Industrial & Engineering Chemistry Research, Vol.58, No.19, 8195-8205, 2019
Generalization of Scheduling Models for Batch Plants and Pipeless Plants
Unlike in batch operations where significant progress has been made, there is no adequate research work done on optimal operations of pipeless plants. Pipeless plants were initially explored as an alternative production setup with potential applications in production of paints, adhesives, lubricants, food, pharmaceuticals, and fine chemicals. The concept of pipeless plants requires a renewed momentum to explore if they could become competitive. Unlike in the literature, where researchers have presented different aspects of scheduling models for pipeless plants through different models, in this work the focus is to develop a unified/generalized approach whereby the advancements in the batch scheduling literature can be readily extended to pipeless plants. We explore how batch scheduling models can be extended and generalized for the optimal operation of pipeless plants with fixed design, layout, and routing aspects. The important modifications and additions made to the batch scheduling models for handling scheduling of pipeless plants include: (i) modifications in state task network representation to bring in vessel, product, and equipment identities, leading to a "generalization" of models for batch and pipeless plants; (ii) treatment of moveable vessels as states, and not as separate resources, which are consumed only by the first unit and subsequently released only by the last unit in the production sequence for each product; and (iii) elimination of material balances and batch size aspects to simplify the models and reduce the problem size. Emphasis has been laid on development of a systematic procedure for effective STN representation for pipeless plants leading to unification of scheduling models for batch plants and pipeless plants. Two scheduling models are presented with varying problem size and complexity to explore the best suitable model for the optimal operation of pipeless plants. The proposed models have been validated on two benchmark problems, drawn from the literature, involving multiproduct and multipurpose plant features.