Optimal scheduling of an integrated batch and continuous process system
Master Thesis
2000
Permanent link to this Item
Authors
Supervisors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher
University of Cape Town
Department
License
Series
Abstract
An industrial problem has been posed involving the determination of the optimal scheduling and sequencing of a batch process system. The plant incorporates both batch and continuous units and is further complicated by the existence of a nonlinear relationship between the processing rate and the lot size in the continuous plant. The primary goal of this project is to investigate the application of a rigorous mathematical approach to determine the optimal operating policy of a complex processing system. In broad terms, scheduling involves the allocation of a limited number of resources to ensure the completion of a set of tasks, in an optimal way. The problem is formulated as a mathematical model through the use of mixed-integer programming. The difficulty associated with this type of work is in describing the plant using constraints to develop a mathematical model which both accurately reflects the plant and renders the problem solvable. The model must incorporate the connectivity of the plant and prevent resource-task allocation conflicts. A survey of the pertinent literature has been conducted and is reviewed in Chapter 2. Here the principles of batch scheduling are addressed. The focus of the literature review is on the formulation of such problems into a mathematical representation. A description of two fundamentally different approaches for the formulation of short-term scheduling problems is presented. Other aspects of the review include the commercially available software used in the solution of these problems, campaign planning and on-line scheduling. The work in this thesis initially focuses on the scheduling of a simplified version of the industrial plant. The simplified version consists exclusively of batch processing units. The motivation for this was to develop a formulation which incorporated the unique characteristics of the full problem but was computationally easier to solve. A number of scenarios were conducted which show both the flexibility of the formulation and the ability of the formulation to reschedule tasks when faced with different operating conditions. In the discussion of these scenarios, issues such as the computational efficiency and the implications of the results are addressed. The next step was to add a continuous plant upstream to the simplified batch process. The continuous plant exhibits a nonlinear relationship between the processing rate and the efficiency of operation. An approximation of the non-linearity is proposed and further scenario studies are carried out. Finally, the full problem is tackled but due to the great computational expense required to solve the problem an alternative method is proposed. This method is based on a series of random-type scenarios, which serve as an alternative as well as benchmark to the solutions obtained via rigorous optimization.
Description
Bibliography: leaves 136-139.
Keywords
Reference:
MacRosty, R. 2000. Optimal scheduling of an integrated batch and continuous process system. University of Cape Town.