Chemical Engineering Science. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 1972, Vol. 27, pp. 215 1-2161. Pergamm Press. Printed in Great Britain zyxwvutsrqponmlkjihgfedcbaZYXWVUTSR Coolant flow rate vs. agitator speed as manipulated variables for the temperature control of a continuous stirred tank reactor H. R. DEMO, 0. A. IGLESIAS, I. H. FARINA, E. WILLIS and M. DE SANTIAGO Departamento de Ingenieria Quimica, Universidad National de la Plata, 1 esq. 47 la Plata, Argentina zyxwvutsrqpon (Received 1 April 1971; accepted 25 February 1972) Abstract-The performance of a CSTR temperature feedback proportional control when coolant flow rate and agitator speed are being used as control variables, is analyzed. The reactor and jacket dynamics are included. The non-linear equations are integrated numerically and the state plane for each control variable is obtained. The’ analysis shows that, given certain conditions, agitator speed seems to be better than coolant flow rate for temperature control purposes. 1. INTRODUCTION TEMPERATURE control of a continuous stirred tank reactor is of great importance in some exothermic chemical reactions as, for example, in polymerization processes where a runaway reaction would spoil the reactor in a few seconds 111. Aris and Amundson[2] laid the theoretical foundations for the study of CSTR dynamics and control in 1958. They analyzed the reactor control problem where the coolant flow-rate is the control variable, and made a comparison between temperature and concentration as measured variables. The following important conclusions were derived; (a) concentration is not a useful measured variable in most cases; (b) in most cases it is possible to get stability with proportional control (plus some amount of derivative and integral action) on measurements of temperature deviation; (c) an integral controller cannot stabilize the system. However, they did not consider the dynamic behavior of the coolant system. An important assumption was that the overall heat transfer coefficient varies with coolant flow-rate accord- ing to the relationship, UA = C,F,“.8. (l)f tThe meaning of the symbols is explained in the Nomen- clature. Weber and Haniott [3] performed experi- mental work simulating a zero order reaction. They used an auxiliary system where heat was generated proportionally to the temperature, assuming operation in the vicinity of the steady- state. Therefore, they made an experimental local linearization of the chemical reaction. Thus they verified the applicability of empirical rules of controller parameter adjusting and obtained a reasonably good temperature control. In this case the control variable was again the coolant flow-rate. Weber[4] suggested that when the overall heat transfer coefficient depends mainly on agitator speed, this speed should be used as the control variable. Aris and Amundson considered the extreme condition in which the overall heat transfer coefficient is only a function of the coolant flow- rate. The purpose of this paper is to consider the more realistic case, where the overall heat transfer coefficient depends only on the agitator speed. Also we will compare coolant flow-rate with agitator speed as control variables in temperature proportional control. 2. THE DYNAMIC MODEL The mathematical model considered in this work takes into account not only the reactor dynamics but the coolant jacket dynamics as well. Mass and energy conservation on the reactor give the following equations. Mass 2151