An integrated multimodal transportation model for ...docs.trb.org/prp/17-02769.pdf · An integrated multimodal transportation model for ... examined the cost comparison between truck

1 Park, Szmerekovsky, Osmani, and Aslaam

An integrated multimodal transportation model for switchgrass-based bioethanol supply

chain with a case study based on North Dakota

Yong Shin Park1*, Joseph Szmerekovsky2, Atif Osmani3 and N. Muhammad Aslaam4

1First Author: Yong Shin Park (*Corresponding Author)

Affiliation: Transportation and Logistics Program, North Dakota State University

Address: 1320 Albrecht Blvd, Fargo, ND 58105, USA

Phone: +1 (701) 231-7767

Fax: +1 (701) 231-1945

Email: [email protected]

2Second Author: Joseph Szmerekovsky

Affiliation: Department of Management and Marketing, North Dakota State University

Address: NDSU Department 2420, PO Box 6050, Fargo, ND 58108, USA

Phone: +1 (701) 231-8128

Fax: +1 (701) 231-7508


3Third Author: Atif Osmani

Affiliation: Department of College of Business, Minnesota State University

Address: 1104 7th Ave South, Moorhead, MN 56563, USA

Phone: +1 (218) 477-2489

Fax: +1 (218) 447-2238


4 Fourth Author: N. Muhammad Aslaam

Affiliation: Transportation and Logistics Program, North Dakota State University

Address: 1320 Albrecht Blvd, Fargo, ND 58105, USA

Phone: +1 (701) 231-7767

Fax: +1 (701) 231-1945


CONTENTS Number of Words

Abstract 129

Text 3506

Reference 1135

Figure and Table Figure: 4, Table: 5, Total 9 (250 * 9 = 2250)

Total 7020

mailto:[email protected]

mailto:[email protected]


ABSTRACT

This study formulates a mixed integer linear programming (MILP) model that integrates

multimodal transport into the switchgrass-based bioethanol supply chain (MTSBSC). The two

transport modes are truck and rail. The objective of this study is to minimize the total cost for

cultivation/harvesting, infrastructure, storage process, bioethanol production, and transportation.

Strategic decisions, including the number and location of intermodal facilities and biorefineries,

and tactical decisions, such as amount of biomass shipped, processed, and converted into

bioethanol are validated using the state of North Dakota as a case study. It was found that

multimodal transport scenario is more cost effective than single mode of transport (truck), which

results in cheaper bioethanol cost. A sensitivity analysis was conducted to demonstrate the impact

of key factors on MTSBSC decision and bioethanol cost.


INTRODUCTION

Due to worldwide global warming, energy security, multiple societal issues, and increasing oil

demand, there has been great interest in the development of cellulosic biofuel using renewable

biomass feedstock from wood, forest residues, and agricultural residues, which are supreme

alternatives for transportation fuel. According to the U.S. Energy Information Administration

(EIA), in 2015, the United States (U.S) consumed about 7.08 billion barrels of petroleum products

(i.e., an average of about 20 million barrels per day), which account for 21% of worldwide

consumption (1). The transportation industry is the most dominant sector for nation’s petroleum

consumption accounting for 56% of total U.S. fuel use (2). Bioethanol is one type of cellulosic

biofuel, and corn is the major source of current bioethanol as a first generation renewable resource

in U.S. However, there is much debate about first generation biofuel associated with global food

security due to biofuel production directly from food crops (3). As an alternative, lignocellulosic

biomass feedstock is a promising source for producing bioethanol. Switchgrass is one type of

lignocellulosic biomass which is regarded as one of the best second generation renewable energy

resource (4). Many researchers have made great effort on lignocellulosic biomass based bioethanol

supply chain design with a primary focus on minimizing the total system cost by prescribing a

strategic (i.e. location of biomass storage and size of new refinery) and tactical (i.e. amount of

biomass shipped and processed) supply chain plan (5-10). Some other studies developed a model

that maximizes profit (11,12) and/or minimizes risk associated with biomass supply chain

investment (9). A number of studies have extended previous models by considering a multi-period

model to deal with spatial and temporal dimension for long term strategic plan for biomass supply

chain (14,15). Multiple types of biomass feedstock were addressed for forest (12), urban waste

(12), and other agricultural biomass (13,17,18). Recent studies have contributed to sustainability

issues to investigate the environmental impact and regulation (19-23).

A typical biofuel supply chain plan should simultaneously consider the determination of

location of feedstock area, harvesting method, storage, biorefineries, transport of biomass and

biofuel, and biofuel production (17). Making financially optimal decisions is a key strategy in

biomass supply chain. Locating storages close to biorefineries will reduce unit transportation costs,

but might increase the transportation costs if the storages are far away from the harvesting

/collecting area. Biomass can be directly shipped to a preprocessing plant or sent to an intermodal

hub or storage facility from the harvesting/collecting area. Storage serves as a warehouse to store

biomass and manage inventories. Intermodal hubs also play an important role in consolidating

freight load of multiple mode of transportation (i.e. truck, rail, and ship) in supply chain networks

(18). Each transportation mode will impact supply chain costs as well (3). The truck is known to

be the most economical mode for short haul shipment and rail car is the cheaper mode for long

haul shipment. Rail car can handle more tons of cargo at a lower cost than truck, which is also

more energy efficient transportation mode (19). Multimodal transport, which is a combination of

at least two different modes of transport offers more flexibility, is cheaper, and more efficient

transportation mode. It enhances commercial viability, and should be integrated into cellulosic

biofuel supply chain design (20). However, existing literature related to cellulosic bioethanol

supply chain design, assumes the truck is the only transport mode, although the multimodal

transportation option is very attractive for geographic dispersion of demand and supply chain of

biofuel (21).

To the best of authors knowledge, only limited number of studies are found in the literature

for the application of multimodal transport into bioethanol supply chain design (8, 17, 18, 20, 23,

29, 27, 24). This study is motivated by Ekşioǧlu et al.(25), which addressed the impact of


intermodal facilities for decision support system (DSS) for corn-based biofuel supply chain design.

This study also determines the minimum cost of biofuel delivery with different levels of production

capacity and transportation cost. However, lack of investigation of biomass storage location when

it is integrated with intermodal facility on biofuel supply chain, is missing part from previous study.

Also, there is limited work that integrate multimodal transport into switchgrass-based bioethanol

supply chains (MTSBSC). William et al. (26) examined the cost comparison between truck and

rail transport mode for downstream switchgrass-based bioethanol supply chain throughout the

United States. Other works including Zhu and Yaq (27), You et al. (15), An et al. (28) and Zhang

et al. (3), considered only truck transport mode in switchgrass-based bioethanol supply chain as a

whole.

Based on the commonly identified supply chain aspects from reviewed literatures, this

paper proposes a MILP (Mixed-integer linear programming) that will investigate the cost effective

MTSBSC. The goal of this study is to minimize total system cost including, marginal rental cost,

cultivation cost, harvesting cost, infrastructure capital cost, transportation cost across the entire

supply chain over one year planning horizon time period. Proposed supply chain structure of

MTSBSC model are shown in Fig 1. Switchgrass biomass is harvested and transported to storage

located at intermodal facility directly by trucks. The switchgrass biomass stored at truck yard is

shipped to biorefineries by truck and biomass stored at rail yard is transported by rail. Then,

bioethanol produced from biorefineries is delivered to demand zone via truck.

FIGURE 1 Switchgrass-based multimodal bioethanol supply chain structure.

Key features of this study includes: (1) two transport modes using both truck and rail will

demonstrate the applicability of the model for case of North Dakota throughout the entire

switchgrass-based bioethanol supply chain from feedstock to end user; (2) Prior literatures lacks

the investigation of switchgrass biomass storage at intermodal facility. Either round bales or


square bales stored under tarp storage system will be built nearby rail spurs or along state highway

at the intermodal facility, which have not been addressed in any other study except a study by

Zhang et al. (17), their study placed forest wood storage at nearby rail spurs along with Class A

highways to alleviate the impact of the spring breakup period on truck flow, which is also

applicable to our case study.

PROBLEM STATEMENT AND MODEL FORMULATION

The main objective of this paper is to build a MTSBSC model that aids the design and operation

management of the bioethanol supply chain network. The MTSBSC design problem consists of

locating a set of intermodal hubs, selecting suitable biorefineries among existing location, and to

determining the route of biomass and bioethanol flows.

Two sets of decision, such as strategic and tactical decision, are made simultaneously. The

strategic decisions are mainly on the location of intermodal storage, number of intermodal storage,

biorefinery location, harvesting area assigned to particular intermodal storage or to biorefineries,

and storage that are assigned to a particular biorefineries. Tactical decisions are amount of biomass

harvested and shipped through the multimodal supply chain network, biomass stored, and amount

of bioethanol produced.

The objective of this study is to seek a minimum cost-strategy of the total switchgrass-

based bioethanol supply chain that integrates both truck and rail transportation modes by

determining various supply chain logistics decision variables. Before describing the model in

detail, notations of subscript indices, input parameters, and decision variables used in model

formulation are presented in the table 1.

The objective function Eqs. (1) - (13) minimizes the annual total supply chain cost

including switchgrass marginal rental cost 𝐶𝑟𝑒𝑛𝑡, cultivation cost 𝐶𝑐𝑢𝑙𝑡, harvesting cost 𝐶ℎ𝑎𝑟𝑣,

storage cost at intermodal facility 𝐶𝑠𝑡𝑜𝑟, switchgrass and bioethanol transportation cost 𝐶𝑡𝑟𝑎𝑛𝑠,

bioethanol production cost 𝐶𝑝𝑟𝑜𝑑, intermodal facility capital investment cost 𝐶𝑖𝑛𝑡𝑐𝑎𝑝, biorefinery

capital investment cost 𝐶𝑏𝑟𝑐𝑎𝑝. Transportation cost 𝐶𝑡𝑟𝑎𝑛𝑠 in Eqs (9) – (13), consists of four

terms: transport cost from switchgrass harvesting area to biorefinery 𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑏, transport cost from

switchgrass harvesting area to intermodal storage 𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑖, transport cost from intermodal storage

to biorefinery 𝑐𝑡𝑟𝑎𝑛𝑠,𝑖𝑏 , and transport cost from biorefinery to demand zone 𝑐𝑡𝑟𝑎𝑛𝑠,𝑏𝑑. In particular,

Eqs. (12) indicates both truck and rail transportation mode used in model formulation. All

variables except binary variables are non-negative continuous.

Minimize 𝐶𝑟𝑒𝑛𝑡 + 𝐶𝑐𝑢𝑙𝑡 + 𝐶ℎ𝑎𝑟𝑣 + 𝐶𝑠𝑡𝑜𝑟 + 𝐶𝑡𝑟𝑎𝑛𝑠 + 𝐶𝑝𝑟𝑜𝑑 + 𝐶𝑖𝑛𝑡𝑐𝑎𝑝 + 𝐶𝑏𝑟𝑐𝑎𝑝 (1)

𝐶𝑟𝑒𝑛𝑡 = ∑ ∑ 𝑐𝑖 × 𝑄𝑖𝑡

𝑖𝜖𝐼𝑡𝜖𝑇

(2)

𝐶𝑐𝑢𝑙𝑡 = ∑ ∑ 𝑣𝑖 × 𝑄𝑖𝑡


(3)

𝐶ℎ𝑎𝑟𝑣 = ∑ ∑ ℎ𝑖 × 𝑄𝑖𝑡


(4)


𝐶𝑠𝑡𝑜𝑟 = ∑ ∑ 𝑐𝑖𝑠𝑡𝑜𝑟 × 𝑄𝑗𝑡


(5)

𝐶𝑝𝑟𝑜𝑑 = ∑ ∑ 𝑐𝑏𝑝 × 𝑄𝑘𝑡


(6)

𝐶𝑖𝑛𝑡𝑐𝑎𝑝 = ∑ ∑ 𝑓𝑖𝑐 × 𝑋𝑗

𝑗𝜖𝐽

(7)

𝑡𝜖𝑇

𝐶𝑖𝑛𝑡𝑐𝑎𝑝 = ∑ ∑ 𝑓𝑏𝑐 × 𝑌𝑘

𝑗𝜖𝐽𝑡𝜖𝑇

(8)

𝐶𝑡𝑟𝑎𝑛𝑠 = 𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑏 + 𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑖 + 𝑐𝑡𝑟𝑎𝑛𝑠,𝑖𝑏 + 𝑐𝑡𝑟𝑎𝑛𝑠,𝑏𝑑 (9)

𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑏 = ∑ ∑ ∑

𝑘𝜖𝐾

( 𝑐𝑡𝑟𝑢𝑐𝑘,𝑚𝑐


× 𝑑𝑠𝑏 + 𝑐𝑡𝑟𝑢𝑐𝑘,𝑙𝑢) × 𝑄𝑠𝑏𝑡 (10)

𝑐𝑡𝑟𝑎𝑛𝑠,𝑠𝑖 = ∑ ∑ ∑

𝑗𝜖𝐽



× 𝑑𝑠𝑖 + 𝑐𝑡𝑟𝑢𝑐𝑘,𝑙𝑢) × 𝑄𝑠𝑖𝑡 (11)

𝑐𝑡𝑟𝑎𝑛𝑠,𝑖𝑏 = ∑ ∑ ∑ ∑

𝑚𝜖𝑀

𝑘𝜖𝐾

{( 𝑐𝑡𝑟𝑢𝑐𝑘,𝑚𝑐

𝑗𝜖𝐽𝑡𝜖𝑇

× 𝑑𝑖𝑏 + 𝑐𝑡𝑟𝑢𝑐𝑘,𝑙𝑢) × 𝑄𝑖𝑏𝑚𝑡}

+{( 𝑐𝑟𝑎𝑖𝑙,𝑚𝑐 × 𝑑𝑖𝑏 + 𝑐𝑟𝑎𝑖𝑙,𝑓𝑐 ) × 𝑄𝑖𝑏𝑚𝑡} (12)

𝑐𝑡𝑟𝑎𝑛𝑠,𝑏𝑑 = ∑ ∑ ∑

𝑒𝜖𝐸


𝑘𝜖𝑘𝑡𝜖𝑇

× 𝑑𝑠𝑖 + 𝑐𝑡𝑟𝑢𝑐𝑘,𝑙𝑢,𝑏) × 𝑄𝑏𝑑𝑡 (13)

The model constrains are presented in Eqs. (14) - (20). Constraint (14) assures that amount

of switchgrass harvested at area i does not exceed the marginal land availability. Constraint (15)

is the feedstock flow conservation constraint that amount of biomass transported from harvesting

area to intermodal storage and refinery is same as what is actually available in feedstock area

during time period t. Constraint (16) impose a flow conservation on intermodal storage. Constraint

(17) is a logic constraint, stating that there is no flow through intermodal storages unless one is

open. Constraint (18) is a flow conservation constraint for refineries. Constraint (19) ensure that a

maximum of one biorefinery can be chosen at each location. Constraint (20) is another logic

constraint that there is no biofuel production unless one is open. Constraint (21) ensure that during


any time period t, the volume of bioethanol from biorefineries to each demand zone must be greater

or equal to the biofuel requirement for each demand zone.

𝑄𝑖𝑡

≤ 𝑎𝑖𝑡 ∀i ∈ I, ∀t ∈ T (14)

𝑄𝑖𝑡 = ∑ 𝑄𝑠𝑖𝑡 +

𝑖𝜖𝐼

∑ 𝑄𝑠𝑏𝑡

𝑗𝜖𝐽

(15)

∑ 𝑄𝑠𝑖𝑡 +

𝑖𝜖𝐼

(1 − 𝛿) × 𝑆𝑗,𝑡−1 = 𝑆𝑗𝑡 + ∑ ∑

𝑚𝜖𝑀

𝑄𝑖𝑏𝑚𝑡

𝑘𝜖𝐾

∀j ∈ J, ∀k ∈ K, ∀t ∈ T (16)

𝑆𝑗𝑡 ≤ ∑

𝑗𝜖𝐽

𝑝𝑗 × 𝑋𝑗 ∀j ∈ J, ∀t ∈ T (17)

∑ ∑ ∑

𝑚𝜖𝑀

( 𝑄𝑠𝑏𝑡 + 𝑄𝑖𝑏𝑚𝑡) × 𝜃 = ∑ ∑

𝑒𝜖𝐸

𝑄𝑏𝑑𝑡

𝑘𝜖𝐾

𝑗𝜖𝐽𝑘𝜖𝐾

∀j ∈ J, ∀k ∈ K, ∀m ∈ M, ∀t ∈ T (18)

∑

𝑝𝜖𝑃

𝑌𝑘𝑝 ≤ 1 (19)

𝑆𝑘𝑡 ≤ ∑

𝑘𝜖𝐾

𝑏𝑘 × 𝑌𝑘 ∀k ∈ K, ∀t ∈ T (20)

∑ 𝑄𝑘𝑡 ≥

𝑘𝜖𝐾

𝑑𝑡 ∀e ∈ E, ∀t ∈ T (21)


TABLE 1 Notations Used in Model Development

Symbol Description Symbol Description

Indices/sets 𝒉𝒊 Harvesting cost of switchgrass ($/ha)

i Switchgrass supply points 𝒄𝒊𝒔𝒕𝒐𝒓 Unit storage cost at storage yard at

intermodal facilities ($/ton)

j Intermodal facility locations 𝒄𝒃𝒓𝒔𝒕𝒐𝒓 Unit storage cost at biorefineries ($/ton)

k Biorefinery locations 𝒄𝒃𝒑 Bioethanol production cost at refineries

($/gallon)

q Capacity level of biorefineries 𝒄𝒕𝒓𝒖𝒄𝒌,𝒍𝒖 Truck loading and unloading cost ($/ton)

e Bioethanol demand points 𝒄𝒕𝒓𝒖𝒄𝒌,𝒎𝒄 Truck variable mileage cost ($/ton-mile)

m Transport mode 𝒄𝒓𝒂𝒊𝒍,𝒇𝒄 Rail fixed cost ($)

t Modeling horizon of 1 year with time periods 𝒄𝒓𝒂𝒊𝒍,𝒎𝒄 Rail variable mileage cost ($/ton-mile)

Input parameters used in model development 𝒄𝒕𝒓𝒖𝒄𝒌,𝒍𝒖 Truck loading and unloading cost ($/ton)

𝑪𝒓𝒆𝒏𝒕 Marginal land rental cost ($) 𝒅𝒔𝒊 Transport distance from supply area to

intermodal facilities (mile)

𝑪𝒄𝒖𝒍𝒕 Biomass cultivation cost ($) 𝒅𝒔𝒃 Transport distance from supply area to

biorefineries (mile)

𝑪𝒉𝒂𝒓𝒗 Biomass harvesting cost ($) 𝒅𝒊𝒃 Transport distance from intermodal facilities

to biorefineries (mile)

𝑪𝒕𝒓𝒂𝒏𝒔 Biomass transport cost ($) 𝒅𝒃𝒅 Transport distance from biorefineries to

demand points (mile)

𝑪𝒊𝒏𝒕𝒄𝒂𝒑 Intermodal facility capital cost ($) 𝜹 Biomass deterioration rate (%)

𝜽 Bioethanol conversion rate (gallon/ton)

𝑪𝒃𝒓𝒄𝒂𝒑 Biorefinery capital cost ($) 𝒅𝒕 Biofuel demand in period t (gallon)

𝑪𝒔𝒕𝒐𝒓 Biomass storage cost ($) Decision variable used in model development

𝑪𝒑𝒓𝒐𝒅 Biofuel production cost ($) 𝑿𝒋 = 1 if an intermodal facility is opened at

location j; 0 otherwise (Binary)

𝒂𝒊 Maximum marginal biomass availability (ton) 𝒀𝒌𝒑 = 1 if a biorefinery is opened at location k

with capacity level p; 0 otherwise (Binary)

𝒄𝒕𝒓𝒂𝒏𝒔,𝒔𝒃 Transport cost of biomass from supply area to

biorefineries ($/ton-mile) 𝑸𝒊𝒕 The quantity of biomass harvested at supply

area i (ton)

𝒄𝒕𝒓𝒂𝒏𝒔,𝒔𝒊 Transport cost of biomass from supply area to

intermodal facility ($/ton-mile) 𝑺𝒋𝒕 The quantity of biomass stored at intermodal

facility (ton)

𝒄𝒕𝒓𝒂𝒏𝒔,𝒊𝒃 Transport cost of biomass from intermodal facilities

to biorefineries ($/ton-mile)

𝒄𝒕𝒓𝒂𝒏𝒔,𝒃𝒅 Transport cost of biofuel from biorefineries to

demand points ($/ton-mile) 𝑺𝒌𝒕 The quantity of biomass stored at biorefinery

(ton)

𝒑𝒋 Storage capacity (ton) 𝑸𝒔𝒊𝒕 The quantity of biomass transported from

supply area to intermodal facility (ton)

𝒇𝒊𝒄 Annualized intermodal facility fixed capital cost ($) 𝑸𝒔𝒃𝒕 The quantity of biomass transported from

supply area to biorefinery (ton)

𝒃𝒌 Biorefinery capacity (gallon) 𝑸𝒊𝒃𝒎𝒕 The quantity of biomass transported from

intermodal facility to biorefinery by transport

mode m during time t (ton)

𝒇𝒃𝒄 Annualized biorefinery fixed capital cost ($) 𝑸𝒃𝒅𝒕 The quantity of biofuel transported from

biorefinery to demand point (gallon)

𝒄𝒊 Annual rental cost of marginal land in i ($/ha) 𝑸𝒌𝒕 The quantity of biofuel produced at

biorefinery (gallon)

𝒗𝒊 Cultivation cost of switchgrass ($/ha)


CASE STUDY

The model proposed above is applied to a case study of switchgrass-based bioethanol supply chain

in North Dakota in order to validate our model, and in response to state policy that promotes the

use of multimodal transportation in delivering switchgrass-based alternative transportation fuel.

North Dakota is an ideally suited region for commercial cultivation of switchgrass with a lot of

potential for use of switchgrass-based bioethanol in the future (3).

FIGURE 2 Intermodal storage, biorefinery candidates in North Dakota. (Note: Intermodal

facility #1: Fairmount, #2: Williston, #3: Tioga, #4: Minot, #5: Bowbells, #6: Devils lake, #7: Grand forks, #8:

Dickinson, #9: Bismark, #10: Carrington, #11: Valley city, #12: Casselton, #13: Fargo, #14: Hankinson, and #15:

Enderlin; Refinery candidates ‘R’ stand for Red Trail Energy, ‘B’ for Blue Flint Ethanol, ‘D’ for Dakota Spirit, ‘T’

for Tharaldson Ethanol, and ‘G’ for Guardian Hankinson)

Harvesting area

All 53 counties in North Dakota are considered to be in support of switchgrass. Switchgrass yield

rate is assumed to be a linear function of the North Dakota annual rainfall, which can be used to

estimate the amount of switchgrass supplied from supply zone (3). Two types of bale including


square and round are considered. Harvesting areas of switchgrass are defined using county

boundaries in ArcGIS platform. To integrate feedstock data onto transportation network data, it

is assumed that the centroid of each county’s polygon is feedstock supply area, which is auto-

generated and identified in ArcGIS map. The associated feedstock parameters including marginal

rental cost (varies by county) (29), cultivation cost ($85.0/ton) (30), harvesting cost (round bale

=$48.2/ha, square bale =$27.9/ha) (31), marginal land availability for each county (29) are

collected.

Intermodal storage

There is only one intermodal facility used for freight transportation in North Dakota. With

increasing agricultural demand and oil delivery, more intermodal option may enhance traffic and

customer service for agricultural and energy industry. Fifteen intermodal facility candidates (#1

~ #15; including existing intermodal facility at Minot) were selected based on North Dakota

strategic freight analysis report from Upper Great Plains Transportation Institute (UGPTI) (32).

Bale storages with tarp system are located at yard where both railway and highway are available

using ArcGIS. The capacity of storage is set as 125,000 tons regardless of locations (17). The

storage cost is set at $21.7/ton, which includes any expense incurred to maintain inventory and

storage (31). Dry matter loss for both types of bale is assumed to be 2% (33). Fixed intermodal

facility capital investment cost is set at $470,597 (32).

Biorefinery

Currently, five corn-based biorefineries including Blue Flint Ethanol (65 MGY), Dakota Spirit (70

MGY), Guardian Hankinson (132 MGY), Red Trail Energy (50 MGY), and Tharaldson Ethanol

(153 MGY) are available in North Dakota, which are presented in Fig 2. It is assumed that with

advanced biofuel conversion technology, multiple types of feedstock could be converted to

bioethanol at refineries. Therefore, these five biorefineries are used as switch-grass based

bioethanol production candidates in this study. A conversion factor of 85 gallons of bioethanol per

ton of biomass is used (34). The capital cost of biorefinery consists of fixed and variable capital

cost (10). Each biorefinery has a different fixed cost and the variable cost is proportional to size

of refinery (12). To determine the fixed capital cost for each biorefinery, cost scaling factor of 1.6

was multiplied by the size of biorefinery (35). Therefore, a medium level of annualized fixed

capital cost is interpolated. The fixed capital cost is $27 million for 65 MGY biorefinery, $28

million for 70 MGY biorefinery, $42.8 million for 132 MGY biorefinery, $22 million for 50 MGY

biorefinery, and $46.8 million for 153 MGY biorefinery. The variable cost is 0.64/gallon

regardless of the fixed capital cost variation (36).

Transportation data

The multimodal transportation network is presented in Fig 2. In this study, transportation networks

including local, rural, urban roads and highways, and railways are considered. It is assumed that

the centroid of each harvesting area is the origin of biomass supply chain. The longitudes and

latitudes of intermodal facility and biorefinery are identified in ArcGIS. The shortest path based

on Dijkstra’s algorithm from origins to destinations are calculated using the OD cost matrix

application in ArcGIS network analysis. In terms of cost associated with truck and railway,

loading/unloading cost of truck is $5/ton (20) and the variable mileage cost is $0.1/ton-mile for

round bale and $0.12/ton-mile for square bale (31). Rail variable mileage cost is $0.02/ton-mile

(26) and fixed cost of rail is $ 6.54/ton (17).


Bioethanol demand

Cities that provide E85 ethanol are considered to be demand centers, and 18 cities were chosen in

this study. Fig.2 shows the geographic distribution of these eighteen cities. The total annual

bioethanol demand is set at 30 MGY, according to official portal for North Dakota State

Government (37).

RESULTS AND DISCUSSION

The optimal system results and comparison with single mode

The minimum cost strategy to integrate multimodal transportation model into switchgrass-based

supply chain suggests that four intermodal facilities (#3, 4, 6 and 14) are required and two

biorefineries (D and R) should be selected. The optimized total cost for the supply chain is $237

million. The total system cost breakdown is presented in Fig 3. It is found that cultivation cost has

the highest contribution, accounting for 36.56% of total cost, followed by production cost,

accounting for 19.48%. The optimal assignment and its flow pattern of biomass to intermodal

storages and biorefineries are analyzed in Table 2. From intermodal storage to biorefinery, only

rail transport mode is used because rail haulage cost is cheaper for long distance shipment. Truck

is used from harvesting area to intermodal facility and from biorefineries to demand center,

because it is the only possible mode for some segments which originate from harvesting area or

end at markets (20).

In order to compare multimodal solutions and single mode solutions, the model was re-run

by eliminating rail transport mode from the model. The comparison made in Table 3 shows that

single mode solution is about $76 million more expensive than multimodal solution. The

multimodal solution indicates that the optimal bioethanol delivered cost is $1.904/gallon, which is

cheaper than single model solution ($2.663/gallon).

FIGURE 3 Total cost breakdown for switchgrass-based multimodal bioethanol supply

chain.


TABLE 2 Optimal Assignment of Biomass Flow to Intermodal Storages and Biorefineries

Counties assigned to intermodal storages Note

Tioga (#3) Burke, Divide, Mountrail, Williams

Minot (#4) Bottineau, McHenry, Pierce, Renville, Rolette, Sheridan, Ward

Devils Lake (#6) Cavalier, Pembina, Ramsey, Towner, Walsh, Rolette

Hankinson (#14) Richland, Sargent

N/A

Counties assigned to biorefineries

D Barnes, Cass, Dickey, Eddy, Foster, Grand Forks, Griggs, Kidder,

LaMoure, Logan, McIntosh, Nelson, Ransom, Steele, Stutsman, Traill, Wells

RAdams, Billings, Bowman, Burleigh, Dunn, Emmons, Golden Valley, Grant,

Hettinger, McKenzie, McLean, Mercer, Morton, Oliver, Sioux, Slope, Stark

N/A

Intermodal Storage assigned to biorefineries

D Hankinson (#14)

R Tioga (#3), Minot (#4), Devils Lake (#6), Hankinson (#14)

Rail is the only mode

to ship biomass from

intermodal storage to

biorefinery

TABLE 3 Cost Comparison for Single mode and Multi-mode

Cost breakdown Single mode Multi-mode

Transportation cost $ 107,592,261.18 $ 82,601,443.23

Bioethanol delivered cost $ 2.663 $ 1.904

Total supply chain cost $ 313,716,741.38 $ 237,253,908.70

Sensitivity analysis

This section discusses results from several sensitivity analyses and analyzed the factors that are

significant to the switchgrass-based multimodal bioethanol supply chain. Sensitivity analysis for

key inputs include conversion rate of switchgrass feedstock to bioethanol, biomass feedstock

availability, different levels of bioethanol demand, and all the unit cost factors - marginal rental

cost, cultivation cost, harvesting cost, transportation cost, storage cost, bioethanol production cost,

and capital investment cost.

Influence of biomass availability and conversion rate on bioethanol cost and location

As a baseline case, the conversion rate is 85 gallons/ton and 13 million tons are available from 53

counties in North Dakota (3). It is assumed that the conversion rate decreases from 85 gallons/ton

to 55 gallons/ton in increments of 5 gallons/ton, and the availability of switchgrass increases by a

total of 5% from current volume of availability. Table 4 presents the change of bioethanol delivered

cost and location decision of intermodal facility and biorefineries by biomass availability and

conversion rate. The highest bioethanol cost is $2.977/gallon with the baseline biomass

availability and the conversion rate of 55 gallons/ton, and the lowest bioethanol cost is $1.850 with

5% increase in biomass availability and conversion rate of 80 gallons/ton. From the results, it was


found that bioethanol cost increases dramatically at the lowest biomass availability and at the

lowest conversion rate, which is because the long haulage shipment occurs at low biomass

availability (17). Intermodal storage location and biorefineries location change over different

biomass availabilities and conversion rates. As can be seen from the results, intermodal storage

location #3 and #6 show up as optimal candidate locations in most case scenarios in addition to

#4, which is currently operating in North Dakota. A higher biomass availability and more

conversion rate result in more requirements for intermodal facility to be opened. In terms of

biorefinery location, most of the case scenarios show that biorefineries ‘D’ and ‘R’ are candidate

sites which should increase their capacity level to handle multiple types of feedstock to convert

biomass into bioethanol and minimize total cost.

TABLE 4 Bioethanol Cost and Location Decision by Biomass Availability and Conversion

Rate

Conversion rate (gallon/ton)

Biomass availability (%) Bioethanol delivered cost ($/gallon)

Baseline 80 75 70 65 60 55

Baseline 1.904 2.016 2.352 2.346 2.520 2.524 2.951

1% 1.885 2.001 2.373 2.355 2.540 2.712 2.931

2% 1.863 1.987 2.384 2.339 2.489 2.729 2.912

3% 1.862 1.972 2.100 2.359 2.505 2.674 2.977

4% 1.863 1.958 2.040 2.353 2.465 2.664 2.873

5% 1.856 1.850 1.970 2.388 2.448 2.637 2.855

Biomass availability (%) Intermodal facility location

Baseline 80 75 70 65 60 55

Baseline 3,4,6,14 3,4,6 3,4,6 3,4,6 3,4,6 3,6,8 3,6,8

1% 3,4,6,14 3,4,6 3,4,6 3,4,6 3,4,8 3,4,6,8 3,6,8

2% 3,4,6,10,15 3,4,6 3,4,6 3,4,6 3,4,6 3,4,6 3,4,6

3% 3,4,6,14 4,6,7 3,4,6 3,4,6 3,4,6 3,4,6,8 3,4,6

4% 3,4,6,10,11,14 3,4,6,14 3,4,6,14 3,4,6 3,4,6 3,4,6 3,6

5% 3,4,6,14 3,4,6,14 3,4,6 3,4,6 3,4,6 3,4,6 3,6,8

Biomass availability (%) Refinery location

Baseline 80 75 70 65 60 55

Baseline D,R D,R D,R D,R D,R B,T B,T

1% D,R D,R D,R D,R D,R B,T B,T

2% D,R D,R D,R D,R D,R D,R B,T

3% D,R D,R D,R D,R D,R B,T D,R



Bioethanol delivered cost change over different bioethanol demand and conversion rate

In addition to the bioethanol delivered cost change with different scenario analyses of biomass

availability and conversion rates, the bioethanol delivered cost with different annual level of


bioethanol demand (MGY) versus conversion rate (gallons/ton) are investigated. It is assumed that

bioethanol demand increases from current annual level of demand of 30MGY up to 45MGY (a

total of 50% increase). Fig 4 presents the resulting bioethanol delivered cost changes by bioethanol

demand and conversion rate. When bioethanol demand is fixed, the bioethanol delivered cost

increases with the increase in conversion rate (Fig 4. a). When the conversion rate remains the

same, the delivered cost of bioethanol also increases, meaning that higher bioethanol delivered

cost occurs with the increasing demand of bioethanol and decreasing conversion rate (Fig 4.b).

The experimental results from Tables 4 and 5, and Fig 4 indicates that relationship between

both biomass availability and conversion rate, and bioethanol demand and conversion rate are

major factors affecting the bioethanol delivered cost. Higher biomass availability means that the

intermodal storage and biorefinery would be supplied from harvesting areas that are close by,

therefore incurs lower shipment cost, resulting in lower unit cost of bioethanol (17). Lower

conversion rate with higher demand implies higher bioethanol production cost, which would

increase transport cost and unit cost of bioethanol.

FIGURE 4 Bioethanol delivered cost by bioethanol demand change and conversion rate.

Influence of different unit cost factors on bioethanol cost

In order to find the most influential unit cost factors on bioethanol delivered cost, each unit cost

was increased or decreased by 10% for sensitivity analysis to investigate the overall switchgrass-

based bioethanol multimodal supply chain system, as presented in Table 5. The results show that

bioethanol cost is not dependent on rental cost, cultivation cost, or harvesting cost. The most

influential unit cost factor is truck transportation cost for biomass, which accounts for 1.42%

increase and 5.62% decrease of optimal value of bioethanol cost ($1.904/gallon). The second most

influential factor is rail transportation cost, accounting for 0.95% increase and 3.53% decrease.

Capital investment cost and bioethanol production cost are third and fourth influential factors.


TABLE 5 Sensitivity Analysis for Delivered Bioethanol Cost

Unit cost factor (10% increase or decrease) Bioethanol cost ($/gallon) Percentage change (%)

Rental cost +10% 0 0

- 10% 0 0

Cultivation cost +10% 0 0

- 10% 0 0

Harvesting cost +10% 0 0

- 10% 0 0

Truck transportation cost +10% 1.931 1.42%

- 10% 1.797 -5.62%

Rail transportation cost +10% 1.922 0.95%

- 10% 1.837 -3.52%

Storage cost +10% 1.899 -0.26%

- 10% 1.894 -0.53%

Production cost +10% 1.917 0.68%

- 10% 1.906 0.11%

Capital cost +10% 1.920 0.84%

- 10% 1.841 -3.31%

SUMMARY AND CONCLUSION

This study presented a Mixed-integer linear programming (MILP) model for integrating

multimodal transport (truck and rail) into switchgrass-based bioethanol supply chain design

(MTSBSC). The model was applied to case study of North Dakota. This research demonstrates

how proposed model can be adopted to make strategic and tactical decision for bioethanol supply

chain. Experimental results indicate that multimodal solution is more cost effective than single

mode solution in terms of total system cost and bioethanol delivered cost. Also, there is interaction

between bioethanol conversion rate and biomass availability as well as conversion rate and

bioethanol demand in supply chain decision for biorefinery and intermodal storage. Higher

biomass availability results in lower unit cost of bioethanol. On the other hand, higher demand of

bioethanol increases bioethanol cost. From the sensitivity analysis, transportation costs are the

most influential factor on bioethanol delivered cost followed by capital investment and production

cost. The storage cost shows no impact on bioethanol cost. This impact should be identified in

future research by considering biomass inventory change over time. This study optimizes

MTSBSC using single objective of economic performance. The current study can be extended by

considering multiple objectives that incorporates environmental impact of MTSBSC.


REFERENCES

1. U.S. Energy Information Administration (EIA), Petroleum & Other Liquids,

http://www.eia.gov/petroleum/data.cfm. Accessed Nov. 3, 2016.

2. U.S. Energy Information Administration (EIA), Energy Use for Transportation,

http://www.eia.gov/Energyexplained/?page=us_energy_transportation. Accessed Nov. 3, 2016.

3. Zhang, J., Osmani, A., Awudu, I., and Gonela, V. An Integrated Optimization Model for

Switchgrass-based Bioethanol Supply Chain. Applied Energy, Vol. 102, 2013, pp. 1205–1217.

4. Sokhansanj, S., Mani, S., Turhollow, S., Kumar, A., Bransby, D., Lynd, L., and Laser, M.

Large-scale Production, Harvest and Logistics of Switchgrass (Panicum virgatum L.) - Current

Technology and Envisioning a Mature Technology. Biofuels, Bioproducts and Biorefining, Vol.

3, No. 2, 2009, pp. 124–141.

5. Cundiff, J. S., Dias, N., and Sherali, H. D. A Linear Programming Approach for Designing a

Herbaceous Biomass Delivery System. Bioresource Technology, Vol. 59, No.1, 1997, pp. 47–55.

6. Akgul, O., Zamboni, A., Bezzo, F., Shah, N., and Papageorgiou, L. G. Optimization based

Approaches for Bioethanol Supply Chains. Industrial & Engineering Chemistry Research, Vol.

50, No. 9, 2010, pp. 4927–4938.

7. Alex Marvin, W., Schmidt, L. D., Benjaafar, S., Tiffany, D. G., and Daoutidis, P. Economic

Optimization of a Lignocellulosic Biomass-to-Ethanol Supply Chain. Chemical Engineering

Science, Vol. 67, No.1, 2012, pp. 68–79.

8. Mansoornejad, B., Chambost, V., and Stuart, P. Integrating Product Portfolio Design and

Supply Chain Design for the Forest Biorefinery. Computers and Chemical Engineering, Vol. 34,

No. 9, 2010, pp. 1497–1506.

9. Dal Mas, M., Giarola, S., Zamboni, A., and Bezzo, F. Capacity Planning and Financial

Optimization of the Bioethanol Supply Chain under Price Uncertainty. Computer Aided

Chemical Engineering, Vol. 28, 2010, pp. 97–102.

10. Huang, Y., Chen, C. W., and Fan, Y. Multistage Optimization of the Supply Chains of

Biofuels. Transportation Research Part E: Logistics and Transportation Review, Vol. 46, No. 6,

2010, pp. 820–830.

11. Walther, G., Schatka, A., and Spengler, T. S. Design of Regional Production Networks for

Second Generation Synthetic Bio-fuel - A Case Study in Northern Germany. European Journal

of Operational Research, Vol. 218, No. 1, 2012, pp. 280–292.

12. Parker, N., Tittmann, P., Hart, Q., Nelson, R., Skog, K., Schmidt, A., and Jenkins, B.

Development of a Biorefinery Optimized Biofuel Supply Curve for the Western United States.

Biomass and Bioenergy, Vol. 34, No.11, 2010, pp. 1597–1607.

http://www.eia.gov/petroleum/data.cfm

http://www.eia.gov/Energyexplained/?page=us_energy_transportation


13. Osmani, A and Zhang, J. Stochastic Optimization of a Multi-feedstock Lignocellulosic-

based Bioethanol Supply Chain under Multiple Uncertainties. Energy, Vol. 59, 2013, pp. 157–

172.

14. Dal-Mas, M., Giarola, S., Zamboni, A., and Bezzo, F. Strategic Design and Investment

Capacity Planning of the Ethanol Supply Chain under Price Uncertainty. Biomass and

Bioenergy, Vol. 35, No.5, 2011, pp. 2059–2071.

15. You, F., Graziano, D. J., and Snyder, S. W. Optimal Design of Sustainable Cellulosic Biofuel

Supply Chains : Multiobjective Optimization Coupled with Life Cycle Assessment and Input –

Output Analysis. AIChE Journal, Vol. 58, No.4, 2012, pp. 1157–1180.

16. Gebreslassie, B.H, Slivinsky, M, Wang, B, and You, F. Life Cycle Optimization for

Sustainable Design and Operations of Hydrocarbon Biorefinery via Fast Pyrolysis, Hydrotreating

and Hydrocracking. Computers & Chemical Engineering. Vol. 50, 2013, pp. 71–91.

17. Zhang, F., Johnson, D. M., and Wang, J. Integrating Multimodal Transport into Forest-

delivered Biofuel Supply Chain Design. Renewable Energy, Vol. 93, 2016, pp. 58–67.

18. Sarder, M., Adnan, Z., and Miller, C. Biomass Transportation Model for Intermodal

Network. Journal of Supply Chain Management, Vol. 2, No. 2, 2013, pp. 7–18.

19. Rodrigue, J.-P., Comtois, C., and Slack, B. The Geography of Transport Systems. Routledge.,

New York, 2013.

20. Xie, F., Huang, Y., and Eksioglu, S. Integrating Multimodal Transport into Cellulosic

Biofuel Supply Chain Design under Feedstock Seasonality with a Case Study based on

California. Bioresource Technology, Vol. 152, 2014, pp. 15–23.

21. U.S. Department of Energy (DOE). Ethanol Production and Distribution.

http://www.afdc.energy.gov/fuels/ethanol_production.html. Accessed November 3, 2016

22. Akgul, O., Shah, N., and Papageorgiou, L. G. Economic Optimisation of a UK Advanced

Biofuel Supply Chain. Biomass and Bioenergy, Vol. 41, 2012, pp. 57–72.

23. Hajibabai, L., and Ouyang, Y. Integrated Planning of Supply Chain Networks and

Multimodal Transportation Infrastructure Expansion: Model Development and Application to the

Biofuel Industry. Computer-Aided Civil and Infrastructure Engineering, Vol. 28, No.4, 2013, pp.

247–259.

24. Mahmudi, H., and Flynn, P. C. Rail vs Truck Transport of Biomass. Applied Biochemistry

and Biotechnology, Vol. 129, No.1-3, 2006, pp. 88–103.

25. Ekşioglu, S. D., Zhang, S., and Li, S. Analyzing the Impact of Intermodal Facilities to the

Design of Supply Chains for Biorefineries, In Transportation Research Record: Journal of the

http://www.afdc.energy.gov/fuels/ethanol_production.html


Transportation Research Board, No. 2191, Transportation Research Board of the National

Academies, Washington, D.C., 2011, pp. 114-151.

26. Morrow, W. R., Griffin, W. M., and Matthews, H. S. Modeling Switchgrass derived

Cellulosic Ethanol Distribution in the United States. Environmental Science and Technology,

Vol. 40, No. 9, 2006, pp. 2877–2886.

27. Zhu, X., and Yao, Q. Logistics System Design for Biomass-to-bioenergy Industry with

Multiple Types of Feedstocks. Bioresource Technology, Vol. 102, No. 23, 2011, pp. 10936–

10945

28. An, H., Wilhelm, W. E., and Searcy, S. W. A Mathematical Model to Design a

Lignocellulosic Biofuel Supply Chain System with a Case Study based on a Region in Central

Texas. Bioresource Technology, Vol. 102, No. 17, 2011, pp. 7860–7870.

29. U.S. Department of Agriculture (USDA). 2012 Census Publications.

https://www.agcensus.usda.gov/Publications/2012/Full_Report/Census_by_State/North_Dakota/i

ndex.asp. Assessed Aug. 11, 2016.

30. Jamie L. Estimated Production Cost Budgets for Biomass : Switchgrass,‘Highlander’

Eastern Gamagrass , Indiangrass , and Big Bluestem. U.S. Whitten Plant Materials Center. Plant

Materials Technical Note No. 102, Department of Agriculture, 2007.

31. Larson, J. a., Yu, T.-H., English, B. C., Mooney, D. F., and Wang, C. Cost Evaluation of

Alternative Switchgrass Producing, Harvesting, Storing, and Transporting Systems and Their

Logistics in the Southeastern USA. Agricultural Finance Review, Vol. 70, No. 2, 2010, pp. 184–

200.

32. Berwick, M., Bitzan, J., Chi, J., and Lofgren, M. North Dakota Strategic Freight Analysis:

The Role of Intermodal Container Transportation in North Dakota. Upper Great Plains

Transportation Institute, 2002.

33. Shinners, K. J., and Binversie, B. N. Fractional Yield and Moisture of Corn Stover Biomass

Produced in the Northern US Corn Belt. Biomass and Bioenergy, Vol. 31, No. 8, 2007, pp. 576–

584.

34. Stuart, Ben J. Liquid Transportation Fuels from Coal and Biomass: Technological Status,

Costs, and Environmental Impacts. The Ohio Journal of Science, Vol 108, No. 5, 2008, pp. 114-

116.

35. Wallace, R., Ibsen, K., AcAllon, A., and Yee, W. Feasibility Study for Co-locating and

Integrating Ethanol Production Plants from Corn Starch and Lignocellulosic Feedstocks. No.

NREL/TP-510-37092, National Renewable Energy Lab, 2005.

https://www.agcensus.usda.gov/Publications/2012/Full_Report/Census_by_State/North_Dakota/index.asp

https://www.agcensus.usda.gov/Publications/2012/Full_Report/Census_by_State/North_Dakota/index.asp


36. Haque, M., & Epplin, F. M. Cost to produce switchgrass and cost to produce ethanol from

switchgrass for several levels of biorefinery investment cost and biomass to ethanol conversion

rates. Biomass and Bioenergy, Vol. 46, 2012, pp. 517–530.

37. North Dakota State Government (ND.gov). Biomass/Biofuels.

http://ndstudies.gov/energy/level1/module-5-biofuels-geothermal-recovered/biomassbiofuels.

Assessed Aug. 9, 2016.

http://ndstudies.gov/energy/level1/module-5-biofuels-geothermal-recovered/biomassbiofuels

Documents

An integrated multimodal transportation model for ...docs.trb.org/prp/17-02769.pdf · An integrated multimodal transportation model for ... examined the cost comparison between truck