Design and application of advanced control methods to gas turbines and networked systems

  • Junxia Mu

    Student thesis: Doctoral Thesis

    Abstract

    This thesis deals with the design and application of modern control techniques to a Rolls Royce aircraft gas turbine engine and networked systems. It is motivated by the need to fully exploit recent advances in control engineering and investigate the suitability of various control methods to gas turbine engines and networked systems.

    The main contributions of the first part of the thesis relate to the gas turbine engine control. Due to the nonlinearities of the gas turbine engines, the rate limiter and saturation constraints on the fuel feed, the aim is to illustrate the potential of a global nonlinear controller to cover the engine operating range. Several nonlinear control methods, gain-scheduling PID controller, approximate model predictive control (AMPC) and nonlinear model predictive control (NMPC), are presented along with the corresponding control algorithms. Since the parameters in a gain-scheduling PID controller change with the operating range, the need is apparent for a global nonlinear controller to cover its operating range. AMPC and NMPC are then demonstrated to be capable of providing a global nonlinear controller for the engine and can be used in the place of the gain scheduling PID controller. It is shown that AMPC is more preferable than NMPC if computational time is at a premium.

    The main theme of the second part of the thesis is the design and application of the networked predictive control (NPC) to compensate for the network delay and data packet dropout in both forward and backward channels for networked systems. NPC using both modified model predictive control and generic polynomial method is presented along with the corresponding control algorithms. For both approaches, the system stability for a fixed network delay is presented and an analytical stability criterion is obtained. This provides some guidelines on how to choose the NPC parameters in the case of random network delay. The performance of NPC can be further improved by using a robust NPC (RNPC). To validate the performance using the proposed control methods, a servo motor system is then used for both Intranet and Internet based simulations and practical experiments. A networked control test rig along with the network delay measurement method is used for real-time implementation. It is shown that both NPC and RNPC can efficiently compensate for the network delay and data packet dropout in both channels.

    This thesis provides basis for the real-time implementation of advanced control methods in gas turbine engines. While this work was applied to a gas turbine engine, these techniques can be applied to a range of nonlinear control systems. The work on the networked predictive control presented in this thesis can provide basis for further research relating to this area.
    Date of AwardJun 2005
    Original languageEnglish
    SupervisorG-P Liu (Supervisor)

    Keywords

    • Gas-turbines

    Cite this

    '