In a fleet of over one hundred units there are four spare engines. Should an age based overhaul policy be used? Scheduling overhaul based on age seems appropriate due to the need to level the load in the engine rebuild shop. Given the expected failure rate, a rebuild interval of, say 10000 hours, and a known overhaul time distribution, we can, using age-reliability modeling tools, set the number of spares and schedule engine overhauls such that there are few field failures, and, at the same time prevent the rebuild shop from becoming over-saturated beyond its capacity to deliver the needed availability to the operation.
If, on the other hand, the age based maintenance policy is replaced by Condition Based Maintenance, (CBM) it is feared that control would be lost. CBM could indicate more than four engines requiring overhaul at a given moment. This situation can happen non-homogeneously. Our shop resources would become saturated in one time period and underused at another. More importantly, operations could find themselves without any replacement engines stocked to meet the demand, thereby incurring downtime.
“TBM or CBM” for a fleet of exchangeable components such as engines, is a common question but it is the “wrong” question. Phrased that way it misleads us into thinking that the answer will be one or the other of the options, “CBM” or “TBM”.
This is really rarely the case. The question that we must ask, rather, is how to use CBM to gradually improve the average life of an engine. What is the optimal mix of CBM and TBM? The current age based strategy (TBM) should be considered as the baseline or the “safest” strategy because it has evolved from experience or from the engine manufacturer’s recommendations, which are usually conservative.
However, CBM will sometimes overrule the TBM decision.[1] This will happen when we are reasonably confident (i.e. careful) based on observational data that an engine with a lower number of hours is in worse condition than an engine with a higher number of hours. The performance (in terms of availability and cost) of our decisions (a mixture of CBM and TBM) will be analyzed over a period of time. This may allow us to extend (carefully) the overhaul interval.
The gradual transformation from the baseline TBM decision model towards a greater number of decisions based on CBM can occur as more confidence is gained in CBM. This confidence is indeed measurable as the standard deviation about the conditional mean time to failure (RULE) as reported by the software (EXAKT).
Growing confidence in a maintenance organization’s ability to make good CBM decision will depend on two factors:
- The skill to identify and record in the CMMS instances of (RCM) failure modes and to record their life-endings as either “Failure” or “Suspension”.
- The availability of condition monitoring data “reflective” of failure modes that actually occur.
We can gauge the development of our skills described in Factor 1 above. Confidence is measurable as the standard deviation of the conditional density function. It is a measure of the amount of spread around the mean. The mean is reported by the software as the Remaining Useful Life Estimate (RULE). This average or “mean” of the conditional density function is also known as the “condtiional” MTTF.
Additionally, after a period of time, a method (the EXAKT cost function) tracks equipment availability and cost. Eventually we may find a relationship between a fleet’s engine availability and the degree of use of CBM for overhaul decisions. A KPI such as the number of work orders issued as a result of time based maintenance divided by the number resulting from CBM will quantify the CBM “energy” expended.
With regard to workshop loading:
- Using TBM instead of CBM can’t guarantee a constant level of work load for the engine rebuild shop. Instead, the failure rate, which depends on the usage of the whole fleet, is the main factor for the work load. For example, if you use the trucks in the way that a large group of engines have about the same age (and assuming they are working in the same condition), then no matter whether TBM or CBM is used, it’s very likely that you will have a very high work load in the rebuild shop at some point in time, and very low work load at other times.
- TBM is usually more conservative than CBM, and thus TBM will on average require more maintenance, generating more work load in the rebuild shop.
- If you can control the work load in the rebuild shop via TBM, you can do the same via CBM, provided CBM is applicable (reliably detects potential failure) and effective (is worthwhile). In TBM, you set your overhaul interval, and in CBM, you set your condition variable control limits.
- A LRCM program, wherein life cycles ending in failure are well distinguished from those ending in suspension, will continuously improve the applicability and effectiveness of CBM. This will lead to the gradual increase in CBM based overhaul decision making.
© 2011, Murray Wiseman. All rights reserved.
- [1]We might think of an equipment fleet as a hockey or soccer team. The coach monitors the players on the field (or on the ice). He replaces a tired player with one who has been resting on the bench. The coach would not remove a player who is in good “condition” only because he has been playing longer. Rather he would replace one whom he observes to be exhausted.↩