If production comes to a halt, companies usually put this down to the machines and readjust them. Our approach is different – we look at Industry-4.0 processes from a material perspective. That’s because differences in material quality often have an adverse impact on processing in standard processes because the machines are unable to compensate for these differences.
Understanding the material
We assume that every material has its own DNA which shows the best way to process the material in order to create the best-possible product. Our aim is to analyse this DNA and to make production processes much more reliable and flexible.
To achieve this, we are using state-of-the art technologies and simulation models to examine all the relevant factors, beginning at the molecular level of the material right through to the fine adjustment of the machine. Three main things are needed here: We have to learn to understand the DNA of the material to be processed. We need sensitive, smart sensor technology that can detect even the smallest differences in material properties and we need flexible and adaptable machine concepts.
Autodidactic production processes
Before the material is processed, it is captured by sensors and its actual status is sent to the database. The material data is compared there with models for materials and machines so that the machine settings for the production process can be determined. At the end of the production process, sensors capture the product made and sent the analysis data back to the database. This data is then compared with the quality parameters defined for the product.
The data collected – from the material before production begins, right through to the finished product – is saved and analysed in the database. This creates a self-learning process that continuously further develops the stored material-machine models.
Realtime analyses and flexible prototypes
We work with databases that fully index and save unstructured data. This means that there are no delays in the production process when measured data is compared with the models. Data is analysed in realtime. We are also building prototypes of machines that can spontaneously adapt to the requirements of the material.