Aspects of economic evaluation of process control systems—I: Influence of statistical nature of disturbances on system profitability

Abstract

There is sufficient evidence of the potential profitability of process control system development to make it important to ask whether a correct proportion of investment is devoted to process control, to assure the maximum contribution to profit. Selection of any particular level of investment ought only to be urged, however, if the basis for the profitability evaluations is reliable. While the level of technical achievement in control engineering has been very high, corresponding economic work has been neither extensive nor advanced enough to command similar confidence, and more is needed to assure reliable outcomes. For example, one of the many factors influencing profitability which must not be left out of account if process control system evaluation is to be reliable is the operating environment of potential disturbances to process operation which must be suppressed if control is to be effective. One aspect of this, with which the present work deals is the influence of statistical combination on the magnitude of individual disturbances. It is the process control system duty to suppress disturbances, and it is shown that there are situations, in which, with the relative magnitudes of disturbances as observed in practice, effectiveness of control can be attenuated, so that profitability of the process control system, or components of the system is reduced below the apparent attainable level. This may be met when process and control system operation are influenced by many variables, the values assumed by these variables fluctuates significantly, and where there are numerous stages of process operation and control, and interaction between them. Three main general points may be made about the outcome of the present stage of work, and its possible impact on design and operation of systems for maximum profitability. 1. (1) On a potentially short term basis, it emphasises factors in structure, environment and operation of process control systems which must be taken in account, with particular reference to disturbance characteristics. 2. (2) In respect of more fundamental work, there remain further steps in analysis of systems with large numbers of interactions, operating stages and variables, and more complexity of structure, with particular reference to the problems of combination of statistical variance. 3. (3) Effective work on both the foregoing lines requires more effective access to operating and structural data on process and control systems, and more thorough analysis and incorporation of cost elements in such data has hitherto been achieved. These are illustrated by graphical presentation of the analysis for a simple example, the control of combustion efficiency for steam generation. This indicates that quantitative statistical analysis must be included at the stage of design to ensure that the contribution to profitability does not fall needlessly below the high marginal level shown by many individual cases of control development, which may typically be three to five times a typical ‘target’ rate for plant and equipment investment generally. Particular attention is essential to an understanding of the nature and amount of process disturbances to ensure that a basic economic duty for a control system actually exists.