I have a small assignment for you maintenance managers and reliability engineers (the bulk of our readership). Import the task list from a typical scheduled preventive maintenance (PM) event into a spreadsheet under the column header "Task." Then, create a second column header titled "Value Proposition" and a third column called "Parts and Labor Cost." Then, one-by-one, start classifying each task into one of the following six categories and assign an estimated parts and labor cost in annualized dollars (e.g. multiply the cost of quarterly costs by four and divide the cost of biannual tasks by two). Take a hard look and ask yourself if you're getting the kind of return on investment you'd expect from preventive maintenance activities that are intended to extend the life of equipment and assure the dependability of manufacturing systems.
Be honest and ask yourself if your PM plan is working for you or against you.
Category No. 1: Non-value added - failure-causing replace/ rebuild. Many PM tasks still call for the technician to perform calendar-based changeout of parts, subassemblies or entire assemblies despite the fact that these components don't fail as a function of time.
In their book "Practical Machinery Management for Process Plants", Heinz Bloch and Fred Geitner report the typical Weibull shape parameters for various industrial parts and subassemblies. In essence, the Weibull shape defines the relationship between failure rate and time to determine if the risk decreases over time, increases over time or remains constant.
Let's explore the time-based rebuild-or-replace strategy for a pump that operates 8,000 hours per year. A centrifugal pump, according to the data from Bloch and Geitner, has a typical life of 35,000 hours and exhibits a running Weibull shape of 1.2, meaning that the failure rate increases slightly over time. However, pumps exhibit a run-in Weibull shape of 0.5, suggesting that there is an increased risk following rebuild or replacement, which is primarily attributable to all the things that can go wrong when a machine is removed, taken apart, put back together and then reinstalled.
Figure 1 exhibits the risk profile for the pump in terms of failure rate over its theoretical life, expressed in percent. As the chart illustrates, the scheduled biannual rebuild increases the average failure rate over the pump's life cycle. This occurs because of early life failures. Just about the time we've worked through the time period during which the risk of early life failures is highest, we schedule a rebuild and the process starts over again!
Category No. 2: Non-value added - failure-causing basic care. Basic care is essential to the health of machines. Lubrication, proper fastener management, balance, alignment and other proactive measures extend the life of equipment.
Lubrication is among the most commonly executed basic care measures in most industrial plants - it's essential to mechanical reliability. I frequently see PMs that read, for example, "lubricate the conveyor." This seems to make sense, right? Let's break this down further. A conveyor typically includes a motor, couplings, a gearbox, head and tail roll bearings, idler bearings, etc. Most of the tasks require different lubricants, different intervals and, frequently, different tools. Lumping everything together into a single PM leads to overlubrication of some components, under-lubrication of other components, suboptimal lubricant selection (e.g. a multi-purpose grease for the motor, coupling and bearings) and lubricant mixing.
While it's clear that we need to lubricate all of the components, we must treat the tasks individually for them to add value.
Category No. 3: Non-value added - ambiguous. Inspections are among the most common tasks found in a maintenance plan.
Of course, we want to inspect our machines to identify opportunities to respond to ensure that proper conditions are maintained and to respond expeditiously to problems before they have a chance to cause a functional failure. Too often, we see ambiguity in inspection assignments. For instance, "check pressure," "check temperature," "inspect electrical system," etc. What do these tasks mean to you? Without a clear definition of what OK looks like, inspections are empty and valueless. It may be a picture, it may be a number, it may be a text description, but in any event, the PM task must clearly define the expected state and/or define what constitutes an abnormal condition.
Replace "check temperature" with something like "confirm running temperature is between 100 and 105 degrees Fahrenheit and report deviations to engineering" and you've got an executable PM.
Category No. 4: Value added, but wrong interval. Frequently, tasks are appropriately selected and clearly defined, but the interval is improperly assigned. For example, a PM to confirm that the temperature falls within a specified range is a great inspection, but if we're only looking once a year or once a quarter, we're really not applying process control. Temperature should be measured and trended with a high degree of frequency.
For some systems, continuous monitoring is justified. At a minimum, check it every day for continuously operating equipment - every shift is even better.
Grease lubrication is another PM task that is frequently performed at the wrong interval. Sadly, when a machine fails for causes unknown, we frequently react in a knee-jerk fashion. Under the gun to do something to avoid recurrence of the failure, we often wrongly increase the frequency of lubrication. In some cases, it's justified; but frequently, it just makes matters worse. Unfortunately, within a very short period of time, the new PM interval becomes the de facto standard, and nobody recalls the reason why.
Category No. 5: Value added, but improperly assigned. I'm a fan of operator-driven reliability. I've always believed that operations should own the reliability of manufacturing processes because they oversee all of the affecting functions - including upstream and downstream supply chain. A large percentage of PMs are inspections. Basic, non-intrusive inspections should be carried out by operations personnel. They live with the machines and are in the best position to identify abnormalities. For instance, as the operator of your own vehicle, you're in the best position to identify an unusual sound. Unless the sound is clearly abnormal, a mechanic can't hear minor changes in the way your vehicle sounds his first time behind the wheel.
Make operator inspections clear and straightforward, preferably reduced to "yes or no" questions. Trend the data and respond quickly to reported variations. If maintenance technicians ignore the results of operator inspections, they'll quit doing them.
Category No. 6: Value added - essential, clear, properly assigned and performed at the correct interval. Enough said! This is the goal. Convert scenarios one through five into this state and your preventive maintenance process is working for you, not against you (Figure 2).
Preventive maintenance is among the most common root causes leading to the need to perform corrective maintenance. It need not be. Evaluate your PMs and eliminate tasks that fail to add value or actually create failure. Eliminate the waste and ambiguity and properly assign the tasks at the proper interval, and avoid the temptation to knee-jerk react to failures by simply adding new PMs to the system or increasing the frequency with which tasks are executed without proper cause analysis. You'll find yourself spending less money on preventive maintenance and, at the same time, increase the reliability of your manufacturing systems.
Research suggests that those organizations that spend the least on maintenance typically enjoy the greatest level of manufacturing reliability. Take a hard look at your PM system. Where do you stand?