The Modeling and Simulation of Combined Discrete/Continuous Process
THE MODELLING AND SIMULATION OF
COMBINED DISCRETE/CONTINUOUS
PROCESSES
by
Paul Inigo Barton, M.Eng, A.C.G.I.
May 1992
A thesis submitted for the degree of Doctor of Philosophy of the University of London and for the Diploma of Membership of the Imperial College
Department of Chemical Engineering
Imperial College of Science, Technology and Medicine
London SW7 2BY
To Amy
2
Abstract
Currently available dynamic simulation packages are mainly suitable for the continuous simulation of large industrial processing systems. In practice, however, few processes can be considered to operate in a completely continuous manner because discrete changes a ect most operations to a greater or lesser extent. Even in a conventional `continuous' process, start-up and shut-down operations, or the application of digital controllers will result in discrete changes superimposed on the predominantly continuous behaviour. Similarly, batch and semi-continuous processes always experience frequent discrete control actions in order to maintain operation in a dynamic, often cyclic, mode. Simulation of systems with these discrete components requires a more sophisticated tool - one that can perform simulations of a combined discrete/continuous nature.
This thesis considers the issues involved in the development of a general-purpose software package for the modelling and simulation of combined discrete/continuous processing systems of arbitrary complexity. The key requirements for such a package are analysed, and a new simulation language based on three distinct categories of entities, models, tasks and processes, is introduced. Model entities describe the continuous physico-chemical mechanisms governing the time dependent behaviour of unit operations, including any discrete changes resulting from these mechanisms, while task entities describe the external control actions or disturbances imposed on a system. A process entity represents a complete dynamic
COMBINED DISCRETE/CONTINUOUS
PROCESSES
by
Paul Inigo Barton, M.Eng, A.C.G.I.
May 1992
A thesis submitted for the degree of Doctor of Philosophy of the University of London and for the Diploma of Membership of the Imperial College
Department of Chemical Engineering
Imperial College of Science, Technology and Medicine
London SW7 2BY
To Amy
2
Abstract
Currently available dynamic simulation packages are mainly suitable for the continuous simulation of large industrial processing systems. In practice, however, few processes can be considered to operate in a completely continuous manner because discrete changes a ect most operations to a greater or lesser extent. Even in a conventional `continuous' process, start-up and shut-down operations, or the application of digital controllers will result in discrete changes superimposed on the predominantly continuous behaviour. Similarly, batch and semi-continuous processes always experience frequent discrete control actions in order to maintain operation in a dynamic, often cyclic, mode. Simulation of systems with these discrete components requires a more sophisticated tool - one that can perform simulations of a combined discrete/continuous nature.
This thesis considers the issues involved in the development of a general-purpose software package for the modelling and simulation of combined discrete/continuous processing systems of arbitrary complexity. The key requirements for such a package are analysed, and a new simulation language based on three distinct categories of entities, models, tasks and processes, is introduced. Model entities describe the continuous physico-chemical mechanisms governing the time dependent behaviour of unit operations, including any discrete changes resulting from these mechanisms, while task entities describe the external control actions or disturbances imposed on a system. A process entity represents a complete dynamic