Modern automated process control systems that use programmable logic controllers use software controllers based on the proportional integral-differential control law, the PID controller. In most cases, this regulator is sufficient for conducting the technological process. It has high performance with configurable and sufficient quality of regulation. However, using a PID controller for non-linear, poorly defined, multi-connected objects with a long delay time can lead to unstable control quality indicators, accumulation of errors, and ultimately to a deterioration in product quality. One of the most promising methods of control is Model Predictive Control - MPC. The method base on predictive models of control objects. The quality of the controller's control depends on how well the system dynamics described by the model used to design the controller. In most cases, MPC-based control use to handle optimal control problems on the Manufacturing Execution System-MES. However, thanks to the development of microprocessors and increased CPU performance, it becomes possible to apply the principles of control with a predictive model at a lower level, and perform real-time operational control in optimal modes. The work presents the algorithm of MPC controller. The control object is a SISO object with a nonlinear characteristic and a long transition process. Studies of the developed MPC regulator showed that the quality of regulation, compared to the PID regulator, increased by more than 20%, the time to get to set point decreased, and there was almost no over-regulation. The most effective application of the MPC controller is seen in processes with long transitions and with a significant delay time.
Modeling and Control of Crane Overload Protection During Marine Lifting Operation Based on Model Predictive Control
