| Model predictive control (MPC) is a class of optimization control algorithms which predicts the future outputs with a plant model. It has been widely used in industrial field, regarded as an effective control strategy especially for linear systems. However, for some industrial processes with heavy non-linear elements, including nonlinear model and nonlinear constraints, it is difficult to obtain satisfactory control results by using a linear MPC. Therefore, nonlinear model predictive control (NMPC) has been widely investigated in recent years.With the progress of our society and technology, a variety of controller based on embedded platforms came into being. FPGA (Field Programmable Gate Array) is developed quickly as a class of programmable logic devices. It combines the advantages of traditional software and hardware design approaches, which doesn't only ensure high system performance and improve flexibility and adaptability of design, but also greatly reduce the scale of hardware. Due to these characteristics, FPGA has become an effective hardware platform to achieve MPC control. Therefore, FPGA is used to design and implement nonlinear predictive controller in this paper.The main research work of this dissertation can be summarized as the following four aspects:1) Algorithm design of the NMPC optimization problem. Analyze online optimization problem of NMPC and propose new feasible algorithms based on existing theory, which can ensure feasibility of the solution of the optimization in each sampling interval, even if the computing resources are limited.2) Software & hardware design and implementation on embedded system. Select the appropriate embedded platform and analyze the corresponding cost of time and resources in each part of NMPC process. In order to make full use of FPGA hardware resources and parallel computing capacity, the tasks of software and hardware should be well arranged.3) Simulation and testing of the NMPC controller. By closed-loop test between the embedded controller and MATLAB, it is proved that the controller can real-time achieve fast and effective results.4) Field support software for FPGA based predictive controller. The software is designed to solve the problems in finding the optimal parameters and updating parameters to the controller in field application environment. With the functions of parameter design, simulation and data transmission, it can show the control effect of the parameters intuitively and revise the parameters easily. |
