City Research Online

Abstract

A number of algorithms exist which can be applied to large scale or decentralised systems where the aim is to determine controller set-points such that the system is operated under optimum conditions. These algorithms are generally applied in a two-layer hierarchical structure, where the lower layer performs direct regulatory control, and the upper layer has the task of determining the set-points of the regulatory controllers to obtain optimum steady-state performance.

In this thesis the performance of such algorithms is examined with specific reference to their application to a laboratory pilot scale vaporiser which is composed of two interconnected subsystems and is controlled from within a distributed computer system. The first subsystem consists of a vaporiser tube and associated details while the second subsystem Is a hot water tank which provides the heating medium for the vaporisation of the freon. Each of the above subsystems has its own local regulatory and optimising decision units. Coordination is performed at supremal level with the aim of trying to ensure that overall system objectives and constraints are satisfied.

Three different types of algorithms are examined: the interaction prediction or direct method; the interaction balance or price method; and the method of integrated system optimisation and parameter estimation. The first two methods are also assessed with various types of feedback measurement information from the process.

The first stage of this work examines the various algorithms and assesses their applicability through off-line simulation of the vaporiser and tank systems. For this a mathematical model is developed and is used with two different control configurations. The second stage of the work consisted of constructing a local microcomputer control system for the hot water tank enabling local feedback control of the hot water tank to take place independently of the already existing local feedback control of the vaporiser. Finally, the algorithms with the best performance indices were selected for implementation in on-line computer control using the distributed computer network.

Publication Type:	Thesis (Doctoral)
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Departments:	School of Science & Technology > Engineering School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads