The history of industry is one that comes in waves. The first wave brought us steam power, the second one mass production and the third one automation. As you could read in a previous blog post, we entered the era of Industry 4.0 in which the Internet of Things plays an important role. Technologies such as data gathering, cloud computing and artificial intelligence are central. Today we focus on automatic calibration as a potential game changer.
In the factory of the future you’ll find “smart” machines that can adjust themselves. They are not just connected to each other, the data they generate is re-used practically immediately.
Today we already have very powerful machines that can perform complex tasks, but they are inflexible, as they can usually only do one thing. One input coming in and one output going out, as soon as you change the input, speed or other (environmental) parameters, you often have to shut down the machine and tweak or even rebuild it. Tweaking the machine requires high levels of skill and experience in relation to the flexibility and complexity of the machine.
In the future, machines will be able to handle multiple products, generate multiple outputs, but still remain easy to operate at the same time. The latter is especially important: machines that can handle different inputs usually have many sensitive settings or parameters that allow it to operate. An operator is required to have an extreme amount of knowledge - which is not scalable.
The solution in such a case can be automatic calibration: smart machines that capture data using sensors and have programmable settings. By combining the generated data with machine learning algorithms and automatic calibration, you can calculate the right settings for a machine to handle a specific product. There is a constant feedback loop that checks which settings lead to a machine process that handles the product perfectly, without having to code anything. An operator can quite literally push a “calibrate now” button, this calibration can be automatically triggered when performance declines, environmental or other factors change or on set times.
In the visual above, the difference is shown between the traditional machines with an operator and the machines with automatic calibration. Modular machines relying on operators will bring a lot of overhead and non-reproducible results. As mentioned earlier, this process isn’t scalable as knowledge sharing will become difficult in between plants because the quality depends on the knowledge of the operator.
Using automatic calibration, this complex task can be done for the operator and insights into the machine performance and optimisation are easily stored and replicated across multiple machines. The operator oversees the operation, but doesn’t influence the process significantly.
Automatic calibration is useful in many scenario’s. The most obvious one is of course once you start using your machine, but it doesn’t end there. After maintenance, machines tend to operate slightly differently. Not only the machines can change, but also the requirements of the customer. Automatic calibration is sometimes permanently active, when fully integrated in a continuous process, triggered when the performance is declining.
Implementing automatic calibration of course implies some prerequisites. For example, your machine must be set up to be modular and sensors that measure all kinds of things are a bare minimum. Collected data must in turn be able to be processed, something that can happen locally, or in the (private) cloud. Output of the process is often fed to the PLC’s that handle every movement or action taken by a machine. Lastly, it’s a big plus when machines can handle failures and don’t crash or have a long downtime.
If you’d like to discover if and how you can use automatic calibration, we can recommend reaching out to discuss your specific challenges!