Self-optimizing thermal correction

Besides static and dynamic errors, thermal influences significantly limit the accuracy of machine tools. State of the art correction approaches mainly base on correlative coherences between manufacturing task and displacement. They are proven to be very effective for the trained manufacturing scenarios. However, if the actual scenario differs from the trained cases, their benefit is decreasing. Hence, correlative measures are not suitable for autonomous self-optimizing under changing conditions due to their limited flexibility and automatically adaptable model representations are required. Our alternative approach for thermal correction and compensation bases on structure models for the thermal and thermo-elastic behavior of the machine tool. Due to the incorporated physical behavior, they are adaptable to the current machine state as well as the manufacturing condition and therewith are convenient for autonomous self-optimization. Their applicability is realized by an order reduction of FE-model or by the utilization of models with concentrated properties. They form a digital image of the thermo-elastic functional chain. This Digital Twin is computed in the machine controller in ""thermal real-time"". The accuracy of the model is determined by the uncertainties of the model parameters. These parameters have to be adjusted during commissioning of the machine. By cyclically updating the model parameters during machining the models are readjusted and maintain their accuracy. Measurement principles, such as thermography, photogrammetry and the utilization of thermo-sensors were investigated and will be presented. Experiments were carried out for a parallel kinematic test setup ""MiniHex"". The correction approach is integrated in an industrial controller system ""Beckhoff TwinCAT 3"". Further research will focus on establishing self-optimizing machining systems, with decentralized cooling systems or PCM-storage systems on the component layer.