TBM Can’t Stand Alone
Klaus M. Blache | April 1, 2021
Q: What’s wrong with time-based maintenance?
A: Time-based maintenance (TBM) is running out of time. Maintenance thinking has gone through many changes since the 1978 Nolan and Heap report for United Airlines. The report compiled information on the conditional probability of failure versus operating age for numerous electrical and mechanical airline-industry components. Its findings challenged the commonly held belief that component age and likelihood of failure were directly related. Instead, as we now know, only 11% of items reviewed were age related and 89% were random. It also pointed to two mistaken beliefs about how we traditionally scheduled PM.
Most assumed that:
• Components start with an inherent/as-installed reliability that will deteriorate with age.
• Likelihood of occurrence (statistics) can be used to predict when things fail to enable timely interventions.
As stated in the report, the earlier belief was that, because mechanical parts wear out, the reliability of any equipment is directly related to operating age. It was incorrectly thought that the more frequently PM was performed, the better protected assets were against likelihood of failure. The study found that many types of failures could not be prevented, regardless of the amount of maintenance performed. This led to numerous aircraft designed-in redundancies in key functions.
Because of these findings, high levels of time-based maintenance do not make sense. This kind of maintenance uses time as the main trigger for replacing or fixing an asset at a set interval or usage frequency, regardless of its condition. If you replace a component, or do a complete overhaul every year, it at least is on the schedule, but most likely is done too frequently. Examples include changing a filter every three months, replacing batteries in smoke alarms, seasonal checks on HVAC, lubricating a pump every 3,000 hr., and tightening screws/bolts.
The time-based maintenance goal, of course, is to prevent failures before they occur. But since it’s carried out, whether it’s needed or not, it’s not cost effective. Keep in mind that failure is a process over time, i.e., wear, misalignment, contamination. Once a failure begins, it can take days or years until the actual failure event occurs. In principle, you would want to do any time-based maintenance before the P (point where you can identify a potential failure) on the P-F curve. In practice, because of caution or because you always did it that way, maintenance is usually performed too often. This also introduces new “infant mortality” likelihood. Once you get past point P, it’s time to use technologies such as ultrasound, vibration or oil analysis, lubricant sampling, thermography, motor-circuit analysis, or corrosion monitoring.
There are two ways to approach your PMs.
Time-based maintenance (TBM):
• Easily scheduled based on set-time interval
• Minimal training needed
• Doesn’t consider asset condition based on use
• Usually results in too much or not enough maintenance.
Run-time maintenance (RTM), also called metered maintenance:
• Based on asset utilization, such as number of hours running, number of cycles, or number of miles
• More efficient than TBM.
You can also do a combination of the two. For example, every two months or 500 hr.
of run time.
There are situations when time-based maintenance is needed, especially when there are critical safety functions or regulatory requirements. Examples include checking fire extinguishers, boilers, and elevators. If you have parts with a known MTBF (mean time between failure), TBM can be considered if higher levels of maintenance are not cost effective.
A basic model for servicing is to ask if the asset/issue is high risk or a critical asset. If it’s neither, then you need to analyze what should be breakdown maintenance (very low risk) or preventive maintenance (low risk). If it’s high risk or a critical asset, you should be using predictive and condition-based maintenance. Those assets in the middle (medium risk) should be decided based on a ROI basis.
As I work with companies, I’ve found time-based maintenance ranging from 25% to 80% of total maintenance. I use 20% as a best-practice recommendation. Time-based is better than reactive maintenance and is easy to implement into a CMMS. If you have assets that continuously run, there are some possible benefits. Primary disadvantages are:
• increased overall costs in labor, parts, and downtime (if it affects production)
• maintenance too often introduces greater human error (lack of precision maintenance)
• as mentioned previously, there is no accounting for asset usage and condition
• maintenance, if not done often enough, leads to more downtime.
The airline study and following efforts resulted in huge reductions in maintenance, coupled with improved safety. As stated in the report, the initial study by special teams of industry and FAA personnel developed the first scheduled-maintenance program based on the principles of reliability-centered maintenance.
Time-based maintenance cannot alone be an effective way to manage maintenance with the many non-age-related failure modes (89% per Nolan and Heap). It should be used as part of a comprehensive reliability and maintenance process.
Reliability-centered maintenance (RCM) is also more difficult to perform outside of the airline industry (or similar highly regulated organizations) without mature and robust processes in place. The best-practice target should be about 10% reactive maintenance, 25% predictive technologies and condition-based maintenance (finding potential problems), 65% predictive maintenance (30% performing the work orders to fix what you find with PdM/CBM, 20% time-based maintenance, and 15% all other PMs).
The ultimate goal is design-in reliability and maintainability and ongoing continuous improvement, because you want to eliminate the reasons for all unnecessary maintenance. Learn how to decide when to use what type of maintenance to enhance your competitive advantage. EP
Based in Knoxville, Dr. Klaus M. Blache is director of the Reliability & Maintainability Center at the Univ. of Tennessee, and a research professor in the College of Engineering. Contact him at firstname.lastname@example.org.