable CBM program results in optimized time-based maintenance and reduces surprises in the field.
Similarly, the as-left readings and comments for all maintainable components, supported with photos, will help the reliability engineer calculate component MTBF, uncover hidden or unknown failure modes, calculate P-F intervals with higher confidence, and optimize preventive maintenance and CBM frequencies, resulting in higher reliability and availability of assets.
So, what does all this mean? Given the importance of your site’s maintenance checklists, consider digitizing them. Today’s digital technology allows technicians to create and fill out a maintenance checklist on the go, as it can typically be done through a mobile device, whether connected to the internet or not. Readings, observations, and photos can be captured and synced with a server, enabling policies and key performance indicators to be created and trended over time—ultimately boosting reliability efforts to the next level. MT
For more information about this and other asset-management topics, visit meridium.com.