Springback optimization for CNC tube bending machine based on an artificial neural networks (ANNs)

Predicting the springback angle has become the major production problem among tube benders. Springback is where the tube on a mandrel-less rotary draw bending tends to bounce back after being bent when the clamps are released. Accurately predicting the springback angle is crucial for effective tube bending. Machine learning (ML), a popular prediction approach, was applied to functions such as prediction or function approximation, pattern classification, clustering, and forecasting. To achieve this, the springback angle values from 27 experiments were collected and used as input into artificial neural networks (ANNs) in one area of ML. This research was conducted to study the optimization of the springback angle when bending ASTM A-210 Gr. A1 seamless tube with an outside diameter of 44.45 mm, using the 4 input factors Wall Thickness, Bending Radius, Dwell Time, and Bending Angle. The results showed that all factors significantly influence the springback angle in the tube bending process; different prediction methods were analyzed by comparing the results using different activation functions. The results showed that the optimal neural network architecture is 4-98-1; these results were achieved using the Sigmoid function, giving the lowest mean squared error (MSE) = 0.001892. The resulting coefficient of determination (R2 ) = 99.42%, the ReLU function R2 = 98.99%, the TanH function R2 = 98.53%, and the Identity function, which was 79.53%. It was also found that the best prediction of the springback angle using the best regression equation, with R2 = 82.32%, was better than the prediction using the 65 neurons with the Identity function R2 = 79.53%, a 2.79% difference in favor of the regression equation.