Implementation of a simulation model for the scheduling of a multi-product dairy plant

Abstract

Dynamic scheduling of multi-product plants is one of the most difficult groups of problems of high economic importance. The well developed, conventional solutions use various methods of mathematical programming (e.g. MILP), formulated as the optimization of a simplified model. However, the specific technological details and the possible multiple use of the model require also detailed dynamic model based sub-optimal methods. Considering the inherent coupling between structural and functional characteristics of dynamic scheduling, Direct Computer Mapping (DCM) based Programmable Structures of Research Group on Process Network Engineering are experimentally applied for the generation of the dairy plant model. The generation starts from the process network representation of the flow-sheet and from one state and one transition meta-prototype. The flow-sheet comprises storages of raw materials, intermediate materials and end-products, as well as the recipes, the in parallel working multi-functional process units with the numerous time-driven and event-driven constraints. Model generation is followed by the programming of the actual prototypes and parameterization of the model. In the present work the generated process model and the simulation tests will be discussed, focusing on the combined use of the transitions for the description of the material balance, as well as for the interpretation of the time- and event-driven rules, determining the sequence of operations.