A scheme of on-line optimizing control is presented for a nonadiabatic fixed bed reactor which experiences a highly exothermic reaction. The control scheme has been devised to perform adaptive control of bed temperature and on-line steadv K. S. LEE Experiments were conducted with a pilot scale fixed bed reactor where partial oxidation of n-butane to maleic anhydride occurs. The reactor system was interfaced with a microcomputer for real-time control and optimization. With an objective function consisting of a net profit by producing maleic anhydride plus a penalty term on high bed temperature, it was clearly shown that the reaction conditions were driven to the expected optimum region. SCOPEThe rising costs of energy and raw materials have afforded a strong incentive to chemical engineers to develop a well-posed control scheme which gives greater economic return in operating a process. Optimizing control has been accepted as one of the promising ways to fulfill the above requirement and now attracts much attention (Latour, 1979a,b).As was reviewed by Arkun and Stephanopoulos (1980), various optimizing control methods have been proposed up to now. Many of the schemes devised, however, have been based on the steady state behavior of a process (e.g., Box and Draper, 1969;Edler et al., 1970;Haimes and Wismer, 1972;Roberts and Williams, 1981). Two major drawbacks of these methods are that they are slow, and that if the process is subject to persistent disturbances which prevent it from reaching any steady state, they will not be sufficient.Bamberger and Isermann (1978) presented a new optimizing control method where a nonlinear dynamic model of a process is identified using a recursive identification method. Their scheme has many outstanding features, such as noise insensitivity and fast convergence to the optimum point. These features were illustrated by experimentally implementing the method on a thermal pilot process. This idea recently was extended to a large-scale system with many interconnected subprocesses by Garcia and Morari (1981).In this paper, we are concerned with on-line optimizing control of a fixed bed reactor with exothermic reaction. Aside from economic benefit, the study of on-line optimizing control of the reactor is well-justified because modeling difficulty and time-varying characteristics due to catalyst deactivation do not allow advance determination of long-term optimum operating conditions.In a fixed bed reactor with exothermic reaction, control of bed temperature is required in addition to on-line optimization. In this work, we propose an on-line optimizing control scheme which can achieve the above two requirements simultaneously in a fully adaptive manner. This scheme is developed by combining an adaptive control loop for bed temperature with the optimizing control loop of Bamberger and Isermann (1978). In the optimizing control loop, identification of a reactor model and steady state optimization are conducted. The objective function for optimization is taken to be the sum of a net prof...

An extended method for the design of discrete-time multivariable adaptive control is presented. Unlike most adaptive control algorithms, future control inputs are, in this method, predetermined by using the long-term predictor. With this concept, applications of discrete-time multivariahle adaptive controls, which have been applied to only a limited class of practical problems owing to the unboundedness of the control inputs, can be extended to some types of non-minimum phase process. In addition, it is shown that a more generalized control objective, such as minimizing a sum of quadratic errore plus a penalty on control action, can be easily adopted in the design of adaptive controllers by using the concept of a. long-term predictor.

No abstract

A method is developed for on-line noninteracting control of a nonlinear multivariable process to handle constraints on the control variables based on the derivative decoupling control approach. An extension of the proposed modified derivative decoupling control method to load changes (unknown disturbance) is also treated. This modified method is investigated for noninteracting control of laboratory-scale mixing tanks in series both by digital simulation and by experiment using an on-line microcomputer. Liquid levels and temperature in tanks mixing hot and cold water inflow streams are controlled. The results are comparedwith those obtained by instantaneously optimal control and by the inverse Nyquist array technique. Both simulated and experimental performances of the modified derivative decoupling controller are found to be better in comparison with those obtained by the controller based on the inverse Nyquist array. Further, experiments showed that the modified derivative decoupling controller can reduce interaction to a negligible level.

No abstract