Optimising Maintenance Strategies

“When optimising maintenance strategies often the focus is on setting an optimum interval for Inspection or PM tasks.

It is also worth stepping back and challenging the actual task itself and rather than say using visual inspection routes to look, listen, smell, touch and feel, it may be worth considering what other techniques are available such as vibration analysis, or infrared, or acoustic devices that can decrease the frequency of inspection by increasing the warning time or P-F interval and also the likelihood of actually detecting the warning signs.

This tip might sound obvious, but when working with onsite groups, they often can fall into groupthink where they do not step back to challenge existing practices. This is one area where an external facilitator has value to company RCM teams.

Tip provided by Mick Drew, Director ARMS Reliability Engineers

Reliability Block Diagrams

“One of the reasons why RCM studies stall is that too much analysis takes place before benefits are realized. This really stresses the stakeholders and project sponsors in being able to continue investing resources into a study before benefits are realized. So when initiating an RCM study as part of a reliability improvement initiative, select a small ‘bite-size’ system that has the potential to provide a significant impact to the business. This way, the stake holders will be more likely to buy into the process and support further studies. One way to identify high-impact areas is to construct high level Reliability Block Modelling (RBD) and use downtime records to identify the ‘bottle-neck’ areas.

This is powerful when simple Pareto charts do not provide guidance, and so useful for complex plants that may have multiple levels of redundancy or buffers. ”

Tip provided by Mick Drew, Director ARMS Reliability Engineers

Improving Equipment History

1. An accurate equipment history is valuable to identify bad actors; emerging trends, longer term systemic problems, and create failure mode data sets to forecast likely future failures.
2. Reliability Engineers are often frustrated with the poor quality of data recorded and often have to resort to anecdotal information captured through facilitated reviews.
3. In order to improve the quality of data from busy field personnel, they need to “buy in” to the value of recording accurate data.
4. Adding summary text comments usually gives better information than dropdown lists but it still needs some structure.
5. Text comments entered on work orders after completing a job needs to describe what work was done, what failure modes were evident, what were the effects of failure, and what is recommended by the technician in the future.
6. The downside of text box has been the time necessary to type the comments, but with increasing keyboard skills across the enterprise, this factor is less of an issue.
7. Analysing text comments in database applications is another difficulty, but a check list following the text box, or regular review of feedback by a Reliability Engineer, helps overcome this difficulty.
8. If the checklists are constructed to be consistent with the RCM analysis, the feedback data can provide valuable updates.
9. If the technicians are involved in the RCM analysis and in particular, regularly perform an RCM update, then they will have greater understanding of the value of their comments, and greater “buy in” to the resulting changes to strategy.
10. This matches the concept that in a proactive system, technicians roles move from “fixing” equipment to managing reliability of processes.

Tip provided by ARMS Reliability Engineers

Reliability Centred Maintenance (RCM) Tip

When adding failure modes to the FMEA part of an RCM study ensure that the failure mode is the physical cause of the loss of function. After all it is the physical failure mode behaviour that we seek to change when deciding what maintenance to do.

So for the failure modes associated with no flow:

• don’t say pump stopped, say broken shaft or seized bearing, or collapsed impellor or similar. In other words if you can relate the physical mode of failure to the loss of function, maintenance technicians will know exactly what happened and the maintenance task can be quite specific.

Building an effective FMEA provides an excellent troubleshooting guide for both operators and maintainers, and a side benefit is that it can be used to build reporting codes in work order management systems.

Tip provided by Mick Drew, Director ARMS Reliability Engineers

The Function Statement

The FMEA part of conducting an RCM study is vitally important to the quality and acceptance of the study by other stakeholders.

If you find your functional failure statement simply “fails to perform the function” your study may be hard for others to relate to. In order to relate to operators and improve the quality of your analysis:

• Ensure functional statements relate to the purpose of the item actually being required.

Next your functional failure statement needs to represent the operational problems regarding the loss of function.

Example boiler feed water pump.

Function Statement: Don’t say function is to pump water instead say function is to deliver feedwater to boiler at x litres per sec.

Functional Failure Statement: Don’t say fails to pump water or fails to deliver, instead say, low flow or no flow, or intermittent flow.

Operators and maintainers are more likely to relate to the analysis, if the terminology is the same as what they would use in reporting problems.

Tip provided by Mick Drew, Director ARMS Reliability Engineers

Preventive vs. Corrective Maintenance – Clearing The Confusion

Historically, some confusion exists on how to classify maintenance tasks as preventive (PM) or corrective (CM). In its simplest form, PM tasks are inspection and servicing actions that are performed on a scheduled basis, and CM tasks are unexpected actions that require an unscheduled response. However, the confusion arises primarily in two commonly encountered areas.

The first deals with predictive and condition directed PM tasks (PdM and CD respectively) which measure and trend equipment “health” parameters for signs of incipient failure modes. Such measurements are accomplished on a scheduled basis – and pre-planned actions are also scheduled to occur when the incipient failure condition progresses to a critical stage, thus preventing outright failure. Notice that this form of PM is pre-planned and scheduled in its entirety, including the final action to preclude outright failure. This scenario is preventive maintenance from start to finish – i.e. the maintenance actions to preclude an outright failure are part of the preventive – not corrective - scenario.

A second issue deals with RCM-based decisions to deliberately run-to-failure (RTF). The scheduled strategy here is to do nothing until actual failure occurs if safety, uptime and economics of restoration are not compromised. Even though the actual restoration tasks cannot be scheduled, it is nevertheless pre-planned. From the outset, restoration of the failed item is pre-planned to occur and, as above, this scenario is likewise preventive maintenance from start to finish.

In both of the above issues many people tend to categorize the restoration action as corrective. Not so – it is preventive! When these two issues are not properly treated and recorded as preventive (not corrective) actions, the metrics that are measured and reported become distorted and management is apt to conclude a distorted view about the overall effectiveness of their PM program.

Tip provided by Mac Smith
Author, RCM Gateway to World Class Maintenance
http://www.jmssoft.com

To overcome the problem of not implementing the results of an RCM study, make sure the study includes the following outputs:

Maintenance tasks grouped to a maintenance plan

Maintenance plan that identifies associated equipment and/or functional locations, with frequency and duration of tasks.

Ensure all tasks in the plan have work instruction documents so tradesman know what needs to be done.

For all tasks required by the plan, document what the failure modes are that the tasks address, so that when undertaken the maintainers know what to look for.

All maintenance plans should have a resource requirement prediction before loading to CMMS so you know you have enough people.

Make sure you also predict the level of unplanned or breakdown work likely to occur to guard against having to redeploy resources from important planned work to urgent breakdown work.

Make sure the plan will deliver plant availability required by the business plan.

Make sure the plan meets the corporate risk profile.

Make sure there are spares available to undertake both planned and unplanned work.

Finally, make sure you have a budget necessary to do both planned and unplanned work.

Tip provided by Michael Drew
ARMS Reliability

Send mail to mace@macecg.com.au with questions or comments about this web site.
Last modified: 01/15/11