Question: In your recent post “How can I tell if my PM program's a turkey?” you wrote: "Remember that most PMs should be inspections that are focused on finding equipment in the act of failure, i.e. the P-F curve. If the craftspeople are fixing everything on the PM and not generating corrective work orders, there is most likely very little backlog to plan and schedule. This lack of corrective work impedes proactive work management and drives down the levels of craft efficiency achievable. Intrusive PMs can increase the level of self-induced failures by disturbing otherwise stable assets."
Are you saying that corrective actions taken during a PM should be recorded on separate work orders? And if so, why?
Allen, reliability engineer, Iowa
Answer: Hello, Allen. Consider that the goal of planning and scheduling work is to eliminate avoidable delays (looking for information or parts, starting work that's not a priority now, or waiting on equipment to be available for work). We want to reduce the level of reactivity and drive improvements in craft efficiency. With craft efficiency, we are not trying to make people work harder; rather, we want to give them the things they need to do their work. Not having the time or materials to do their job is a typical technician complaint.
All that said, we want to move from unplanned work, which is typically reactive, to proactive planned and scheduled work. So a metric we use is planned vs. unplanned (emergency or urgent work that happens in the current week). We look for a target of 90% planned and 10% unplanned or reactive work.
On the P-F curve, we have the point of installation or restoration – some period before we can find the potential for failure (the "P"), and the point of functional failure when the asset no longer meets the user requirements (the "F"). We know that finding things in the act of failure and having a window of time in which to plan corrective work, get the materials without expediting them, and schedule the work for the optimum time is the least-expensive route for dealing with the majority of failures. Most failures are random, not time-based.
The inspections/repairs that you mention are being executed on a time-based approach when the technician is identifying the "F," not the "P," from above. The techs are doing the PM, finding the failure, and executing the repair. The "do it now" approach prevents the planning and scheduling of work. We accomplish less work during the same period; parts not stocked are expedited; more stocked parts are required; other PMs get deferred; costs rise; and the organization is way more reactive than necessary.
With inspection PMs, technicians should be looking for the "P," as I indicated. When finding the "P," they should generate a corrective work order that is then executed via the work execution process, i.e. planning and scheduling. The ‘P” was an established point. Say the pump is capable of producing 120 GPM and the user wants a minimum of 100 GPM. Assume establishing the "P" at 107 GPM when the impeller has degraded to that point that allows us to plan, schedule, and execute the work at optimum time and costs. When a technician is performing the inspection and finds a value less than 107 GPM, he or she has two choices.
- If the value is in a range from, say, 107 down to 105 GPM, then a corrective work order can be generated, as we have discussed. The CMMS work type should be set to ‘CM from PM’ as an example to show that the item was found on the PM for later use to demonstrate PM program value. This approach is the ideal situation.
- If the flow is closer to the point of failure "F," and the pump will fail at some point to meet the user requirements before the next repair window, then the technician can decide to replace the item now (fix it now in the PM window).
When deciding to fix it now, there is a rule of thumb that most organizations overlook: Set a time limit for fixing things with the PM work order. If the repair takes less than 20 or 30 minutes, capture the labor and materials on that PM work order. In the case of the impeller, the time required will be much greater. If the time exceeds the value that you set (30 minutes in this example), then the technician must write an emergency work order to fix it, as an emergency is what it is – the technician found the failure. At this point, the repair is unplanned and reactive. As is typical with reactive work, the technician spends more time chasing the parts and getting information to do the work, and costs are higher than if the work were planned and scheduled. Not to mention, the technician likely can’t complete the other work assigned for the day.
Another benefit of generating corrective work orders from the PM is that it will help determine the actual time required to execute the PM tasks on a repetitve basis. If you're fixing broken items when doing the PMs, the time required will fluctuate greatly each time. We can never put together an accurate work schedule with this approach.
Poorly designing the PM tasks to where technicians find the "F" on what should be inspection PMs means that many PMs require asset downtime. These PMs reduce asset availability and reduce production’s willingness to find line downtime for PMs and corrective work because you can never tell them how long the PM activities will take to execute. Your business processes for work execution should have all of this documented as training tools and ways of working.
What other insights might you have on this topic that did I not mention? Send me an email at the address below and I will respond. Please post your comments so everyone can all learn.
Jeff Shiver, CMRP
If you have problems in the fields of maintenance, reliability, planning and scheduling, MRO storerooms, or leadership as examples, please contact Jeff Shiver with your question(s) here.