An AI process control system with simulation database and adaptive filter for V-bending

Describes an artificial intelligence (AI) V-bending process control system with a numerical simulation database and adaptive filter which was proposed and developed to achieve production with high accuracy and flexibility. The punch force-stroke curve (F-S curve) which includes process information in a compound manner is stored in the database as expertise and applied to evaluate and control the process. An FEM code was used to simulate the V-bending process to obtain the F-S curve during loading and the springback value during unloading of the process. An online adaptive filter was applied to modify the simulated F-S curve and the simulated springback value. Furthermore, the concept of a multi-regional filter is proposed to improve filtering accuracy. The modified F-S curves and springback values are stored in the database as pseudo-experimental ones, and used in the V-bending process control with an intelligent process control system. The AI control system of the V-bending process was evaluated using four kinds of materials as workpieces. Results show the FEM simulation database with online adaptive filtering is very effective for precision control. A high accurate V-bending process was achieved without the trial of V-bending tests.