Traditional maintenance is built primarily on a reactive and preventive approach. When something breaks, you fix it. When it’s time to change the oil, you change it.

The smart maintenance model includes reactive and preventive approaches, but goes further with remote, condition-based monitoring, predictive maintenance, and cognitive maintenance. This post, crafted from an excerpt from Microsoft’s Smart Maintenance eBook, explores each approach in detail, which is right for your business, and when.

Reactive Maintenance

Reactive maintenance works well for tools and items that are part of the supply chain, but aren’t likely to cause disruption if they go offline. Every plant or manufacturing facility has items like these that fall outside the rigors of a more advanced maintenance program.

Use reactive maintenance with items that:

  • Are small
  • Are unlikely to fail
  • Are redundant
  • Have a low cost for downtime

What you need to make it work:

  • Workers trained to spot the failure as soon as it happens
  • Back-up parts and inventory to ensure redundancy is maintained

Preventive Maintenance

The preventive approach, which has been around for decades, might be the first maintenance methodology based on data. Changing the oil in vehicles every 3,000 miles, for example, is based on evidence showing that a lot of engine problems can be avoided if the oil is used for only a certain number of miles. With data showing the 3,000-mile mark to be optimal under normal conditions, we can create a preventive maintenance schedule.

As the foundation that other maintenance approaches build on, preventive maintenance means fixing and maintaining before failure can happen.

Use preventive maintenance with items that:

  • Are in heavy use
  • Are expensive to replace
  • Have many moving parts that require inspection and/or regular maintenance
  • Are critical to the supply chain

What you need to make it work:

  • A schedule for maintenance that’s built into the supply chain timeline so there are no surprises or disruptions when a machine goes offline
  • A preventive method that’s right for each piece of machinery or part: time-based maintenance for a compressor every 15 days, for example, or usage-based for electrical components after every production cycle
  • A maintenance team dedicated to maintaining the schedule and inventory necessary for upcoming inspections

Remote, condition-based monitoring

This approach refines preventive maintenance by implementing wireless sensors that relay data to a maintenance manager. Now instead of performing preventive inspections on a monthly schedule, for example, maintenance can be performed whenever the data says it’s necessary.

With the power of sensors and data collection, preventive maintenance becomes a sophisticated, more accurate, and efficient practice. Integrating sensors and data collection also lays the groundwork for more advanced maintenance approaches and turns machinery and parts into Internet of Things (IoT) devices so they can be monitored from anywhere.

Use remote condition-based monitoring with items that:

  • Have random failures with no discernible pattern.
  • Are not subjected to wear
  • Have measurable activities, such as vibration, temperature, water, or air flow, pressure, or audio

What you need to make it work:

  • The ability to gather data from your factory or equipment via sensors
  • A platform or dashboard to collect the data and deliver notifications
  • Training for your employees so they can properly respond to work orders

Predictive Maintenance

Accurate predictions rely on quality data. Predictive maintenance brings together data and technology to accurately inform the maintenance schedule.

With the groundwork laid for remote condition-based monitoring, we’re ready to advance into smart maintenance. Up to this point, the maintenance approaches described have fit a specific need, but they’re limited in their usefulness. The digital feedback loop that’s part of smart maintenance means we can be predictive, anticipating equipment failure or maintenance needs based on both historical data and near real-time data. Then we can act to prevent failure before it happens.

Use predictive maintenance when:

  • Your company has shifted its maintenance culture to a proactive mindset
  • Equipment is subject to wear-out
  • Replacement parts or servicing products, such as lubricant, are readily available
  • The failure pattern of equipment is known

What you need to make it work:

  • Comprehensive training so that everyone understands how the predictive maintenance program works, why it’s important to the company, and what their role is in making it successful
  • A technology partner to help bring all the elements together, from sensors to data collection, dashboards, and data analysis

Cognitive Maintenance

Here we’ve reached the pinnacle of the smart maintenance model. Cognitive maintenance means your program is able to think ahead with much more specificity and accuracy than the predictive maintenance model can.

As the most technologically advanced approach to maintenance, cognitive maintenance helps ensure that equipment is in good working order. But it also helps optimise your workforce, production, sales, and customer satisfaction, by eliminating downtime and increasing throughput.

Cognitive maintenance is best for companies that:

  • Have high production capacity or a large volume of equipment in high use
  • Are already embracing digital transformation in other areas with strong support from leadership
  • Understand the value of building business continuity through intelligent, unified systems

What you need to make it work:

  • Comprehensive training so that everyone understands how the cognitive maintenance program works, why it’s important to the company, and what their role is in making it successful
  • A technology partner to help bring all the elements together, from sensors to data collection, dashboards, and data analysis

View the next blog in the series on ‘How to Get Your Maintenance IoT Project Up and Running’.