Simulation and queueing network modeling of single-product production campaigns

Abstract

A two-stage approach to modeling single-product campaigns in specialty chemical production facilities has been developed. In the first stage a simple analytical queueing network model is formulated. In some instances this approximate model is accurate enough for its intended applications (e.g. debottlenecking, capacity planning or scheduling). If it is not, the results indicate which portions of the process must be subjected to more detailed modeling, and a general-purpose simulation language is used for that purpose. In an illustrative application of the approach, a queueing network model was developed for a single-product campaign in an existing multiproduct pharmaceutical plant, and a SLAM II simulation model was subsequently used to model finite intermediate storage and non-Markovian arrival and service times. The final model was validated and used to explore alternative production procedures to reduce the campaign duration and yield additional production capacity.