278600 Multiscale Strategic Planning Model for the Design of Integrated Ethanol and Gasoline Supply Chain

Thursday, November 1, 2012: 4:05 PM
325 (Convention Center )
Federico Andersen, Chemical Engineering, Planta Piloto de Ingenieria Quimica (PLAPIQUI), CONICET- Universidad Nacional del Sur , Bahia Blanca, Argentina, Maria Soledad Diaz, Process Systems Engineering, Planta Piloto de Ingenieria Quimica (PLAPIQUI), CONICET - Universidad Nacional del Sur, Bahia Blanca, Argentina and Ignacio E. Grossmann, Department of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA

Multiscale Strategic Planning Model for the Design of Integrated Ethanol and Gasoline Supply Chain

Federico Andersen1, Soledad Diaz1 and Ignacio Grossmann2,

(1)  Chemical Engineering, Planta Piloto de Ingenieria Quimica (PLAPIQUI), Universidad Nacional del Sur - CONICET, Bahia Blanca, Argentina

(2)  Dept. of Chemical Engineering, Carnegie Mellon University, Pittsburgh, PA 15213

Ethanol continues to receive attention as a major energy source from biomass that can help to address the increasing needs of liquid fuels that are predicted over the next few decades. Research and industrial efforts have made of ethanol as one of the safest (both economically and environmentally) alternatives to conventional fuels, at least in the short and medium term. However several challenges remain, which have not yet been fully solved and developed. One of the main concerns is related to the implementation of higher blends of ethanol with gasoline and the utilization of alternative production processes for lignocellulosic ethanol. These concerns are based on the RFS Schedule under the Energy Independence and Security Act of 2007, which points to the utilization of 21 of 36 billion gallons of advanced biofuel by 2022. The widely used raw material to produce E10 within the US is corn, but due to increasing demands, alternative raw materials such as lignocellulosic material must be considered for the production of fuel grade ethanol. Facing the need to start the commercialization of higher ethanol blends, new challenges with respect to bioethanol production can be identified. Therefore in this transition from E10 to E30 it would be necessary not only to consider the addition of new raw materials, but also to allow for the distribution of different ethanol blends at the same time in the same gas station.

Because of the need to provide the market with multiple fuel blends, the business of gas stations has become a multi-blend market, where the demand of each blend depends on the driver's preference and legal requirements. This preference is mainly related to the price of the blend and the miles per gallon of fuel that can be driven, which is also different among vehicles. This feature requires an accurate forecast of demand for every blend and poses a new challenge to estimate and satisfy this demand with minimal cost, opening up a new market for multi-product gas stations, where blender pumps are used to obtain the desired blend mixing E85 with E10 at the gas station itself. Since most of the current gasoline distribution takes place through single-product gas stations and the investments to retrofit all of them into multi-product would be huge, we have modeled the coexistence of both of these types of gas stations and the prospect to move towards the multi-product ones in a gradual way, considering the possibility to enhance or reduce the total quantity of gas stations.

In this work we formulate the integrated ethanol and gasoline supply chain, taking into account the stages from the harvesting sites, through production sites for ethanol, petroleum refineries, blending stages and up to the retail of different blends in gas stations. The general problem can be stated as follows. Given a superstructure that combines all these stages and different means of transportation which connect the nodes, the integrated ethanol and gasoline supply chain can be represented with a multiperiod model and a time horizon of 20 years. Given also an initial capacity for ethanol plants and gasoline distribution center and data of type and quantity of existing gas stations; and given a forecast of demand for different blends over the entire time horizon, the main goal is to determine several major decisions in order to minimize cost. Some of these decisions involve: whether to install or not small; medium or large ethanol plants and gasoline distribution centers (taking into account economy of scale for the investments), and timing both for retrofits of different types of gas stations that involve blending pumps. The costs components included in the analysis are: investment capital cost (where economy of scale has been considered); raw materials; production; transportation; storage and distribution cost.

Since the industries and players of the bioethanol supply chain are widely spread over the US, the coordination over different geographical locations is a key feature for the optimization of the supply chain. Another key feature is based on different time scales required in the model. This fact arises from the inherent difference between the lifetime of investments in ethanol plants, which is on the order of years, and the replenishment of the gas stations that occurs every week or every two weeks. Because of both features the need to cope with two different formulations of the model can be identified. Each formulation contains a different level of detail of some stages, especially in the downstream supply chain (at gas stations level).

Within the more general formulation we have considered an aggregated strategic planning model in order to identify the regions of the US where more investments are needed and the optimal configuration of the network. Within the second formulation, even though the main modeling effort has been devoted to the operation of gas stations and the selection of blending pumps, the global characteristic of a supply chain model is also being reflected. Hence, this multiscale strategic planning model contains a high level of details about gas stations in comparison to other stages of the supply chain.

To illustrate the application of the proposed MILP models, we have considered over a geographical region a set of feasible retrofits between the single-product and multi-product gas stations, and decrease of capacity in the period that they are being retrofitted. Also we have included the possibility to satisfy high ethanol fuel demand with lower ethanol fuels.

Integer variables are associated to active gas stations; required retrofits; new gas stations and those that should be dropped off. Due to the size of the multiscale strategic planning model, we have applied a decomposition scheme to efficiently solve the problem using the Lagrangean decomposition strategy.

1- Datta et al., 2011. Ethanol – the primary renewable liquid fuel. J Chem Technol Biotechnol 86: 473 – 480.

2- Martin M., Grossmann I. 2011. Systematic synthesis of sustainable biorefineries: A review.

3- Neiro S., Pinto. J. 2004. A general modeling framework for the operational planning of petroleum supply chains. Computers and Chemical Engineering 28, 871-896

4- Rusell D., Ruamsook K., Thomchick E. 2009. Ethanol and the Petroleum Supply Chain of the Future: Five Strategic Priorities of Integration. Transportation Journal. 48-1

5- You F., Grossmann I. 2008. Mixed-Integer Nonlinear Programming Models and Algorithms for Large-Scale Supply Chain Design with Stochastic Inventory Management. Ind. Eng. Chem Res., 47, 7802-7817

6- You F., Wang B., 2011. Life  Cycle Optimization of Biomass-to-Liquids Supply Chains with Distributed-Centralized Processing Networks. Industrial & Engineering Chemistry Research. 50(17): 10102-10127

Extended Abstract: File Not Uploaded
See more of this Session: Supply Chain Optimization II
See more of this Group/Topical: Computing and Systems Technology Division