Machining cycle time prediction: Data-driven modelling of machine tool feedrate behavior with neural networks

Accurate prediction of machining cycle times is important in the manufacturing industry. Usually, Computer-Aided Manufacturing (CAM) software estimates using commanded feedrate from toolpath file basic kinematic settings. Typically, methods do not account for geometry or tolerance and therefore underestimate considerably. Removing need machine-specific knowledge, this paper presents a data-driven time method by building neural network model each machine tool axis. In study, datasets composed feedrate, nominal acceleration, measured were used to train model. Validation trials representative industrial thin-wall structure component on commercial center showed that estimated with more than 90% accuracy. This models have capability learn behavior complex system predict times. Further integration will be critical implantation digital twins Industry 4.0.