“With AI, lasers will become even more efficient, easier to operate and more adaptable”


these are exciting times! We expect the coronavirus pandemic to act as a kind of catalyst: those who are well prepared will be perfectly placed to exploit the huge opportunities that lie ahead. Soon we’ll see whether we have laid the right foundations for the future in our joint projects.’ Trumpf has also recognised the benefits


of collaboration to achieve its goals from the point of view of selecting AI provision partners. The company is now the newest strategic partner and investor to Nnaisense, a provider of artificial neural networks, deep learning, reinforcement learning, artificial evolution and general purpose AI. Jürgen Schmidhuber, chief scientist and co-founder at Nnaisense said: ‘For many years we have been working on many client engineering and research projects and have developed new approaches combining control theory and AI, as well as highly scalable, evolutionary reinforcement learning methods. ‘We very much look forward to


collaborating with this innovation leader in sheet metal fabrication machinery and industrial lasers. Nnaisense will support Trumpf in its latest laser technologies, applying our AI expertise to create industry- leading solutions.’


AI now enables laser systems to be operated using voice commands


Quality control Nnaisense works with customers to produce largely bespoke AI solutions. ‘Off- the-shelf AI solutions are often too generic,’ said Schmidhuber, ‘resulting in lost time and investment.’ The backbone of the firm’s solutions is its NNAI engine, which can be customised for three application areas: inspection, modeling and control. The benefits of these three areas for the


user are, said Schmidhuber, that automated inspection can help to ensure efficient quality control. Modelling can help to predict the dynamics of a process, sampling data from the actual process to learn a predictive model bespoke to that application. When it comes to control, intelligent automation allows the sensory-motor loop to be closed in a way that goes beyond traditional control engineering, applying deep reinforcement learning to adapt neural network controllers through safe and efficient interaction with a learned process model. This makes the process much safer for users. Nnaisense is well versed in bringing AI


into laser processing, having collaborated with Electro Optical Systems (EOS) to incorporate intelligent monitoring into its additive manufacturing processes. EOS concentrates on selective laser melting and selective laser sintering. During this process, the distribution of laser energy within the layer is a key factor determining the material properties of the part. Harald Krauss, from the innovation


team at EOS, revealed: ‘One of the most important things you have to consider when


WWW.LASERSYSTEMSEUROPE.COM | @LASERSYSTEMSMAG


thinking about a part’s exposure strategy is heat conduction in the part. The selective laser melting process introduces a lot of heat from the laser beam, and this has to be dissipated through the part and the system. It will depend on the geometry conditions and part shapes in the build process, so you have to adapt your processing strategy there. It is important, to avoid defects, to use monitoring systems to optimise output.’ Nnaisense assisted by developing a deep


network model that can accurately predict a heat map based on job parameters. It can detect process anomalies when sensor readings deviate from predicted behaviour, and control laser intensity to avoid defects and optimise material properties.


Feel the heat Florian Trifterer, senior researcher at Nnaisense, said that ‘what is supposed to be produced is defined by build instructions. These are translated into detailed laser path instructions which are executed by the machine, and the observed thermal image is captured from the observed thermal radiation captured from the just-finished layer. It is here that we can see if any heat is uniformly distributed, which could be caused by a number of things, for example, the way the laser moved. The difficulty is being able to differentiate genuine spurious defects from systematic effects.’ This is where AI comes into play.


Trifterer said: ‘What if we could predict the systematic effects, based on the building structures?’ This is exactly what Nnaisense


g SUMMER 21 LASER SYSTEMS EUROPE 9


Trumpf


Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30