High Performance Control of AC Drives with Matlab / Simulink Models

7.6 Model Predictive Control (MPC)

In the last 20 years, model predictive control (MPC) has achieved a significant level of acceptability and success in practical process control applications, and has been mainly applied in power plants and petroleum refineries. With the development of the modern micro-controller, digital signal processors, and field programmable gate Arrays, MPC applications have been found in a variety of areas including chemicals, food processing, automotive and aerospace applications, and power electronics and drives [60–65].

MPC algorithms use an explicit process model to predict the future response of a process or plant. A cost function represents the desired behavior of the system. An optimization problem is formulated, where a sequence of future actuations is obtained by minimizing the cost function. The first element of the sequence is applied and all calculations are repeated for every sample period. The process model therefore plays a decisive role in the controller. The chosen model must be able to capture the process dynamics to precisely predict future outputs and be simple to implement and understand. Linear models have been widely used, as they can be easily obtained by system identification technique, or by linearization of first-principles non-linear models. The cost function, in consequence, is quadratic and the constraints are in the form of linear inequalities. For such quadratic programming (QP) problem, active set methods (AS), and interior ...