How to determine the best maintenance policy

David Berger explores the differences between use-, failure-, and condition-based maintenance. By David Berger, P.Eng., contributing editor.

David Berger, a Certified Management Consultant (C.M.C.) registered in Ontario, Canada, is a Principal of Western Management Consultants, based in the Toronto office

Any company that has physical assets must maintain them over time, in order to maximize their return on those assets. Obviously, running all your assets to failure does not yield the highest return, but then neither does 100% preventive maintenance — that is, no assets run to fail. Comparing these two strategies in terms of cost, 100% preventive maintenance requires maintenance at regular intervals in order to prevent failure. This can be costly compared to simply replacing or repairing the asset upon failure. However, there are other factors that must be considered, such as the probability and impact of failure, the performance and reliability of the asset, and the age and condition of the asset.

Maintenance policy options

There are only three ways that maintenance can be triggered, regardless of asset type or its attributes. These correspond to the only three maintenance policies possible:

1. Use-based maintenance (UBM): This policy refers to maintenance triggered by time, meter, or event. For example, use-based maintenance on your car engine might be defined as changing the oil every three months, every 3,000 miles, or after every fourth tank fill-up. This corresponds to time, meter, or event driven UBM.

2. Fail-based maintenance (FBM): By running an asset to failure, you are adopting a fail-based maintenance policy for that asset. In the car example above, not changing the oil at all and running the engine to failure implies adopting an FBM policy.

3. Condition-based maintenance (CBM): The only other maintenance policy possible is condition-based maintenance — triggering maintenance when the condition of an asset exceeds an upper control limit, drops below a lower control limit, or trends in a given manner — for example, the slope is rapidly increasing. In addition to the actual maintenance work, a CBM policy necessitates regular condition inspections, from on-line real-time to manual inspections at regular intervals. Following the car example, choosing CBM over the other two policies would require regular inspections of the oil and sophisticated lubrication analysis to determine the level and characteristics of particulates. A cheaper alternative would be to visually inspect the oil to determine how black and dirty it is; however, this method is not as accurate in optimizing engine performance and lifecycle cost.

The three policies above replace the many words used to categorize maintenance work done. These include preventive, corrective, demand work, proactive, elective, predictive, daily/routine, scheduled, reactive, and many more. Ask three people in your company what these terms mean, including what triggers maintenance work, and you will undoubtedly get three different answers.

The most often used but least understood of these terms is “preventive maintenance.” If you ask your employees whether or not an inspection is preventive maintenance, most would say it is. This is false because an inspection is not maintenance at all. Nothing is repaired or replaced. It is a requirement of CBM in order to track when maintenance is triggered. If an inspection of the oil shows that replacement is required to prevent failure, then condition-based maintenance, not UBM, is performed. Use-based maintenance implies that the oil is simply replaced at a given interval, regardless of inspection results. Similarly, if upon inspection it is discovered that an asset has failed — for example, a light bulb is out — then the trigger for maintenance is failure, and it is classified as FBM.

The confusion lies in that companies mix the three policies on a given work order, thereby making it difficult to analyze asset history to determine the optimal maintenance policy. Keeping careful track of policies on a work order allows planners and asset owners to better understand for each policy, the relationship amongst probability of failure, impact of failure, asset performance, reliability, lifecycle cost, aging, and other factors.

Factors to consider in choosing the maintenance policy

Sometimes the best policy is somewhat obvious, such as FBM for changing a light bulb in an area where there are many other light bulbs and few people. However, for many assets or components, the choice of maintenance policy is not obvious. Even for the same asset, a light bulb, there are several factors that may favor a policy other than FBM.

Risk: One of the most important factors to consider in determining the optimal maintenance policy is the probability and impact of failure, or risk. For example, suppose the light bulb is in a safety-related device such as an emergency exit sign. Similarly, what if there is only one light bulb used to illuminate a short walkway through a park used by university students. In these two examples, UBM is probably a better option since FBM is too risky from a safety or potential liability perspective, and CBM too costly relative to UBM.

Maintenance costs: Each maintenance policy has an associated ongoing and one-time setup cost. CBM is typically the most expensive as it requires both regular inspection and maintenance work orders, which in turn may require costly production downtime. As well, there may be one-time setup and equipment costs (sensors, monitoring equipment, analytical tools), or ongoing service costs such as vibration or lubrication analysis performed by a third party. UBM is typically cheaper than CBM to set up but is usually more costly on an ongoing basis because maintenance is done at regular intervals to avoid failure, regardless of condition. Note that maintenance work orders such as repair/replace are typically more expensive than inspection work orders. FBM is usually the least expensive in terms of maintenance costs if the impact of failure is low.

Benefits: The value in selecting the best maintenance policy is not just lower maintenance cost. Other potential benefits to consider are lower asset lifecycle cost and greater asset availability, reliability, performance, and quality of output. In some cases, the greater benefit of one maintenance policy over the other lies in the cost avoidance. For example, avoiding failure through a more expensive CBM policy may be warranted to avoid delay or loss in production, loss in reputation leading to reduced sales, or reduced health, safety and environmental damage/costs. Most maintenance departments tend to fixate on maintenance costs, but these other factors may be far more substantive in the short and long terms.

Thus, for each asset or component, starting with the most critical, ask four questions.

  • What does this asset/component do?
  • What happens if it fails, from negligible to catastrophic impact, in terms of financial, environmental, health and safety?
  • What is the probability of failure?
  • What is the best maintenance policy that balances the ongoing and one-time cost of the policy, the risk, and associated benefits?

Once the optimal policy is determined, the CMMS is used to build work plans for all three policies, plan/schedule UBM work and CBM inspections, and forecast CBM and FBM.

Other factors: The optimal maintenance policy should also consider factors such as:

  • the remaining useful life of the asset or its parent (the plant will be shut down next year, so it is not cost-effective to replace a related asset)
  • the asset’s degradation curve (suppose there is greater risk of failure at the beginning and end of a particular asset’s life, so adjust frequency of CBM inspections accordingly)
  • aging factors that cause an asset to degrade faster or slower (an asset running in a hot and humid environment at peak load for an extended period requires more frequent UBM).