Predictive maintenance is a strategy that uses condition monitoring and analysis equipment to automate performance checks of production assets. The process optimizes maintenance intervention so technicians’ time can be spent on high-value repairs and upkeep, instead of walking “routes” and checking machines that don’t need attention. An example of predictive maintenance is using vibration monitoring data to analyze the performance of a rotating machine, such as a pump or motor, and predict when maintenance is required based on changes in the vibration patterns.

By tracking equipment performance data trends and predicting when maintenance is required, you can increase equipment uptime, reduce maintenance costs, and avoid unplanned downtime.

Data used for predictive maintenance comes from vibration sensors, temperature sensors, pressure sensors, and other equipment. Vibration and temperature sensors can detect issues such as bearing wear and equipment component misalignment, while temperature sensors can help detect overheating or cooling issues. Pressure sensors can detect improper pressure levels in hydraulic or pneumatic systems.

IIoT, or the Industrial Internet of Things, has unlocked the ability to implement predictive maintenance on a large scale. This is because a wealth of information from various sensors and other points can now be made available. And getting this data up to the cloud for monitoring has never been easier. Predictive maintenance is arguably the "killer app" for IIoT.

Implementing a predictive maintenance pilot is not complex. You will need the following:

    1. Sensors to collect machine health data.

    2. A gateway to connect sensors to the cloud.

    3. A web dashboard where data can be viewed from anywhere.

Once a pilot proves the process of adding a condition monitoring system, for predictive maintenance, the program can be easily scaled. However, preventative maintenance may never go away entirely; it depends on the cost/benefit equation, which is unique to every business. In certain instances, businesses may choose to run some less critical assets until they breakdown.

A power company might use predictive maintenance to monitor a gas turbine, so technicians can plan the optimal time to perform maintenance. Or a manufacturer might use predictive maintenance to set benchmarks for asset performance, so alerts and warnings can be set up to notify the maintenance team that an asset needs immediate intervention.

For a specific example, we can look at a tire manufacturer that uses a massive gearbox to transmit power from a motor to a mixer for liquid rubber. These gearboxes are critical to their operation, yet if one experiences catastrophic failure, there is a two-year lead time to secure a replacement. Because of this need for a healthy gearbox, some tire manufacturers choose to employ ten or more vibration sensors on important rotating components such as input shafts, output shafts, couplings, bearings, gears, and others. If there is a problem with a component, such as a bearing, the maintenance team is notified so it can perform maintenance before serious problems develop.