Biomass-To-Bioenergy and Biofuel Supply Chain Optimization

Home , Bioenergy, Biomass to liquid

Computers and Chemical Engineering 66 (2014) 36–56

Contents lists available at ScienceDirect

Computers and Chemical Engineering

j ournal homepage: www.elsevier.com/locate/compchemeng

Biomass-to-bioenergy and biofuel supply chain optimization:

Overview, key issues and challenges

a a,∗ b

Dajun Yue , Fengqi You , Seth W. Snyder

Northwestern University, Evanston, IL 60208, USA

Argonne National Laboratory, Argonne, IL 60439, USA

a r a

t b

i c l e i n f o s t r a c t

Article history: This article describes the key challenges and opportunities in modeling and optimization of biomass-to-

Received 19 July 2013

bioenergy supply chains. It reviews the major energy pathways from terrestrial and aquatic biomass to

Received in revised form

bioenergy/biofuel products as well as power and heat with an emphasis on “drop-in” liquid hydrocarbon

14 November 2013

fuels. Key components of the bioenergy supply chains are then presented, along with a comprehensive

Accepted 17 November 2013

overview and classiﬁcation of the existing contributions on biofuel/bioenergy supply chain optimization.

Available online 18 December 2013

This paper identiﬁes fertile avenues for future research that focuses on multi-scale modeling and opti-

mization, which allows the integration across spatial scales from unit operations to bioreﬁnery processes

Keywords:

and to biofuel value chains, as well as across temporal scales from operational level to strategic level.

Supply chain modeling

Biofuels Perspectives on future biofuel supply chains that integrate with petroleum reﬁnery supply chains and/or

Bioenergy carbon capture and sequestration systems are presented. Issues on modeling of sustainability and the

Mathematical programming treatment of uncertainties in bioenergy supply chain optimization are also discussed.

1. Introduction production capacity of advanced biofuels is less than 1 billion gal-

lon worldwide. It is foreseeable that the bioenergy industry will be

Biomass-derived liquid transportation fuels and energy prod- undergoing a rapid expansion in the coming decade. Many sustain-

ucts have been proposed as part of the solution to climate change able and robust biomass-to-bioenergy supply chains, which link the

and our heavy dependence on fossil fuels, because the biomass sustainable biomass feedstock and the ﬁnal fuel/energy products,

feedstock can be produced renewably from a variety of domestic need to be designed and developed for lower costs, less environ-

sources, and the production and use of bioenergy/biofuel prod- mental impacts and more social beneﬁts (Elia, Baliban, & Floudas,

ucts have potentially lower environmental impacts than their 2012; Hosseini & Shah, 2011; Sharma, Ingalls, Jones, & Khanchi,

petroleum counterparts (An, Wilhelm, & Searcy, 2011a; Gold & 2013). To accelerate the transition towards the large-scale and sus-

Seuring, 2011; Marquardt et al., 2010). Consequently, many coun- tainable production and use of biofuels and bioenergy products,

tries have set national biofuels targets and provide incentives and a signiﬁcant challenge in research is to systematically design and

supports to accelerate the growth of bioenergy industry. For exam- optimize the entire bioenergy supply chains from the biomass feed-

ple, in the U.S. the Renewable Fuels Standard (RFS), part of the stock production to the biofuel/bioenergy end-use across multiple

Energy Independence and Security Act (EISA) of 2007, establishes spatial scales, from unit operations to bioreﬁnery processes and to

an annual production target of 36 billion gallons of biofuels by the entire value chain, as well as across multiple temporal scales,

2022 (EISA, 2007). Currently most biofuels in the U.S. are made from strategic to operational levels, in a cost-effective, robust and

from corn starch that might have negative implications in terms sustainable manner (Daoutidis, Marvin, Rangarajan, & Torres, 2013;

of both food prices and production (McNew & Grifﬁth, 2005; Marquardt et al., 2010). It is the objective of this paper to iden-

Rajagapol, Sexton, Hochman, Roland-Holst, & Zilberman, 2009). tify the key research challenges and opportunities in modeling and

To avoid adverse impacts on food supply, EISA further speciﬁes optimization of biomass-to-bioenergy supply chains using Process

that 16 out of the 36 billion gallons of renewable fuels produced Systems Engineering (PSE) tools and methods, and to chart a path

in 2022 should be advanced biofuels made from non-starch feed- for addressing these challenges.

stocks, such as cellulosic or algal biomass. Yet, the current annual In this paper, we classify the existing contributions on bioenergy

supply chain optimization, elucidate the research challenges, and

demonstrate that multi-scale modeling and optimization approach

∗ can play a leading role to address these challenges. We ﬁrst review

Corresponding author. Tel.: +1 847 467 2943; fax: +1 847 491 3728.

E-mail address: [email protected] (F. You). the major biomass-to-biofuels pathways, with a speciﬁc emphasis

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 37

on producing “drop-in” liquid hydrocarbon fuels from terrestrial The second group of terrestrial biomass feedstocks, the cellu-

and aquatic biomass, which may blend with or displace gasoline, losic biomass, can avoid adverse impacts on food supply, because

diesel, jet fuel, kerosene or home heating oil. We then describe they are non-starch, non-edible and non-food feedstocks. Cellulosic

the key components of the biomass-to-biofuels supply chains and biomass feedstocks can be obtained from a number of sources, such

their major characteristics, along with a comprehensive overview as agricultural residues, forest residues and energy crops. Agricul-

and classiﬁcation of the existing contributions on bioenergy sup- tural residues are typically plant parts left in the ﬁeld after harvest

ply chain optimization. We further demonstrate the important role (e.g., corn stover), as well as the secondary residues like manure and

of multi-scale modeling and optimization, which allows the inte- food processing wastes. Forest residues are leftover wood or plant

gration across multiple temporal and spatial scales. The potential material from logging operations, forest management, and land-

research vistas on future bioenergy supply chains that integrate clearing, as well as secondary residues like mill wastes. Dedicated

with petroleum reﬁnery supply chains and/or carbon capture and energy crops (e.g., poplar, switchgrass, Miscanthus) are typically

sequestration systems are presented next. Key issues on modeling fast-growing trees and perennial grasses speciﬁcally grown for

of sustainability and the treatment of uncertainties in bioenergy energy uses.

supply chain optimization are also discussed. Although the focus

of this paper is on biofuels supply chains, at the end of this paper 2.1.2. Aquatic feedstocks

we discuss the modeling and optimization for supply chains with Aquatic biomass includes a diverse group of primarily

non-fuel end-uses of biomass derivatives, including biopower and aquatic, photosynthetic algae and cyanobacteria ranging from the

biomass-to-heat supply chains and biomass-to-chemicals supply microscopic (microalgae and cyanobacteria) to large seaweeds

chains, which represent other products from biomass and will (macroalgae) (DOE/EERE, 2010b).

impact the supply chain. Many macroalgae, microalgae, and cyanobacteria carry out

photosynthesis to drive rapid biomass growth. Certain strains of

2. Biomass-to-biofuels pathways microalgae make extremely efﬁcient use of light and nutrients. As

they do not have to produce structural compounds such as cellulose

Biomass is unique among renewable energy sources in that it for leaves, stems, or roots, and because they can be grown ﬂoating

can be easily stored until needed and provides a liquid fuel alter- in a rich nutritional medium, microalgae can have faster growth

native for use in today’s transportation system. In the short term, rates than terrestrial crops. In some cases, algae growth rates can

biofuels are the only renewable resources that can address the be an order of magnitude faster than those of terrestrial crop plants.

transportation sector’s heavy dependence on foreign oil without As algae have a harvesting cycle of 1–10 days, their cultivation per-

replacing the vehicle ﬂeet (DOE/EERE, 2013d). In this section, we mits several harvests in a very short time-frame, a strategy differing

review the major pathways from biomass to biofuels, and discuss from that associated with yearly crops. Algae naturally remove and

the main properties of the feedstocks, fuel products and interme- recycle nutrients (e.g., nitrogen and phosphorous) from water and

diates. Fig. 1 shows a number of major conversion pathways from wastewater, and can be used to sequester carbon dioxide from the

terrestrial and aquatic biomass to intermediates and to final biofuel flue gases emitted from fossil fuel-fired power plants. Adding to the

products (DOE/EERE, 2010b, 2011a, 2011b). appeal, some strains of microalgae can grow on land unsuitable for

other established crops, for instance: arid land, land with exces-

2.1. Terrestrial and aquatic biomass feedstocks sively saline soil, and drought-stricken land. This minimizes the

competition on arable land with the cultivation of food crops. More

Biomass is a renewable resource that acquiries carbon by carbon importantly, algae biomass can contain high levels of oil (lipids or

dioxide ﬁxation during their growing cycles. In comparison to fossil triacylglycerides), making it a promising feedstock for the produc-

fuels such as natural gas and coal, which take millions of years to tion of renewable gasoline, diesel, and jet fuel. Algae can produce

form, biomass is easy to grow, collect, utilize and replace quickly one or even two orders of magnitude more oil per unit area than

without depleting natural resources. conventional crops such as rapeseed, palms, soybeans, or jatropha

(DOE/EERE, 2010b; Hasan & Chakrabarti, 2009).

2.1.1. Terrestrial feedstocks The cultivation of macroalgae (e.g., seaweed) is similar to that of

Many types of abundantly available biomass on the land terrestrial plants but in aquatic environment. In practice, macroal-

can be utilized as feedstocks for biofuel production. The 2011 gae are cultivated in large offshore farms, near-shore coastal zones,

report released by U.S. Department of Energy (DOE) and Oak or open pond facilities. In contrast, depending on the type of

Ridge National Laboratory (ORNL) titled “U.S. Billion-Ton Update: microalgae and cyanobacteria, there are various cultivation meth-

Biomass Supply for a Bioenergy and Bioproducts Industry”, details ods, varying both in their advantages and challenges. Broadly

biomass feedstock potential throughout the contiguous United speaking, algae can be cultivated via photoautotrophic or het-

States at a county level of resolution (DOE/EERE, 2011c). The report erotrophic methods. The former requires light and carbon dioxide

examines the U.S. capacity to sustainably produce over 1.6 billion for photosynthesis while the latter requires suitable feedstocks

dry tons of terrestrial biomass annually for conversion to bioen- such as lignocellulosic sugars for growth. The cultivation system

ergy and bioproducts, while continuing to meet existing demand for microalgae and cyanobacteria is either closed bioreactor or open

for food, feed, and ﬁber. The report estimates that the United States pond. Closed bioreactor yields less loss of water and better biomass

could potentially produce approximately 85 billion gallons of biofu- quality, but has scalability problems. Open pond requires lower

els annually—enough to replace approximately 30% of the nation’s capital costs and take advantage of evaporative cooling, but it is

current petroleum consumption. subject to weather and condition changes and inherently difﬁcult

The terrestrial biomass feedstock can be generally categorized to avoid contamination and foreign algae (DOE/EERE, 2010b).

into two groups. The ﬁrst group includes corn grain, sugarcane,

soy bean, oil seed, etc. These feedstocks are rich in sugar or lipids, 2.2. Types and major properties of biofuels

and have high yields after converted into bioethanol or biodiesel.

Currently, most biofuels are made from these feedstocks, due to 2.2.1. Ethanol and other alcohols

the maturity in technologies and lower unit production cost. How- As the ﬁrst generation biofuel technology, biomass-derived

ever, the use of these feedstocks for biofuel production might have ethanol and other alcohols were largely produced from starch-

implications both in terms of world food prices and production. or sugar-based biomass, such as corn and sugarcane. Due to the

38 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Fig. 1. Selected conversion pathways from terrestrial and aquatic biomass to intermediates and to ﬁnal biofuel products.

increasing debate on its adverse impact on global food prices, the Furthermore, the polarity resulting from the high oxygen content

research focus on alcohol-based biofuel is shifted to employing cel- in FAME can cause it to cling to pipe or vessel walls, thus posing

lulosic biomass as feedstock options, which is considered second potential contamination of subsequent batches if transported in

generation biofuel technology. pipeline. Therefore, FAME is also expressly prohibited by most large

Ethanol is extensively distributed and utilized as a fuel blend U.S. pipeline operators (Bunting et al., 2010). In addition, vehicle

component throughout the U.S. Its low freezing point makes it suit- warranties typically limit FAME blending to 5%.

able for use in cold climates. It can be blended with gasoline in

any combination, and it is currently approved as a 10% blend for 2.2.3. Drop-in biofuels

all vehicles and as an 85% blend for ﬂex-fuel vehicles. The United Compared to the ethanol and bio-diesel, advanced hydrocarbon

States Environmental Protection Agency (EPA) has also approved biofuels have the appealing advantage that they are compatible

a 15% blend for all vehicles from model year 2001 and newer, yet with existing infrastructure. Known as drop-in fuels, hydrocarbon

the availability of which is very limited. There are, however, sev- biofuels can serve as petroleum substitutes in existing reﬁner-

eral distribution-related challenges associated with ethanol. It is ies, tanks, pipelines, pumps, vehicles, and smaller engines. These

completely miscible with water and will separate from a gasoline fuels deliver more energy density than ethanol. Moreover, the con-

mixture if enough water is present either in the pipeline or as water version processes of hydrocarbon biofuels can also yield a wide

bottoms in a storage tank. In addition, numerous studies have cited range of bioproducts that can substitute for specialty petrochem-

ethanol’s role in corrosion or stress corrosion cracking of pipeline icals, helping to replace the whole barrel of oil and enhancing the

walls. Due to all these challenges, many major U.S. pipeline opera- economic and environmental sustainability of bioreﬁneries.

tors expressly prohibit ethanol and ethanol–gasoline mixtures in These renewable gasoline, diesel, and jet fuels are essen-

the pipeline. Therefore, the rail car, barge, and tanker truck are tially identical to their existing petroleum-derived counterparts

the principle transportation modes for ethanol, which makes long- in properties, except that they are derived from domestic cel-

distance transportation cost-prohibitive (Bunting, Bunce, Barone, lulosic biomass sources or algal masses. Advanced hydrocarbon

& Storey, 2010). fuels are believed to be the next generation biofuels, which would

As with ethanol, butanol is also an oxygenated fuel that can be help reduce the carbon footprint and potentially, supply and price

blended with gasoline in any combination and requires only minor volatility in all transportation sectors, relieve the energy security

modiﬁcations for use in existing vehicles. The producers of butanol issues, and create related job opportunities for a better social equal-

claim that butanol is compatible with existing oil pipeline infras- ity (DOE/EERE, 2013d; MASBI, 2013).

tructure, but these claims have yet to be validated. In addition,

butanol has four isomers (molecular arrangements), thus opti- 2.3. Intermediates

mization of these isomers from either a manufacturing or fuel

performance standpoint still deserves further research. Note that 2.3.1. Sugars or carbohydrate derivatives

most fermentation routes produce a single isomer as the dominant Sugars can be easily obtained from starch- or sugar-rich crops

product (Bunting et al., 2010). such as corn and sugarcane. However, to avoid the impact on food

prices, the development focus has shifted to deriving sugars and

2.2.2. Bio-diesel other carbohydrate derivatives from lignocellulosic biomass feed-

The fatty ethyl methyl esters (FAME) is often referred to as stocks. Lignocellulose (mainly lignin, cellulose, and hemicellulose)

bio-diesel, because it is used extensively as a diesel supplement is the primary component of plant residues, woody materials, and

due to its similarity to diesel both physically and chemically and grasses. The cell wall structure of these plant matters is partially

due to its ability to be blended with diesel in any combination. comprised of sugar polymers (carbohydrates). Pretreatment and

Three major issues facing FAME bio-diesel distribution are a higher hydrolysis processes are used to break down the cell wall through

cloud point than diesel, lower stability, and the cleaning effect. introduction of enzymes in addition to heat and other chemicals

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 39

in order to convert the carbohydrate portion of the biomass into pumping work. Its polarity due to high oxygen content causes

an intermediate sugar stream. These sugars can serve as interme- pyrolysis oil more likely to cling to pipe and vessel walls. After all,

diate building blocks that can be further fermented or chemically due to its promising potential to replace the petroleum crude, there

catalyzed into a range of advanced biofuels and value-added chem- is signiﬁcant research on stabilizing and upgrading pyrolysis oils

icals. (Bunting et al., 2010).

One major obstacle to developing a cost-competitive conversion

pathway is the efﬁciency of separating and converting cellulosic 3. General structures and key components of

biomass into sugars. Starch and sugars function as food storage for biomass-to-biofuels supply chains

the plant, while lignocellulose is structural and designed to restrict

rapid metabolism. Due to the recalcitrant nature of the cell wall, Development of the U.S. advanced bioenergy industry will lead

it is often difﬁcult and complicated to completely hydrolyze the to the establishment of many manufacturing supply chains of

carbohydrate components into sugars and usable intermediates, hydrocarbon biofuels, which link the sustainable biomass feed-

making this material expensive to convert to biofuels (DOE/EERE, stocks with the ﬁnal biomass-derived gasoline, diesel and jet fuel

2011a; Melero, Iglesias, & Garcia, 2012). products through the following ﬁve major components: biomass

production system, biomass logistics system, biofuel production

2.3.2. Lipids system, biofuel distribution system and biofuel end-use (Gold &

Lipids can be derived from a variety of sources usually through Seuring, 2011; Iakovou, Karagiannidis, Vlachos, Toka, & Malamakis,

physical extraction processes. The most common approach is from 2010). Although these ﬁve elements are similar to those in a

oil seed plants that are abundant in oil content such as rapeseed petroleum-to-fuel supply chain, they could not be investigated

and soybean in temperate climates and palm in tropical climates. with same systematical framework due to the differences in feed-

Another approach is from animal fats (tallow), which are a byprod- stocks and products, and consequently different strategies for

uct of meat production and cooking. Though still under research, production, transportation and storage (An et al., 2011a; Sharma

the largest potential impact is to extract lipids from algae. It is et al., 2013).

reported that certain strains of algae may have order-of-magnitude

higher yield of lipids compared to the vegetable and animal origin 3.1. Feedstock supply and logistics

sources.

Note that the term lipids refers not to a single chemical but Feedstock production and logistics components of the supply

constitute a group of molecules including monoglycerides, diglyc- chain deliver high-quality, stable, and infrastructure-compatible

erides, triglycerides, phospholipids, fatty acids, and others. The feedstocks from diverse biomass resources to bioreﬁneries. For

lipids derived from various biomass sources may signiﬁcantly dif- terrestrial biomass, a main challenge is to develop an efﬁcient

fer in compositions. These fatty acid esters can be used to produce feedstock supply system for cost-effective and time-sensitive col-

bio-diesel via transesteriﬁcation process or converted to hydrocar- lection, preprocessing, storage, and transport, in order to deliver

bon biofuels via hydrotreating, catalytic cracking and metathesis diverse consistently high-quality terrestrial biomass for conver-

processes (Fahy et al., 2005). It is reported that hydrotreated esters sion.

of fatty acids (HEFA) are one of the two biofuels that have received

American Society for Testing and Materials (ASTM) approval for 3.1.1. Seasonal supply and storage

aviation (MASBI, 2013). One of the critical challenges in the operation of biofuel sup-

ply chains is the seasonal nature and annual variability of biomass

2.3.3. Syngas supply. Most biomass resources are plant matters, which need to be

Syngas derived from biomass usually via gasiﬁcation process is planted, cultivated, and harvested, going through a growing cycle.

a gas mixture composed mainly of CO and H2. Syngas is normally Crop residues are usually collected after the harvest of the agri-

used as a feed to manufacture Fischer–Tropsch fuels, which are the cultural crops. For instance, the corn stover in the U.S. Corn Belt

other ASTM approved aviation fuels (MASBI, 2013). Additionally, is mainly harvested from September through November (USDA,

syngas can also be used as a fuel for heat supply. The hydrogen 2010). The wood residues are grown over multiple years, which

is useful for hydrotreating operations, necessary to upgrade fuels makes them less seasonal compared to crop residues. They are

and to remove impurities. If shipped long distances, syngas would usually available all year round. However, the yields may vary in

require a distribution infrastructure similar to that required by different months. For instance, it might be more difﬁcult to collect

natural gas. However, additional modiﬁcations and maintenance the wood residues during snowing seasons. In addition, it is also

are needed to prevent syngas leaks, as the hydrogen component is reported that the harvesting timing and frequency may affect the

prone to leaks and the carbon monoxide component is highly toxic yields of energy crops, thus careful planning and scheduling would

(Bunting et al., 2010). be necessary in order to guarantee the quantity and quality of the

biomass supply (Sokhansanj et al., 2009).

2.3.4. Bio-oil The biomass supply may be discrete due to the seasonality, but

Bio-oil, also known as pyrolysis oil, is a promising bio-crude pro- the demand for transportation fuels is all-year-round. Therefore,

duced from biomass feedstocks undergoing pyrolysis processes, the resulting operational challenge is to manage the biomass stor-

which can be further upgraded via hydroprocessing to produce age in order to maintain a continuous supply for the production at

hydrocarbon biofuels and other bioproducts (Ringer, Putsche, & bioreﬁneries.

Scahill, 2006). Adding to the appeal, most biomass sources can be

pyrolyzed, creating signiﬁcant potential for this conversion route. 3.1.2. Pre-treatment and degradation properties of biomass

There are, however, several signiﬁcant issues associated with its The storage of biomass resources can take place in various

distribution in and compatibility with the crude pipeline infras- modes. As the cheapest option, ambient storage leads to signiﬁcant

tructure. Bio-oil can have a very high oxygen and water content. cost reduction at the storage and handling stage of the biomass

It is highly corrosive and chemically unstable due to the high char supply chain. However, side-effects may include biomass degrada-

and solid contents. In addition, its alkali metals content can lead to tion (e.g. silage), heating value reduction, and potential health risks,

fouling and contribute to catalyst poisoning. The higher viscosity mainly because of the presence of high water content. The degrada-

of pyrolysis oil compared to petroleum crude requires increased tion rate is estimated to be 1% material loss/month for the ambient

40 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

storage (Rentizelas, Tolis, & Tatsiopoulos, 2009b). To reduce the quantity and composition of the sugars stream released depend on

degradation, covered storage in pole-frame structures is often used the feedstock used and the pretreatment technology employed. A

in practice, incurring a 0.5% material loss/month rate. If a higher- hydrolysate conditioning and/or neutralization process is used to

quality biomass supply is required, a closed warehouse with hot adjust the pH of and purify the pretreated material. In hydrolysis,

air drying capability can be employed. The material loss in this sce- the pretreated material, primarily cellulose, is guided through a

nario can be assumed negligible (Cundiff, Dias, & Sherali, 1997). As chemical reaction that releases the readily fermentable sugar, glu-

can be seen, the tradeoff between storage cost and material loss is cose. This step is usually accomplished with enzymes or strong

the focus when determining the selection of storage modes. acids. The conversion from sugars to biofuels can take place in

In addition to the direct storage of raw biomass, pretreatment two ways: biological and chemical processing. In biological pro-

processes on raw biomass materials are also adopted in many prac- cessing, the most common approach is to introduce a fermenting

tices. In this scenario, after the biomass has been harvested, thermal microorganism to the biomass hydrolysates. The continued sac-

and chemical treatments are applied to reduce moisture content, chariﬁcation and fermentation may take a few days to convert all

remove contaminants, and improve feedstock quality, stability, sugars to biofuels or other chemicals of interest. On the other hand,

and processing performance, including torrefaction, pelletization, a catalyst is added in chemical processing to make the reaction

fast pyrolysis, ammonia ﬁber expansion (AFEX), etc. (Bals, Rogers, less energy intensive and thus more efﬁcient than biological fer-

Jin, Balan, & Dale, 2010; Uslu, Faaij, & Bergman, 2008). Ongoing mentation. Furthermore, the yields of different end products and

research is trying to develop an advanced uniform-format feed- intermediate chemicals can be adjusted while using different reac-

stock supply system, aiming to reduce the variability in biomass tions and catalysts. The purpose of the ﬁnal step, product upgrading

(both format and quality). The products of this system are sup- and recovery, is to separate the fuel from the water and residual

posed to be standard, consistent, quality-controlled commodities solids, which could involve any transformation, distillation, sepa-

that can be efﬁciently handled, stored, transported and utilized by ration, and cleanup processes (Biddy & Jones, 2013; Davis, Biddy,

bioreﬁneries. Major technologies may involve milling, densiﬁca- Tan, Tao, & Jones, 2013b; DOE/EERE, 2011a).

tion, torrefaction, blending and pelletization (DOE/EERE, 2013a).

3.2.2. Thermochemical technologies

3.1.3. Logistics

The thermochemical technologies utilize high temperature in

The logistics of biomass plays an important role in the bioenergy

addition to catalysts to change the physical properties and chemical

supply chain, because it integrates time-sensitive feedstock col-

structures of the biomass resources. There are two major thermo-

lection, storage, and delivery operations into efﬁcient, year-round

chemical reaction pathways: gasiﬁcation and pyrolysis (DOE/EERE,

supply systems that deliver consistently high-quality biomass.

2011b).

Unlike fossil fuels, biomass resources have a sparsely spatial distri-

bution. Both the biomass collection and delivery require extensive

efforts in equipment selection, shift arrangement, vehicle routing,

3.2.2.1. Gasiﬁcation. In the gasiﬁcation pathway, the lignocellu-

and ﬂeet scheduling (DOE/EERE, 2013c). Moreover, if independent

losic biomass resources are fed into a gasiﬁer, where they are

◦

pretreatment depots are built to provide uniform-format feed-

thermally decomposed (700–1300 C) with limited or no oxygen,

stocks, of which the location should be cautiously determined in

and then oxidized to yield a raw syngas, which is primarily a

the supply chain design phase. A depot would locate to facilitate

mixture of hydrogen and carbon monoxide. The raw syngas may

the transportation and production, and serve like a small satel-

contain some contaminants, including tars, acid gas, ammonia,

lite system which receives biomass of different types and various

alkali metals, and other particulates. Therefore, a gas cleanup pro-

sources from its neighborhood and then forward the converted

cess is necessary to produce a contaminant-free syngas product. If

uniform-format biomass to a network of much larger supply ter-

necessary, the gas may be further conditioned to lower the sulfur

minals (DOE/EERE, 2013a).

levels and adjust the hydrogen to carbon monoxide ratio. The syn-

gas can then be utilized by the Fischer–Tropsch synthesis, including

3.2. Biomass conversion in advanced bioreﬁnery

a collection of chemical reactions, to produce hydrocarbon liquid

fuel products (DOE/EERE, 2010a; Swanson, Satrio, Brown, Platon,

Conversion of biomass into commercially viable liquid fuels

& Hsu, 2010; Talmadge, Biddy, Dutta, Jones, & Meyer, 2013). Alter-

is another main component of the supply chain. A bioreﬁnery is

natively, methanol and hydrogen can be produced from the syngas

a facility that integrates conversion processes and equipment to

via methanol synthesis and water–gas-shift process, respectively.

convert biomass into biofuels, power and chemicals. The bioreﬁn-

ery is analogous to a petroleum reﬁnery, but could have reduced

3.2.2.2. Pyrolysis. Pyrolysis processes also heat biomass in the

environmental impact and consumes less fossil energy because of

absence of oxygen, but at a relatively lower temperature

its diversiﬁcation in feedstocks and products. Broadly speaking,

◦

(450–600 C). The main product of the pyrolysis process is a crude-

biomass-to-liquid conversion technologies can be classiﬁed into

like liquid known as pyrolysis oil, while gas and charcoal are also

two categories, namely biochemical and thermochemical (Anex

produced. Depending on the length of the reaction time, there

et al., 2010).

are slow pyrolysis, intermediate pyrolysis, and fast pyrolysis. Fast

pyrolysis is often adopted because the yield of pyrolysis oil is

3.2.1. Biochemical technologies

maximized. Oil produced in pyrolysis processing must have par-

The major biochemical conversion processes include pretreat-

ticulates and ash removed by ﬁltration to create a homogenous

ment, hydrolysis, biological or chemical processing, and product

product. The pyrolysis oil is then upgraded to desired hydro-

upgrading and recovery. During pretreatment, biomass materi-

carbon fuels via hydrotreating and hydrocracking processes to

als undergo a mechanical or chemical process to fractionate the

reduce its total oxygen content and to break down long-chain

lignocellulosic complex into soluble and insoluble components.

hydrocarbons (DOE/EERE, 2010a; Jones et al., 2009; Wright, Satrio,

Soluble components include mixtures of ﬁve- and six-carbon sug-

Brown, Daugaard, & Hsu, 2010). Furthermore, National Advanced

ars and some sugars oligomers. Insoluble components include

Biofuels Consortium (NABC) is investigating advanced conversion

cellulosic polymers and oligomers and lignin. The main purpose

technologies such as catalytic fast pyrolysis, hydropyrolysis, and

of pretreatment is to break down the plant cell walls to permit

hydrothermal liquefaction (NABC, 2013).

further deconstruction during the hydrolysis step. Note that the

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 41

3.2.3. Algae harvesting and conversion stations by tanker trucks. This process is reasonably economical in

Microalgal species accumulate lipids such as triacylglycerols areas that have easy access to abundant biomass supply and bio-

(TAG), while cyanobacteria and macroalgae accumulate mostly fuel production facilities, e.g., Midwest regions. However, it is not

carbonhydrates. Microalgae and cyanobacteria are in very small an attractive option for long-distance delivery, especially as volume

∼

size ( 1 to 30 ␮m), in suspension in water culture, which makes increases. One impact of the distribution supply chain, is that the

harvesting more difﬁcult. Harvesting methods like forced ﬂoccu- Midwest has become a net exporter of starch ethanol via barges on

lation, sedimentation, ﬁltration and centrifugation are generally the Mississippi while the West Coast has become an importer of

used. Drying is required to achieve high biomass concentrations. sugarcane ethanol (DOE, 2010; NCEP, 2009; USDA, 2007).

In contrast, harvesting of macroalgae (e.g., seaweeds) is very much

similar to that of terrestrial agricultural crops. Modern mecha-

3.3.2. Features of ‘drop-in’ fuels

nized machines and equipment can be employed in large off-shore

In contrast, drop-in biofuels that are highly compatible, or

macroalgae farms. Washing and dewatering are required. In addi-

totally fungible, with the existing hydrocarbon fuel ﬁnishing, fuel

tion, due to the large size of macroalgae, milling is needed to reduce

handling, distribution, and end-use infrastructure would result

macroalgae to particle sizes that can be more efﬁciently processed

in easier and more widespread market acceptance (DOE/EERE,

(Lundquist, Woertz, Quinn, & Benemann, 2010).

2010b).

Depending on the type of algae and targeting end products, there

Hydrocarbon biofuels can be used directly in the existing

are three algae conversion pathways: conversion of algal extracts,

pipelines, dispensers and vehicle engines without any additional

direct production of biofuels from algae, and processing of whole

effort, because these advanced biofuels are the same in properties

algae (DOE/EERE, 2010b). The conversion from extracts derived

as their petroleum counterparts and compatible with the existing

from algal sources is the typical mode of biofuel production from

infrastructure. Thus, the remaining challenge in the modeling and

algae. The most common type of algal extracts is lipid-based. Cur-

optimization of supply chains for these infrastructure-compatible

rent practices for lipid extraction involve mechanical disruption,

biofuels is to address the full scale integration with the existing

solvent extraction, and supercritical technologies. The extracted

petroleum reﬁnery processes and distribution supply chains. In the

lipids can then be converted into bio-diesel, advanced hydrocarbon

short term, the introduction of advanced biofuels to existing dis-

biofuels, and other high-value chemicals via chemical transester-

tribution systems may require capacity expansion and additional

ﬁcation, enzymatic conversion, and catalytic cracking. The direct

scheduling efforts to coordinate with the fossil fuels. However, it

production refers to heterotrophic fermentation and growth from

is believed that, the advanced hydrocarbon biofuels would become

algal biomass. It has certain advantages in terms of process cost

the major stream and play a more important role in the fuel dis-

because it can eliminate several process steps (e.g., oil extraction)

tribution infrastructures in the future (DOE/EERE, 2013d; MASBI,

and their associated costs. The system is readily scaled up and there

2013).

is an enormous potential to use various ﬁxed carbon feedstocks.

Biofuels that can be produced directly from algae include alcohols,

alkanes, and hydrogen. In addition, processing of whole algae is also 4. Literature review

considered a promising algae biofuel solution. Potential conver-

sion methods include gasiﬁcation and pyrolysis, which are widely There is a large body of literature regarding the modeling

employed for cellulosic biomass, as well as supercritical processing and optimization of biofuel supply chains, covering different

and anaerobic digestion of whole algae. These methods avoid the feedstocks, products, processes, system properties, and model-

costs associated with the extraction processes and are amenable to ing viewpoints. The recent journal and conference papers closely

processing a diverse range of algae (Davis, Aden, & Pienkos, 2011; related to the subject are summarized in Table 1. The articles are

Davis, Biddy, & Jones, 2013a; FAO, 2009). However, extraction is sorted in chronological order. The articles published in the same

still preferred when oil content is high to leverage the entropic year are sorted in alphabetical order by last name of the authors. A

investment by the organism. number of the works may cover multiple aspects, thus occupying

multiple cells in the table, which indicates that a holistic optimiza-

3.3. Fuels distribution and end-use system tion model for modeling and optimization of biofuel supply chains

can be multi-scale, multi-perspective, and multi-criterion.

The last components in the biofuels supply chain are the fuel The column – “food crops to ethanol, biodiesel” – represents

distribution and end-use systems, which ensure that biomass- the biofuel supply chains using ﬁrst-generation biofuel technolo-

derived fuel products can cost-effectively and sustainably reach gies. These supply chains typically use starch- or oil-based food

their market and be used by consumers as a replacement for crops as feedstocks to produce ethanol and biodiesel, which can

petroleum-based fuels. be blended with petroleum-based gasoline and diesel. These tech-

nologies are commercialized and have a signiﬁcant penetration

3.3.1. Infrastructure incompatibility of ﬁrst generation biofuels in the current transportation liquid fuel markets. However, the

Although bio-ethanol and bio-diesel represent the major ﬁrst-generation biofuel technologies are often criticized on their

biomass-derived transportation fuel products in the current mar- implications on food prices and production. The column – “cellu-

ketplace, as mentioned before, bio-ethanol and bio-diesel are losic biomass to ethanol, biodiesel” – represents the biofuel supply

banned by most U.S. pipeline operators due to their polarity chains using second-generation biofuel technologies. These supply

and other corrosion, contamination issues. Therefore, the current chains use crop residues, wood residues, or dedicated energy crops

supply chains for ethanol and bio-diesel require dedicated fuel dis- as feedstocks, thus avoid the potential impact on food markets. Nev-

tribution and blending systems (Bunting et al., 2010). The most ertheless, ethanol and biodiesel are not the ideal substitutes to fossil

common practice is to transport bio-ethanol or bio-diesel from the fuels, because they are not completely compatible with the existing

biofuel production facility to distribution terminals by train, barge, distribution and end-use infrastructure. The column – “cellulosic

or truck to minimize the opportunity for contact with water and biomass to hydrocarbon biofuels” – represents the biofuel supply

dirt, which makes it difﬁcult for shipping larger volumes of biofu- chains using third-generation biofuel technologies. They are able

els. The biofuels are kept in dedicated tanks until blended with the to produce cellulosic-biomass-derived transportation fuels with

corresponding fossil fuels (ethanol with gasoline and FAME with properties that are functionally identical to their petroleum-based

diesel). Once blended, the ﬁnal fuel products are delivered to retail counterpart. At the current stage, the third-generation biofuel

42 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 al. etc.

, , ,

al.

et , al. ,

Frisk,

Rosso, al.

food, et

Shah (2011) (2010) (2003)

Elia Keirstead

cek

, , , Cucek (2010) Sokhansanj

, al. ˇ Elia Cu

(2010b) (2012) Robba,

Tatsiopoulos,

and heat,

, ,

Cucek

Flisberg, Samsatli,

Tolis et , , and

and

al. Smith

(2010) (2012)

et Shah (2009)

and

al. and (2011) (2012a,b) (2012a,b) (2012a,b) (2010)

power,

al. (2009)

et and (2009a) Adjiman,

al. al. al. al. al.

to Mindardi,

Sowlati,

Gan Keirstead, Rentizelas, Leduc Tittmann

, , , , , et et et et et

(2012)

Tolis Sacile Ronnqvist

ˇ ˇ ˇ cek cek cek al.

ˇ ˇ ˇ Frombo Cu Cucek Tatsiopoulos Cu Cu and (2007) Biomass Dunnett, Tittmann Velazquez-Marti Aksoy Frombo, Mobini, et (2011) (2011) Fernandez-Gonzalez (2010) (2012a,b) Pantaleo, (2010) and (2012) and

biofuels

(2012)

Algae Avami ,

, al.

and You

, , al. and et Yue Yue

Wang

and Wang

al.

, ,

Yue

Lee,

al.

et (2011) , , al.

al. (2011b)

(2011) and Aksoy et You

et , Kim , (2012) and et

and et

Realff, Xiao,

Romagnoli

(2010) Tittmann

Kim , al. to El-Halwagi

(2011b) Yue

(2012) (2012) and

Realff,

Dotzauer,

(2012) You (2012) al. Yue Schatka, You

Wang

, and , al. Kim,

Furtner al. al. Obersteiner

(2011) (2010)

, et (2012) and

Kondili, al. al.

(2011)

Yao, et biofuels Bowling, Sharma et et

Kim, Borenstein

Baliban, (2010) al.

and , , et et

(2009) ,

and al.

and

(2011)

biomass et

(2011) al. (2012)

(2011)

al. Kim Gebreslassie

et Wang Sarker,

and

(2011)

, , Elia, Elia Walther, Avami You Elia Walther

Parker

et Natarajan, al. , , , , , , , ,

(2011) (2010) (2013b) (2013b) (2013b)

and Wang et (2011a)

al. al. al. al. al.

Leduc You Parker (2012) Leduc, Elia et (2013b) Cellulosic (2011) (2009) Lee hydrocarbon (2011a) Kaldellis Whittaker, (2012) Tittmann Papapostolou, Gebreslassie, Sharma, Gebreslassie Bowling Gebreslassie Ferreira (2013b) et Ponce-Ortega, (2011) (2011) (2011) (2011a) et et et McCallum, and Spengler (2011) Floudas , , , , , al.

, Fan Fan

, ,

et and al. , and

and

Cucek,

al. You You

al. al. (2009)

, ,

, al. , (2011) (2012) et

Bernardi and and et (2013) (2013)

(2011)

et et

et (2012a,b) Smith

Hart, Benjaafar,

(2011) (2011) ˇ (2011b) cek

al. al. Chen, al.

al. Arora (2010a) Kravanja

(2011) (2010) (2012)

Ouyang ˇ (2012) Bezzo Bernardi Chen

Chen Cu et et al. al.

et and

Avami Eksioglu,

ethanol,

, , et , , El-Halwagi

, , (2012) Varbanov, Parker

and

(2012)

al.

, et et , (2012a,b) Parker Giarola

and

Ponce-Ortega,

and to

Lundgren,

, ,

and

and Parker, al.

Searcy

Onal Gan al. et

cek Huang,

al. and

, Smith

Schmidt, Aksoy et

ˇ chains. (2012) (2012) Cu Shah, (2012) (2012)

, et (2012) (2008) (2008)

Dotzauer

, and and (2011)

Bernardi Bernardi

Klemes, Kang, Daoutidis and , , (2010) (2010)

Leduc,

(2010b)

al. al.

, Klemes, You ˇ cek al. biomass

(2012)

, (2010) and

Leightley, et et al. al. al.

et et

(2011b)

Zamboni,

and ˇ (2010) et

Gan Akgul, Chen Marvin, Santibanez-Aguilar Avami Avami Marvin Cu Tittmann, Tittmann

supply

Lam, (2012) al.

, , , , , , , , , , , et et

al. (2012) (2013) (2012)

et Hwang, (2010)

Wilhelm,

al. al. al.

Cellulosic Cucek Dunnett Kravanja et Giarola (2012) Serna-Gonzalez, Cucek, Leduc An, Bai, Giarola, Santibanez-Aguilar, Gonzalez-Campos, Bernardi Cucek Santibanez-Aguilar Avami You Dunnett Santibanez-Aguilar Bernardi (2012) (2010) Papageorgiou Acharya, et (2011) (2012a,b) (2011) (2010) (2012) Varbanov, Tiffany, et (2011) biodiesel (2011) (2012) Franklin, Fan (2011) Jenkins biofuel

al.

et , al.

Shah ,

and Kostin, optimization et

Shah, , (2011) Zamboni Dal-Mas, and

and

, , and Petrolia Bai

and

al.

(2009a) Zamboni, (2011) An

(2011) Mele, Corsano ethanol,

et , , and (2009)

and

Zhang,

al.

al. to

(2011) (2009) (2009b) (2011)

Bezzo, Li, et

Vecchietti, (2012)

Zamboni,

(2011)

Zamboni,

al. al. ,

Mele Akgul,

Shah, , , (2011)

al. Guillen-Gosalbez,

modeling

et et

crops (2009a)

Jimenez Bezzo

al.

the

Bezzo, (2011) Bezzo and Jimenez Food Mele, Zamboni, Eksioglu, Corsano, Zamboni Mele Mele Dal-Mas Bai Sokhansanj, Guillen-Gosalbez, (2010) Papageorgiou biodiesel Montagna (2009a) (2012) and et Giarola, on

publications

recent

optimization

production

operation design

technologies

models

some

of and

chain

theory

Components Supply Planning Selection Decentralized Multi-objective Sustainability Uncertainties Game Simulation Table Summary

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 43

technologies are relatively mature and the corresponding sup- the biofuel supply chain (Hosseini & Shah, 2011). In this paper, we

ply chain design and optimization is desired for commercial-scale portray a schematic framework of multi-scale modeling and opti-

implementation. The column – “algae to biofuel” – represents the mization for biofuel supply chains in Fig. 2. From bottom to top,

supply chains using fourth-generation biofuel technologies, which four system layers regarding different temporal and spatial scales

produce various liquid fuels and chemicals from lipids and/or typify the level of ecosystem, supply chain, process, and molecule,

hydrocarbons in algae feedstocks. Compared to other columns, respectively.

there is a signiﬁcant lack of studies on these nascent technolo- At the ecosystem layer, the ecosystem and human society con-

gies, mainly because of the scarcity of credible data and references. stitute the environment where various activities in biofuel supply

Processes regarding algae cultivation, harvesting, and conversion chains take place. All these activities would cause impacts and foot-

need to be well classiﬁed and documented before the supply chain prints on the ecosystem throughout the life cycle and value chain

modeling and optimization can be implemented. The last column of the biofuel products, involving biomass cultivation, biomass har-

– “biomass to power, heat, food, etc.” – includes supply chains vesting and logistics, biomass pretreatment and fuel production in

with non-fuel uses of biomass resources, which provide electric- conversion facilities, and fuel distribution and end-use. Modeling

ity and heat supply to surrounding communities or produce foods and optimization tools can be applied to simultaneously evalu-

and other valuable chemical products. As can be seen, the large ate and identify the sustainable solutions with minimum adverse

body of literature in this column indicates the importance of these impacts on the environment and maximum beneﬁts to the soci-

non-fuel biomass utilization pathways. ety. Since multiple conﬂicting objectives are often involved when

By observing the rows of Table 1, we can see that most research optimizing the sustainability of biofuel supply chains, a number

works focused on “supply chain design”, mainly involving the of multi-objective optimization techniques can be applied, such as

transportation network design and bioreﬁnery facility location epsilon-constraint method, goal programming method, etc. Trian-

problems. A few studies integrated the “planning and operation” of gle charts can be used for comparison between different solutions

bioreﬁneries into the supply chain models, thus achieving a higher with different economic, environmental, and social performances.

resolution in both spatial and temporal scales. We note that earlier Pareto frontiers obtained from multi-objective optimization can

works on the design of biofuel supply chains seldom considered the reveal the tradeoffs between conﬂicting objectives and provide a

“selection of technologies”. However, as more and more conver- set of Pareto-optimal solutions for decision-making.

sion options become available, selection of technologies become At the supply chain layer, modeling and optimization tools

a critical component in the recent works on supply chain design can play an important role in optimizing the supply chain net-

and optimization. As will be further discussed in later sections, work structure and improving the various activities involved in the

“decentralized production” is a relatively new concept, mainly installation and operation of biofuel supply chains. A biofuel sup-

emerging from the design of advanced hydrocarbon biofuel supply ply chain usually consists of multiple sites and multiple echelons,

chains, which suggests that the traditional integrated bioreﬁnery which requires coordination across the entire supply chain network

can be decomposed into multiple conversion steps in multiple (You, Grossmann, & Wassick, 2011). At the design stage, super-

geographically distributed production facilities. “Multi-objective structure optimization allows the decision maker to determine the

optimization” is required when conﬂicting objectives are con- optimal network design from a number of choices among candidate

sidered simultaneously. This often involves the tradeoff between suppliers, facility locations, technology options, transport modes,

economic, environmental and social performances of the supply etc. By using mathematical programming approaches, description

chain as well as the risks associated with design and operation. of the various activities, even across multiple sectors, can be easily

Since biofuels are considered as a renewable energy alternative interpreted into equations in the constraints or objective functions.

to fossil fuels, the “sustainability” issues deserve further atten- However, a comprehensive and detailed superstructure, though

tion to avoid introducing adverse impacts on the ecosystem and appealing, may be computationally intractable. Therefore, when

human society. Some works employed the life cycle analysis (LCA) modeling the biofuel supply chain systems, we should take advan-

methodology to assess the environmental impacts (e.g., greenhouse tage of the properties of the supply chain system (e.g., network

gas emissions), while some addressed job creation opportunities structure, spatial scale) and drop off the components that have neg-

brought by the biofuel value chains. While most models considered ligible inﬂuences on the optimization objectives to trade off the

the design and operation of biofuel supply chains under determinis- modeling resolution and computation efﬁciency.

tic environments, some recent works investigated the performance At the process layer, the processes and units at a production

of a biofuel supply chain system in the presence of various “uncer- facility are studied in a more detailed manner. Superstructure-

tainties” (e.g., price volatility, yield uncertainty). Instead of treating based optimization can again be employed for the design, synthesis,

the biofuel supply chain system as a cooperative entity, we can also and retroﬁt of conversion processes, involving the selection

view the biofuel supply chain system as a value chain composed among candidate processing methods, equipment units, catalysts,

of different parties, who are acting selﬁshly and in a noncoopera- resource supply scenarios, etc (Yeomans & Grossmann, 1999).

tive manner. Few existing works utilized “game theory” and spatial Regarding the operation at a bioreﬁnery, three levels of deci-

market equilibrium to model the inter-relationships in biofuel sup- sions need to be optimized, namely planning, scheduling, and

ply chains. In addition to mathematical programming approaches, control. Optimization for planning involves the development of

“simulation models” are also useful tools for the study of biofuel robust forecasting model, multi-year strategy for capacity expan-

supply chains, which can easily capture dynamics in complex sys- sion and process retroﬁt, multi-period targets for purchase, sales

tems and provide process data for the optimization. and production, etc. Optimization for scheduling involves the efﬁ-

cient and timely allocation of equipment units, raw materials, and

human labors to fulﬁll the external and internal orders. Optimiza-

5. Multi-scale modeling and optimization tion for control involves real-time monitoring and adjustment of

process parameters to meet the required quantity and quality of

While most works provide a great insight to speciﬁc ﬁelds, they the products. The three decision levels are closely related, thus also

may fall into a narrow perspective and do not cover sufﬁciently providing enormous opportunities for vertical integration (Chu &

all the aspects of the biofuel supply chain. Therefore, we need a You, 2012; Engell & Harjunkoski, 2012; Wassick et al., 2012).

multi-scale modeling and optimization framework to provide a At the molecule level, modeling and optimization tools can

holistic view and organically integrate the different components of be employed by researchers with diverse background to facilitate

44 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Fig. 2. Illustration of multi-scale modeling and optimization of biofuel supply chains.

their studies on biomass-to-biofuel conversion technologies, both As illustrated by the arrows on the left of Fig. 2, the system

theoretical and experimental. Optimization tools can help to iden- scale addressed in the model grows from top to bottom, whereas

tify the promising reaction pathways that are cost-effective and the practical details considered in the model increases from

sustainable, thus saving unnecessary efforts in experimental tri- bottom to top. The double-headed arrows between adjacent

als. Furthermore, studies in molecular engineering, bioengineering, layers indicate that the modeling layers are never isolated but

and multi-physics simulation can be guided by the system engi- closely inter-connected. Generally speaking, the lower layer

neering approaches, while in return serve as the practical basis for conveys target or requirement to the higher layer, whereas the

upper-level modeling and optimization. higher layer provides detailed data which serve as building

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 45

blocks for the modeling and optimization of the lower Optimization models regarding vehicle routing and ﬂeet schedul-

layer. ing can be employed to cope with these problems. At the logistics

phase, one needs to ﬁgure out the optimal inventory policies

5.1. Multi-scale decisions regarding how much to replenish from the upstream, how much to

deliver to the downstream, how much to utilize for the production

The optimization of biofuel supply chains requires considering of intermediates or fuel products, and what level of inventory to

all the components across the entire biomass-to-biofuels supply keep in the storage units. These problems become more challenging

chain in a comprehensive model that allows the effective inte- as the supply of several types of biomass feedstocks would be sub-

gration of long term strategic design decisions with short term ject to season-to-season changes and limited in certain months of a

operational planning and scheduling decisions, which will be dis- year. Lots of tools and theories developed for generic supply chains

cussed respectively in this section (Sharma et al., 2013). can be incorporated for inventory optimization in biofuel sup-

ply chains, including sophisticated forecasting models, enterprise

5.1.1. Strategic decisions resource planning (ERP) system, material allocation rules, service

The strategic decisions are those decisions that have an inﬂu- level measurements, information ﬂow structure, etc. At the produc-

ence over years and decades, and even beyond the lifetime of the tion phase, major operational decisions involve the coordination of

project. Once a strategic decision is made, it is very unlikely to various material ﬂows and scheduling of multiple processing tasks

be altered in the short term. The ﬁrst and most critical strategic with continuous/batch equipment units in order to guarantee the

decision is the design of biofuel supply chain network. We can timely delivery of requested products both in quantity and qual-

construct a superstructure to optimize the selection of biomass ity. Extensive studies on planning and scheduling in petroleum

suppliers, location of conversion facilities, assignment of customer and chemical production industries can be adapted to study the

serving areas, and transportation links that connect different sites problem. Additionally, in cases that the quality and water con-

and deliver the biomass/biofuel across the supply chain network. tent of biomass feedstocks vary signiﬁcantly, optimization of the

Geographic information system (GIS) can be employed to visualize operational parameters (e.g., temperature, pressure, ﬂow rate) and

the spatial conﬁguration of biomass resources, candidate sites, fuel real-time control trajectory would be required. Advanced control

demand densities, etc., thus facilitating the modeling process (Esri, algorithms and dynamic simulation models exploiting the physi-

2012). The major tradeoff in the design of biofuel supply chain is cal and thermodynamic properties are helpful for handling these

between the costs of transportation and the capital and produc- problems.

tion costs of production facilities, because the transportation costs

account for a signiﬁcant portion in the biofuel supply chain while

the construction of production facilities may not necessarily bene- 5.2. Uniqueness in biofuel supply chains

ﬁt from the economy of scale. Besides, the selection of conversion

technologies and corresponding infrastructures is also an impor- Compared to their petroleum counterparts, some major compo-

tant strategic decision. Because different conversion technologies nents in the supply chain systems for biofuel products are similar

may vary in capital and operational costs, biofuel productivity and in function but different in structure and property. The most signif-

requirement for biomass resources, optimal selection of conversion icant difference is at the raw material acquisition stage. Petroleum

technologies may depend on the level and distribution of biomass emerges from point sources like drill shafts, guarantees perennial

availability and fuel demand, as well as the local restriction and supply and can be hauled over long distance with minimum trans-

policy at the facility candidate sites. Binary variables can be used port costs. However, the raw biomass feedstock, as a low-energy

to represent the selection of processes and conversion pathways density solid rather than a high-energy density liquid, is inher-

at the supply chain level, and the selection of catalysts, equip- ently distributed over a large area and is subject to seasonality,

ment units, operation conditions and processing methods at the thus requiring considerable collection and transportation efforts

process level. When evaluating the economic objectives (e.g., net and large storage volumes (DOE/EERE, 2013c).

present value, annualized total cost), besides the conventional costs From the aspect of supply chain design, the locations of biomass

for construction, materials, and labors, we should also take into processing facilities should be optimized in order to reduce the

account the government incentives (both construction and volu- transportation cost of low-energy density solid biomass resources

metric) and additional credits from selling the by-products (e.g., from surrounding biomass suppliers (e.g., farms, forests, paper

bio-char) (DOE/EERE, 2013b). Another perspective to note is that mills) and to lower the capital and operational costs. The capac-

the conﬁguration of the biofuel supply chain may not be static but ities of the biomass processing facilities should be able to meet the

can evolve over time. With the wide acceptance of biofuel prod- requirement in peak times and might be worth it to be repurposed

ucts and increasing incentives, capacity expansion and technology for part of the year. If a ﬂexible processing facility able to treat

upgrading on top of the existing biofuel supply chain system may different types of biomass resources is built, multi-period planning

be desired in the future. Particularly, it may be possible to adapt the can be utilized to schedule the shifts of biomass suppliers according

multi-period planning models proposed for generic supply chains to the feedstock availability in each month.

to biofuel supply chains. From an operational point of view, the biomass is cost-

prohibitive and unstable for long-distance transport because of its

5.1.2. Operational decisions low energy density and high water content. In the modeling prac-

Operational decisions are those decisions that are adjusted more tice, a circle around the biomass supplier is often speciﬁed as the

frequently in correspondence to the current external and inter- feasible region for building biomass processing facilities, of which

nal conditions, which usually have impacts for no less than a year the radius equals to the “feasible distance” for raw biomass trans-

or even a day. Due to the large number of activities involved portation (Uslu et al., 2008). Considering the low transport density

in the biofuel supply chain, optimization models for operational of biomass resources, both volume and weight limits may be spec-

decisions vary signiﬁcantly in scale, complexity and formulation. iﬁed for the capacity of transportation links (You, Tao, Graziano, &

At the biomass acquisition phase, operational decisions involve Snyder, 2012). In addition, degradation of biomass resources also

how to allocate the operators to harvesting machines, how to makes the storage more difﬁcult to handle. A linear monthly rate of

arrange the working shifts, how to assign the harvesting operation material loss is usually assumed, varying in the storage technology

areas, how to convey the harvested biomass to storage units, etc. used. Correspondingly, we can incorporate in our model a mass

46 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Table 2

balance on biomass inventory accounting for the material loss at

Total discounted capital, production, and transportation costs of the three designs

each biomass processing facility and in each month.

under different side length of the square world.

40 km 60 km 135 km 200 km

5.3. Distributed vs. centralized strategy

Centralized $290 MM/y $530 MM/y $2224 MM/y $5156 MM/y

In the case of petroleum reﬁnery, there is the clear rule of $1.94/GEG $1.58/GEG $1.31/GEG $1.38/GEG

Distributed $457 MM/y $786 MM/y $2610 MM/y $5126 MM/y

“economy of scale”. Increasing aggregation of petroleum feed-

$3.06/GEG $2.34/GEG $1.54/GEG $1.38/GEG

stocks, typically by pipeline or ocean tanker, only increases costs

Two-stage $298 MM/y $528 MM/y $1927 MM/y $4031 MM/y

marginally. With high pressure and temperature requirements,

$2.30/GEG $1.81/GEG $1.30/GEG $1.24/GEG

increasing production scale follows a 0.6–0.8 scaling factor driv-

ing economies of scale. However, this is not automatically true

for bioreﬁnery (Dunnett, Adjiman, & Shah, 2008). The advantages

of larger plants might be offset by the increased costs of hauling transportation cost, which can be signiﬁcantly reduced through

more low density biomass feedstocks to the point of processing two-stage design. However, we note that the centralized design

for longer distance (Shastri, Rodriguez, Hansen, & Ting, 2012). It still yields the minimum unit cost, because of the relatively low

is also reported that transportation takes a much larger piece in conversion efﬁciency of the two-step processing of the two-stage

the cost pie of biofuel supply chains than that of petroleum sup- design. When considering a 135 km 135 km square region, the

ply chains. Therefore, to reveal the tradeoffs between the costs of two-stage design has both the minimum total cost and minimum

transportation and process construction/operations, we consider unit cost. Moreover, the difference between the costs of the central-

an illustrative example in this section. ized design and distributed design becomes smaller. Interestingly,

We assume a superstructure of the hydrocarbon biofuel sup- in a 200 km 200 km square region, the distributed design becomes

ply chain in a hypothetical 4 × 4 matrix region as shown in Fig. 3a. more cost-effective than the centralized design. The reason is that

Each green square area represents a farm, of which the yield of the higher transportation cost resulting from the larger area offsets

biomass is evenly distributed. There are ﬁve candidate sites for the capital saving of the economy of scale in the centralized design.

locating the conversion facilities, as represented by blue circles. The

only demand zone, represented by the pink circle, is at the center

5.4. Competitions between different parties

of the square region. In this superstructure, two conversion path-

ways are considered: (1) directly converting biomass to liquid fuels

In most existing literature, the entire biofuel supply chain is con-

in an integrated bioreﬁnery (represented by yellow plant) and (2)

sidered as entity centralized system. This might be true if all parties

preconverting biomass into bio-oil in fast pyrolysis plants (repre-

involved in the biofuel supply chain are cooperative, which happens

sented by red plant) and then upgrading the bio-oil to liquid fuels

in the case that the farms, production facilities, and distribution

in a gasiﬁcation plant (represented by blue plant). For clarity, we

systems are all acquired by the same organization. However, more

also illustrate the two conversion pathways in Fig. 3b.

often, the parties are non-cooperative, thus causing competition

On the basis of the superstructure, we consider three types

in the price and use of materials. In such scenarios, game theory

of supply chain designs, namely centralized, distributed and

is a powerful tool for the study of mathematical models concern-

two-stage. In the centralized scenario, all the biomass resources

ing conﬂict and cooperation between intelligent rational parties

collected from the 16 farms are shipped to a large, integrated gasi-

in the supply chain system. In game theory, the interactions and

ﬁcation plant located in the center of the square region for the

competitions between different parties can be modeled as various

conversion to liquid fuels. In the distributed scenario, four inte-

types of games depending on their behaviors, including coopera-

grated gasiﬁcation plants of a smaller size are built at the center

tive or non-cooperative games, symmetric and asymmetric games,

of four corners of the square region, each processing the biomass

zero-sum and non-zero-sum games, simultaneous and sequential

resources from the four adjacent farms. The liquid fuels are then

games, etc. (Dutta, 1999; Myerson, 1997).

shipped to the demand zone in the center of the square region

A typical example of the non-cooperative biofuel supply chain

from the four integrated bioreﬁneries. The third scenario – two-

is the competition for biomass resources with agricultural sector

stage design – includes four fast pyrolysis plants located at the

at the biomass acquisition stage (Bai, Ouyang, & Pang, 2012). For

center of four corners and an upgrading facility at the center of the

instance, if corn is considered as the feedstock for biofuel produc-

square region. The biomass resources collected from the four adja-

tion, the biofuel company might offer certain prices to procure

cent farms are preconverted to bio-oil in the pyrolysis plants. Then

feedstocks and the farmers have the option to ship and sell their

the bio-oil is shipped to the central upgrading facility for further

corns to local grain markets or nearby biomass processing facili-

conversion to liquid fuels.

ties. In this case, the biofuel company can optimize their offered

Since our focus is on the tradeoff between the costs of

prices to maximize its proﬁt, while the set of spatially distributed

transportation and process construction/operations, the major

farmers would react accordingly to maximize their individual prof-

differences between the three designs are in facility locations, pro-

its. The problem can be regarded as a Stackelberg leader-follower

duction capacities, and technology selections. All the data and a

game (von Stackelberg, 1934) and formulated as a bi-level supply

detailed discussion are provided in the work by You and Wang

chain optimization problem.

(2011). Here in this paper, we merely present the major results.

Another approach is to introduce proper contract mechanisms

As summarized in Table 2, the centralized layout might be the

to leverage the interests of different parties, among which the

best option for the 40 km × 40 km square region, because of the

price transfer policy is an intuitive and efﬁcient method to cre-

economy of scale and higher efﬁciency of integrated conver-

ate a fair, optimized proﬁt distribution in the biofuel supply chain

sion. Although with the highest transportation cost and slightly

(Gjerdrum, Shah, & Papageorgiou, 2001). The transfer price is the

higher capital cost than does the centralized design, the two-stage

selling price at the material supplier and the purchasing cost at the

design can be still considered a viable approach, because of its

material receiver, which is also the major variable in this optimiza-

low operating cost resulting from the credit from charcoal, the

tion problem. To avoid unfair proﬁt distribution between parties

major byproduct in preconversion processes. If in a 60 km × 60 km

in the biofuel supply chain, a nonlinear Nash-type (Nash, 1950)

square region, the two-stage design becomes the least costly

objective function can be employed.

option, because the larger area implies a higher growth rate of the

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 47

Fig. 3. Illustration of centralized vs. distributed supply chain networks. (a) Comparison of three designs of the biofuel supply chain. (b) Illustration of centralized-distributed

pathways. (For interpretation of the references to color in this ﬁgure legend, the reader is referred to the web version of this article.)

5.5. International trade of biomass/biofuels converted in exporting countries. Furthermore, when considering

the transportation mode for biomass/biofuel shipments, pipeline

Although current focus is on the use of domestic biomass transportation, though requiring considerable upfront costs, may

resources for biofuel production, international trades of emerge as a solution that is economically viable in the long run, in

biomass/biofuels may also provide win–win opportunities to addition to truck, train and barge (DOE, 2010).

both exporting and importing countries (Kaditi & Swinnen, 2007). Another interesting problem inherent in international trades of

Some countries, especially developing countries, have large vacant biomass/biofuel supply chains is associated with the multi-national

land, lower cost of labor and equipment, from which importing or cross-region ﬁnancing activities. When the materials are shipped

can help domestic biofuel industries to improve competitiveness across the country/state borders, currency exchange, custom tariffs

of biofuel products, to create more stable biomass supplies, and to or different state taxes may incur (Lamers, Junginger, Hamelinck, &

contribute to economic development and technology innovation in Faaij, 2012). All these factors may lead to additional costs for inter-

exporting countries (Junginger et al., 2008). For instance, Uslu et al. national or inter-region trades, thus should be taken into account in

(2008) provided a comprehensive analysis of the biomass/biofuel the economic evaluation of the biofuel supply chain project. Adding

supply chain from Latin America to the European Union. Addi- to this point, carbon trading platforms would also contribute to the

tionally, as the largest wood pellets exporter in the world, Canada economic performance and sustainability of international biofuel

holds considerable potential for international biomass/biofuel supply chains. Since countries may have deﬁcits/surplus in their

trades with the biofuel industry in the U.S. (NREL, 2013b). respective carbon emission quota, incorporating importing scenar-

Long-distance transportation is inevitable in international ios into the optimization model can help to achieve efﬁcient use of

trades of biomass/biofuels. Since raw biomass resources are usually resources while to meet the environmental restrictions. Another

too low in energy transport density, there is a universal consensus alternative is to trade renewable fuel credits on the international

that raw biomass needs to undergo pretreatment and densiﬁca- market and enable local optimization of supply chains. This method

tion processes to improve the energy density and biomass stability. could provide sufﬁcient incentive to increase biofuel production

Therefore, the remaining problem is to determine which pretreat- above local mandates.

ment method, at what point in the supply chain, with which

conversion technology and what transportation mode would give 5.6. Integration with other supply chains

the optimal delivery performance for international biofuel sup-

ply chains. International trading options include direct import of 5.6.1. Integration with petroleum supply chains

biofuels products produced in exporting countries, or import of Conventionally, bioreﬁneries are considered independent of

intermediate energy carriers (e.g., torrefaction pellets, pyrolysis oil) petroleum supply chains. However, with the development of

48 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Fig. 4. Entry points for biomass resources into existing petroleum infrastructure.

advanced hydrocarbon biofuels, great opportunities emerge for can be constructed to incorporate the biofuel supply chain into

biofuel industry to take advantage of the existing infrastructure existing petroleum supply chain systems with binary variables rep-

of petroleum supply chains (mainly reﬁnery units and pipelines), resenting the selection of biomass/biofuel insertion points, choice

thus saving considerable amounts of capital and operation costs. As of conversion technologies, existence and capacity level of biomass

shown in Fig. 4, three possible insertion points for biofuel supply pretreatment units, as well as other retroﬁtting decisions associ-

chain to entry the petroleum infrastructure include (Bunting et al., ated with catalysts, processing methods, process heat integration,

2010; DOE/EERE, 2013d), hydrogen demand, etc. Furthermore, as additional material streams

introduced from the biofuel supply chain are involved in the inte-

• grated production system, more efforts on planning are required for

Insertion point (a): A bio-crude that can be co-processed with

the purchase, sales, transportation, and storage, but more impor-

conventional crude oil

• tantly, on the scheduling in raw material processing and fuel

Insertion point (b): Reﬁnery-ready intermediates that are com-

production to coordinate the additional material streams in order

patible with speciﬁc reﬁnery streams for further processing at the

reﬁnery to achieve maximum utilization of reﬁnery units and maximum

• overall proﬁt from both biofuel and petroleum products.

Insertion point (c): A near-ﬁnished fuel or blend stock that will

be minimally processed at the reﬁnery.

5.6.2. Integration with carbon mitigation

For insertion point (a), bio-crude (e.g., pyrolysis oil) can be CO2 capture and sequestration (CCS) is a major means to reduce

mixed with petroleum crude oil, then sent to crude distillation greenhouse gas emissions, especially from centralized energy

units, and ﬁnally converted to fuel products via a series of upgrading conversion sites (e.g., power stations, petroleum reﬁneries). Con-

units. For insertion point (b), bio-crude is directly sent to upgrading ventional practice for carbon sequestration is to remove CO2

units after certain treatment to remove oxygen and other con- from the atmosphere and deposit in a reservoir (e.g., soil, ocean)

taminants. For insertion point (c), there are two options, namely (DOE/NETL, 2010). However, the recent development of algae-

blending biofuels with conventional fossil-fuels in the petroleum based biofuel technology creates new options for CO2 sequestration

reﬁnery and then shipping them to the customers using existing as well as the opportunities to take advantages of the carbon con-

pipelines, or delivery of biofuels directly to distribution centers and tent for the growth of algae biomass, thus increasing the atom

blending there. Note that, minimum capital costs are still required efﬁciency and leading to CO2 capture and utilization (CCU).

to retroﬁt petroleum reﬁnery units to be compatible with these There is a large body of literature investigating the potential

insertion points (Tong, Gleeson, Rong, & You, 2013; Tong, Gong, of using algae for carbon mitigation, but research on how to con-

Yue, & You, 2013). Depending on regional air emissions standards, nect the carbon-sources with algae-based carbon-sinks is relatively

blending of a ﬁnished biofuel might require speciﬁc adjustments of limited. Some researchers suggest building the algae cultivation

the petroleum fuel to meet requirements for Reformulated Blend- facilities close to the carbon-sources (Wilcox, 2012). However, this

stock for Oxygenate Blending (RBOB). option requires considerable amounts of capital costs and may not

From the modeling perspective, there are abundant existing be optimal with a large number of carbon-sources that are located

tools for the design, planning and scheduling of petroleum supply geographically apart. In this scenario, optimization on sitting of

chains due to its maturity, whereas there is currently a signiﬁcant algae cultivation sites and design of CO2 pipeline transport network

lack of supporting tools for the decision-making in emerging bio- may come into play. When the total amount of CO2 needed to be

fuel supply chains. Therefore, with the resemblance and difference transported is rather small, simple source-sink matching approach

between the two types of supply chains in mind, we should take may be applicable at the early stage of CCU development. How-

advantage of the existing tools and exploit the synergies among ever, in scenarios with a large amount of CO2 to be transported

the various activities in the integrated system. A superstructure and a great number of carbon-sources to be matched, source-sink

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 49

matching approach appears less efﬁcient and superstructure-based process-based LCA methodology, following the four-phase proce-

optimization approaches for the design of CO2 pipeline transport dure suggested by ISO.

network may become more appealing.

The major costs of the CO2 pipeline transport network involve

6.1.1.1. Goal and scope deﬁnition. In this phase, project description,

the capital costs for installing the pipelines and pump stations,

goal of study, boundary of system and deﬁnition of functional unit

as well as the costs associated with operation and energy con-

are stated. Depending on the goal of study, the system boundary

sumption (Zhou, Liu, & Li 2013). It is the pressure provided at the

can be “cradle-to-grave”, “cradle-to-gate” or “gate-to-gate”. In the

pump stations that drives the CO2 to ﬂow through the pipelines,

study of biofuel supply chains, the “cradle-to-gate” boundary is

which has to be above a certain value to maintain the supercritical

often adopted, including the life cycle stages from biomass culti-

or dense-phase liquid states of CO2 to ensure transportation efﬁ-

vation, harvesting, pretreatment, through transportation, storage

ciency, meanwhile cannot exceed certain limits in order to avoid

and conversion, to the gate of ﬁnished-product distribution centers.

risks of explosion and other safety issues. Therefore, the tradeoff

Moreover, the deﬁnition of functional unit is also critical, based on

between the number of pump stations and the distance of CO2

which, all the calculations will be performed and normalized. For

transportation needs to be established. An insight on the design of

instance, for a biofuel supply chain that produces biomass-derived

CO2 pipeline network, is that one needs to avoid valleys and depres-

gasoline, diesel and jet fuel, the “gasoline gallon equivalent (GGE)”

sions because a pipeline failure could starve a region of breathable

– a unit based on energy content – can be selected as the functional

air. Binary variables can be introduced to represent the existence

unit (Wikipedia, 2013b).

of a pump station at a certain candidate node, the existence of a

pipeline along a certain segment, as well as the selection of the

diameters of the pipelines with discrete standard sizes. Continuous

6.1.1.2. Inventory analysis. This second stage of LCA involves the

variables can be used to represent the pressure rise, ﬂow rate of CO2,

compilation and quantiﬁcation of life cycle inventory (LCI) associ-

and the amount of CO2 available at carbon-sources and the capacity

ated with each process/stage within the life cycle boundary. LCI is

at algae cultivation sites. Existing works on optimization of natu-

simply a list of material inputs and outputs for a given product sys-

ral gas pipeline networks may be good references for the design

tem throughout its life cycle, including the cumulative extraction

and operation of CO2 pipeline transport networks (Baumrucker &

of resources from the environment (e.g., primary energy, mineral

Biegler, 2010; Gopalakrishnan & Biegler, 2013).

resources) as well as the cumulative emissions to the environment

(e.g., gas emissions, liquid and solid wastes). LCI for a lot of pro-

6. Modeling sustainability issues in bioenergy and biofuel

cesses can be found in commercial LCA databases such as Ecoinvent

supply chains

(2008), GaBi (2011), GREET (2012), and openLCA (2012). However,

for new processes that are not documented, cumulative LCI should

As a renewable alternative to conventional fossil-fuels, the sus-

be calculated according to energy and mass balance, using the unit

tainability issues associated with biofuels deserve our attention.

process raw data and LCI of background processes. Sometimes, on-

In order to reduce adverse impacts on ecosystem and improve the

site measurements may be necessary.

overall economic proﬁtability and social beneﬁts, systematic mod-

eling and optimization frameworks are required to simultaneously

assess and identify the sustainable solutions for the design and 6.1.1.3. Impact assessment. In this phase, the LCI obtained from the

operation of biofuel supply chains (You et al., 2012). In this section, previous phase is translated according to certain damage assess-

we will introduce how modeling and optimization tools can be uti- ment models. These models can generate a uniﬁed or a series of

lized to address the sustainability issues in biofuel supply chains, environmental performance indicators, which are easily under-

concerning environment, society, and economy, respectively. standable and user-friendly numbers. The most widely-used LCA

metrics include GWP (IPCC, 2007), Eco-indicator 99 (Goedkoop

6.1. Environmental impacts & Spriensma, 2001), IMPACT 2002+ (Humbert, Schryver, Bengoa,

Margni, & Jolliet, 2012), etc. GWP focuses on greenhouse gas (GHG)

Accompanied by the rise in environmental awareness and emissions causing the global warming effects, while Eco-indicator

tightening of environmental policies, the study of environmen- 99 and IMPACT 2002+ evaluate the environmental impacts in more

tal sustainability has received increasing attention in the past comprehensive categories (e.g., human health, ecosystem quality

decades. Among the various approaches, life cycle assessment (LCA) and resources). We note one critical problem that has been recently

methodology stands out to be the most successful tool for evaluat- highlighted is about the water footprint of biofuel production sys-

ing and analyzing the environmental impacts of product systems tems (Chiu, Suh, Pﬁster, Hellweg, & Koehler, 2012; Gerbens-Leenes

(Azapagic, 1999; Azapagic & Clift, 1999; Fava, 1994). The classical & Hoekstra, 2011), which should also be explicitly dealt with in bio-

LCA framework is based on stages and processes, which is well reg- fuel supply chain design (Bernardi, Giarola, & Bezzo, 2012; Bernardi,

ulated by ISO standards and speciﬁc methodology guidelines (ISO, Giarola, & Bezzo, 2013; Cuˇ cek,ˇ Klemes,ˇ & Kravanja, 2012a,b).

2006). In addition, input–output models and hybrid methods are

also promising tools for LCA, of which Eco-LCA as an example, is

6.1.1.4. Interpretation. In the ﬁnal phase, the LCA results are ana-

based on an integrated ecological-economic model of the U.S. econ-

lyzed to provide a set of conclusions and recommendations, usually

omy and accounts for various ecosystem goods and services (Bakshi

in the form of written reports. In this regard, when we consider

et al., 2013). We note that the uncertainties in environmental evalu-

the biofuel supply chain, the goal of LCA is to provide criteria

ation can have signiﬁcant inﬂuences on the reliability of LCA-based

and quantitative measurements for comparing different network

decisions. These uncertainties may stem from the limited knowl-

layouts and operation alternatives. However, one of the critical

edge about the physical processes under study and the normative

drawbacks of classical LCA framework is that it lacks a system-

choices regarding scenarios and mathematical models (Huijbregts,

atic approach for generating such alternatives and identifying the

Heijungs, & Hellweg, 2004).

best one in terms of environmental performance (Grossmann &

Guillén-Gosálbez, 2010). Therefore, most researchers merely use

6.1.1. Four-phase life cycle assessment

LCA as a post-optimization tool for the evaluation of environmental

Here in this paper, considering its relative maturity and wider

sustainability.

acceptance, we would like to brieﬂy introduce the classical

50 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Fig. 5. Life cycle optimization modeling framework.

6.1.2. Life cycle optimization chains in the United States (NREL, 2012). To evaluate the number

To circumvent the aforementioned limitation, the life cycle opti- of jobs that will accrue to the state or local region from a project,

mization framework was proposed in recent years (Gebreslassie, JEDI performs an input–output multiplier analysis, where a mul-

Slivinsky, Wang, & You, 2013a; Gebreslassie, Waymire, & You, tiplier is a simple ratio of total systemic change over the initial

2013b; Wang, Gebreslassie, & You, 2013; Yue & You, 2013b). change (e.g., employment) resulting from a given economic activity

This general modeling framework couples optimization tools with (e.g., startup or termination of a project). The multipliers are esti-

classical LCA methodology, allowing us not only to evaluate the mated through economic input–output (EIO) models, which were

environmental impacts of diverse design and operation alterna- originally developed to trace supply linkages in the economy, and

tives but also to identify the optimal solution (Fig. 5). When other to quantify the effects of change of expenditure within a regional

criteria (e.g., economic and social) are considered at the same time, economy in multiple industry sectors. In the input-out analysis,

the life cycle optimization framework is also able to perform multi- the employment effects from the construction and operation of a

objective optimization. In this case, the results are often presented biofuel supply chain can be classiﬁed into three categories: direct,

as a set of Pareto-optimal solutions, consisting of a Pareto frontier indirect and induced.

which reveals the tradeoffs in decision under multiple objectives.

•

One can choose among these Pareto-optimal solutions based on the Direct effect: The immediate or on-site effect created by expen-

preference for the design and operation of potential biofuel supply diture. For example, in constructing a bio-reﬁnery plant, direct

chains. effects may include the on-site contractors and hired crews in

the civil engineering team. In addition, direct effects may include

6.1.3. Functional-unit-based evaluation the jobs for building the process equipment units at bio-reﬁnery

For most economic objectives, we often try to optimize the total plants.

•

quantity, i.e., maximizing total proﬁt or minimizing total cost. How- Indirect effect: The increase in economic activity that occurs

ever, for environmental criteria, the “environmental impact per when contractors, vendors, or manufacturers receive payment

functional unit” is more critical (Yue, Kim, & You, 2013b). This is for goods or services and in turn are able to pay others who sup-

because, for example, one may care more about the carbon footprint port their business. For instance, indirect effects may include the

of producing one gallon of biomass-derived gasoline, rather than banker who ﬁnances the contractor, the accountant who keeps

the annual total CO2 emission of the entire biofuel supply chain. the contractor’s books, and the steel mills, electrical manufactur-

Therefore, the environmental impact evaluated based on functional ers and other suppliers that provide the necessary materials.

•

unit is, in fact, the indicator of how “green” the supply chain is. How- Induced effect: This refers to the change in wealth that occurs or

ever, one arising issue is that functional-unit-based environmental is induced by the expenditure of those persons who are directly

objective function can involve fractional terms, thus introducing or indirectly employed by the project.

nonlinearity into the model. Advanced solution methods includ-

ing parametric algorithm (You, Castro, & Grossmann, 2009; Zhong The total employment effect from a single expenditure can

& You, 2014), reformulation–linearization method (Yue, Guillén- be calculated by summing up the three kinds of effect, using

Gosálbez, & You, 2013a), as well as several general-purpose MINLP regional-speciﬁc multipliers and personal expenditure patterns.

solvers and global optimizers can be utilized to resolve the issue State-by-state multipliers for employment, economic activity, and

(Floudas, 1999; Grossmann, 2002). personal expenditure patterns can be derived from the IMPLAN

Professional model (IMPLAN, 2013). In this way, the changes

6.2. Social factor in employment brought about by investment in the design and

operation of a biofuel supply chain can be matched with the

6.2.1. Job creation corresponding multipliers for each industry sector. Both the one-

The most important perspective in the social dimension for the time impacts resulting from the construction phase and the

design and operation of biofuel supply chains may be the employ- yearly impacts resulting from the annual operations along the

ment effect, which can be measured by the number of accrued project lifetime should be considered in the measure of social

local jobs (full-time equivalent for a year) in a regional econ- impacts.

omy (Bamuﬂeh, Ponce-Ortega, & El-Halwagi, 2013; Lira-Barragán,

Ponce-Ortega, Serna-González, & El-Halwagi, 2013; You et al., 6.2.2. Other perspectives

2012). The Jobs and Economic Development Impact (JEDI) model Besides job creation, the social dimension still involves many

developed at National Renewable Energy Laboratory (NREL) can be other perspectives such as human rights, labor practices, decent

used and integrated into the aforementioned multi-objective opti- work conditions, society wellness, product responsibility, etc.

mization framework, to systematically evaluate the social impacts Despite of the existence of several conceptual frameworks, the

associated with the construction and operations of biofuel supply quantitative evaluation methodology for such social perspectives

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 51

is still immature (UNEP, 2009). One of the promising approaches

under research right now is to broaden the scope of LCA framework

to include social inventory results. However, by the time of writing

this review, the choice and formulation of indicators as well as the

position towards the use of generic data are still under debate in

the social LCA community (Benoît et al., 2010; Dreyer, Hauschild, &

Schierbeck, 2006; Jørgensen, Bocq, Nazarkina, & Hauschild, 2008;

Weidema, 2006). Although no appropriate software tools for social

LCA are currently available, there are several pilot developments

including openLCA (openLCA, 2012), Earthster (GreenDelta, 2013),

etc.

6.3. Economic factor

The idea of economic sustainability is to promote the use of

available resources in a way that is both efﬁcient and responsible,

and likely to provide long-term beneﬁts. In addition, an alternative

feedstock can reduce risk from disruption of a highly-concentrated

correlated market. In the case of design and operation of a biofuel

supply chain, it calls for using resources (e.g., capital, land, labor,

research) wisely so that the business continues to healthily function

over a number of years, while consistently returning a risk-adjusted

return.

Several metrics can be used as the economic objectives in the

optimization model for biofuel supply chains. The most common

Fig. 6. Multi-scale uncertainties with selected optimization approaches.

ones are net present value (Wikipedia, 2013c) and annualized total

cost (Wikipedia, 2013a). Both metrics account for the time value

7.1. Strategic uncertainty

of money over the lifetime of the project (usually 15–20 years for

biofuel supply chains) by using discounted cash ﬂow analysis. In

The most signiﬁcant type of strategic uncertainty for biofuel

addition, it is believed that minimizing the cost associated with per

supply chains might associate with the government incentives and

unit of biofuel products would lead to a more cost-effective supply

policies (MASBI, 2013). Due to the relatively high production cost of

chain, thus promoting the competitiveness and market acceptance

biofuel products and their limited market acceptance at the current

of the biofuel products. Besides, as a renewable alternative, the

stage of development, the incentives and supporting policies serve

efﬁciency of energy and resource use in biofuel supply chains is

as a main driver of the biofuel industry, making the biofuel products

also an important indicator. We can minimize the cumulative fos-

competitive with their petroleum counterpart both in terms of eco-

sil energy demand or the consumption of certain type of critical

nomic cost and sustainable performance. However, the incentives

resources (e.g., labor, materials) associated with per unit of biofuel

and policies may be changed after the biofuel supply chain is built,

products, in order to achieve the maximum utilization of available

due to the changes in economic, political or technical environments,

resources (Althaus et al., 2010). Responsiveness is another impor-

thus causing deviation between the realized proﬁt return and the

tant indicator affecting the economic performance of the biofuel

original anticipation. Furthermore, political uncertainty increases

supply chain, because to what extent the order of customers can

perceived market risk, increasing ﬁnancing costs to account for

be fulﬁlled on time is a touchstone for determining the quality of

risk-adjusted returns. Scenario analysis would be straightforward

the design and operation of a supply chain. By using the stochastic

to address the uncertainty in incentives and policies by inves-

inventory theory, we can employ the guaranteed service time from

tigating the performances in optimistic, neutral and pessimistic

the gate of distribution centers to biofuel retailers or individual cus-

scenarios. Alternatively, stochastic programming can be employed

tomers as one of the optimization objectives (You & Grossmann,

to maximize the proﬁt expectation among a large number of

2008a, 2011a). Again, since multiple objectives may be involved,

possible scenarios. Facility disruption under catastrophic events

multi-objective optimization techniques can be utilized to simul-

(e.g., earthquake, hurricane, plant explosion, or terrorist action) is

taneously optimize the objectives of interest.

another type of strategic uncertainty that might cause the paral-

ysis or malfunctioning of the biofuel supply chain system. Metrics

7. Uncertainty analysis and risk management for biofuel

regarding the resilience under destructive disruptions can be incor-

supply chains

porated into the optimization model as one of the objectives at the

supply chain design phase to enhance the robustness of the supply

Uncertainties are ubiquitous, of various types and varying

chain design, though usually with additional costs. Alternatively, in

degrees, and emerging from all stages and activities in the bio-

cases that probabilities of disruption can be estimated from histor-

fuel supply chain (Awudu & Zhang, 2012). Corresponding to the

ical data or operating experiences, stochastic programming would

multi-scale decisions, we can also broadly categorize the uncer-

be a great tool to handle facility disruption. Besides, technology

tainties into strategic and operational levels, which are usually

evolution is an important type of strategic uncertainty. Consider-

handled with by optimization models of different scales and formu-

ing that many biofuel technologies are relatively young, the process

lations (Lee, 2012). The most widely employed approaches include

conversion rate, energy efﬁciency, and material handling costs

scenario analysis, stochastic programming, robust optimization,

are expected to improve in the future. Therefore, one may need

fuzzing programming, chance-constraint optimization, stochastic

to determine whether to employ mature technologies with less

inventory theory, etc (Sahinidis, 2004). A summary is provided

improving space or nascent technologies with considerable poten-

in Fig. 6, including a selected number of uncertainty sources and

tial, which makes stochastic programming (Birge and Louveaux,

promising approaches for optimization under uncertainty.

1997) a proper tool for solving this issue. The last type of strategic

52 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

uncertainty we would like to mention corresponds to the cli- intermittent sources like wind and solar energy (EPRI, 2010). In

mate and weathers. Different from petroleum resources, the annual addition, there has been a long history of the use of biomass for

biomass yield is largely dependent on the sunlight, precipitation, heat supply. In modern practices, electric power and heat are often

climate, as well as pest threats. For biomass, there are potential generated simultaneously, namely cogeneration or combined heat

catastrophic risks at both the seasonal and multi-seasonal scales. and power (CHP) (Shipley, Hampson, Hedman, Garland, & Bautista,

Robust optimization might be the appropriate tool for the design 2008).

and operation of biofuel supply chains to guarantee the quantity Compared to the biofuel supply chain system, the upstream

and quality of biomass supplies. of the biomass-to-power-and-heat supply chain is almost the

same, including biomass harvesting, logistics, and pretreatment

7.2. Operational uncertainty into pellet, chip, or pyrolysis oil. Therefore, the aforementioned

optimization tools for the upstream of the biofuel supply chain also

Compared to the strategic uncertainties, the operational uncer- apply in this case. Furthermore, the downstream of the biomass-to-

tainties are encountered thus need to be taken care of on a more power-and-heat supply chain is identical to the current distribution

frequent basis. The most recognized type of operational uncer- system for power and heat. Biopower can be easily integrated with

tainty is the volatility in prices and costs, namely purchasing prices the grid and serve the loads in remote areas, while heat can be car-

of biomass resources at farms, bills for utilities (e.g., natural gas, ried by hydrosystems to serve nearby facilities. Hence, the major

electricity, stream, etc.) and costs associated with various activi- modeling challenge is at the production phase.

ties (e.g., labor, transport fuel, etc.). Note that pricing uncertainties According to the technology characterizations developed

are different between biomass and petroleum. Petroleum feed- by NREL and EPRI (Electric Power Research Institute), there

stock and product prices are correlated. However, biomass products are three primary conversion routes for CHP applications of

(biofuels) are coordinated to petroleum fuels, while biomass feed- biomass—direct combustion, gasiﬁcation, and coﬁring. Moreover,

stock costs are more correlated to climate and weather. In addition, there are numerous types of boiler technologies, e.g., spreader

because of the variability in biomass feedstocks and the lack of stocker, ﬂuidized-bed boiler, cyclone boiler, etc., which vary in cost,

experienced workers, the uncertainty in process yields associated efﬁciency, technology maturity, as well as sustainability perfor-

with disturbances in operating conditions, errors in operation, etc. mances (Bain, Amos, Downing, & Perlack, 2003; Wiltsee, 2000).

can lead to unqualiﬁed biofuel products and delay in order deliv- Binary variables can be used to facilitate the optimal selection

ery. We suggest probability theory and statistical tools can help to of these routes and technologies in the supply chain system, in

explore the uncertainty dynamics. Whereas, robust optimization terms of economic, environmental, and/or social metrics. In sce-

may provide assistance to ensure satisfactory performances of bio- narios where biomass is co-ﬁred with coal, scheduling tools can

fuel supply chains under volatile market environment and unstable help to coordinate the two material streams. Additionally, in sce-

process outcomes. Another two types of operational uncertainty – narios where existing fossil-based power plants are to be retroﬁtted

supply delays and demand ﬂuctuation – are common topics in the and renovated for CHP from biomass, multi-period planning tools

operations research community. Supply delays, also known as lead would contribute to the optimization of project evaluation, con-

time variability, often occur during batch production and mate- struction planning, resource allocation, etc. Superstructure-based

rial transportation, while demand ﬂuctuation is closely related to optimization always plays an important role in the process design,

the local consumption behavior and regional economy structure. no matter for bioreﬁneries or biomass-based CHP facilities.

There is no doubt that accurate forecasting models are critical

to make biofuel companies well-informed of the future market. 8.2. Biomass-to-chemicals supply chain

Metrics such as conditional value-at-risk (Rockafellar & Uryasev,

2002) and downside risk (Eppen, Martin, & Schrage, 1989), coupled The petrochemical industry makes a wide range of products

with the corresponding models and algorithms (You, Wassick, & from fossil fuels. These products include plastics, ﬁbers, solvents,

Grossmann, 2009), can also be employed to control and manage the and other products that are ubiquitous in our daily life. Actually,

risks at the optimal design and operation of biofuel supply chains. In the same or similar products can, for the most part, be potentially

addition, optimization models based on stochastic inventory theory made from biomass as well (NREL, 2013a; PNNL, 2013). The U.S.

(Graves & Willems, 2000, 2005; You & Grossmann, 2008b) might Department of Energy have identiﬁed a list of “Top 10” building

be the most promising approach, because they can explicitly opti- block or platform chemicals in 2004 from more than 300 candidates

mize the working and safety inventory to hedge against the supply (Werpy & Petersen, 2004), which is recently revisited by Bozell

delay and demand ﬂuctuation, whilst guarantee the required ser- and Petersen (2010). At the time when the report was produced,

vice level (You & Grossmann, 2011b; Yue & You, 2013a; You, Pinto, most platform chemicals are sugar-based. However, with the devel-

Grossmann, & Megan, 2011). opment of advanced biofuel technologies, it becomes appealing

to produce high-value-added low-volume bioproducts accompany

8. Supply chains of non-fuel end-uses for biomass with the high-volume but low-value hydrocarbon biofuels. Differ-

derivatives ent from petroleum feedstock, biomass has much higher oxygen

content than hydrocarbons fuels. Therefore, its conversion to fuels

Besides the production of biofuels for the transportation sector, requires release of the oxygen content either as CO2 or other oxy-

biomass sources also provide a sustainable alternative for produc- genated species, which might lead to distinct synergeis to produce

ing electric power and heat. Furthermore, valuable co-products and highly-oxygenated biobased chemicals along with drop-in biofuels.

byproducts can be obtained accompany with the production of bio- From fossil sources, these compounds would have been produced

fuels, which leads to additional proﬁt for the bioreﬁnery industry. by oxydizing reduced hydrocarbons (Petersen, Bozell, & White,

2013).

8.1. Biopower and biomass-to-heat supply chain To enhance the overall proﬁtability and productivity and

stimulate expansion of the bioreﬁning industry, multi-scale mod-

Among the renewable power generation options, biopower eling and optimization can play an important role in integrating

offers the advantage of being dispatchable, that is, the biomass can the production of biofuel and advanced bioproducts. Systematic

be stored for the use of power plant operation to serve the load. approaches can be employed to make the assessment, screening

This avoids many of the grid integration challenges associated with and identiﬁcation of promising platform chemicals and reaction

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 53

pathways more efﬁcient. Furthermore, process synthesis and inte- Akgul, O., Zamboni, A., Bezzo, F., Shah, N., & Papageorgiou, L. G. (2011). Optimization-

based approaches for bioethanol supply chains. Industrial & Engineering

gration and multi-physics process simulation can help to lower the

Chemistry Research, 50, 4927–4938.

overall energy intensity of unit operations in bioreﬁneries, maxi-

Aksoy, B., Cullinan, H., Webster, D., Gue, K., Sukumaran, S., Eden, M., & Sammons,

mize the use of all feedstock components, byproducts and waste N. (2011). Woody biomass and mill waste utilization opportunities in Alabama:

Transportation cost minimization, optimum facility location, economic feasibil-

streams, and use economies of scale, common processing opera-

ity and impact. Environmental Progress & Sustainable Energy, 30, 720–732.

tions, materials, and equipment to drive down all production costs.

Althaus, H.-J., Bauer, C., Doka, G., Dones, R., Frischknecht, R., Hellweg, S., Humbert,

Since the additional material streams related to advanced bioprod- S., Jungbluth, N., Köllner, T., Loerincik, Y., Margni, M., & Nemecek, T. (2010). In

R. Hischier, & B. Weidema (Eds.), Implementation of life cycle impact assessment

ucts may be similar to those in the petroleumchemical supply

methods. EcoInvent Centre, Swiss Centre for Life Cycle Invetories.

chains, we can also leverage the tools on planning and schedul-

An, H. J., Wilhelm, W. E., & Searcy, S. W. (2011a). Biofuel and petroleum-based fuel

ing to coordinate the use of resources in production and optimize supply chain research: A literature review. Biomass & Bioenergy, 35, 3763–3774.

An, H. J., Wilhelm, W. E., & Searcy, S. W. (2011b). A mathematical model to design a

the logistic activities in the downstream distribution system. More-

lignocellulosic biofuel supply chain system with a case study based on a region

over, the upstream of the biomass-to-chemicals supply chain is the

in Central Texas. Bioresource Technology, 102, 7860–7870.

same as the biofuel supply chain, thus the aforementioned model- Anex, R. P., Aden, A., Kazi, F. K., Fortman, J., Swanson, R. M., Wright, M. M., Satrio,

J. A., Brown, R. C., Daugaard, D. E., Platon, A., Kothandaraman, G., Hsu, D. D.,

ing and optimization methods for biomass harvesting and logistics

& Dutta, A. (2010). Techno-economic comparison of biomass-to-transportation

also apply in this case.

fuels via pyrolysis, gasiﬁcation, and biochemical pathways. Fuel, 89(Supplement

1), S29–S35.

Avami, A. (2012). A model for biodiesel supply chain: A case study in Iran. Renewable

9. Conclusion

& Sustainable Energy Reviews, 16, 4196–4203.

Awudu, I., & Zhang, J. (2012). Uncertainties and sustainability concepts in biofuel

supply chain management: A review. Renewable & Sustainable Energy Reviews,

Although ethanol and bio-diesel are the most worldwide

16, 1359–1368.

recognized biomass-derived liquid transportation fuel products

Azapagic, A. (1999). Life cycle assessment and its application to process selection

currently, it is foreseeable that in the near future advanced hydro- design and optimisation. Chemical Engineering Journal, 73, 1–21.

Azapagic, A., & Clift, R. (1999). The application of life cycle assessment to process

carbon biofuels would experience a rapid growth due to the

optimisation. Computers & Chemical Engineering, 23, 1509–1526.

political expectation, economic promise, environmental require-

Bai, Y., Hwang, T. S., Kang, S., & Ouyang, Y. F. (2011). Biofuel reﬁnery location and

ment, and social beneﬁts. As reviewed in this paper, cellulosic and supply chain planning under trafﬁc congestion. Transportation Research Part B:

algae-based hydrocarbon biofuels have numerous advantages over Methodological, 45, 162–175.

Bai, Y., Ouyang, Y. F., & Pang, J. S. (2012). Biofuel supply chain design under compet-

ethanol and bio-diesel in compatibility with existing distribution

itive agricultural land use and feedstock market equilibrium. Energy Economics,

infrastructure, implication on food price and land competition, and

34, 1623–1633.

potential for substituting various high-value petrochemicals. The BainF R.L., Amos, W. A., Downing, M., & Perlack, R. L. (2003). Biopower techni-

cal assessment: State of the industry and technology. In NREL/TP-510-33123.

multi-scale modeling and optimization framework can preserve

National Renewable Energy Laboratory.

a holistic view and facilitate informed systematical researches in

Bakshi, B. R., Fiksel, J., Baral, A., Bolinger, J., Mishchenko, O., Singh, S., Ukidwe,

ﬁelds of molecular engineering, unit operation, process design, N., Urban, B., & Zhang, Y. (2013). Eco-LCA: Ecologically-based life cycle assess-

ment. OH: Process Systems Engineering Group, and Center of Resilience, The

supply chain management, and sustainability assessment. Oppor-

Ohio State University., http://resilience.eng.ohio-state.edu/eco-lca/. Accessed

tunities and challenges emerge from multiple spatial and temporal

on: 7/17/2013.

scales in the biofuel supply chain systems are discussed. Per- Bals, B., Rogers, C., Jin, M., Balan, V., & Dale, B. (2010). Evaluation of ammonia ﬁbre

expansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass har-

spectives on centralized-distributed network design, multi-player

vested in different seasons and locations. Biotechnology for Biofuels, 3, 1.

interaction, and integration with other supply chain systems are

Bamuﬂeh, H., Ponce-Ortega, J., & El-Halwagi, M. (2013). Multi-objective opti-

revealed in this paper. Additionally, we classify decisions into mization of process cogeneration systems with economic, environmental

and social tradeoffs. Clean Technologies and Environmental Policy, 15,

strategic and operational levels according to their inﬂuences on

185–197.

the temporal scale. We also investigate issues associated with

Baumrucker, B. T., & Biegler, L. T. (2010). MPEC strategies for cost optimization of

biofuel supply chains in the threefold sustainability—economy, pipeline operations. Computers & Chemical Engineering, 34, 900–913.

environment, and society, as well as the uncertainty and risks. Non- Benoît, C., Norris, G., Valdivia, S., Ciroth, A., Moberg, A., Bos, U., Prakash, S.,

Ugaya, C., & Beck, T. (2010). The guidelines for social life cycle assessment of

food end-uses of biomass derivatives are also introduced in this

products: Just in time!. The International Journal of Life Cycle Assessment, 15,

paper, identifying the opportunities for power and heat supply and 156–163.

advanced bioproducts. Bernardi, A., Giarola, S., & Bezzo, F. (2012). Optimizing the economics and the car-

bon and water footprints of bioethanol supply chains. Biofuels. Bioproducts and

Bioreﬁning, 6, 656–672.

Acknowledgement Bernardi, A., Giarola, S., & Bezzo, F. (2013). Spatially explicit multiobjective optimiza-

tion for the strategic design of ﬁrst and second generation bioreﬁneries including

carbon and water footprints. Industrial & Engineering Chemistry Research, 52,

The authors gratefully acknowledge ﬁnancial support from the 7170–7180.

Initiative for Sustainability and Energy at Northwestern University Biddy, M., & Jones, S. (2013). Catalytic upgrading of sugars to hydrocarbons technol-

ogy pathway. In NREL/TP-5100-58055, PNNL-22319. National Renewable Energy

(ISEN) and Argonne National Laboratory via a Northwestern-

Laboratory, Paciﬁc Northwest National Laboratory.

Argonne Early Career Investigator Award for Energy Research. The

Birge, J. R., & Louveaux, F. V. (1997). Introduction to stochastic programming. Springer.

submitted manuscript has been created by UChicago Argonne, LLC, Bowling, I. M., Ponce-Ortega, J. M., & El-Halwagi, M. M. (2011). Facility location and

supply chain optimization for a bioreﬁnery. Industrial & Engineering Chemistry

Operator of Argonne National Laboratory (“Argonne”). Argonne, a

Research, 50, 6276–6286.

U.S. Department of Energy Ofﬁce of Science laboratory, is operated

Bozell, J. J., & Petersen, G. R. (2010). Technology development for the production

under Contract No. DE-AC02-06CH11357. The U.S. Government of biobased products from bioreﬁnery carbohydrates—The US Department of

Energy’s ‘Top 10’ revisited. Green Chemistry, 12, 539–554.

retains for itself, and others acting on its behalf, a paid-up nonex-

Bunting, B., Bunce, M., Barone, T., & Storey, J. (2010). Fungible and compatible biofuels:

clusive, irrevocable worldwide license in said article to reproduce,

Literature search, summary, and recommendations. Oak Ridge: Oak Ridge National

prepare derivative works, distribute copies to the public, and Laboratory, Fuels, Engines, and Emissions Research Center.

Chen, C. W., & Fan, Y. Y. (2012). Bioethanol supply chain system planning under

perform publicly and display publicly, by or on behalf of the Gov-

supply and demand uncertainties. Transportation Research Part E: Logistics and

ernment.

Transportation Review, 48, 150–164.

Chen, X., & Onal, H. (2012). An economic analysis of the future U.S. bio-

fuel industry, facility location, and supply chain network. Available at SSRN:

References http://dx.doi.org/10.2139/ssrn.2084313.

Chiu, Y.-W., Suh, S., Pﬁster, S., Hellweg, S., & Koehler, A. (2012). Measuring ecological

Akgul, O., Shah, N., & Papageorgiou, L. G. (2012). Economic optimisation of a UK impact of water consumption by bioethanol using life cycle impact assessment.

advanced biofuel supply chain. Biomass & Bioenergy, 41, 57–72. The International Journal of Life Cycle Assessment, 17, 16–24.

54 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Chu, Y., & You, F. (2012). Integration of scheduling and control with online Engell, S., & Harjunkoski, I. (2012). Optimal operation: Scheduling advanced control

closed-loop implementation: Fast computational strategy and large-scale global and their integration. Computers & Chemical Engineering, 47, 121–133.

optimization algorithm. Computers & Chemical Engineering, 47, 248–268. Eppen, G. D., Martin, R. K., & Schrage, L. (1989). A scenario approach to capacity

Corsano, G., Vecchietti, A. R., & Montagna, J. M. (2011). Optimal design for sustain- planning. Operations Research, 37, 517–527.

able bioethanol supply chain considering detailed plant performance model. EPRI. (2010). Biopower generation: Biomass issues, fuels, technologies, and research,

Computers & Chemical Engineering, 35, 1384–1398. development, demonstration, and deployment opportunities. Electric Power

Cuˇ cek,ˇ L., Klemes,ˇ J. J., & Kravanja, Z. (2012). A review of footprint analysis tools for Research Institute.

monitoring impacts on sustainability. Journal of Cleaner Production, 34, 9–20. Esri. (2012). ArcGIS—Mapping and spatial analysis for understanding our world. ,

Cucek, L., Lam, H. L., Klemes, J. J., Varbanov, P. S., & Kravanja, Z. (2010). Synthesis of http://www.esri.com/software/arcgis.

regional networks for the supply of energy and bioproducts. Clean Technologies Fahy, E., Subramaniam, S., Brown, H. A., Glass, C. K., Merrill, A. H., Murphy, R. C., Raetz,

and Environmental Policy, 12, 635–645. C. R. H., Russell, D. W., Seyama, Y., Shaw, W., Shimizu, T., Spener, F., van Meer,

Cuˇ cek,ˇ L., Varbanov, P. S., Klemes, J. J., & Kravanja, Z. (2012). Total footprints-based G., VanNieuwenhze, M. S., White, S. H., Witztum, J. L., & Dennis, E. A. (2005). A

multi-criteria optimisation of regional biomass energy supply chains. Energy, comprehensive classiﬁcation system for lipids. European Journal of Lipid Science

44, 135–145. and Technology, 107, 337–364.

Cundiff, J. S., Dias, N., & Sherali, H. D. (1997). A linear programming approach for FAO. (2009). Algae-based biofuels: A review of challenges and opportunities for devel-

designing a herbaceous biomass delivery system. Bioresource Technology, 59, oping countries. Rome, Italy: Food and Agriculture Organization of the United

47–55. Nations.

Dal-Mas, M., Giarola, S., Zamboni, A., & Bezzo, F. (2011). Strategic design and invest- Fava, J. A. (1994). A techical framework for life-cycle assessment. Pensacola: SETAC

ment capacity planning of the ethanol supply chain under price uncertainty. Foudation.

Biomass & Bioenergy, 35, 2059–2071. Ferreira, L., & Borenstein, D. (2011). Normative agent-based simulation for supply

Daoutidis, P., Marvin, W. A., Rangarajan, S., & Torres, A. I. (2013). Engineering biomass chain planning. Journal of the Operational Research Society, 62, 501–514.

conversion processes: A systems perspective. AIChE Journal, 59, 3–18. Flisberg, P., Frisk, M., & Ronnqvist, M. (2012). FuelOpt: A decision support system for

Davis, R., Aden, A., & Pienkos, P. T. (2011). Techno-economic analysis of autotrophic forest fuel logistics. Journal of the Operational Research Society, 63, 1600–1612.

microalgae for fuel production. Applied Energy, 88, 3524–3531. Floudas, C. A. (1999). Deterministic global optimization: Theory, methods and applica-

Davis, R., Biddy, M., & Jones, S. (2013). Algal lipid extraction and upgrading to hydro- tions. Boston: Kluwer Academic Publishers.

carbons technology pathway. In NREL/TP-5100-58049, PNNL-22315. NREL. Frombo, F., Mindardi, R., Robba, M., Rosso, F., & Sacile, R. (2009). Planning woody

Davis, R., Biddy, M., Tan, E., Tao, L., & Jones, S. (2013). Biological conversion of sug- biomass logistics for energy production: A strategic decision model. Biomass &

ars to hydrocarbons technology pathway. In NREL/TP-5100-58054, PNNL-22318. Bioenergy, 33, 372–383.

National Renewable Energy Laboratory, Paciﬁc Northwest National Laboratory. GaBi. (2011). Software and database contents for life cycle engineering. PE Interna-

DOE. (2010). Report to congress, dedicated ethanol pipeline feasibility study. U.S. tional., http://www.gabi-software.com.

Department of Energy. Gan, J. B., & Smith, C. T. (2011). Optimal plant size and feedstock supply radius: A

DOE/EERE. (2010a). Biopower technical strategy workshop summary report. Denver, modeling approach to minimize bioenergy production costs. Biomass & Bioen-

CO (2009): U.S. Department of Energy, Ofﬁce of Energy Efﬁciency & Renewable ergy, 35, 3350–3359.

Energy. Gebreslassie, B. H., Slivinsky, M., Wang, B., & You, F. (2013). Life cycle optimization for

DOE/EERE. (2010b). National algal biofuels technology roadmap. U.S. Department of sustainable design and operations of hydrocarbon bioreﬁnery via fast pyrolysis

Energy, Ofﬁce of Energy Efﬁciency & Renewable Energy. hydrotreating and hydrocracking. Computers & Chemical Engineering, 50, 71–91.

DOE/EERE. (2011a). Biochemical conversion: Using hydrolysis, fermentation, and catal- Gebreslassie, B. H., Waymire, R., & You, F. (2013). Sustainable design and synthesis of

ysis to make fuels and chemicals. U.S. Department of Energy, Ofﬁce of Energy algae-based bioreﬁnery for simultaneous hydrocarbon biofuel production and

Efﬁciency & Renewable Energy. carbon sequestration. AIChE JournalV 59, 1599–1621.

DOE/EERE. (2011b). Thermochemical conversion: Using heat and chemistry to make Gebreslassie, B. H., Yao, Y., & You, F. (2012). Design under uncertainty of hydrocar-

fuels and bioproducts. U.S. Department of Energy, Office of Energy Efficiency & bon biorefinery supply chains: Multiobjective stochastic programming models,

Renewable Energy. decomposition algorithm and a Comparison between CVaR and downside risk.

DOE/EERE. (2011c). U.S. billion-ton update: Biomass supply for a bioenergy and AIChE Journal, 58, 2155–2179.

bioproducts industry. In R. D. Perlack, & B. J. Stokes (Eds.), ORNL/TM-2011/224 Gerbens-Leenes, W., & Hoekstra, A. Y. (2011). The water footprint of biofuel-based

(p. 227). Oak Ridge, TN: Oak Ridge National Laboratory. transport. Energy & Environmental Science, 4, 2658–2668.

DOE/EERE. (2013a). Advanced uniform-format feedstock supply system. U.S. Depart- Giarola, S., Zamboni, A., & Bezzo, F. (2011). Spatially explicit multi-objective

ment of Energy, Office of Energy Efficiency & Renewable Energy. Available at: optimisation for design and planning of hybrid first and second generation

http://www1.eere.energy.gov/bioenergy/biomass feedstocks.html#Advanced bioreﬁneries. Computers & Chemical Engineering, 35, 1782–1797.

Uniform Format Feedstock Supply (accessed on 7/9/2013). Gjerdrum, J., Shah, N., & Papageorgiou, L. G. (2001). Transfer prices for multienter-

DOE/EERE. (2013b). Federal and state laws and incentives. Department of prise supply chain optimization. Industrial & Engineering Chemistry Research, 40,

Energy, Ofﬁce of Energy Efﬁciency & Renewable Energy. Available at: 1650–1660.

http://www.afdc.energy.gov/laws/ (accessed on: 7/19/2013). Goedkoop, M., & Spriensma, R. (2001). Eco-indicator 99: A damage oriented method

DOE/EERE. (2013c). Feedstock supply and logistics: Biomass as a commodity. U.S. for life cycle impact assessment; methodology report. PRé, Product Ecology Con-

Department of Energy, Ofﬁce of Energy Efﬁciency & Renewable Energy. sultants.

DOE/EERE. (2013d). Replacing the whole barrel: To reduce U.S. dependence on oil. U.S. Gold, S., & Seuring, S. (2011). Supply chain and logistics issues of bio-energy produc-

Department of Energy, Ofﬁce of Energy Efﬁciency & Renewable Energy. tion. Journal of Cleaner Production, 19, 32–42.

DOE/NETL. (2010). DOE/NETL carbon dioxide capture and storage RD&D roadmap. U.S. Gopalakrishnan, A., & Biegler, L. T. (2013). Economic nonlinear model predictive

Department of Energy and National Energy Technology Laboratory. control for periodic optimal operation of gas pipeline networks. Computers &

Dreyer, L., Hauschild, M., & Schierbeck, J. (2006). A Framework for Social Life Cycle Chemical Engineering, 52, 90–99.

Impact Assessment (10 pp). The International Journal of Life Cycle Assessment, 11, Graves, S. C., & Willems, S. P. (2000). Optimizing strategic safety stock place-

88–97. ment in supply chains. Manufacturing & Service Operations Management, 2,

Dunnett, A., Adjiman, C., & Shah, N. (2007). Biomass to heat supply chains applica- 68–83.

tions of process optimization. Process Safety and Environmental Protection, 85, Graves, S. C., & Willems, S. P. (2005). Optimizing the supply chain conﬁguration for

419–429. new products. Management Science, 51, 1165–1180.

Dunnett, A., Adjiman, C., & Shah, N. (2008). A spatially explicit whole-system model GreenDelta. (2013). Earthster 2 turbo. In Tools & consulting for sustain-

of the lignocellulosic bioethanol supply chain: An assessment of decentralised ability. GreenDelta. Available at: http://www.greendeltatc.com/Earthster-2-

processing potential. Biotechnology for Biofuels, 1. Turbo.200.0.html (accessed on: 7/16/2013).

Dutta, P. K. (1999). Strategies and games: Theory and practice. Massachusetts Institute GREET. (2012). The greenhouse gases, regulated emissions, and energy use in trans-

of Technology. portation model. Argonne National Laboratory., http://greet.es.anl.gov/.

Ecoinvent. (2008). The ecoinvent database V2.0. In Technical Report. Swiss Centre for Grossmann, I. (2002). Review of nonlinear mixed-integer and disjunctive program-

Life Cycle Inventories., http://www.ecoinvent.ch/. ming techniques. Optimization and Engineering, 3, 227–252.

EISA. (2007). Energy Independence and Security Act of 2007. Public Law 110–140, 110th Grossmann, I. E., & Guillén-Gosálbez, G. (2010). Scope for the application of math-

Congress. EISA. ematical programming techniques in the synthesis and planning of sustainable

Eksioglu, S. D., Acharya, A., Leightley, L. E., & Arora, S. (2009). Analyzing the design processes. Computers & Chemical Engineering, 34, 1365–1376.

and management of biomass-to-bioreﬁnery supply chain. Computers & Industrial Hasan, M. R., & Chakrabarti, R. (2009). Use of algae and aquatic macrophytes as feed in

Engineering, 57, 1342–1352. small-scale aquaculture: A review. Rome, Italy: Food and Agriculture Organization

Eksioglu, S. D., Li, S., Zhang, S., Sokhansanj, S., & Petrolia, D. (2010). Analyzing impact of the United Nations.

of intermodal facilities on design and management of biofuel supply chain. Hosseini, S. A., & Shah, N. (2011). Multi-scale process and supply chain mod-

Transportation Research Record, 144–151. elling: From lignocellulosic feedstock to process and products. Interface Focus,

Elia, J. A., Baliban, R. C., & Floudas, C. A. (2012). Nationwide energy supply chain anal- 1, 255–262.

ysis for hybrid feedstock processes with signiﬁcant CO2 emissions reduction. Huang, Y. X., Chen, C. W., & Fan, Y. Y. (2010). Multistage optimization of the supply

AIChE Journal, 58, 2142–2154. chains of biofuels. Transportation Research Part E: Logistics and Transportation

Elia, J. A., Baliban, R. C., Xiao, X., & Floudas, C. A. (2011). Optimal energy supply net- Review, 46, 820–830.

work determination and life cycle analysis for hybrid coal, biomass and natural Huijbregts, M., Heijungs, R., & Hellweg, S. (2004). Complexity and integrated

gas to liquid (CBGTL) plants using carbon-based hydrogen production. Comput- resource management: Uncertainty in LCA. The International Journal of Life Cycle

ers & Chemical Engineering, 35, 1399–1430. Assessment, 9, 341–342.

D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56 55

Humbert, S., Schryver, A. D., Bengoa, X., Margni, M., & Jolliet, O. (2012). IMPACT NABC. (2013). National advanced biofuels consortium process strategies. National

2002+: User guide draft for version Q2.21. Quantis., http://www.quantis- Advanced Biofuels Consortium. Available at: http://www.nabcprojects.org/

intl.com/index.php. process strategies.html (accessed on: 7/19/2013).

Iakovou, E., Karagiannidis, A., Vlachos, D., Toka, A., & Malamakis, A. (2010). Waste Nash, J. F., Jr. (1950). The bargaining problem. Econometrica, 18, 155–162.

biomass-to-energy supply chain management: A critical synthesis. Waste Man- NCEP. (2009). Task force on biofuels infrastructure. National Commission on Energy

agement, 30, 1860–1870. Policy.

IMPLAN. (2013). IMPLAN professional model. IMPLAN Group LLC. Available at: NREL. (2012). JEDI models: Jobs and economic development impact models. NREL.

http://implan.com/V4/Index.php (accessed date: 7/16/2013). NREL. (2013a). Bioproducts. National Renewable Energy Laboratory. Available at:

IPCC. (2007). Glossary in fourth assessment report: Climate change 2007. , http://www.nrel.gov/learning/re bioproducts.html (accessed on 7/15/2013).

http://www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4 syr appendix.pdf. NREL. (2013b). Energy analysis—International trade of wood pellets. In NREL/BR-

ISO. (2006). ISO 14040: Environmental management-life cycle assessment-principles 6A20-56791. National Renewable Energy Laboratory.

and frame work. International Standard: ISO Standards. openLCA. (2012). A new open source software for life cycle assessment. GreenDeltaTC

JonesF S.B., Holladay, J. E., Valkenburg, C., Stevens, D. J., Walton, C. W., Kinchin, C., GmbH., http://www.openlca.org.

Elliott, D. C., & Czernik, S. (2009). Production of gasoline and diesel from biomass via Papapostolou, C., Kondili, E., & Kaldellis, J. K. (2011). Development and implementa-

fast pyrolysis, hydrotreating and hydrocracking: A design case. Paciﬁc Northwest tion of an optimisation model for biofuels supply chain. Energy, 36, 6019–6026.

National Laboratory. PNNL-18284. Parker, N., Tittmann, P., Hart, Q., Nelson, R., Skog, K., Schmidt, A., Gray, E., & Jenkins,

Jørgensen, A., Bocq, A., Nazarkina, L., & Hauschild, M. (2008). Methodologies for B. (2010). Development of a bioreﬁnery optimized biofuel supply curve for the

social life cycle assessment. The International Journal of Life Cycle Assessment, Western United States. Biomass & Bioenergy, 34, 1597–1607.

13, 96–103. Petersen, G., Bozell, J., & White, J. (2013). Selecting targets. In Catalytic process devel-

Junginger, M., Bolkesjø, T., Bradley, D., Dolzan, P., Faaij, A., Heinimö, J., Hektor, opment for renewable materials. Wiley-VCH Verlag GmbH & Co. KGaA.

B., Leistad, Ø., Ling, E., Perry, M., Piacente, E., Rosillo-Calle, F., Ryckmans, Y., PNNL. (2013). Bio-based product research. Paciﬁc Northwest National Laboratory.

Schouwenberg, P.-P., Solberg, B., Trømborg, E., Walter, A. d. S., & Wit, M. d. Available at: http://www.pnnl.gov/biobased/ (accessed on 7/15/2013).

(2008). Developments in international bioenergy trade. Biomass and Bioenergy, Rajagapol, D., Sexton, S., Hochman, G., Roland-Holst, D., & Zilberman, D. D. (2009).

32, 717–729. Model estimates food-versus-biofuel trade-off. California Agriculture, 63.

Kaditi, E., & Swinnen, J. (2007). International biomass trade—What are Rentizelas, A. A., Tatsiopoulos, I. P., & Tolis, A. (2009). An optimization model for

sound policy approaches to guarantee a fair trade of biomass with- multi-biomass tri-generation energy supply. Biomass & Bioenergy, 33, 223–233.

out further depletion of natural resources in exporting countries? In Rentizelas, A. A., Tolis, A. J., & Tatsiopoulos, I. P. (2009). Logistics issues of biomass:

Agrinergy Workshop CEPS, Brussels, November 20, 2007,. Available at: The storage problem and the multi-biomass supply chain. Renewable & Sustain-

http://agrinergy.ecologic.eu/download/kaditi agrinergy trade.pdf. able Energy Reviews, 13, 887–894.

Keirstead, J., Samsatli, N., Pantaleo, A. M., & Shah, N. (2012). Evaluating biomass Ringer, M., Putsche, V., & Scahill, J. (2006). Large-scale pyrolysis oil production: A

energy strategies for a UK eco-town with an MILP optimization model. Biomass technology assessment and economic analysis. In NREL/TP-510-37779. National

& Bioenergy, 39, 306–316. Renewable Energy Laboratory.

Kim, J., Realff, M. J., & Lee, J. H. (2011). Optimal design and global sensitivity analysis Rockafellar, R. T., & Uryasev, S. (2002). Conditional value-at-risk for general loss

of biomass supply chain networks for biofuels under uncertainty. Computers & distributions. Journal of Banking & Finance, 26, 1443–1471.

Chemical Engineering, 35, 1738–1751. Sahinidis, N. V. (2004). Optimization under uncertainty: State-of-the-art and oppor-

Kim, J., Realff, M. J., Lee, J. H., Whittaker, C., & Furtner, L. (2011). Design of biomass tunities. Computers & Chemical Engineering, 28, 971–983.

processing network for biofuel production using an MILP model. Biomass & Santibanez-Aguilar, J. E., Gonzalez-Campos, J. B., Ponce-Ortega, J. M., Serna-Gonzalez,

Bioenergy, 35, 853–871. M., & El-Halwagi, M. M. (2011). Optimal planning of a biomass conversion sys-

Lamers, P., Junginger, M., Hamelinck, C., & Faaij, A. (2012). Developments in interna- tem considering economic and environmental aspects. Industrial & Engineering

tional solid biofuel trade—An analysis of volumes, policies, and market factors. Chemistry Research, 50, 8558–8570.

Renewable and Sustainable Energy Reviews, 16, 3176–3199. Sharma, B., Ingalls, R. G., Jones, C. L., & Khanchi, A. (2013). Biomass supply chain

Leduc, S., Lundgren, J., Franklin, O., & Dotzauer, E. (2010). Location of a biomass based design and analysis: Basis, overview, modeling, challenges, and future. Renew-

methanol production plant: A dynamic problem in northern Sweden. Applied able and Sustainable Energy Reviews, 24, 608–627.

Energy, 87, 68–75. Sharma, P., Sarker, B. R., & Romagnoli, J. A. (2011). A decision support tool for strate-

Leduc, S., Natarajan, K., Dotzauer, E., McCallum, I., & Obersteiner, M. (2009). Opti- gic planning of sustainable bioreﬁneries. Computers & Chemical Engineering, 35,

mizing biodiesel production in India. Applied Energy, 86, S125–S131. 1767–1781.

Leduc, S., Starfelt, F., Dotzauer, E., Kindermann, G., McCallum, I., Obersteiner, M., & Shastri, Y. N., Rodriguez, L. F., Hansen, A. C., & Ting, K. C. (2012). Impact of distributed

Lundgren, J. (2010). Optimal location of lignocellulosic ethanol reﬁneries with storage and pre-processing on Miscanthus production and provision systems.

polygeneration in Sweden. Energy, 35, 2709–2716. Biofuels Bioproducts & Bioreﬁning-Biofpr, 6, 21–31.

Lee, J. H. (2012). Energy supply chain optimization: A challenge for control engineers. Shipley, A., Hampson, A., Hedman, B., Garland, P., & Bautista, P. (2008). Combined

In V. Kariwala, L. Samavedham, & R. D. Braatz (Eds.), Eighth IFAC international heat and power—Effective energy solutions for a sustainable future. In ORNL/TM-

symposium on advanced control of chemical processes (pp. 361–370). Furama 2008/224. Oak Ridge National Laboratory.

Riverfront, Singapore: International Federation of Automatic Control. Sokhansanj, S., Mani, S., Turhollow, A., Kumar, A., Bransby, D., Lynd, L., & Laser, M.

Lira-Barragán, L. F., Ponce-Ortega, J. M., Serna-González, M., & El-Halwagi, M. M. (2009). Large-scale production, harvest and logistics of switchgrass (Panicum

(2013). Synthesis of integrated absorption refrigeration systems involving eco- virgatum L.)—Current technology and envisioning a mature technology. Biofuels,

nomic and environmental objectives and quantifying social beneﬁts. Applied Bioproducts and Bioreﬁning, 3, 124–141.

Thermal Engineering, 52, 402–419. Swanson, R. M., Satrio, J. A., Brown, R. C., Platon, A., & Hsu, D. D. (2010). Techno-

Lundquist, T. J., Woertz, I. C., Quinn, N. W. T., & Benemann, J. R. (2010). A realistic economic analysis of biofuels production based on gasiﬁcation. In NREL/TP-6A20-

technology and engineering assessment of algae biofuel production. Berkeley, CA: 46587. National Renewable Energy Laboratory.

Energy Biosciences Institute, University of California. Talmadge, M., Biddy, M., Dutta, A., Jones, S., & Meyer, A. (2013). Syngas upgrading

Marquardt, W., Harwardt, A., Hechinger, M., Kraemer, K., Viell, J., & Voll, A. (2010). to hydrocarbon fuels technology pathway. In NREL/TP-5100-58052, PNNL-22323.

The biorenewables opportunity—Toward next generation process and product National Renewable Energy Laboratory, Paciﬁc Northwest National Laboratory.

systems. AIChE Journal, 56, 2228–2235. Tatsiopoulos, I. P., & Tolis, A. J. (2003). Economic aspects of the cotton-stalk biomass

Marvin, W. A., Schmidt, L. D., Benjaafar, S., Tiffany, D. G., & Daoutidis, P. (2012). logistics and comparison of supply chain methods. Biomass & Bioenergy, 24,

Economic optimization of a lignocellulosic biomass-to-ethanol supply chain. 199–214.

Chemical Engineering Science, 67, 68–79. Tittmann, P. W., Parker, N. C., Hart, Q. J., & Jenkins, B. M. (2010). A spatially explicit

MASBI. (2013). Fueling a sustainable future for aviation. Midwest Aviation Sustainable techno-economic model of bioenergy and biofuels production in California. Jour-

Biofuels Initiative., www.masbi.org. nal of Transport Geography, 18, 715–728.

McNew, K., & Grifﬁth, D. (2005). Measuring the impact of ethanol plants on local Tong, K., Gleeson, M. J., Rong, G., & You, F. (2013). Optimal design of

grain prices. Review of Agricultural Economics, 27, 164–180. advanced ‘drop-in’ hydrocarbon biofuel supply chain integrating with

Mele, F. D., Guillen-Gosalbez, G., & Jimenez, L. (2009). Optimal planning of supply existing petroleum reﬁneries under uncertainty. Biomass & Bioenergy,

chains for bioethanol and sugar production with economic and environmental http://dx.doi.org/10.1016/j.biombioe.2013.10.023, accepted.

oncerns. In J. Jezowski, & T. Thullie (Eds.), 19th European symposium on computer Tong, K., Gong, J., Yue, D., & You, F. (2013). Stochastic programming

aided process engineering (26) (pp. 997–1002). Amsterdam: Elsevier Science approach to optimal design and operations of integrated hydrocarbon bio-

Bv. fuel and petroleum supply chains. ACS Sustainable Chemistry & Engineering,

Mele, F. D., Kostin, A. M., Guillen-Gosalbez, G., & Jimenez, L. (2011). Multiobjec- http://dx.doi.org/10.1021/sc400267t, accepted.

tive model for more sustainable fuel supply chains. A case study of the sugar UNEP. (2009). Guidelines for social life cycle assessment of products. In United

cane industry in Argentina. Industrial & Engineering Chemistry Research, 50, Nations Environment Programme. United Nations.

4939–4958. USDA. (2007). Ethanol transportation backgrounder. U.S. Department of Agriculture.

Melero, J. A., Iglesias, J., & Garcia, A. (2012). Biomass as renewable feedstock in USDA. (2010). Field crops usual planting and harvesting dates. United States

standard reﬁnery units. Feasibility opportunities and challenges. Energy & Envi- Department of Agriculture, National Agricultural Statistics Service. Agricultural

ronmental Science, 5, 7393–7420. Handbook Number 628.

Mobini, M., Sowlati, T., & Sokhansanj, S. (2011). Forest biomass supply logistics for Uslu, A., Faaij, A. P. C., & Bergman, P. C. A. (2008). Pre-treatment technologies,

a power plant using the discrete-event simulation approach. Applied Energy, 88, and their effect on international bioenergy supply chain logistics. Techno-

1241–1250. economic evaluation of torrefaction, fast pyrolysis and pelletisation. Energy, 33,

Myerson, R. B. (1997). Game theory: Analysis of conﬂict. Harvard University Press. 1206–1223.

56 D. Yue et al. / Computers and Chemical Engineering 66 (2014) 36–56

Velazquez-Marti, B., & Fernandez-Gonzalez, E. (2010). Mathematical algorithms to You, F., & Grossmann, I. E. (2011b). Stochastic inventory management for tacti-

locate factories to transform biomass in bioenergy focused on logistic network cal process planning under uncertainties: MINLP models and algorithms. AIChE

construction. Renewable Energy, 35, 2136–2142. Journal, 57, 1250–1277.

von Stackelberg, H. (1934). Marktform und gleichgewicht. Wien & Berlin: Springer. You, F., Grossmann, I. E., & Wassick, J. M. (2011). Multisite capacity, production and

◦

VI, 138 S. 8 : J. Springer. distributionplanning with reactor modiﬁcations: MILP model, bilevel decompo-

Walther, G., Schatka, A., & Spengler, T. S. (2012). Design of regional production sition algorithm vs. lagrangean decomposition scheme. Industrial & Engineering

networks for second generation synthetic bio-fuel—A case study in Northern Chemistry Research, 50, 4831–4849.

Germany. European Journal of Operational Research, 218, 280–292. You, F., Tao, L., Graziano, D. J., & Snyder, S. W. (2012). Optimal design of

Wang, B., Gebreslassie, B. H., & You, F. (2013). Sustainable design and synthe- sustainable cellulosic biofuel supply chains: Multiobjective optimization cou-

sis of hydrocarbon bioreﬁnery via gasiﬁcation pathway: Integrated life cycle pled with life cycle assessment and input–output analysis. AIChE Journal, 58,

assessment and technoeconomic analysis with multiobjective superstructure 1157–1180.

optimization. Computers & Chemical Engineering, 52, 55–76. You, F., & Wang, B. (2011). Life cycle optimization of biomass-to-liquid supply

Wassick, J. M., Agarwal, A., Akiya, N., Ferrio, J., Bury, S., & You, F. (2012). Address- chains with distributed—Centralized processing networks. Industrial & Engineer-

ing the operational challenges in the development, manufacture, and supply of ing Chemistry Research, 50, 10102–10127.

advanced materials and performance products. Computers & Chemical Engineer- Yue, D., Guillén-Gosálbez, G., & You, F. (2013). Global optimization of

ing, 47, 157–169. large-scale mixed-integer linear fractional programming problems: A

Weidema, B. (2006). The integration of economic and social aspects in life reformulation–linearization method and process scheduling applications.

cycle impact assessment. The International Journal of Life Cycle Assessment, 11, AIChE Journal, 59, 4255–4272.

89–96. Yue, D., Kim, M. A., & You, F. (2013). Design of sustainable product systems and

Werpy, T., & Petersen, G. (2004). Top value added chemicals from biomass volume supply chains with life cycle optimization based on functional unit: General

I—Results of screening for potential candidates from sugars and synthesis gas. U.S. modeling framework, mixed-integer nonlinear programming algorithms and

Department of Energy. case study on hydrocarbon biofuels. ACS Sustainable Chemistry & Engineering, 1,

Wikipedia. (2013a). Equivalent annual cost. In Wikipedia, the free encyclopedia. 1003–1014.

Available at: http://en.wikipedia.org/wiki/Equivalent annual cost (accessed on: You, F., Pinto, J. M., Grossmann, I. E., & Megan, L. (2011). Optimal distribution-

7/16/2013). inventory planning of industrial gases: ii. minlp models and algorithms

Wikipedia. (2013b). Gasoline gallon equivalent. In Wikipedia, the free encyclopedia. for stochastic cases. Industrial & Engineering Chemistry Research, 50,

Available at: http://en.wikipedia.org/wiki/Gasoline gallon equivalent (accessed 2928–2945.

on: 7/16/2013). You, F. Q., Wassick, J. M., & Grossmann, I. E. (2009). Risk management for a global

Wikipedia. (2013c). Net present value. In Wikipedia, the free encyclopedia. Available supply chain planning under uncertainty: Models and algorithms. AIChE Journal,

at: http://en.wikipedia.org/wiki/Net present value (accessed on: 7/16/2013). 55, 931–946.

Wilcox, J. (2012). The role of algae in carbon capture. In Carbon capture sci- Yue, D., & You, F. (2013a). Planning and scheduling of ﬂexible process networks under

ence + business media LLC. Springer (Chapter 7). uncertainty with stochastic inventory: MINLP models and algorithm. AIChE Jour-

Wiltsee, G. (2000). Lessons learned from existing biomass power plants. In NREL/SR- nal, 59, 1511–1532.

570-26946. National Renewable Energy Laboratory. Yue, D., & You, F. (2013b). Sustainable scheduling of batch processes under economic

Wright, M. M., Satrio, J. A., Brown, R. C., Daugaard, D. E., & Hsu, D. D. (2010). Techno- and environmental criteria with minlp models and algorithms. Computers &

economic analysis of biomass fast pyrolysis to transportation fuels. In NREL/TP- Chemical Engineering, 54, 44–59.

6A20-46586. National Renewable Energy Labratory. Zamboni, A., Bezzo, F., & Shah, N. (2009). Spatially explicit static model

Yeomans, H., & Grossmann, I. E. (1999). A systematic modeling framework of super- for the strategic design of future bioethanol production systems.

structure optimization in process synthesis. Computers & Chemical Engineering, 2. Multi-objective environmental optimization. Energy & Fuels, 23,

23(6), 709–731. 5134–5143.

You, F., Castro, P. M., & Grossmann, I. E. (2009). Dinkelbach’s algorithm as an efﬁ- Zamboni, A., Shah, N., & Bezzo, F. (2009). Spatially explicit static model for the strate-

cient method to solve a class of MINLP models for large-scale cyclic scheduling gic design of future bioethanol production systems. 1. Cost minimization. Energy

problems. Computers & Chemical Engineering, 33, 1879–1889. & Fuels, 23, 5121–5133.

You, F., & Grossmann, I. E. (2008a). Design of responsive supply chains under demand Zhong, Z., & You, F. (2014). Globally convergent exact and inexact parametric

uncertainty. Computers & Chemical Engineering, 32, 3090–3111. algorithms for solving large-scale mixed-integer fractional programs and appli-

You, F., & Grossmann, I. E. (2008b). Mixed-integer nonlinear programming models cations in process systems engineering. Computers & Chemical Engineering, 61,

and algorithms for large-scale supply chain design with stochastic inventory 90–101.

management. Industrial & Engineering Chemistry Research, 47, 7802–7817. Zhou, C., Liu, P., & Li, Z. (2013). A superstructure based mixed-integer programing

You, F., & Grossmann, I. E. (2011a). Balancing responsiveness and economics in pro- approach. In A. Kraslawski, & I. Turunen (Eds.), 23rd European symposium on

cess supply chain design with multi-echelon stochastic inventory. AIChE Journal, computer aided process engineering Lappeenranta, Finland.

57, 178–192.