Control Theory And Application Of Combined Integrating System

In recent years, a new kind of processes is obtained via process modeling based on analysis of process mechanism. Such kind of processes is composed of two or more integrating components, which is called combined integrating process. Combined integrating processes exist extensively in process industry. However, up to now, little attention has been paid to such kind of processes. Currently, most combined integrating processes are controlled by conventional control strategies, such as PID controller, Smith predictor and internal model control, etc, which results in acceptable performance. However, as modern process industry demands improvement in product quality, control accuracy, energy saving and environmental protection, conventional control strategies are no longer suitable. Thus, it is necessary and urgent to research control strategies exclusively for combined integrating processes.The inclusion of component like in combined integrating processes makes it difficult to carry out analysis of control strategy and system performance, especially the analysis of closed loop system stability, of combined integrating processes. This dissertation focuses on the research of theory and application related to the control of combined integrating processes. The main work is outlined as follows:(1) Definition and description of combined integrating processes Describe mechanism based modeling of combined integrating processes that exist in process industry. Propose a complete definition of combined integrating processes according to which it can be known that combined integrating processes are open loop stable, which is different from commonly known integrating processes. Since combined integrating processes function as low pass filters, it is widely used in signal processing and control areas. We introduce and analyze several applications of combined integrating process.(2) Theoretical foundation of stability analysis of combined integrating processes Most combined integrating processes are multiple time delayed processes in essence. This dissertation investigates the stability analysis of time delayed system with both uncertain coefficients and uncertain delays. The stability analysis is composed of two main parts: One part is about the determination of all possible stability boundaries in parameter space. The other part is to determine the actual stability regions in parameter space, which is based on the results of first part. An approach based on the concept of "building block", which is used for determination of stability regions in delay space, is employed to determine stability boundaries in coefficient-delay space. Moreover, the algorithm is optimized to improve numerical efficiency. Using Kronecker Summation method, a necessary condition that uncertain coefficients that make system marginally stable is obtained. Based on features of Routh array of polynomials, the sufficient and necessary condition of uncertain coefficients on stability boundaries must satisfy is obtained. As to Routh array based method, the computational complexity is analyzed and it has been proved that the number of kernel hyperplanes is upperbounded by a Fibonacci number.When using the above mentioned two methods, all uncertain delays should be transformed to uncertain coefficients via Rekasius transformation.(3) Control strategy of combined integrating processesWhen the desired closed loop system is also a kind of combined integrating process, a kind of anti-delay quasi-repetitive controller and a kind of anti-delay quasi-PI controller are proposed for the control of combined integrating processes. To analyze the robust stability of closed loop system, the stability analysis method proposed in this dissertation is applied to determine the permissible varying range of process parameters that guarantee stability of closed loop system. Considering the good performance of combined integrating processes, a kind of anti-delay quasi-PI/PID controller is designed for low order time delay processes whose desired closed loop processes are combined integrating processes. The proposed controller can be easily implemented and tuned. Simulation results show the effectiveness of proposed controller.(4) Disturbance rejection of combined integrating processesBased on IMC principle, a kind of disturbance rejection controllers is designed which can reject both constant disturbance and sinusoidal disturbance. To guarantee good disturbance rejection performance without degrading setpoint tracking response, Astrom's modified Smith predictor configuration is employed. A signal reconstruction procedure is proposed to extract a sinusoidal signal having same frequency with sinusoidal disturbance from system's output which contains sinusoidal signal as well as constant component and noise. Based on the reconstructed signal, a kind of unscented Kalman filter is designed to estimate its frequency. Simulation results show that the proposed method can fast track the frequency of sinusoidal with high accuracy.(5) Practical applicationWe apply the proposed the advanced algorithms for control of two kinds of combined integrating processes: outlet moisture of drying process of cut tobacco and threshing and re-drying process of tobacco. Practical application shows that the proposed advanced control algorithms promote the control accuracy, leading to improvement in product quality and energy efficiency.