The Challenges of Machine Learning Power Valve Development

At the moment we cannot expect to know for example the volume flows in our system and often I think you only have some pressures that you're measuring because pressure sensors are quite robust they are also not that expensive. And so you need to work with this very reduced information or with thisvery reduced set of information, which is a great challenge for us. So who are we working with here to overcome these challenges? We work with some great people from the technical University of Munich namely with the chairs of automation and information system - Professor Fugelheuser and Professor Luman. With them we are developing this concept for a machine learning power valve that can use these reduced data streams for intelligent control.