Guaranteed Cost And Robust H_∞Networked Control For Fuel Cells

ABSTRACT:In recent years, as the concept of smart grid emerges and develops, higher requirement to the power system has been put forward. And the weakness of the traditional centralized power system has become more obvious. In order to increase the diversification of energy supply source and enhance the robustness of the grid, renewable energy and distributed energy resources (DERs) technology becomes an increasing concern. Fuel cell is an environment friendly, high efficiency, modularized energy resource and those advantages make it an important power supply technology in DERs. However, dispersion of energy resources will inevitably lead to the decentralization of control, which makes it a challenge for engineers to design a control scheme. In this context, being of certain practical significance, a networked control scheme for fuel cell system is designed in this dissertation.Firstly, a short survey of NCS is presented, which introduces the basic theory, existing problems and current research situation of NCSs application to smart grid. Background of distributed energy generation and the significance of the fuel cell network control are also explained. Modeling approaches of NCSs with stochastic short delay is summarized and several different NCS models have been established according to their different node driven modes.Secondly, considering the situation that not all states of the fuel cell system can be measured and the controlling input is constrained in actual control systems, an input constrained guaranteed cost controller with an observer is proposed for networked control system in the presence of short network-induced delay. Then the controller is used in the fuel cell system and obtains satisfactory effect, which demonstrates the effectiveness of the algorithm.Thirdly, a switched system including two subsystems is established for networked control system with time-delay, packet loss and disturbance, according to the packet loss. Each subsystem is modeled with algebraic method respectively, and the uncertainty of the time-delay in the system is converted to the uncertainty of parameter in the linear discrete system model. The corresponding state feedback controllers are designed for both two subsystems using the proposed control law designing method based on linear matrix inequality. This system is applied to the stack control of fuel cell, and a desired result is achieved. Finally, the work in this dissertation is summarized, and recommendations for further research work are also given.