*G. Lloyds Raja, Ahmad Ali Department of Electrical Engineering, Indian Institute of Technology Patna, India *lloyd.raja@gmail.com, ali@iitp.ac.in Series Cascade Control: An Outline Survey Abstract— Series cascade control structure (SCCS) is widely used in process industries as unity feedback configuration fails to reject the load disturbances before the controlled variable deviates from the setpoint. Several variants of the conventional SCCSs have been reported in the literature to obtain improved closed-loop performance as compared to the conventional structure. For processes with large time delay, cascade control may not yield satisfactory closed-loop performance. To overcome this problem, several modified Smith predictor based SCCS have been reported in the literature. In this manuscript, various series cascade control strategies are briefly reviewed and their advantages and disadvantages are discussed. Suitable tuning strategies for a class of stable, unstable and integrating process models are recommended in order to help the user in selecting the appropriate control strategy. This review covers only model-based series cascade control strategies in which controller parameters are expressed in terms of known process model parameters. Keywords— Series cascade control structure; Smith predictor; closed-loop performance; direct synthesis; IMC. I. INTRODUCTION In conventional unity feedback schemes, the rejection of disturbances does not happen until the controlled variable deviates from the setpoint. This motivated the control engineers to use series cascade control structure (SCCS) in which an intermediate sensor and controller are used to reject the disturbances before they affect the controlled variable. The general block diagram of SCCS is shown in Figure 1.  1 and  2 are the transfer functions of primary and secondary process models whereas  1 and  2 represent the primary and secondary controllers, respectively. ‘u 2 ’ denotes the output of the secondary controller. Output of the secondary loop is denoted by y 2 whereas y 1 denotes the output of the primary loop. d 1 and d 2 are the disturbances entering the input and output of the secondary process models. Setpoint of primary loop is represented by r 1 and output of the primary controller serves as the set-point (r 2 ) for the secondary loop. Cascade control is more effective when the secondary loop is faster than the primary loop and major disturbances enter the secondary loop. Temperature control of natural draft furnace [1] is a well-known practical scenario where SCCS is used. Franks and Worley [2] have used SCCS in a chemical reactor and reported improved closed-loop performance as compared to unity feedback configuration. Integral time absolute error (ITAE) performance measure was used to quantify the improvement achieved by cascade control. McMillan [3] has reported that improved closed-loop performance is achieved by SCCS for stable, unstable and integrating process models. The criterion for conditional stability was derived by Yu et al. [4]. Krishnaswamy et al. [5] have used the PI and P controllers in the primary and secondary loops. Furthermore, the authors have reported that the SCCS yields improved regulatory performance for stable first order plus time delay (FOPTD) process models when disturbances enter the secondary loop at d 1 . r1 y1 d1 d2 - Gp2 Gp1 r2 y2 - Gc2 Gc1 Secondary loop Primary loop u2 e(t) Figure 1. Conventional series cascade control structure Conventional SCCS consists of two controllers and two nested loops as shown in Figure 1. The tuning strategies reported in [2-24] have used conventional SCCS or its variants to control stable, integrating and unstable process models. If a long time delay exists in the primary loop, cascade control scheme fails to yield satisfactory closed- loop performance. By incorporating Smith predictor in the primary loop of SCCS, several modified structures have been reported till date for stable, integrating and unstable [25-37] process models. In process industries, the dynamics of secondary process is usually stable whereas that of the primary process may be stable, unstable or integrating in nature [25-37]. In this manuscript, an attempt has been made to briefly review the various series cascade control structures for stable, unstable and integrating process models. Moreover, the advantages and limitations of various series cascade control strategies are discussed. After carefully analyzing the simulation studies given in literature, suitable control strategies are recommended for a class of process models. The paper is organized as follows: Conventional SCCS and its variants for stable, unstable and integrating process models with small time delay are discussed in section-II. Modified Smith predictor based cascade control strategies for stable, integrating and unstable process models with large time delay are discussed in section-III. Section-IV presents the criteria for choosing an appropriate control strategy for a particular process model. Concluding remarks are given in section-V. II. SCCS AND ITS VARIANTS A. For stable process models Majority of the tuning strategies reported in literature have considered the following first order plus time delay 2017 Indian Control Conference (ICC) January 4-6, 2017. Indian Institute of Technology, Guwahati, India 978-1-5090-1795-9/16/$31.00 ©2017 IEEE 409