In the world of smart factories, Artificial Intelligence (AI), and Industry 4.0, it can quickly be confusing to see the difference between advances in technology. Even in the world of machine maintenance, we’ve seen robots, machine learning, and other technologies and service models appear in the span of the last few decades.
It is hence important for us to differentiate between all the terms floating around in the realm of maintenance, to ensure that manufacturers stay on the same page as technology providers. After all, if you shortened it, the abbreviation ‘PM’ could now refer to Preventive Maintenance, Predictive Maintenance, or Prescriptive Maintenance! It’s time to break down the differences between the three.
What Is Preventive Maintenance?
Preventive maintenance is the model of maintenance that everyone is familiar with: the time-based type that depends on the recommended maintenance timelines provided by manufacturers.
Partial or full maintenance is regularly performed at specific intervals in the hopes of preventing early machine breakdown by creating scheduled downtime for maintenance. The time intervals chosen would have been decided by the engineers that built it, based on their projection of the product lifespan. It involves a lot of planning, record-keeping, and scheduling, that must be followed, and routine inspection has to be carried out regardless of whether the machine is functioning at optimal level or not.
Preventive Maintenance is still the most popular type of maintenance, as it permeates all sectors, from automotive to consumer goods to aviation and aerospace.
While this helps with planning, preventive maintenance can be disruptive as maintenance can occur too often with no increase in productivity and cause a lot of money lost due to the scheduled downtime.
What Is Predictive Maintenance?
Predictive maintenance is a technique that is currently heavily promoted as the next step in maintenance: it involves incorporating Industry 4.0 sensors and other monitoring strategies to gauge when machines actually require maintenance. It is condition-based and tends to rely on Internet of Things (IoT) devices that can work in real-time, hence providing up-to-date conditions of the actual machine.
These devices will continuously or periodically monitor actual conditions of the machine on the shop floor, and the data is supposed to be available remotely, with cloud-based devices and interoperability factoring strongly into devices created. These devices will then provide analytics as to the time before maintenance is required.
Promising that it will only perform maintenance when it is required, hence causing less downtime and improving productivity, there have been some positive results from predictive maintenance projects, with examples that show that predictive maintenance can bring up to $17 million in cost savings within a few years.
What Is Prescriptive Maintenance?
Prescriptive Maintenance is similar to Predictive Maintenance but goes one step further in trying to automate the maintenance process. Instead of just monitoring and providing recommendations, the aim is for machine learning and AI techniques to allow the machine to make its own decision as to maintenance steps.
The machines and devices will collect data as it runs, and will provide multiple recommendations such as reducing current productivity to allow for longer time before maintenance. Prescriptive and Predictive Maintenance are commonly mistaken to mean one and the same, but instead of just predicting the outcome, as per prescriptive maintenance, predictive maintenance will use machine learning and logic to tell you what can be done, and keep updating as it goes.
As expected, this requires real-time data that compares different machines throughout the factory, and is best implemented throughout the factory to provide for accurate results. If you need to understand it more clearly, you can try out IBM Watson’s ‘Industry 4.0 Model Factory’ game, which we covered in our Industry 4.0 news a few months ago.
Learn more about Industry 4.0 and Smart Manufacturing terms by downloading our Industry 4.0 glossary!