Simple optimization for dual composition control

Optimization has become the ultimate objective for designers of distillation control systems. This is as it should be, since optimization promises to adjust the operation of distillation columns for maximum profit, even as the cost of energy and the values of product streams continually vary. In pursuit of this elusive objective, much effort has been focused on developing elaborate computer models of various separations to aid in determining optimum set points. Many of the models are capable of excellent results. However, they are also quite expensive, especially if they are developed by outside sources. Beyond their initial cost, they are difficult and expensive to calibrate and maintain, their assumptions and limitations may not be well-understood by operating personnel, and plant conditions inevitably change. Consequently, all too often, they lose their accuracy and fall into disfavor and disuse. This article describes an algorithm that uses a simple algebraic model to predict optimum composition set points for dual composition control of two-product columns. Although this technique will not equal the accuracy of more elaborate models, it can reliably predict whether an optimum lies within the range of allowable compositions, and whether it is sharp or broad. Where an optimum does exist, roughmore » economic data can give a first approximation of the economic incentive for applying more powerful analysis techniques. An optimum that lies beyond the allowable operating region indicated the need for constraint controls that can maintain the separation at its operating limits.« less