Model Predictive Control for Complex Dynamic Systems

With increasing complexities and dynamics, modern aerospace systems are facing important challenges in automating and optimizing the operation of a variety of systems, which, for example, may include energy management in propulsion / power systems and flight control systems. Classical PID (proportional-integral-derivative) controls were traditionally used to manage many processes; however, they are becoming insufficient to deal with the issues of higher order of complexity and nonlinearity, multiple control variables or lots of disturbances in the processes. Model predictive control (MPC) can provide promising features with which these problems can be overcome. This approach is to proactively adjust control actions based upon internal model prediction and optimization strategies. It has the advantage over traditional PID controllers that it can optimize the operation for multiple objectives at the same time and guarantees that the path towards the control objectives is generally optimal. This paper presents the initial work targeted toward FPGA-based real-time MPC implementation for energy management of a hybrid power system. The concept of MPC, operation principle, and real-time implementation issues will be discussed in detail. Finally, based on simulation results, the effects of different design parameters on system performances will be elaborated.