Sascha Herrmann

Abstract

In this PhD project we investigate process scheduling problems for multi-product chemical plants operated in continuous production mode. Planning problems in the process industries are characterised by a heightened complexity, which is due to numerous technological particularities of chemical production processes, such as specific requirements in the storage of processed commodities or the necessity of sequence-dependent changeovers of the processing units. The occurrence of continuous material flows in connection with limited storage capacity for intermediate products constitutes a distinguishing feature of continuous-mode production.

Although continuous-mode production is accorded great economic significance in the areas of the chemical and pharmaceutical industries, the overwhelming part of the process scheduling literature is devoted to discrete batch production. For the case of continuous-mode production, mixed-integer linear and nonlinear programming approaches have been proposed in the literaure, which for real-world size instances however quickly reach their limitations.

That is why we pursue a two-stage decomposition approach with an operations planning and an operations scheduling stage. At the operations planning stage, we generate the set of the operations to be executed on the processing units, together with their respective operating conditions such as durations, processing rates, and proportions of charge materials and outputs. At the second stage, the operations are then scheduled on the processing units of the plant subject to storage-capacity and material-availability constraints.

In order to take account of uncertainty through random processing times and yields, concepts of robust and reactive planning as well as online optimization developed for batch process scheduling are transferred to the case of continuous-mode production. The resulting scheduling procedures are evaluated on case studies taken from literature and practice. Specific recommendations for the use of appropriate planning methods are worked out for various uncertainty scenarios.