Artificial Neural Networks to Obtain the Machining Parameters for a Compliant Part in a Single Re-Machining Step

In contemporary manufacturing systems, the efficiency of the production line is crucial for maintaining competitiveness in the current market. Therefore, minimizing re-machining operations becomes imperative to optimize production, reduce costs, and utilize resources more effectively. This necessity is particularly apparent in machining processes involving sizable components that require numerous hours, and at times, even days to accomplish. This study employs two Artificial Neural Networks with one layer and different numbers of neurons to determine the required material removal quantities for the production of cast iron columns with different lengths. The algorithm was tested on a 3-axis CNC machine in a real industrial scenario in the company PAMA S.p.A., a member of the European AIDEAS project consortium. The goodness of the solution has been demonstrated by the high score value in the prediction of the removal parameters (score = 0.998) and confirmed by the experience of the production manager of the company