Asset maintenance strategies have a bearing on multiple aspects of industrial operations, including revenues, quality, and safety. That’s why, predicting asset failure has been top of mind for businesses since the 1980s. Predictive maintenance finds its origins in the Just-in-Time paradigm of Japanese manufacturing, and was further fleshed out by C.H. Waddington in the 1990s. Since then, predictive maintenance has been used as an umbrella term for a multitude of maintenance strategies that serve to eliminate the need for reactive maintenance.

In asset maintenance, the intervals at which your teams schedule maintenance is a key determinant of your maintenance costs and their effectiveness. So, how many types of predictive maintenance approaches are there? Is one better than the others? 

Four types of predictive maintenance strategies

Here are four maintenance strategies that are interchangeably called predictive maintenance.

Predetermined maintenance

When a manufacturer-provided maintenance schedule is applied to an asset, it is called predetermined maintenance. These schedules are typically created by the manufacturer based on their MTTF observations and failure mechanisms that have been observed in the past.

This type of maintenance should not be applied to assets that are a part of risk-sensitive systems. Moreover, this maintenance strategy is not very effective because it fails to account for the deployment conditions of a particular asset, and is devised by observing a large sample – in which, your asset could be an outlier.

Corrective maintenance

This maintenance strategy seeks to rectify a problem that precludes an asset from functioning normally. Corrective maintenance could be planned or unplanned. The maintenance plan devised in this strategy could be based on multiple run-to-failure (RTF) observations, or on condition-based monitoring. But when corrective maintenance is unplanned, it will lead to unexpected costs in your budget.

Unplanned corrective is best leveraged for non-strategic assets, like lightbulbs in some parts of a production facility. The effectiveness of planned corrective maintenance, on the other hand, rests upon the strategy that is used to devise the plan.

Preventive maintenance

In this strategy, maintenance work is performed on a static schedule. This schedule is usually created based on historical observations, or risk-mitigation measures. Therefore, the maintenance processes are performed irrespective of the asset condition at that time.

While preventive maintenance is leveraged for critical components, it can add steep costs to your budget because it fails to account for the real-time conditions of assets. In other words preventive maintenance presents an optimization opportunity for businesses.

Condition-based maintenance

This type of maintenance leverages condition monitoring to determine maintenance schedules. Condition-based maintenance will usually rely on historical data to benchmark the condition of a piece of equipment in real time. Various types of sensors are used to stream data that reflects an asset’s condition.

Condition-based maintenance is the most cost-effective approach to scheduling maintenance, as it can predict equipment failure with near-complete accuracy. Maintenance activity can then be scheduled to maximize the useful life of an asset while minimizing maintenance costs.

Dig deeper when implementing predictive maintenance programs

So, which strategy should you leverage when implementing predictive maintenance at your facility? Best-of-breed solutions will leverage artificial intelligence (AI) and machine learning (ML) technologies to inform an optimal maintenance strategy. An ideal approach is to understand the underlying techniques that are leveraged by your predictive maintenance strategy. Here are a few of them.

Vibration analysis

This is the most commonly used technique that is used to monitor the condition of assets, and it can be applied to a number of assets with rotating parts. Because vibrations are caused by a number of parts (gears or bearings), small changes in vibration signature can be used to detect misalignments, imbalance, wear, or looseness. Changes in vibration signatures over time can be used, not only to schedule repair work orders, but also to understand the root cause of sub-optimal performance.

Oil analysis

This technique makes use of computer vision and viscosity sensors to understand the condition of the oil that is used in a machine. Because oil can perform various functions in the machinery (lubrication, or thermal functions), it can be inspected for debris, contaminants, or oil wear. Oil analysis can be very useful to schedule oil replacement in critical machinery, and serves as a crucial approach for detecting wear of inaccessible moving parts. 

Temperature technology

Temperature technology encompasses a set of techniques that are used to build heatmaps of systems, or log temperatures of components in an isolated fashion. Temperature sensing can be especially useful for detecting hot spots in electrical systems, and can help replace fuses before they reach their capacity, or to locate a faulty termination in a circuit.  

Acoustic monitoring

While acoustic monitoring is a type of vibration analysis technique, it has been observed to perform better than the latter in predicting failures of some types of equipment. However, it is also more expensive, as it makes use of sensitive equipment to build an acoustic fingerprint of a system, over which AI algorithms are applied to spot failure patterns. 

Motor-circuit analysis

Motor circuit analysis is a technique that is used to measure the health of a motor. It detects electrical undulations between the rotor and the stator to mitigate the risk of contaminants and ground faults. This technique can also detect major failures like rotor eccentricity, broken bars, or coupling problems by analyzing the current signature. 

To sum it up…

Predictive maintenance spans a number of maintenance strategies that have been used to eliminate reactive maintenance and unprecedented failures in industrial operations. However, not all of these strategies are equally powerful and cost-effective. When implementing a predictive maintenance program at your organization, it is crucial to understand how effective a technology is for your particular use case.

UptimeAI is a predictive maintenance platform that leverages leading predictive maintenance approaches to help you power a reliable plant with the most cost-effective maintenance strategy. Reach out to us today to learn more.

Subscribe to our newsletter