Create Comprehensive System for Asset Management

Doing the Right Work

Doing the right work really depends on two factors: (1) is it the right maintenance strategy, and (2) is there a formal scheduling process? By the later, I mean automated, resource-leveled weekly maintenance schedule based on backlog ranking as applied by WO matrix. If it's not automated, then the supervisor is telling the scheduler what work to place on the schedule. I'm not sure any human can fairly select the right work when there are large backlogs (e.g. > 500 work orders).

Defects Can Be Inserted Anywhere in the Process

If there was a way to discover defects (think of them as hidden failures) in the plant or system, then you could prevent the work from initiating in the first place. From the designer of the equipment, to the manufacturer, to the installer, and eventually, the operator/maintainer. Defects can be inserted anywhere in the life-cycle. You can be good at maintaining assets but the real money is in eliminating the defects so the work never really happens.

-- This is called defect elimination.

Setting up a Defendable PM Program

Each PM is linked to either an asset or location. A PM can have one or more job plans. But where did this PM come from? Perhaps it was uploaded from the legacy system. But where did that come from? How do you know this PM (maintenance strategy) actually addresses a failure mode?

Few can say they have a perfect system of maintenance strategies. For example, you can have PMs but they can't be traced back to any failure modes they are meant to address. Further, the plant conditions (and assets themselves) have changed over time.

-- but, even with this approach, you can't PM your way to reliability.

You Can have 100% PM Compliance but be Doing the Wrong Work

Whoa! That would be most unfortunate. I guess we can forget this metric.

If you had performed RCM/PMO analysis for your systems and assets, you would know exactly what maintenance strategies is required. But few have ever performed this analysis. And, if you did, where did you store this data-set? Answer: Usually it is in an external Excel file somewhere. And when a new failure mode is discovered, no one went back and updated the Excel file. Or, when a new asset is installed, ....[same answer].

You Also Could be Doing the Wrong Work if it was not on Weekly Schedule

Sometimes work is performed which was not on the weekly maintenance schedule. I'm taking about self-inflicted reactive as opposed to true emergency/urgent work. Stuff happens. Maybe the WO prioritization scheme was too narrow and failed to identify higher importance. Maybe the supervisor performing subjective selection failed to "select your work". Maybe the schedule was poorly put together. Maybe the owner of the work failed to properly prioritize or status the work. In any case, a proper Level--Plan--Schedule process should be followed. Having an automated routine to automatically level a backlog, enables the reviewers to run it multiple times for accuracy purposes, and then issue the schedule.

-- most organizations do not have a resource leveled weekly schedule

Without Feedback Loops you could be Missing Opportunities for Improvement

Root cause analysis (RCA) can identify a maintenance strategy. Chronic failure analysis can also identify a maintenance strategy. And (formal/validated) work order feedback can identify a maintenance strategy. Unfortunately, the majority of all organizations are not able to extract bad actors from their CMMS and drill-down on failure modes.

-- feedback is the best way to ensure continuous improvement, i.e. maintenance strategy refinement

Are you Configuring the System to Optimize Decision Making?

By configuring the system, we mean, designing the CMMS for reliability engineering. Most CMMS products focus on asset problem code. But is this simplistic approach adequate for the reliability engineer? Also, consider that maintenance is performed at the component level. And this is where the failure mode definition begins.This is the type of detail is needed on a WO.

-- most organizations fail to realize the potential of failure discovery and analysis within the CMMS

Three Key Points for Optimizing the CMMS for RCM

[A] Failure mode is the language of RCM per SAE JA-1011/12. But have we configured the CMMS work order to capture the failure mode? Per the RCM Blitz definition, failure mode is 3 pieces: failed component, component problem, and cause code. This needs to be 3 separate fields on the screen. Link the failed component to the ITEM.COMMODITY.

[B] If the results from an RCM/PMO analysis could be kept inside the CMMS [new application] then you can establish immediate validation of the failure mode.

[C] If a failure analytic is designed, and a reliability team runs this report in a monthly meeting to pull top 10 bad actor list, then the team can dynamically drill down on the 3-piece failure mode.... on screen.

Imagine the Possibilities

A work order was coded with functional failure flag marked as "Yes". At job completion, the maintenance technician identified the primary component which caused the failure -- and entered this value. The component problem code had a static domain where he chose the appropriate word. The cause code came from a 3-level hierarchy [This can be stored in a 3-level FC hierarchy] where he found the right answer.

Upon record save, an automation script checked to see if this failure mode existed for that asset. The RCM analysis failure mode app had the exact same validated field format (for the 3-piece failure mode). If it did not, then the script routed this data to reliability engineer for analysis. If he concurred as valid, then he goes to new RCM FM application and inserts new failure mode AND maintenance strategy. The script then sends a notification to PM coordinator to set up new PM and job plan.

But what if...the failure mode exists? Then why did this functional failure occur. Now we have a brand new set of questions to ask and answer. Again the reliability engineer is involved. I might start by going back to the work order itself and study that data. Plus I might talk with maintenance technician and operator.

-- -- -- this is the way a CMMS should be used