With the introduction of industrial IoT, a production line nowadays is far more than just a mechanical piece of engineering. Having sensors to measure temperature, humidity, or sometimes even an infrared scanner, a manufacturing machine suddenly becomes a source of big data. Maybe this is not your first thought when you were considering buying a new machine, but having a strategy in place to use your machine’s data can be vital for the decisions you make in the end.
Why your data counts
Gathering data within manufacturing is important for many reasons. The most crucial aspect however is the ability to analyse it for predictive maintenance. A joint study by Emerson and Wall Street Journal found that unplanned downtime costs industrial manufacturers an estimated $50 Billion per year. Equipment failure is the cause of 42% of this unplanned downtime.
Increased productivity
There is a lot to gain in this field. Predictive maintenance is by far the best option to reduce costs even further. Of course, situations are widely different, but on average, introducing predictive maintenance can increase your productivity by an estimated 30 percent.
Human inspection
Currently, many factories rely on manual human-machine inspection to keep their expensive machinery up and working. Though this is a good thing to do, it has its limits. Human inspection is always a snapshot and only catches so much of the complete situation. Furthermore, good inspection personnel is scarce and it is pricey to carry out.
Benefits predictive maintenance
Automated predictive maintenance is in many ways the opposite. It’s real-time and based on the history of your data it can predict when and which part of the machine needs inspection. This will not only reduce the amount of downtime but also extends the lifespan of the machine since it reduces the chance of a breakdown.
Difficulties
As simple as it may sound, implementing predictive maintenance is definitely not easy. There are several reasons:
- No factory is the same. Every kind of production needs it own specified approach. This means that ground expertise is vital in finding the right solution.
- Data can easily be ‘polluted’. Even the slightest changes in your equipment surroundings can mean a big difference to the algorithm that currently collects data.
- In order to provide an accurate prediction, you need a lot of (historical) data
Strategy before buying
This third point is exactly the reason why you have to consider predictive maintenance before you start buying a new machine or setting up a new production line.
For example: what are you going to do with all the data you gather? Saving data about temperature or humidity might sound trivial, but it’s not. When a sensor measures something every 5 seconds, and does this 24 hour a day, a machine sensor can easily count up to two or three gigabytes per day. Are you going to store all of this? And where? And is this really necessary?
Even just to answer these questions it is important to at least consider the option of predictive maintenance beforehand. Even if you are not going to use it right now, it’s important to save the data when you want to introduce it later on. Also making sure you are collecting ‘clean’ data is something you should ensure right from the start.
At LINKIT we know all about predictive maintenance and the choices that are coming with it. So, if you need advice to setup a data strategy for your machinery, you can always get in touch with one or our experts for more information. Or download our free whitepaper to discover more about increasing your business value by introducing predictive maintenance and/or quality inspection.
