Current control systems for the shot peening operation merely rely on old technologies, which often require repetitive processes to obtain pre-validated Almen systems to guide industrial productions. These designs for the manufacturing paradigm are not efficient for complicated workflows in modern manufacturing operation. Thus, in this study, we propose a practical model-based control system to address the issues; especially for a smarter and automated shot peening machine. In particular, the closed-loop control system development utilizes a model-based proportional-integral-derivative (PID) control technology and extreme gradient boosting (XGBOOST) machine learning algorithm. The control system includes an internal process model, a proxy model, a model-based PID controller, and pressure sensors with a low-pass filter for feedback control. The developed control system is integrated into a physical shot peening machine for on-site control validation and demonstration. In both in-silico and on-site control demonstrations, the obtained control performance is stable, robust, and reliable for different operational conditions. The measurement intensities are very close to targeted setting intensities. All the differences are smaller than the industrial threshold of (±0.01 mmA). It implies that the control system can use in industrial peening operations without the need for Almen system development for operational guidance. In other words, the control system can significantly reduce the total cost of the actual production by eliminating the cost, time, and labor of the iterative trials to build the Almen system.
An Internal Model-Based PID Control for Smart Shot Peening Operation
