The electronic computer offers the possibility of achieving complete and precise control of physical systems. An application of particular interest is the computer control of industrial processes to specified criteria of performance; e.g., control of the process such that product of uniform, specified quality is produced under conditions of minimum cost.
A computer control application to a batch-type process is presented. The method of control, based on repetitive computer action and fast-time scaling makes it possible to obtain a specified end point consistent with a prescribed criterion of optimum performance.
Dynamic analysis of the computer control system is difficult because of the nonlinear and sampled-data characteristics of the computer element. A method for approximate determination of system behavior based on perturbation techniques is described. It is shown, in particular, that the effect of a disturbance signal on the processing path varies in almost direct proportion with the sampling period. Thus, by making the sampling period small relative to the duration of the process, simplifying approximations to the mathematical model may be employed in the derivation of the computer-control equations.
The study is augmented by data on computer control of a three-component-batch chemical process. The results of the dynamic analysis are compared qualitatively with the experimental results of predictor and optimizing control runs.