Modeling and control of a fixed bed reactor

A comprehensive study of a fixed bed reactor carrying out the water-gas shift reaction, from the mathematical modeling to controller applications, is presented. The development of the mathematical model for a distributed parameter system in this dissertation is oriented toward assisting control system development. It represented a major part of the work in this research. A mathematical model based on mass and energy balances of the reactor system is developed in detail. This partial differential equation model is then discretized by the orthogonal collocation procedure. Dynamic and steady state simulations show the effects of changes in input variables on the reactor output variables and the sensitivity of the state variables to such changes, and thus guides to effective control system configurations.; The development of effective multivariable control strategies depends on the choice of manipulated and controlled variables. Since fixed bed reactors are distributed parameter systems, the selection of sensor locations is a complicated problem associated with sensor sensitivity and interaction. A strategy based on Singular Value Analysis is used to determine the selection of sensor locations and control loop pairing for SISO designs.; In order to deal with the complex behavior of a fixed bed reactor, i.e. nonlinearity, large time delays, and inverse response, the Dynamic Matrix Control or Quadratic Dynamic Matrix Control (DMC/QDMC) algorithm is implemented. Control of a distributed fixed bed reactor, using the single-input single-output (SISO) DMC or QDMC algorithm, is demonstrated. The control algorithm seems to be very effective in dealing with difficult dynamic process characteristics; nonlinearity, large time delays, and inverse response. DMC or QDMC algorithm shows excellent control system performances when the reactor is subjected to changes in set point or faces a disturbance. The multi-input multi-output (MIMO) QDMC algorithm is also successfully implemented on the reactor. The QDMC algorithm yields excellent performances in set point tracking and disturbance rejection. The results show the loop interaction handling capability of the MIMO QDMC algorithm, as well as its ability to deal with the inverse response, and large time delays.