Choose the right PM task frequency
Having a background in mechanical engineering, I have always appreciated the level of analysis necessary for many conventional engineering tasks. For example, when asked to consider a bearing carrying a particular load at a particular speed, the engineer’s job is not over after determining in which direction the load is being applied.
An engineer should determine the amount of applied load and make the calculations to determine the size and type of bearing needed. In this sense, would it be enough for a reliability engineer to look at an asset’s failure history and simply point out that time-based maintenance should be performed? Would it not be better for the engineer to recommend condition-based monitoring or calculate the optimum replacement interval if wear-out is the predominant failure mode?
Unfortunately, I encounter many organizations that ignore, fail to consider, or simply don’t record the failure history of their equipment and are forced into defaulting to time-based replacements and guessing at the maintenance interval. What would happen if electrical engineers didn’t measure or inquire about the incoming voltage and simply specified a 120 Vac motor? What if a structural engineer never asked where the building was going to be built and as a result failed to consider seismic or wind loads of the particular area? As reliability and maintenance engineers, we need to embrace the scientific aspects of our profession and treat our daily activities with the same regimen as our other engineering peers.
To this point, I’d like to focus on the determination of a maintenance plan. A maintenance plan is the actions taken to mitigate failure modes and ensure asset reliability. Although building a maintenance plan is only one part of the reliability and maintenance engineer’s job requirements, it is an important one, and a strategic maintenance plan can deliver immediate results.
There are two main components of an effective maintenance plan: the type of maintenance task and the task interval. I will use three main maintenance task types. For readers who are familiar with RCM, you will recognize these types as On-Condition Tasks, Hard-Time Tasks, and Failure-Finding Tasks.
On-Condition Tasks are periodic tasks, such as inspections, checks, or tests, designed to detect a potential failure and allow correction before it becomes a functional failure. This means periodically looking at the component and comparing it to a standard that ensures the component can still fulfill its function.
If you are familiar with the P-F curve, an On-Condition Task is one that looks for the P point on the curve with the intent that the detection gives adequate time to have the failure corrected before the F point on the curve. Constructing a P-F curve requires recording the results of the check and plotting the result versus the elapsed time from installation at which the check was conducted.
If enough measurements are taken, a fairly consistent curve can be developed for each failure mode. Making sure that the data is gathered carefully and consistently will aid in increasing the quality of the P-F curve.
Picking the proper frequency to make the check or inspection can be a little tricky. The basis for the frequency is some fraction of the P-F interval. In most situations, picking a frequency that is 1/3 of the P-F interval will be adequate, but in more critical situations using 1/5 of the P-F interval will give better confidence that the check will be made and the negative condition found and corrected.
On-Condition Tasks are preferred over any other task type. They allow for most, if not all, of the useful life of a component to be used while still avoiding the failure. This category includes the traditional predictive maintenance (PdM) tasks. Keep in mind that PdM does not always have to involve sophisticated detection equipment. Using your human senses, gauges, or simple tools can also detect conditions that indicate a potential failure has occurred.
Hard-Time Tasks are periodic tasks, such as removal, rebuilding, or reworking, that are designed to prevent a functional failure. This means that regardless of the condition, the equipment will be worked on when a calendar date, number of miles, or number of cycles has been reached. The intention is to remove, rebuild, or repair the equipment just before the probability of failure (hazard) reaches an unacceptable level and return the equipment to its original condition. Constructing a hazard profile for the equipment involves collecting life data (times to failure) for the failure mode in question, as well as calculating the probability of failure for potential replacement frequencies, and plotting the probabilities versus time. Remember that Hard-Time Tasks are only applicable to failure modes that result in an increasing hazard profile at the end of the equipment’s life.
Selection of the task frequency will depend on the shape of the hazard profile and at what point the curve begins to rise. Picking a point that is directly before the rise in slope will result in the most useful life, but may result in occasional failure due to slight differences between curve profiles on each installation.
In contrast, picking a point far away from the rising point of the curve will result in fewer unplanned failures, but will shorten the useful life of the equipment. It is preferred to pick a point that is close to the rising point of the curve, but how close will depend on your confidence in the consistency of the curve.
Hard-Time Tasks have the dubious honor of being the most selected maintenance task type, but probably also the most misunderstood. A number of studies have shown that the majority of equipment does not exhibit an increasing failure rate at the end of life. This would mean that a lot of the time-based maintenance replacements and rework are potentially being done to mitigate failure modes that are not suitable for this type of maintenance task.
Failure-Finding Tasks are periodic tasks, redundant equipment checks, or emergency system checks that are designed to detect a functional failure that has occurred, but is not evident. Evident failures are ones that will be noticed by operations during the normal operation of the equipment, while hidden failures are ones that would not be noticed. Examples of hidden failures are seizing of stand-by pumps and sensor fouling on smoke alarms. The intention of these tasks is to check on equipment to uncover functional failures before a need for the equipment arises and only then is the failure detected. These checks can be anything from test-running the equipment to observe the operation to simulating a hazardous condition to ensure the safety alert equipment is operational.
The primary reason for selecting the appropriate frequency to perform Failure-Finding Tasks is to ensure the availability requirements are met. If the frequency of inspection is too great, the functional failures that are occurring to the equipment are discovered when operated versus during a check. Consideration must be given to the mean time to failure (MTTF) of the equipment as well. For equipment with very large MTTFs, the inspection frequency can be extended, while equipment with small MTTFs will need to be checked more frequently.
I encourage anyone responsible for the creation and upkeep of a preventive maintenance program to embrace the engineering aspects of this work and give it the attention that it deserves.
Bill Barto is a reliability engineering subject matter expert with Life Cycle Engineering (LCE). He can be reached at bbarto@LCE.com.