Control problems come in different degrees of complexity. You might want to keep the speed of a motor constant independent of load, for instance. There is one control signal, one measurement and dynamics of the system are well known. In this case you can program the functionality. Classical control theory offers good solutions for many situations where programming isn’t possible anymore. However, sometime even control theory reaches its limit, whenever the system is highly non-linear, the effect happens very delayed our the system has an unknown dynamics.
We plan to enable the customer to solve complex control problems in a structured and plannable way with a highly automated workflow. Resulting in predictable time-effort, reachable performance and risk.
In China, we implemented a use case in the area of Autonomous Control and Process Optimization already. In the complex production of monocrystalline silicon ingots, relevant decision-making processes are dependent on the know-how of the machine operator and the implemented control strategy. A single control error can lead to low quality of the ingots and restart of the entire production step. Thanks to Artificial Intelligence (AI)-enhanced control the production could be optimized. Our software enables a fully automated heating power control, which leads to up to 8 % faster production and up to 65% improved quality.