Sometimes you have to destroy things in the lab so that no one gets hurt later: how hot can a spinning tire get before it bursts, how much weight can a screw bear before it gives out… these are the things you’d rather find out under controlled conditions than in real life. And that’s why destructive testing is still crucial at times. But digitalization has initiated such wide-ranging developments in this area that it’s not just a matter of finding out how a material reacts when you knock on it anymore. In Germany, testXpo is an entire trade fair dedicated to this topic. Germanedge product partner QDA SOLUTIONS will be there, investigating how neural networks can lead to advancements surrounding predictive maintenance, among other subjects.
Given the evolution of digitalization-driven, non-destructive processes, it seems almost archaic that a product has to be destroyed in order to establish how best to produce it. Nevertheless, destructive testing will remain necessary as long as there are people who could potentially be harmed by product defects. In other words: you have to know what happens to the materials so that nothing happens to the people interacting with them.
Of course, destructive testing is not always about protecting human lives. If a yogurt cup is too thin and bursts in a shopping bag, no one gets hurt. If a machine applies too much adhesive to an adhesive film and causes needlessly high costs, only the figures suffer. Still, the interesting thing here is the common thread: no matter which aspect of production we consider, people are and remain the determining factor. Which process is employed how and where is always dependent on human need. Nevertheless, you don’t have to leave everything up to people alone. Why go through countless painstaking repetitions to find the value that indicates a metal rail’s breaking point? Why rip the film off the wall again and again until the optimum ratio of adhesive to bonding strength has been determined?
With software from companies like QDA SOLUTIONS, it’s already possible to test material quality (whether in the lab or during ongoing production), certify it in accordance with a growing jungle of regulations and standards, simultaneously ensure that production takes place within defined parameters, and send timely notifications when deviations become apparent, no matter which level of production they affect, whether it be the machine and operator, the shift supervisor, or higher up. And that’s just the beginning. In cooperation with the Fraunhofer Institute for Manufacturing Engineering and Automation IPA, QDA SOLUTIONS recently developed a big data project that takes data from machines equipped with production monitoring sensors and delivers it to a neural network which then uses the data to predict maintenance requirements. The fundamental question behind predictive maintenance – “When will what happen?” – can be applied both to the material performance and to all steps of the production process. Now you just need to know which of the sometimes thousands of data generated per second are relevant for the evaluation and what exactly they reveal.
Of course, this won’t remain a secret to our customers: QDA SOLUTIONS will have representatives at the aforementioned testXpo trade fair, hosted by the company ZwickRoell in Ulm from October 14th to 17th, 2019. They look forward to providing you with further information on the functions, applications, and developments of their solutions in this area.