An MILP model with continuous volume and time variables has been presented to schedule a pipeline that connects one refinery to several DCs. The issue of inventory management such as supplying daily demand with the shortest possible delay and establishing a settling period for discharged products has been taken into account. Some new aspects of pipeline operation such as the priority for discharging batches at DCs and bound for batch size is included in the model formulation. By solving the model we are able to identify a sequence of products that must be injected to the pipeline, size and pumping duration, start and finish pumping time of each batch, amount of discharged product, their duration, and availability time so that the costs of delays and interfaces are minimized with the highest pipeline utilization. Since the model solution time in long-term planning is very high, an algorithm was presented to iteratively find the solution considering the initial inventory and daily demand at each step. The model and algorithm were tested with randomly generated instances and a real-life case. The value of the objective function and running time is adequate and reasonable. Further work is also needed to solve the challenges of longer scheduling horizons and more complex operational aspects, all with full optimality in reasonable time. The uncertainty about future demand for petroleum derivatives leads us to believe that a stochastic model is required to address this situation.
Algorithm for long-term scheduling of multiproduct pipelines
Industrial and Engineering Chemistry Research ; 50 , 24 ; 13899-13910
2011
12 Seiten, 8 Bilder, 14 Tabellen, 15 Quellen
Aufsatz (Zeitschrift)
Englisch
Factors affecting long-term integrity of transmission pipelines
Tema Archiv | 2000
|Long term performance of fusion bond epoxy coated pipelines
Tema Archiv | 2004
|Multiproduct software sets optimal pump use
Tema Archiv | 2006
|