Study of Power Plant Main Steam Temperature System Based on Smith-LADRC Controller

Against such characteristics as large delay, nonlinear, time-varying and uncertainty of controlled object, the Smith-LADRC cascade control system is proposed in this paper. The principle of the LADRC is briefly introduced firstly, and then the LADRC-P control system with Smith predictor is designed and the closed loop system is studied in MATLAB simulation. In addition, control effect of the new structure and Smith-P cascade control structure is compared, and the robust performance is analyzed as well. Results show that when using different control strategies for the same load, performance under Smith-LADRC control is better than that under Smith-P control. The new control structure shows excellent control performance, outstanding adaptability and robustness against model uncertainties and external disturbances. Introduction In thermal power plant, boiler superheated steam outlet temperature (main steam temperature) is one of the important parameters affecting the security and efficiency of unit [1]. Considering the main steam temperature control system is a typical large inertia, large delay, nonlinear and time-varying system, the system has a complex structure and difficult to control [2]. Therefore it is very important to control the main steam temperature accurately. Many scholars apply various control strategies to main steam temperature control system. Conventional PID cascade control method often fails to obtain satisfactory results. Predicative control strategies, such as Smith method [3], DMC and internal mode control (IMC) are much dependent on precise models of the plants. And many advanced control methods, such as fuzzy control [4], artificial neural network and adaptive control require either information on the system states or an efficient on-line identifier, thus may be difficult to apply in practice. ADRC method was first applied for systems with large time-delay in [5], which proposed a new method for these systems. Active disturbance rejection control(ADRC) is an object-model independent control method which was first proposed by Prof. Han in 1998 for rejecting disturbance of a nonlinear system [6]. Literature [7] simplified the ADRC design procedure by considering its ‘linear’ version and then proposed a linear active disturbance rejection control(LADRC).The final tuning parameters for a LADRC are reduced to 3, which greatly simplifies the process of tuning and help develop the LADRC idea and make it an applicable control strategy. By using this method the perfect performance can be obtained for multi-variable control systems with small time-delay, but for main steam temperature control system with large time-delay, the control performance will decline apparently. For the large time-delay plants, Smith Predictor is well used to overcome the effect of large time delay. Found on the above situation, in this paper a new method for main steam temperature control is presented which combines LADRC-P cascade control with Smith Predictor and the framework of the LADRC-P two-loop cascade control system based on Smith Predictor is also constructed to get better control performance. Finally the effectiveness of the control program is simulated by MATLAB. 3rd International Conference on Machinery, Materials and Information Technology Applications (ICMMITA 2015) © 2015. The authors Published by Atlantis Press 905 Design and Tuning of LADRC The structure of a second-order LADRC is shown in Figure.2. Fig.1. Structure of LADRC Consider a generalized second-order system given by:   , , , , y f y y u t bu       (1) where y and u are system output and input respectively, ω is the external disturbance and b is a constant. The entire f is the uncertainty of the system, which is the combination of the unknown internal dynamics of the system and external disturbance. And an extended state observer(ESO) is given by:   t u y y f z y z y z , , , , , , 3 2 1        (2) Assume that   t u y y f , , , ,    is differentiable and let   h t u y y f  , , , ,    . The above equation(2) is equivalent to:        z C y h E u B z A z