Unexpected failure of machinery is never a welcomed option for any plant, especially in the case of critical assets. Productivity may take a big hit when an asset goes down for emergency fixes. Could a developing problem have been detected earlier? Could the failure have been prevented or even predicted? An operation’s maintenance strategy — or lack of one — holds the key.

Among typical strategies, reactive maintenance can represent an absence of any organized maintenance planning and control, although for selected equipment this may be appropriate. Preventive maintenance involves planned, scheduled, calendar-based, and/or time-based maintenance practices. Predictive maintenance describes a range of technologies with the goal to detect developing machinery faults at an early stage before they constitute a problem, thereby predicting the need for specific and targeted maintenance activities.

Which maintenance strategy, or combination of strategies, will benefit a particular operation? The decision-making process begins with an understanding of the differences among the strategies and then ranking and prioritizing assets from business critical to noncritical. Business critical assets will be those whose failure would be considered intolerable and unacceptable, while the failure of relatively noncritical assets would have relatively less impact on an operation.

All of these steps can position an operation squarely on the road to developing a viable and practical road map for optimized maintenance.

Reviewing the strategies

Technologies for detecting developing machinery faults at an early stage include sensors and specialized devices to monitor various parameters, such as vibration, and analytical and data management software to capture, trend, diagnose and report timely information on the operating conditions of assets.

The term reactive maintenance often will be used in place of the term breakdown maintenance, usually connoting an absence of any organized maintenance strategy and, therefore, inherently becoming a reactive process. Inevitably, most plants will be subject to some degree of breakdown maintenance. But this approach across an operation can carry extremely high costs.

For example, based on industry research, the direct cost of machinery repairs undertaken during breakdowns can be at least three times greater than the cost of planned repairs. In turn, production outage time required for the completion of an emergency repair can be up to five times greater than costs associated with planned repairs.

While breakdown maintenance may be an aspect of a reactive approach, the two are not synonymous. Reactive maintenance can reflect a conscious decision to adopt a Run-to-Failure approach for selected machines, at least for relatively noncritical assets with a low impact on failure risk and where the maintenance cost would exceed replacement cost.

Preventive maintenance by its nature will be time-based, either according to the calendar or to hours run, and undertaken in the hope of avoiding machinery breakdowns. Time intervals for preventive maintenance will be determined by manufacturers’ recommendations, past operating history and/or regulatory requirements.

Such planned maintenance schedules can help reduce failure risks in many cases, but the approach can present problems. The maintenance schedules often will require that machines performing satisfactorily be shut down for maintenance. Unfortunately, sometimes those machines will be returned to service in a “worse than before” condition, whether because of improper procedures or the use of old or inferior components when new, quality replacement parts are unavailable.

Predictive maintenance applies objective technology to quantify machine condition before failure modes can escalate. Changes in the operating conditions of equipment can help foretell when failure initially begins and whether conditions are worsening. Equipped with ample warning, operations can get ahead of the curve on problems and take proactive measures to prevent catastrophic machinery failure.

Technologies for detecting developing machinery faults at an early stage include sensors and specialized devices to monitor various parameters, such as vibration, and analytical and data management software to capture, trend, diagnose and report timely information on the operating conditions of assets. Then, remedial maintenance initiatives can be taken before it is too late.

3 maintenance strategies

Reactive maintenance — An absence of any organized maintenance planning and control, also known as breakdown maintenance

Preventive maintenance — Planned, scheduled, calendar-based and/or time-based maintenance practices

Predictive maintenance — Applies objective technology to quantify machine conditions before failure modes can escalate

Prescribing maintenance tasks

So which of the maintenance strategies will make sense for a particular facility on the road to optimized reliability? More likely than not, a mix of the strategies will be engaged, largely because each asset will vary in terms of criticality.

For the majority of industrial operations, the reality is that it would be impractical and costly to monitor the health of each and every machine, especially those deemed noncritical for core production. This reality suggests that before any maintenance strategy is implemented, facilities should take a step back to rank and prioritize assets based on the business goals of each. Then, the proper maintenance activities can be assigned.

Ultimately, this process of ranking the criticality of assets can help in performing the right maintenance on the right equipment with the right resources for the right reasons.

For noncritical assets, sometimes relatively simple and cost-effective maintenance tasks can be selected and assigned. These tasks can include visual inspection, cleaning, lubrication checks and minor adjustments when necessary, among others.

A multiyear review of the industrial landscape, based on an evaluation of more than 100 Reliability Centered Maintenance (RCM) projects involving thousands of types of assets, shows the natural divide between critical and noncritical machinery and further makes a strong case for assigning the right maintenance strategy for the right asset.

This evaluation found that on average, 35 percent of assets across industries were classified as critical and 65 percent as noncritical (although variations will occur within specific industries). And the implementation of a mix of maintenance strategies became evident: 70 percent of assets in the survey were assigned preventive maintenance tasks, 16 percent were assigned predictive maintenance activities, and 13 percent were allowed to run to failure.

As a start for any facility, qualitative methods can be employed to establish and rank equipment criticality at system and tag level without considering individual machine failures. This can be accomplished using a points-based system that targets the severity and frequency of asset failures in safety, environment and production to arrive at a relative criticality ranking.

In addition, an RCM process can be pursued as a core tool for the development of a maintenance strategy facility-wide. This process is often recommended for detecting and preventing asset failures where criticality is high and confidence in existing maintenance is low. As a result, RCM can be highly effective in preparing a road map of appropriate maintenance tasks, frequencies, and maintenance time and staff skills mix.

Once assets have been viewed either as business critical or noncritical, facilities can decide how to handle them from a maintenance strategy perspective.

For a particular noncritical asset, is there a reason to maintain based on an evaluation of criteria? If so, relatively simple and cost-effective maintenance tasks can be selected and assigned. These tasks can include visual inspection, cleaning, lubrication checks and minor adjustments when necessary, among others. Each task should be assigned a frequency and person responsible for implementation. If no compelling reason to regularly maintain an asset from a production and business perspective exists, then it can be appropriate to run the machine without maintenance attention until corrective maintenance is required.

For critical assets, a predictive maintenance program usually will be recommended to head off potential problems and help keep equipment up and running as intended without interruption. The rollout should identify failure causes or symptoms for each asset, prescribe proactive tasks to mitigate these causes, and assign frequency and responsibility for oversight.

Realizing benefits

The benefits of a predictive maintenance program for the most critical assets can be substantial and tangible.

As a real-world example, a customer implemented a predictive maintenance program for assets identified as critical and, in just the first six months, the measurement of vibration in a transfer compressor detected a faulty bearing. By alerting the plant’s maintenance team to replace the bearing before it failed, the plant saved thousands of dollars in repair costs that would have accrued. Additionally, the program uncovered similar bearing faults on a process pump. Had that bearing run to failure, up to 12 hours in unplanned downtime would have been necessary and associated costs and production disruption would have followed.

The overall payoff: The facility’s predictive maintenance program tailored for assets deemed critical helped saved multiple days of unplanned downtime and many tens of thousands of dollars in related costs.

Getting started on the road to the appropriate mix of maintenance strategies can be a daunting exercise enterprise-wide, sometimes even at the early stages of classifying assets as business critical or noncritical. For a big assist, one of the best courses of action before embarking is to enlist the expertise of a provider fully equipped with the experience and technologies for the job. Such partnering can pay dividends in developing a viable road map toward optimized maintenance at every level.


Michael J. Trainor, CMRP, is manager of asset reliability consulting at SKF USA Inc. He has worked globally for more than 10 years studying client needs and delivering asset reliability improvement projects. He can be reached at michael.j.trainor@skf.com or 484-886-6311. Visit skfusa.com for more information.