Modern manufacturing equipment frequently has a huge number of setting parameters. These parameters are often only defined once for the various product types. However, production conditions are constantly changing. Material parameters fluctuate, as also the environmental conditions, such as temperature or humidity. The process window with its various parameter combinations is often too large to determine the optimum parameters manually.
Using data-based machine learning methods, we find optimal setting parameters for automated production processes for the different product variants. This can be done once or continuously in a control loop. Setting parameters, ambient conditions and measured values acquired at a high frequency rate are all taken into account.
To analyze the parameters, we implement a combination of machine learning methods and statistical evaluation algorithms. Based on the acquired data, we identify the necessary parameter settings to achieve optimal equipment effectiveness. If applicable, new data can be collected if no or too little data is available.