Transactions of the Institute of Measurement and Control

Linkens, D.A.; Ting, C.H.; Mahfouf, M.

doi: 10.1177/014233129401600503pmid: N/A

A control engineer is required to design a system to meet the specified desired performance with guaranteed reliability and safety. Reliability and safety are particularly important in the field of process control. However, not only in process control but also in other industrial applications, reliability and safety are emphasised as well as the performance of the controller. The goal of a supervisory structure is to establish the health of a system and design an alarm system for the user in the hope of preventing undesirable consequences. The component is called fault detection and diagnosis (FDI), and a process model-based algorithm is used in this study to monitor the operation of a system. Nevertheless, because stability is a primary requirement of a control system, another requirement of supervision is to guarantee the stability as well as the health of the plant. This is the theme of this paper which combines both control and supervision using GPC.

Rossiter, J.A.; Pratt, R.W.; Smith, K.J.

doi: 10.1177/014233129401600504pmid: N/A

Structural tests are carried out in the automotive and aerospace industries to assess fatigue and vibration properties of engineering structures. Very often the multivariable nature of these tests is ignored and a sequential loop closure scheme is used with PID controllers in each loop. Earlier work has focused on the application of continuous-time Nyquist Array and Characteristic Locus methods to such tests. This paper extends that work by considering a discrete design approach to the Characteristic Locus method. The design process for a two-input two-output system and subsequent implementation on a beam and vibrator rig is described. With modest computational hardware, excellent results are obtained and shown to be superior to those obtained from the implementation of a discretised continuous-time Characteristic Locus controller.

Tse, N.S.F.; Robb, F.F.

doi: 10.1177/014233129401600505pmid: N/A

This Research Report draws the attention of control engineers to some of the results obtained from the application of chaos theory to modelling complex behaviour, The example chosen is a typical management accounting system in a manufacturing firm, and the model includes not only the formal control devices, but also the responses af the human operators, the managers, to the variance signals from that system.Management accounting has become an eclectic discipline and so over time it has called upon a wide variety of methodologies to help to address its problems. It has, perhaps reluctantly, accepted the notion that statistical analysis and the assessment of probabilities can overcome, in part at least, the ambiguity of many managerial situations. It would be expected to be very eager to consider ways and means of reducing reliance on such methods and of increasing the explanaiory power of some of the tools already at its disposal within its own core discipline; for example, the setting of standards for budgetary control.This paper reports recent research into the application of 'Dynamical Systems Theory'as a means of explaining apparently random cost behaviour in a standard cost context. It suggests that attention should be directed to the effects of different standard setting policies on the stability of the behaviour of managerial control systems.