BO

Da

b

a

ARR1AA

KSBBMM

1

uafsupStspEa2foRTtis

0h

Computers and Chemical Engineering 66 (2014) 36–56

Contents lists available at ScienceDirect

Computers and Chemical Engineering

j our na l ho me pa g e: www.elsev ier .com/ locate /compchemeng

iomass-to-bioenergy and biofuel supply chain optimization:verview, key issues and challenges

ajun Yuea, Fengqi Youa,∗, Seth W. Snyderb

Northwestern University, Evanston, IL 60208, USAArgonne National Laboratory, Argonne, IL 60439, USA

r t i c l e i n f o

rticle history:eceived 19 July 2013eceived in revised form4 November 2013ccepted 17 November 2013vailable online 18 December 2013

a b s t r a c t

This article describes the key challenges and opportunities in modeling and optimization of biomass-to-bioenergy supply chains. It reviews the major energy pathways from terrestrial and aquatic biomass tobioenergy/biofuel products as well as power and heat with an emphasis on “drop-in” liquid hydrocarbonfuels. Key components of the bioenergy supply chains are then presented, along with a comprehensiveoverview and classification of the existing contributions on biofuel/bioenergy supply chain optimization.This paper identifies fertile avenues for future research that focuses on multi-scale modeling and opti-

eywords:upply chain modelingiofuelsioenergyathematical programmingulti-scale modeling

mization, which allows the integration across spatial scales from unit operations to biorefinery processesand to biofuel value chains, as well as across temporal scales from operational level to strategic level.Perspectives on future biofuel supply chains that integrate with petroleum refinery supply chains and/orcarbon capture and sequestration systems are presented. Issues on modeling of sustainability and thetreatment of uncertainties in bioenergy supply chain optimization are also discussed.

. Introduction

Biomass-derived liquid transportation fuels and energy prod-cts have been proposed as part of the solution to climate changend our heavy dependence on fossil fuels, because the biomasseedstock can be produced renewably from a variety of domesticources, and the production and use of bioenergy/biofuel prod-cts have potentially lower environmental impacts than theiretroleum counterparts (An, Wilhelm, & Searcy, 2011a; Gold &euring, 2011; Marquardt et al., 2010). Consequently, many coun-ries have set national biofuels targets and provide incentives andupports to accelerate the growth of bioenergy industry. For exam-le, in the U.S. the Renewable Fuels Standard (RFS), part of thenergy Independence and Security Act (EISA) of 2007, establishesn annual production target of 36 billion gallons of biofuels by022 (EISA, 2007). Currently most biofuels in the U.S. are maderom corn starch that might have negative implications in termsf both food prices and production (McNew & Griffith, 2005;ajagapol, Sexton, Hochman, Roland-Holst, & Zilberman, 2009).o avoid adverse impacts on food supply, EISA further specifies

hat 16 out of the 36 billion gallons of renewable fuels producedn 2022 should be advanced biofuels made from non-starch feed-tocks, such as cellulosic or algal biomass. Yet, the current annual

∗ Corresponding author. Tel.: +1 847 467 2943; fax: +1 847 491 3728.E-mail address: you@northwestern.edu (F. You).

098-1354/$ – see front matter © 2013 Elsevier Ltd. All rights reserved.ttp://dx.doi.org/10.1016/j.compchemeng.2013.11.016

© 2013 Elsevier Ltd. All rights reserved.

production capacity of advanced biofuels is less than 1 billion gal-lon worldwide. It is foreseeable that the bioenergy industry will beundergoing a rapid expansion in the coming decade. Many sustain-able and robust biomass-to-bioenergy supply chains, which link thesustainable biomass feedstock and the final fuel/energy products,need to be designed and developed for lower costs, less environ-mental impacts and more social benefits (Elia, Baliban, & Floudas,2012; Hosseini & Shah, 2011; Sharma, Ingalls, Jones, & Khanchi,2013). To accelerate the transition towards the large-scale and sus-tainable production and use of biofuels and bioenergy products,a significant challenge in research is to systematically design andoptimize the entire bioenergy supply chains from the biomass feed-stock production to the biofuel/bioenergy end-use across multiplespatial scales, from unit operations to biorefinery processes and tothe entire value chain, as well as across multiple temporal scales,from strategic to operational levels, in a cost-effective, robust andsustainable manner (Daoutidis, Marvin, Rangarajan, & Torres, 2013;Marquardt et al., 2010). It is the objective of this paper to iden-tify the key research challenges and opportunities in modeling andoptimization of biomass-to-bioenergy supply chains using ProcessSystems Engineering (PSE) tools and methods, and to chart a pathfor addressing these challenges.

In this paper, we classify the existing contributions on bioenergy

supply chain optimization, elucidate the research challenges, anddemonstrate that multi-scale modeling and optimization approachcan play a leading role to address these challenges. We first reviewthe major biomass-to-biofuels pathways, with a specific emphasis

emica

oadttapogrwsosownbci

2

cnbtrrtdtp

2

dffw

2

crRBbSedefcec

isaCtei

D. Yue et al. / Computers and Ch

n producing “drop-in” liquid hydrocarbon fuels from terrestrialnd aquatic biomass, which may blend with or displace gasoline,iesel, jet fuel, kerosene or home heating oil. We then describehe key components of the biomass-to-biofuels supply chains andheir major characteristics, along with a comprehensive overviewnd classification of the existing contributions on bioenergy sup-ly chain optimization. We further demonstrate the important rolef multi-scale modeling and optimization, which allows the inte-ration across multiple temporal and spatial scales. The potentialesearch vistas on future bioenergy supply chains that integrateith petroleum refinery supply chains and/or carbon capture and

equestration systems are presented next. Key issues on modelingf sustainability and the treatment of uncertainties in bioenergyupply chain optimization are also discussed. Although the focusf this paper is on biofuels supply chains, at the end of this papere discuss the modeling and optimization for supply chains withon-fuel end-uses of biomass derivatives, including biopower andiomass-to-heat supply chains and biomass-to-chemicals supplyhains, which represent other products from biomass and willmpact the supply chain.

. Biomass-to-biofuels pathways

Biomass is unique among renewable energy sources in that itan be easily stored until needed and provides a liquid fuel alter-ative for use in today’s transportation system. In the short term,iofuels are the only renewable resources that can address theransportation sector’s heavy dependence on foreign oil withouteplacing the vehicle fleet (DOE/EERE, 2013d). In this section, weeview the major pathways from biomass to biofuels, and discusshe main properties of the feedstocks, fuel products and interme-iates. Fig. 1 shows a number of major conversion pathways fromerrestrial and aquatic biomass to intermediates and to final biofuelroducts (DOE/EERE, 2010b, 2011a, 2011b).

.1. Terrestrial and aquatic biomass feedstocks

Biomass is a renewable resource that acquiries carbon by carbonioxide fixation during their growing cycles. In comparison to fossiluels such as natural gas and coal, which take millions of years toorm, biomass is easy to grow, collect, utilize and replace quicklyithout depleting natural resources.

.1.1. Terrestrial feedstocksMany types of abundantly available biomass on the land

an be utilized as feedstocks for biofuel production. The 2011eport released by U.S. Department of Energy (DOE) and Oakidge National Laboratory (ORNL) titled “U.S. Billion-Ton Update:iomass Supply for a Bioenergy and Bioproducts Industry”, detailsiomass feedstock potential throughout the contiguous Unitedtates at a county level of resolution (DOE/EERE, 2011c). The reportxamines the U.S. capacity to sustainably produce over 1.6 billionry tons of terrestrial biomass annually for conversion to bioen-rgy and bioproducts, while continuing to meet existing demandor food, feed, and fiber. The report estimates that the United Statesould potentially produce approximately 85 billion gallons of biofu-ls annually—enough to replace approximately 30% of the nation’surrent petroleum consumption.

The terrestrial biomass feedstock can be generally categorizednto two groups. The first group includes corn grain, sugarcane,oy bean, oil seed, etc. These feedstocks are rich in sugar or lipids,nd have high yields after converted into bioethanol or biodiesel.

urrently, most biofuels are made from these feedstocks, due tohe maturity in technologies and lower unit production cost. How-ver, the use of these feedstocks for biofuel production might havemplications both in terms of world food prices and production.

l Engineering 66 (2014) 36–56 37

The second group of terrestrial biomass feedstocks, the cellu-losic biomass, can avoid adverse impacts on food supply, becausethey are non-starch, non-edible and non-food feedstocks. Cellulosicbiomass feedstocks can be obtained from a number of sources, suchas agricultural residues, forest residues and energy crops. Agricul-tural residues are typically plant parts left in the field after harvest(e.g., corn stover), as well as the secondary residues like manure andfood processing wastes. Forest residues are leftover wood or plantmaterial from logging operations, forest management, and land-clearing, as well as secondary residues like mill wastes. Dedicatedenergy crops (e.g., poplar, switchgrass, Miscanthus) are typicallyfast-growing trees and perennial grasses specifically grown forenergy uses.

2.1.2. Aquatic feedstocksAquatic biomass includes a diverse group of primarily

aquatic, photosynthetic algae and cyanobacteria ranging from themicroscopic (microalgae and cyanobacteria) to large seaweeds(macroalgae) (DOE/EERE, 2010b).

Many macroalgae, microalgae, and cyanobacteria carry outphotosynthesis to drive rapid biomass growth. Certain strains ofmicroalgae make extremely efficient use of light and nutrients. Asthey do not have to produce structural compounds such as cellulosefor leaves, stems, or roots, and because they can be grown floatingin a rich nutritional medium, microalgae can have faster growthrates than terrestrial crops. In some cases, algae growth rates canbe an order of magnitude faster than those of terrestrial crop plants.As algae have a harvesting cycle of 1–10 days, their cultivation per-mits several harvests in a very short time-frame, a strategy differingfrom that associated with yearly crops. Algae naturally remove andrecycle nutrients (e.g., nitrogen and phosphorous) from water andwastewater, and can be used to sequester carbon dioxide from theflue gases emitted from fossil fuel-fired power plants. Adding to theappeal, some strains of microalgae can grow on land unsuitable forother established crops, for instance: arid land, land with exces-sively saline soil, and drought-stricken land. This minimizes thecompetition on arable land with the cultivation of food crops. Moreimportantly, algae biomass can contain high levels of oil (lipids ortriacylglycerides), making it a promising feedstock for the produc-tion of renewable gasoline, diesel, and jet fuel. Algae can produceone or even two orders of magnitude more oil per unit area thanconventional crops such as rapeseed, palms, soybeans, or jatropha(DOE/EERE, 2010b; Hasan & Chakrabarti, 2009).

The cultivation of macroalgae (e.g., seaweed) is similar to that ofterrestrial plants but in aquatic environment. In practice, macroal-gae are cultivated in large offshore farms, near-shore coastal zones,or open pond facilities. In contrast, depending on the type ofmicroalgae and cyanobacteria, there are various cultivation meth-ods, varying both in their advantages and challenges. Broadlyspeaking, algae can be cultivated via photoautotrophic or het-erotrophic methods. The former requires light and carbon dioxidefor photosynthesis while the latter requires suitable feedstockssuch as lignocellulosic sugars for growth. The cultivation systemfor microalgae and cyanobacteria is either closed bioreactor or openpond. Closed bioreactor yields less loss of water and better biomassquality, but has scalability problems. Open pond requires lowercapital costs and take advantage of evaporative cooling, but it issubject to weather and condition changes and inherently difficultto avoid contamination and foreign algae (DOE/EERE, 2010b).

2.2. Types and major properties of biofuels

2.2.1. Ethanol and other alcoholsAs the first generation biofuel technology, biomass-derived

ethanol and other alcohols were largely produced from starch-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

aquat

irlg

caaaSatecmbewtttd&

bmctbmpmp

2

bddTc

Fig. 1. Selected conversion pathways from terrestrial and

ncreasing debate on its adverse impact on global food prices, theesearch focus on alcohol-based biofuel is shifted to employing cel-ulosic biomass as feedstock options, which is considered secondeneration biofuel technology.

Ethanol is extensively distributed and utilized as a fuel blendomponent throughout the U.S. Its low freezing point makes it suit-ble for use in cold climates. It can be blended with gasoline inny combination, and it is currently approved as a 10% blend forll vehicles and as an 85% blend for flex-fuel vehicles. The Unitedtates Environmental Protection Agency (EPA) has also approved

15% blend for all vehicles from model year 2001 and newer, yethe availability of which is very limited. There are, however, sev-ral distribution-related challenges associated with ethanol. It isompletely miscible with water and will separate from a gasolineixture if enough water is present either in the pipeline or as water

ottoms in a storage tank. In addition, numerous studies have citedthanol’s role in corrosion or stress corrosion cracking of pipelinealls. Due to all these challenges, many major U.S. pipeline opera-

ors expressly prohibit ethanol and ethanol–gasoline mixtures inhe pipeline. Therefore, the rail car, barge, and tanker truck arehe principle transportation modes for ethanol, which makes long-istance transportation cost-prohibitive (Bunting, Bunce, Barone,

Storey, 2010).As with ethanol, butanol is also an oxygenated fuel that can be

lended with gasoline in any combination and requires only minorodifications for use in existing vehicles. The producers of butanol

laim that butanol is compatible with existing oil pipeline infras-ructure, but these claims have yet to be validated. In addition,utanol has four isomers (molecular arrangements), thus opti-ization of these isomers from either a manufacturing or fuel

erformance standpoint still deserves further research. Note thatost fermentation routes produce a single isomer as the dominant

roduct (Bunting et al., 2010).

.2.2. Bio-dieselThe fatty ethyl methyl esters (FAME) is often referred to as

io-diesel, because it is used extensively as a diesel supplement

ue to its similarity to diesel both physically and chemically andue to its ability to be blended with diesel in any combination.hree major issues facing FAME bio-diesel distribution are a higherloud point than diesel, lower stability, and the cleaning effect.

ic biomass to intermediates and to final biofuel products.

Furthermore, the polarity resulting from the high oxygen contentin FAME can cause it to cling to pipe or vessel walls, thus posingpotential contamination of subsequent batches if transported inpipeline. Therefore, FAME is also expressly prohibited by most largeU.S. pipeline operators (Bunting et al., 2010). In addition, vehiclewarranties typically limit FAME blending to 5%.

2.2.3. Drop-in biofuelsCompared to the ethanol and bio-diesel, advanced hydrocarbon

biofuels have the appealing advantage that they are compatiblewith existing infrastructure. Known as drop-in fuels, hydrocarbonbiofuels can serve as petroleum substitutes in existing refiner-ies, tanks, pipelines, pumps, vehicles, and smaller engines. Thesefuels deliver more energy density than ethanol. Moreover, the con-version processes of hydrocarbon biofuels can also yield a widerange of bioproducts that can substitute for specialty petrochem-icals, helping to replace the whole barrel of oil and enhancing theeconomic and environmental sustainability of biorefineries.

These renewable gasoline, diesel, and jet fuels are essen-tially identical to their existing petroleum-derived counterpartsin properties, except that they are derived from domestic cel-lulosic biomass sources or algal masses. Advanced hydrocarbonfuels are believed to be the next generation biofuels, which wouldhelp reduce the carbon footprint and potentially, supply and pricevolatility in all transportation sectors, relieve the energy securityissues, and create related job opportunities for a better social equal-ity (DOE/EERE, 2013d; MASBI, 2013).

2.3. Intermediates

2.3.1. Sugars or carbohydrate derivativesSugars can be easily obtained from starch- or sugar-rich crops

such as corn and sugarcane. However, to avoid the impact on foodprices, the development focus has shifted to deriving sugars andother carbohydrate derivatives from lignocellulosic biomass feed-stocks. Lignocellulose (mainly lignin, cellulose, and hemicellulose)is the primary component of plant residues, woody materials, and

grasses. The cell wall structure of these plant matters is partiallycomprised of sugar polymers (carbohydrates). Pretreatment andhydrolysis processes are used to break down the cell wall throughintroduction of enzymes in addition to heat and other chemicals

emica

iadci

pbtricm2

2

poaAutrhs

celfbbpoAa

2

auosiarnap(

2

dwhSpTdtIafo

D. Yue et al. / Computers and Ch

n order to convert the carbohydrate portion of the biomass inton intermediate sugar stream. These sugars can serve as interme-iate building blocks that can be further fermented or chemicallyatalyzed into a range of advanced biofuels and value-added chem-cals.

One major obstacle to developing a cost-competitive conversionathway is the efficiency of separating and converting cellulosiciomass into sugars. Starch and sugars function as food storage forhe plant, while lignocellulose is structural and designed to restrictapid metabolism. Due to the recalcitrant nature of the cell wall,t is often difficult and complicated to completely hydrolyze thearbohydrate components into sugars and usable intermediates,aking this material expensive to convert to biofuels (DOE/EERE,

011a; Melero, Iglesias, & Garcia, 2012).

.3.2. LipidsLipids can be derived from a variety of sources usually through

hysical extraction processes. The most common approach is fromil seed plants that are abundant in oil content such as rapeseednd soybean in temperate climates and palm in tropical climates.nother approach is from animal fats (tallow), which are a byprod-ct of meat production and cooking. Though still under research,he largest potential impact is to extract lipids from algae. It iseported that certain strains of algae may have order-of-magnitudeigher yield of lipids compared to the vegetable and animal originources.

Note that the term lipids refers not to a single chemical butonstitute a group of molecules including monoglycerides, diglyc-rides, triglycerides, phospholipids, fatty acids, and others. Theipids derived from various biomass sources may significantly dif-er in compositions. These fatty acid esters can be used to produceio-diesel via transesterification process or converted to hydrocar-on biofuels via hydrotreating, catalytic cracking and metathesisrocesses (Fahy et al., 2005). It is reported that hydrotreated estersf fatty acids (HEFA) are one of the two biofuels that have receivedmerican Society for Testing and Materials (ASTM) approval forviation (MASBI, 2013).

.3.3. SyngasSyngas derived from biomass usually via gasification process is

gas mixture composed mainly of CO and H2. Syngas is normallysed as a feed to manufacture Fischer–Tropsch fuels, which are thether ASTM approved aviation fuels (MASBI, 2013). Additionally,yngas can also be used as a fuel for heat supply. The hydrogens useful for hydrotreating operations, necessary to upgrade fuelsnd to remove impurities. If shipped long distances, syngas wouldequire a distribution infrastructure similar to that required byatural gas. However, additional modifications and maintenancere needed to prevent syngas leaks, as the hydrogen component isrone to leaks and the carbon monoxide component is highly toxicBunting et al., 2010).

.3.4. Bio-oilBio-oil, also known as pyrolysis oil, is a promising bio-crude pro-

uced from biomass feedstocks undergoing pyrolysis processes,hich can be further upgraded via hydroprocessing to produceydrocarbon biofuels and other bioproducts (Ringer, Putsche, &cahill, 2006). Adding to the appeal, most biomass sources can beyrolyzed, creating significant potential for this conversion route.here are, however, several significant issues associated with itsistribution in and compatibility with the crude pipeline infras-ructure. Bio-oil can have a very high oxygen and water content.

t is highly corrosive and chemically unstable due to the high charnd solid contents. In addition, its alkali metals content can lead toouling and contribute to catalyst poisoning. The higher viscosityf pyrolysis oil compared to petroleum crude requires increased

l Engineering 66 (2014) 36–56 39

pumping work. Its polarity due to high oxygen content causespyrolysis oil more likely to cling to pipe and vessel walls. After all,due to its promising potential to replace the petroleum crude, thereis significant research on stabilizing and upgrading pyrolysis oils(Bunting et al., 2010).

3. General structures and key components ofbiomass-to-biofuels supply chains

Development of the U.S. advanced bioenergy industry will leadto the establishment of many manufacturing supply chains ofhydrocarbon biofuels, which link the sustainable biomass feed-stocks with the final biomass-derived gasoline, diesel and jet fuelproducts through the following five major components: biomassproduction system, biomass logistics system, biofuel productionsystem, biofuel distribution system and biofuel end-use (Gold &Seuring, 2011; Iakovou, Karagiannidis, Vlachos, Toka, & Malamakis,2010). Although these five elements are similar to those in apetroleum-to-fuel supply chain, they could not be investigatedwith same systematical framework due to the differences in feed-stocks and products, and consequently different strategies forproduction, transportation and storage (An et al., 2011a; Sharmaet al., 2013).

3.1. Feedstock supply and logistics

Feedstock production and logistics components of the supplychain deliver high-quality, stable, and infrastructure-compatiblefeedstocks from diverse biomass resources to biorefineries. Forterrestrial biomass, a main challenge is to develop an efficientfeedstock supply system for cost-effective and time-sensitive col-lection, preprocessing, storage, and transport, in order to deliverdiverse consistently high-quality terrestrial biomass for conver-sion.

3.1.1. Seasonal supply and storageOne of the critical challenges in the operation of biofuel sup-

ply chains is the seasonal nature and annual variability of biomasssupply. Most biomass resources are plant matters, which need to beplanted, cultivated, and harvested, going through a growing cycle.Crop residues are usually collected after the harvest of the agri-cultural crops. For instance, the corn stover in the U.S. Corn Beltis mainly harvested from September through November (USDA,2010). The wood residues are grown over multiple years, whichmakes them less seasonal compared to crop residues. They areusually available all year round. However, the yields may vary indifferent months. For instance, it might be more difficult to collectthe wood residues during snowing seasons. In addition, it is alsoreported that the harvesting timing and frequency may affect theyields of energy crops, thus careful planning and scheduling wouldbe necessary in order to guarantee the quantity and quality of thebiomass supply (Sokhansanj et al., 2009).

The biomass supply may be discrete due to the seasonality, butthe demand for transportation fuels is all-year-round. Therefore,the resulting operational challenge is to manage the biomass stor-age in order to maintain a continuous supply for the production atbiorefineries.

3.1.2. Pre-treatment and degradation properties of biomassThe storage of biomass resources can take place in various

modes. As the cheapest option, ambient storage leads to significantcost reduction at the storage and handling stage of the biomass

supply chain. However, side-effects may include biomass degrada-tion (e.g. silage), heating value reduction, and potential health risks,mainly because of the presence of high water content. The degrada-tion rate is estimated to be 1% material loss/month for the ambient

4 emica

sdiqanct

ptarafJrs(ptbt

3

slsUbeapsttlsum

3

iaceeibte

3

mualSacof

0 D. Yue et al. / Computers and Ch

torage (Rentizelas, Tolis, & Tatsiopoulos, 2009b). To reduce theegradation, covered storage in pole-frame structures is often used

n practice, incurring a 0.5% material loss/month rate. If a higher-uality biomass supply is required, a closed warehouse with hotir drying capability can be employed. The material loss in this sce-ario can be assumed negligible (Cundiff, Dias, & Sherali, 1997). Asan be seen, the tradeoff between storage cost and material loss ishe focus when determining the selection of storage modes.

In addition to the direct storage of raw biomass, pretreatmentrocesses on raw biomass materials are also adopted in many prac-ices. In this scenario, after the biomass has been harvested, thermalnd chemical treatments are applied to reduce moisture content,emove contaminants, and improve feedstock quality, stability,nd processing performance, including torrefaction, pelletization,ast pyrolysis, ammonia fiber expansion (AFEX), etc. (Bals, Rogers,in, Balan, & Dale, 2010; Uslu, Faaij, & Bergman, 2008). Ongoingesearch is trying to develop an advanced uniform-format feed-tock supply system, aiming to reduce the variability in biomassboth format and quality). The products of this system are sup-osed to be standard, consistent, quality-controlled commoditieshat can be efficiently handled, stored, transported and utilized byiorefineries. Major technologies may involve milling, densifica-ion, torrefaction, blending and pelletization (DOE/EERE, 2013a).

.1.3. LogisticsThe logistics of biomass plays an important role in the bioenergy

upply chain, because it integrates time-sensitive feedstock col-ection, storage, and delivery operations into efficient, year-roundupply systems that deliver consistently high-quality biomass.nlike fossil fuels, biomass resources have a sparsely spatial distri-ution. Both the biomass collection and delivery require extensivefforts in equipment selection, shift arrangement, vehicle routing,nd fleet scheduling (DOE/EERE, 2013c). Moreover, if independentretreatment depots are built to provide uniform-format feed-tocks, of which the location should be cautiously determined inhe supply chain design phase. A depot would locate to facilitatehe transportation and production, and serve like a small satel-ite system which receives biomass of different types and variousources from its neighborhood and then forward the convertedniform-format biomass to a network of much larger supply ter-inals (DOE/EERE, 2013a).

.2. Biomass conversion in advanced biorefinery

Conversion of biomass into commercially viable liquid fuelss another main component of the supply chain. A biorefinery is

facility that integrates conversion processes and equipment toonvert biomass into biofuels, power and chemicals. The biorefin-ry is analogous to a petroleum refinery, but could have reducednvironmental impact and consumes less fossil energy because ofts diversification in feedstocks and products. Broadly speaking,iomass-to-liquid conversion technologies can be classified intowo categories, namely biochemical and thermochemical (Anext al., 2010).

.2.1. Biochemical technologiesThe major biochemical conversion processes include pretreat-

ent, hydrolysis, biological or chemical processing, and productpgrading and recovery. During pretreatment, biomass materi-ls undergo a mechanical or chemical process to fractionate theignocellulosic complex into soluble and insoluble components.oluble components include mixtures of five- and six-carbon sug-

rs and some sugars oligomers. Insoluble components includeellulosic polymers and oligomers and lignin. The main purposef pretreatment is to break down the plant cell walls to permiturther deconstruction during the hydrolysis step. Note that the

l Engineering 66 (2014) 36–56

quantity and composition of the sugars stream released depend onthe feedstock used and the pretreatment technology employed. Ahydrolysate conditioning and/or neutralization process is used toadjust the pH of and purify the pretreated material. In hydrolysis,the pretreated material, primarily cellulose, is guided through achemical reaction that releases the readily fermentable sugar, glu-cose. This step is usually accomplished with enzymes or strongacids. The conversion from sugars to biofuels can take place intwo ways: biological and chemical processing. In biological pro-cessing, the most common approach is to introduce a fermentingmicroorganism to the biomass hydrolysates. The continued sac-charification and fermentation may take a few days to convert allsugars to biofuels or other chemicals of interest. On the other hand,a catalyst is added in chemical processing to make the reactionless energy intensive and thus more efficient than biological fer-mentation. Furthermore, the yields of different end products andintermediate chemicals can be adjusted while using different reac-tions and catalysts. The purpose of the final step, product upgradingand recovery, is to separate the fuel from the water and residualsolids, which could involve any transformation, distillation, sepa-ration, and cleanup processes (Biddy & Jones, 2013; Davis, Biddy,Tan, Tao, & Jones, 2013b; DOE/EERE, 2011a).

3.2.2. Thermochemical technologiesThe thermochemical technologies utilize high temperature in

addition to catalysts to change the physical properties and chemicalstructures of the biomass resources. There are two major thermo-chemical reaction pathways: gasification and pyrolysis (DOE/EERE,2011b).

3.2.2.1. Gasification. In the gasification pathway, the lignocellu-losic biomass resources are fed into a gasifier, where they arethermally decomposed (700–1300 ◦C) with limited or no oxygen,and then oxidized to yield a raw syngas, which is primarily amixture of hydrogen and carbon monoxide. The raw syngas maycontain some contaminants, including tars, acid gas, ammonia,alkali metals, and other particulates. Therefore, a gas cleanup pro-cess is necessary to produce a contaminant-free syngas product. Ifnecessary, the gas may be further conditioned to lower the sulfurlevels and adjust the hydrogen to carbon monoxide ratio. The syn-gas can then be utilized by the Fischer–Tropsch synthesis, includinga collection of chemical reactions, to produce hydrocarbon liquidfuel products (DOE/EERE, 2010a; Swanson, Satrio, Brown, Platon,& Hsu, 2010; Talmadge, Biddy, Dutta, Jones, & Meyer, 2013). Alter-natively, methanol and hydrogen can be produced from the syngasvia methanol synthesis and water–gas-shift process, respectively.

3.2.2.2. Pyrolysis. Pyrolysis processes also heat biomass in theabsence of oxygen, but at a relatively lower temperature(450–600 ◦C). The main product of the pyrolysis process is a crude-like liquid known as pyrolysis oil, while gas and charcoal are alsoproduced. Depending on the length of the reaction time, thereare slow pyrolysis, intermediate pyrolysis, and fast pyrolysis. Fastpyrolysis is often adopted because the yield of pyrolysis oil ismaximized. Oil produced in pyrolysis processing must have par-ticulates and ash removed by filtration to create a homogenousproduct. The pyrolysis oil is then upgraded to desired hydro-carbon fuels via hydrotreating and hydrocracking processes toreduce its total oxygen content and to break down long-chainhydrocarbons (DOE/EERE, 2010a; Jones et al., 2009; Wright, Satrio,

Brown, Daugaard, & Hsu, 2010). Furthermore, National AdvancedBiofuels Consortium (NABC) is investigating advanced conversiontechnologies such as catalytic fast pyrolysis, hydropyrolysis, andhydrothermal liquefaction (NABC, 2013).

emica

3

(cshluIsnmtm(

adafarslbfipabaiBacseacpDsi

3

ddtp

3

bkbastcbodecd

D. Yue et al. / Computers and Ch

.2.3. Algae harvesting and conversionMicroalgal species accumulate lipids such as triacylglycerols

TAG), while cyanobacteria and macroalgae accumulate mostlyarbonhydrates. Microalgae and cyanobacteria are in very smallize (∼1 to 30 �m), in suspension in water culture, which makesarvesting more difficult. Harvesting methods like forced floccu-

ation, sedimentation, filtration and centrifugation are generallysed. Drying is required to achieve high biomass concentrations.

n contrast, harvesting of macroalgae (e.g., seaweeds) is very muchimilar to that of terrestrial agricultural crops. Modern mecha-ized machines and equipment can be employed in large off-shoreacroalgae farms. Washing and dewatering are required. In addi-

ion, due to the large size of macroalgae, milling is needed to reduceacroalgae to particle sizes that can be more efficiently processed

Lundquist, Woertz, Quinn, & Benemann, 2010).Depending on the type of algae and targeting end products, there

re three algae conversion pathways: conversion of algal extracts,irect production of biofuels from algae, and processing of wholelgae (DOE/EERE, 2010b). The conversion from extracts derivedrom algal sources is the typical mode of biofuel production fromlgae. The most common type of algal extracts is lipid-based. Cur-ent practices for lipid extraction involve mechanical disruption,olvent extraction, and supercritical technologies. The extractedipids can then be converted into bio-diesel, advanced hydrocarboniofuels, and other high-value chemicals via chemical transester-cation, enzymatic conversion, and catalytic cracking. The directroduction refers to heterotrophic fermentation and growth fromlgal biomass. It has certain advantages in terms of process costecause it can eliminate several process steps (e.g., oil extraction)nd their associated costs. The system is readily scaled up and theres an enormous potential to use various fixed carbon feedstocks.iofuels that can be produced directly from algae include alcohols,lkanes, and hydrogen. In addition, processing of whole algae is alsoonsidered a promising algae biofuel solution. Potential conver-ion methods include gasification and pyrolysis, which are widelymployed for cellulosic biomass, as well as supercritical processingnd anaerobic digestion of whole algae. These methods avoid theosts associated with the extraction processes and are amenable torocessing a diverse range of algae (Davis, Aden, & Pienkos, 2011;avis, Biddy, & Jones, 2013a; FAO, 2009). However, extraction is

till preferred when oil content is high to leverage the entropicnvestment by the organism.

.3. Fuels distribution and end-use system

The last components in the biofuels supply chain are the fuelistribution and end-use systems, which ensure that biomass-erived fuel products can cost-effectively and sustainably reachheir market and be used by consumers as a replacement foretroleum-based fuels.

.3.1. Infrastructure incompatibility of first generation biofuelsAlthough bio-ethanol and bio-diesel represent the major

iomass-derived transportation fuel products in the current mar-etplace, as mentioned before, bio-ethanol and bio-diesel areanned by most U.S. pipeline operators due to their polaritynd other corrosion, contamination issues. Therefore, the currentupply chains for ethanol and bio-diesel require dedicated fuel dis-ribution and blending systems (Bunting et al., 2010). The mostommon practice is to transport bio-ethanol or bio-diesel from theiofuel production facility to distribution terminals by train, barge,r truck to minimize the opportunity for contact with water and

irt, which makes it difficult for shipping larger volumes of biofu-ls. The biofuels are kept in dedicated tanks until blended with theorresponding fossil fuels (ethanol with gasoline and FAME withiesel). Once blended, the final fuel products are delivered to retail

l Engineering 66 (2014) 36–56 41

stations by tanker trucks. This process is reasonably economical inareas that have easy access to abundant biomass supply and bio-fuel production facilities, e.g., Midwest regions. However, it is notan attractive option for long-distance delivery, especially as volumeincreases. One impact of the distribution supply chain, is that theMidwest has become a net exporter of starch ethanol via barges onthe Mississippi while the West Coast has become an importer ofsugarcane ethanol (DOE, 2010; NCEP, 2009; USDA, 2007).

3.3.2. Features of ‘drop-in’ fuelsIn contrast, drop-in biofuels that are highly compatible, or

totally fungible, with the existing hydrocarbon fuel finishing, fuelhandling, distribution, and end-use infrastructure would resultin easier and more widespread market acceptance (DOE/EERE,2010b).

Hydrocarbon biofuels can be used directly in the existingpipelines, dispensers and vehicle engines without any additionaleffort, because these advanced biofuels are the same in propertiesas their petroleum counterparts and compatible with the existinginfrastructure. Thus, the remaining challenge in the modeling andoptimization of supply chains for these infrastructure-compatiblebiofuels is to address the full scale integration with the existingpetroleum refinery processes and distribution supply chains. In theshort term, the introduction of advanced biofuels to existing dis-tribution systems may require capacity expansion and additionalscheduling efforts to coordinate with the fossil fuels. However, itis believed that, the advanced hydrocarbon biofuels would becomethe major stream and play a more important role in the fuel dis-tribution infrastructures in the future (DOE/EERE, 2013d; MASBI,2013).

4. Literature review

There is a large body of literature regarding the modelingand optimization of biofuel supply chains, covering differentfeedstocks, products, processes, system properties, and model-ing viewpoints. The recent journal and conference papers closelyrelated to the subject are summarized in Table 1. The articles aresorted in chronological order. The articles published in the sameyear are sorted in alphabetical order by last name of the authors. Anumber of the works may cover multiple aspects, thus occupyingmultiple cells in the table, which indicates that a holistic optimiza-tion model for modeling and optimization of biofuel supply chainscan be multi-scale, multi-perspective, and multi-criterion.

The column – “food crops to ethanol, biodiesel” – representsthe biofuel supply chains using first-generation biofuel technolo-gies. These supply chains typically use starch- or oil-based foodcrops as feedstocks to produce ethanol and biodiesel, which canbe blended with petroleum-based gasoline and diesel. These tech-nologies are commercialized and have a significant penetrationin the current transportation liquid fuel markets. However, thefirst-generation biofuel technologies are often criticized on theirimplications on food prices and production. The column – “cellu-losic biomass to ethanol, biodiesel” – represents the biofuel supplychains using second-generation biofuel technologies. These supplychains use crop residues, wood residues, or dedicated energy cropsas feedstocks, thus avoid the potential impact on food markets. Nev-ertheless, ethanol and biodiesel are not the ideal substitutes to fossilfuels, because they are not completely compatible with the existingdistribution and end-use infrastructure. The column – “cellulosicbiomass to hydrocarbon biofuels” – represents the biofuel supply

chains using third-generation biofuel technologies. They are ableto produce cellulosic-biomass-derived transportation fuels withproperties that are functionally identical to their petroleum-basedcounterpart. At the current stage, the third-generation biofuel

42

D.

Yue et

al. /

Computers

and Chem

ical Engineering

66 (2014)

36–56Table 1Summary of some recent publications on the modeling and optimization of biofuel supply chains.

Components Food crops to ethanol,biodiesel

Cellulosic biomass to ethanol,biodiesel

Cellulosic biomass tohydrocarbon biofuels

Algae to biofuels Biomass to power, heat, food, etc.

Supply chain design Mele, Guillen-Gosalbez,and Jimenez (2009),Zamboni, Bezzo, and Shah(2009a), Zamboni, Shah,and Bezzo (2009b),Eksioglu, Li, Zhang,Sokhansanj, and Petrolia(2010), Akgul, Zamboni,Bezzo, Shah, andPapageorgiou (2011),Corsano, Vecchietti, andMontagna (2011), Dal-Mas,Giarola, Zamboni, andBezzo (2011), Mele, Kostin,Guillen-Gosalbez, andJimenez (2011), Bai et al.(2012)

Dunnett et al. (2008), Eksioglu,Acharya, Leightley, and Arora (2009),Cucek, Lam, Klemes, Varbanov, andKravanja (2010), Huang, Chen, andFan (2010), Leduc, Lundgren,Franklin, and Dotzauer (2010a),Leduc et al. (2010b), Parker et al.(2010), Tittmann, Parker, Hart, andJenkins (2010), Aksoy et al. (2011),An, Wilhelm, and Searcy (2011b),Bai, Hwang, Kang, and Ouyang(2011), Gan and Smith (2011),Giarola, Zamboni, and Bezzo (2011),Santibanez-Aguilar,Gonzalez-Campos, Ponce-Ortega,Serna-Gonzalez, and El-Halwagi(2011), Akgul, Shah, andPapageorgiou (2012), Avami (2012),Bernardi et al. (2012), Chen and Fan(2012), Chen and Onal (2012), Cucek,Varbanov, Klemes, and Kravanja(2012), Marvin, Schmidt, Benjaafar,Tiffany, and Daoutidis (2012), Youet al. (2012), Bernardi et al. (2013)

Leduc, Natarajan, Dotzauer,McCallum, and Obersteiner(2009), Parker et al. (2010),Tittmann et al. (2010), Aksoyet al. (2011), Bowling,Ponce-Ortega, and El-Halwagi(2011), Elia, Baliban, Xiao, andFloudas (2011), Kim, Realff, andLee (2011a), Kim, Realff, Lee,Whittaker, and Furtner (2011b),Papapostolou, Kondili, andKaldellis (2011), You and Wang(2011), Elia et al. (2012),Gebreslassie, Yao, and You(2012), Walther, Schatka, andSpengler (2012), Yue et al.(2013b)

Tatsiopoulos and Tolis (2003),Dunnett, Adjiman, and Shah(2007), Rentizelas, Tatsiopoulos,and Tolis (2009a), Cucek et al.(2010), Leduc et al. (2010b),Tittmann et al. (2010),Velazquez-Marti andFernandez-Gonzalez (2010),Aksoy et al. (2011), Elia et al.(2011), Gan and Smith (2011),Cucek et al. (2012a,b), Elia et al.(2012), Keirstead, Samsatli,Pantaleo, and Shah (2012)

Planning and operation Sharma, Sarker, and Romagnoli(2011), Avami (2012)

Avami (2012) Frombo, Mindardi, Robba, Rosso,and Sacile (2009), Flisberg, Frisk,and Ronnqvist (2012)

Selection of technologies Cucek et al. (2010), Parker et al.(2010), Tittmann et al. (2010),Santibanez-Aguilar et al. (2011),Avami (2012), Cucek et al. (2012a,b),You et al. (2012)

Parker et al. (2010), Tittmannet al. (2010), Sharma et al.(2011), You and Wang (2011),Gebreslassie et al. (2012), Yueet al. (2013b)

Frombo et al. (2009), Cucek et al.(2010), Tittmann et al. (2010),Cucek et al. (2012a,b), Keirsteadet al. (2012)

Decentralized production Dunnett et al. (2008) Bowling et al. (2011), Kim et al.(2011a), Kim et al. (2011b), Youand Wang (2011), Yue et al.(2013b)

Multi-objective optimization Zamboni et al. (2009a),Mele et al. (2011)

Giarola et al. (2011),Santibanez-Aguilar et al. (2011),Bernardi et al. (2012), Cucek et al.(2012a,b), You et al. (2012), Bernardiet al. (2013)

You and Wang (2011),Gebreslassie et al. (2012), Yueet al. (2013b)

Cucek et al. (2012a,b)

Sustainability Mele et al. (2009), Zamboniet al. (2009a), Corsano et al.(2011), Mele et al. (2011)

Cucek et al. (2010), Giarola et al.(2011), Santibanez-Aguilar et al.(2011), Avami (2012), Bernardi et al.(2012), Cucek et al. (2012a,b), Youet al. (2012), Bernardi et al. (2013)

Elia et al. (2011), You and Wang(2011), Elia et al. (2012), Yueet al. (2013b)

Cucek et al. (2010), Cucek et al.(2012a,b)

Uncertainties Dal-Mas et al. (2011) An et al. (2011b), Gan and Smith(2011), Avami (2012), Chen and Fan(2012), Marvin et al. (2012)

Leduc et al. (2009), Kim et al.(2011a), Gebreslassie et al.(2012), Walther et al. (2012)

Game theory Bai et al. (2012)Simulation models Ferreira and Borenstein (2011) Mobini, Sowlati, and Sokhansanj

(2011)

emica

tpisphtgPnm–wiabn

wtpbrw“saa“eccgosecStthmgbttottvotmp“st

5

mamh

D. Yue et al. / Computers and Ch

echnologies are relatively mature and the corresponding sup-ly chain design and optimization is desired for commercial-scale

mplementation. The column – “algae to biofuel” – represents theupply chains using fourth-generation biofuel technologies, whichroduce various liquid fuels and chemicals from lipids and/orydrocarbons in algae feedstocks. Compared to other columns,here is a significant lack of studies on these nascent technolo-ies, mainly because of the scarcity of credible data and references.rocesses regarding algae cultivation, harvesting, and conversioneed to be well classified and documented before the supply chainodeling and optimization can be implemented. The last column

“biomass to power, heat, food, etc.” – includes supply chainsith non-fuel uses of biomass resources, which provide electric-

ty and heat supply to surrounding communities or produce foodsnd other valuable chemical products. As can be seen, the largeody of literature in this column indicates the importance of theseon-fuel biomass utilization pathways.

By observing the rows of Table 1, we can see that most researchorks focused on “supply chain design”, mainly involving the

ransportation network design and biorefinery facility locationroblems. A few studies integrated the “planning and operation” ofiorefineries into the supply chain models, thus achieving a higheresolution in both spatial and temporal scales. We note that earlierorks on the design of biofuel supply chains seldom considered the

selection of technologies”. However, as more and more conver-ion options become available, selection of technologies become

critical component in the recent works on supply chain designnd optimization. As will be further discussed in later sections,decentralized production” is a relatively new concept, mainlymerging from the design of advanced hydrocarbon biofuel supplyhains, which suggests that the traditional integrated biorefineryan be decomposed into multiple conversion steps in multipleeographically distributed production facilities. “Multi-objectiveptimization” is required when conflicting objectives are con-idered simultaneously. This often involves the tradeoff betweenconomic, environmental and social performances of the supplyhain as well as the risks associated with design and operation.ince biofuels are considered as a renewable energy alternativeo fossil fuels, the “sustainability” issues deserve further atten-ion to avoid introducing adverse impacts on the ecosystem anduman society. Some works employed the life cycle analysis (LCA)ethodology to assess the environmental impacts (e.g., greenhouse

as emissions), while some addressed job creation opportunitiesrought by the biofuel value chains. While most models consideredhe design and operation of biofuel supply chains under determinis-ic environments, some recent works investigated the performancef a biofuel supply chain system in the presence of various “uncer-ainties” (e.g., price volatility, yield uncertainty). Instead of treatinghe biofuel supply chain system as a cooperative entity, we can alsoiew the biofuel supply chain system as a value chain composedf different parties, who are acting selfishly and in a noncoopera-ive manner. Few existing works utilized “game theory” and spatial

arket equilibrium to model the inter-relationships in biofuel sup-ly chains. In addition to mathematical programming approaches,simulation models” are also useful tools for the study of biofuelupply chains, which can easily capture dynamics in complex sys-ems and provide process data for the optimization.

. Multi-scale modeling and optimization

While most works provide a great insight to specific fields, they

ay fall into a narrow perspective and do not cover sufficiently

ll the aspects of the biofuel supply chain. Therefore, we need aulti-scale modeling and optimization framework to provide a

olistic view and organically integrate the different components of

l Engineering 66 (2014) 36–56 43

the biofuel supply chain (Hosseini & Shah, 2011). In this paper, weportray a schematic framework of multi-scale modeling and opti-mization for biofuel supply chains in Fig. 2. From bottom to top,four system layers regarding different temporal and spatial scalestypify the level of ecosystem, supply chain, process, and molecule,respectively.

At the ecosystem layer, the ecosystem and human society con-stitute the environment where various activities in biofuel supplychains take place. All these activities would cause impacts and foot-prints on the ecosystem throughout the life cycle and value chainof the biofuel products, involving biomass cultivation, biomass har-vesting and logistics, biomass pretreatment and fuel production inconversion facilities, and fuel distribution and end-use. Modelingand optimization tools can be applied to simultaneously evalu-ate and identify the sustainable solutions with minimum adverseimpacts on the environment and maximum benefits to the soci-ety. Since multiple conflicting objectives are often involved whenoptimizing the sustainability of biofuel supply chains, a numberof multi-objective optimization techniques can be applied, such asepsilon-constraint method, goal programming method, etc. Trian-gle charts can be used for comparison between different solutionswith different economic, environmental, and social performances.Pareto frontiers obtained from multi-objective optimization canreveal the tradeoffs between conflicting objectives and provide aset of Pareto-optimal solutions for decision-making.

At the supply chain layer, modeling and optimization toolscan play an important role in optimizing the supply chain net-work structure and improving the various activities involved in theinstallation and operation of biofuel supply chains. A biofuel sup-ply chain usually consists of multiple sites and multiple echelons,which requires coordination across the entire supply chain network(You, Grossmann, & Wassick, 2011). At the design stage, super-structure optimization allows the decision maker to determine theoptimal network design from a number of choices among candidatesuppliers, facility locations, technology options, transport modes,etc. By using mathematical programming approaches, descriptionof the various activities, even across multiple sectors, can be easilyinterpreted into equations in the constraints or objective functions.However, a comprehensive and detailed superstructure, thoughappealing, may be computationally intractable. Therefore, whenmodeling the biofuel supply chain systems, we should take advan-tage of the properties of the supply chain system (e.g., networkstructure, spatial scale) and drop off the components that have neg-ligible influences on the optimization objectives to trade off themodeling resolution and computation efficiency.

At the process layer, the processes and units at a productionfacility are studied in a more detailed manner. Superstructure-based optimization can again be employed for the design, synthesis,and retrofit of conversion processes, involving the selectionamong candidate processing methods, equipment units, catalysts,resource supply scenarios, etc (Yeomans & Grossmann, 1999).Regarding the operation at a biorefinery, three levels of deci-sions need to be optimized, namely planning, scheduling, andcontrol. Optimization for planning involves the development ofrobust forecasting model, multi-year strategy for capacity expan-sion and process retrofit, multi-period targets for purchase, salesand production, etc. Optimization for scheduling involves the effi-cient and timely allocation of equipment units, raw materials, andhuman labors to fulfill the external and internal orders. Optimiza-tion for control involves real-time monitoring and adjustment ofprocess parameters to meet the required quantity and quality ofthe products. The three decision levels are closely related, thus also

providing enormous opportunities for vertical integration (Chu &You, 2012; Engell & Harjunkoski, 2012; Wassick et al., 2012).

At the molecule level, modeling and optimization tools canbe employed by researchers with diverse background to facilitate

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

ng an

tttsaanu

Fig. 2. Illustration of multi-scale modeli

heir studies on biomass-to-biofuel conversion technologies, bothheoretical and experimental. Optimization tools can help to iden-ify the promising reaction pathways that are cost-effective andustainable, thus saving unnecessary efforts in experimental tri-

ls. Furthermore, studies in molecular engineering, bioengineering,nd multi-physics simulation can be guided by the system engi-eering approaches, while in return serve as the practical basis forpper-level modeling and optimization.

d optimization of biofuel supply chains.

As illustrated by the arrows on the left of Fig. 2, the systemscale addressed in the model grows from top to bottom, whereasthe practical details considered in the model increases frombottom to top. The double-headed arrows between adjacent

layers indicate that the modeling layers are never isolated butclosely inter-connected. Generally speaking, the lower layerconveys target or requirement to the higher layer, whereas thehigher layer provides detailed data which serve as building

emica

bl

5

acgoc

5

epbdcssaGtd2btatfittmrtaptamppfambtcuubmt

5

fnoidAhaa

D. Yue et al. / Computers and Ch

locks for the modeling and optimization of the lowerayer.

.1. Multi-scale decisions

The optimization of biofuel supply chains requires consideringll the components across the entire biomass-to-biofuels supplyhain in a comprehensive model that allows the effective inte-ration of long term strategic design decisions with short termperational planning and scheduling decisions, which will be dis-ussed respectively in this section (Sharma et al., 2013).

.1.1. Strategic decisionsThe strategic decisions are those decisions that have an influ-

nce over years and decades, and even beyond the lifetime of theroject. Once a strategic decision is made, it is very unlikely toe altered in the short term. The first and most critical strategicecision is the design of biofuel supply chain network. We canonstruct a superstructure to optimize the selection of biomassuppliers, location of conversion facilities, assignment of customererving areas, and transportation links that connect different sitesnd deliver the biomass/biofuel across the supply chain network.eographic information system (GIS) can be employed to visualize

he spatial configuration of biomass resources, candidate sites, fuelemand densities, etc., thus facilitating the modeling process (Esri,012). The major tradeoff in the design of biofuel supply chain isetween the costs of transportation and the capital and produc-ion costs of production facilities, because the transportation costsccount for a significant portion in the biofuel supply chain whilehe construction of production facilities may not necessarily bene-t from the economy of scale. Besides, the selection of conversionechnologies and corresponding infrastructures is also an impor-ant strategic decision. Because different conversion technologies

ay vary in capital and operational costs, biofuel productivity andequirement for biomass resources, optimal selection of conversionechnologies may depend on the level and distribution of biomassvailability and fuel demand, as well as the local restriction andolicy at the facility candidate sites. Binary variables can be usedo represent the selection of processes and conversion pathwayst the supply chain level, and the selection of catalysts, equip-ent units, operation conditions and processing methods at the

rocess level. When evaluating the economic objectives (e.g., netresent value, annualized total cost), besides the conventional costsor construction, materials, and labors, we should also take intoccount the government incentives (both construction and volu-etric) and additional credits from selling the by-products (e.g.,

io-char) (DOE/EERE, 2013b). Another perspective to note is thathe configuration of the biofuel supply chain may not be static butan evolve over time. With the wide acceptance of biofuel prod-cts and increasing incentives, capacity expansion and technologypgrading on top of the existing biofuel supply chain system maye desired in the future. Particularly, it may be possible to adapt theulti-period planning models proposed for generic supply chains

o biofuel supply chains.

.1.2. Operational decisionsOperational decisions are those decisions that are adjusted more

requently in correspondence to the current external and inter-al conditions, which usually have impacts for no less than a yearr even a day. Due to the large number of activities involvedn the biofuel supply chain, optimization models for operationalecisions vary significantly in scale, complexity and formulation.

t the biomass acquisition phase, operational decisions involveow to allocate the operators to harvesting machines, how torrange the working shifts, how to assign the harvesting operationreas, how to convey the harvested biomass to storage units, etc.

l Engineering 66 (2014) 36–56 45

Optimization models regarding vehicle routing and fleet schedul-ing can be employed to cope with these problems. At the logisticsphase, one needs to figure out the optimal inventory policiesregarding how much to replenish from the upstream, how much todeliver to the downstream, how much to utilize for the productionof intermediates or fuel products, and what level of inventory tokeep in the storage units. These problems become more challengingas the supply of several types of biomass feedstocks would be sub-ject to season-to-season changes and limited in certain months of ayear. Lots of tools and theories developed for generic supply chainscan be incorporated for inventory optimization in biofuel sup-ply chains, including sophisticated forecasting models, enterpriseresource planning (ERP) system, material allocation rules, servicelevel measurements, information flow structure, etc. At the produc-tion phase, major operational decisions involve the coordination ofvarious material flows and scheduling of multiple processing taskswith continuous/batch equipment units in order to guarantee thetimely delivery of requested products both in quantity and qual-ity. Extensive studies on planning and scheduling in petroleumand chemical production industries can be adapted to study theproblem. Additionally, in cases that the quality and water con-tent of biomass feedstocks vary significantly, optimization of theoperational parameters (e.g., temperature, pressure, flow rate) andreal-time control trajectory would be required. Advanced controlalgorithms and dynamic simulation models exploiting the physi-cal and thermodynamic properties are helpful for handling theseproblems.

5.2. Uniqueness in biofuel supply chains

Compared to their petroleum counterparts, some major compo-nents in the supply chain systems for biofuel products are similarin function but different in structure and property. The most signif-icant difference is at the raw material acquisition stage. Petroleumemerges from point sources like drill shafts, guarantees perennialsupply and can be hauled over long distance with minimum trans-port costs. However, the raw biomass feedstock, as a low-energydensity solid rather than a high-energy density liquid, is inher-ently distributed over a large area and is subject to seasonality,thus requiring considerable collection and transportation effortsand large storage volumes (DOE/EERE, 2013c).

From the aspect of supply chain design, the locations of biomassprocessing facilities should be optimized in order to reduce thetransportation cost of low-energy density solid biomass resourcesfrom surrounding biomass suppliers (e.g., farms, forests, papermills) and to lower the capital and operational costs. The capac-ities of the biomass processing facilities should be able to meet therequirement in peak times and might be worth it to be repurposedfor part of the year. If a flexible processing facility able to treatdifferent types of biomass resources is built, multi-period planningcan be utilized to schedule the shifts of biomass suppliers accordingto the feedstock availability in each month.

From an operational point of view, the biomass is cost-prohibitive and unstable for long-distance transport because of itslow energy density and high water content. In the modeling prac-tice, a circle around the biomass supplier is often specified as thefeasible region for building biomass processing facilities, of whichthe radius equals to the “feasible distance” for raw biomass trans-portation (Uslu et al., 2008). Considering the low transport densityof biomass resources, both volume and weight limits may be spec-ified for the capacity of transportation links (You, Tao, Graziano, &

Snyder, 2012). In addition, degradation of biomass resources alsomakes the storage more difficult to handle. A linear monthly rate ofmaterial loss is usually assumed, varying in the storage technologyused. Correspondingly, we can incorporate in our model a mass

4 emical Engineering 66 (2014) 36–56

be

5

“smiifomfitpta

pEbloowipsia

otcficgorsfscsctc

tddd(Abeshdlmso

Table 2Total discounted capital, production, and transportation costs of the three designsunder different side length of the square world.

40 km 60 km 135 km 200 km

Centralized $290 MM/y $530 MM/y $2224 MM/y $5156 MM/y$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$3.06/GEG $2.34/GEG $1.54/GEG $1.38/GEG

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

6 D. Yue et al. / Computers and Ch

alance on biomass inventory accounting for the material loss atach biomass processing facility and in each month.

.3. Distributed vs. centralized strategy

In the case of petroleum refinery, there is the clear rule ofeconomy of scale”. Increasing aggregation of petroleum feed-tocks, typically by pipeline or ocean tanker, only increases costsarginally. With high pressure and temperature requirements,

ncreasing production scale follows a 0.6–0.8 scaling factor driv-ng economies of scale. However, this is not automatically trueor biorefinery (Dunnett, Adjiman, & Shah, 2008). The advantagesf larger plants might be offset by the increased costs of haulingore low density biomass feedstocks to the point of processing

or longer distance (Shastri, Rodriguez, Hansen, & Ting, 2012). Its also reported that transportation takes a much larger piece inhe cost pie of biofuel supply chains than that of petroleum sup-ly chains. Therefore, to reveal the tradeoffs between the costs ofransportation and process construction/operations, we considern illustrative example in this section.

We assume a superstructure of the hydrocarbon biofuel sup-ly chain in a hypothetical 4 × 4 matrix region as shown in Fig. 3a.ach green square area represents a farm, of which the yield ofiomass is evenly distributed. There are five candidate sites for

ocating the conversion facilities, as represented by blue circles. Thenly demand zone, represented by the pink circle, is at the centerf the square region. In this superstructure, two conversion path-ays are considered: (1) directly converting biomass to liquid fuels

n an integrated biorefinery (represented by yellow plant) and (2)reconverting biomass into bio-oil in fast pyrolysis plants (repre-ented by red plant) and then upgrading the bio-oil to liquid fuelsn a gasification plant (represented by blue plant). For clarity, welso illustrate the two conversion pathways in Fig. 3b.

On the basis of the superstructure, we consider three typesf supply chain designs, namely centralized, distributed andwo-stage. In the centralized scenario, all the biomass resourcesollected from the 16 farms are shipped to a large, integrated gasi-cation plant located in the center of the square region for theonversion to liquid fuels. In the distributed scenario, four inte-rated gasification plants of a smaller size are built at the centerf four corners of the square region, each processing the biomassesources from the four adjacent farms. The liquid fuels are thenhipped to the demand zone in the center of the square regionrom the four integrated biorefineries. The third scenario – two-tage design – includes four fast pyrolysis plants located at theenter of four corners and an upgrading facility at the center of thequare region. The biomass resources collected from the four adja-ent farms are preconverted to bio-oil in the pyrolysis plants. Thenhe bio-oil is shipped to the central upgrading facility for furtheronversion to liquid fuels.

Since our focus is on the tradeoff between the costs ofransportation and process construction/operations, the majorifferences between the three designs are in facility locations, pro-uction capacities, and technology selections. All the data and aetailed discussion are provided in the work by You and Wang2011). Here in this paper, we merely present the major results.s summarized in Table 2, the centralized layout might be theest option for the 40 km × 40 km square region, because of theconomy of scale and higher efficiency of integrated conver-ion. Although with the highest transportation cost and slightlyigher capital cost than does the centralized design, the two-stageesign can be still considered a viable approach, because of its

ow operating cost resulting from the credit from charcoal, theajor byproduct in preconversion processes. If in a 60 km × 60 km

quare region, the two-stage design becomes the least costlyption, because the larger area implies a higher growth rate of the

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

transportation cost, which can be significantly reduced throughtwo-stage design. However, we note that the centralized designstill yields the minimum unit cost, because of the relatively lowconversion efficiency of the two-step processing of the two-stagedesign. When considering a 135 km × 135 km square region, thetwo-stage design has both the minimum total cost and minimumunit cost. Moreover, the difference between the costs of the central-ized design and distributed design becomes smaller. Interestingly,in a 200 km × 200 km square region, the distributed design becomesmore cost-effective than the centralized design. The reason is thatthe higher transportation cost resulting from the larger area offsetsthe capital saving of the economy of scale in the centralized design.

5.4. Competitions between different parties

In most existing literature, the entire biofuel supply chain is con-sidered as entity centralized system. This might be true if all partiesinvolved in the biofuel supply chain are cooperative, which happensin the case that the farms, production facilities, and distributionsystems are all acquired by the same organization. However, moreoften, the parties are non-cooperative, thus causing competitionin the price and use of materials. In such scenarios, game theoryis a powerful tool for the study of mathematical models concern-ing conflict and cooperation between intelligent rational partiesin the supply chain system. In game theory, the interactions andcompetitions between different parties can be modeled as varioustypes of games depending on their behaviors, including coopera-tive or non-cooperative games, symmetric and asymmetric games,zero-sum and non-zero-sum games, simultaneous and sequentialgames, etc. (Dutta, 1999; Myerson, 1997).

A typical example of the non-cooperative biofuel supply chainis the competition for biomass resources with agricultural sectorat the biomass acquisition stage (Bai, Ouyang, & Pang, 2012). Forinstance, if corn is considered as the feedstock for biofuel produc-tion, the biofuel company might offer certain prices to procurefeedstocks and the farmers have the option to ship and sell theircorns to local grain markets or nearby biomass processing facili-ties. In this case, the biofuel company can optimize their offeredprices to maximize its profit, while the set of spatially distributedfarmers would react accordingly to maximize their individual prof-its. The problem can be regarded as a Stackelberg leader-followergame (von Stackelberg, 1934) and formulated as a bi-level supplychain optimization problem.

Another approach is to introduce proper contract mechanismsto leverage the interests of different parties, among which theprice transfer policy is an intuitive and efficient method to cre-ate a fair, optimized profit distribution in the biofuel supply chain(Gjerdrum, Shah, & Papageorgiou, 2001). The transfer price is theselling price at the material supplier and the purchasing cost at thematerial receiver, which is also the major variable in this optimiza-

tion problem. To avoid unfair profit distribution between partiesin the biofuel supply chain, a nonlinear Nash-type (Nash, 1950)objective function can be employed.

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

F son ofp der is

5

rbbSlcoce(stht

ttttTmctpbi

ig. 3. Illustration of centralized vs. distributed supply chain networks. (a) Compariathways. (For interpretation of the references to color in this figure legend, the rea

.5. International trade of biomass/biofuels

Although current focus is on the use of domestic biomassesources for biofuel production, international trades ofiomass/biofuels may also provide win–win opportunities tooth exporting and importing countries (Kaditi & Swinnen, 2007).ome countries, especially developing countries, have large vacantand, lower cost of labor and equipment, from which importingan help domestic biofuel industries to improve competitivenessf biofuel products, to create more stable biomass supplies, and toontribute to economic development and technology innovation inxporting countries (Junginger et al., 2008). For instance, Uslu et al.2008) provided a comprehensive analysis of the biomass/biofuelupply chain from Latin America to the European Union. Addi-ionally, as the largest wood pellets exporter in the world, Canadaolds considerable potential for international biomass/biofuelrades with the biofuel industry in the U.S. (NREL, 2013b).

Long-distance transportation is inevitable in internationalrades of biomass/biofuels. Since raw biomass resources are usuallyoo low in energy transport density, there is a universal consensushat raw biomass needs to undergo pretreatment and densifica-ion processes to improve the energy density and biomass stability.herefore, the remaining problem is to determine which pretreat-ent method, at what point in the supply chain, with which

onversion technology and what transportation mode would give

he optimal delivery performance for international biofuel sup-ly chains. International trading options include direct import ofiofuels products produced in exporting countries, or import of

ntermediate energy carriers (e.g., torrefaction pellets, pyrolysis oil)

three designs of the biofuel supply chain. (b) Illustration of centralized-distributed referred to the web version of this article.)

converted in exporting countries. Furthermore, when consideringthe transportation mode for biomass/biofuel shipments, pipelinetransportation, though requiring considerable upfront costs, mayemerge as a solution that is economically viable in the long run, inaddition to truck, train and barge (DOE, 2010).

Another interesting problem inherent in international trades ofbiomass/biofuel supply chains is associated with the multi-nationalor cross-region financing activities. When the materials are shippedacross the country/state borders, currency exchange, custom tariffsor different state taxes may incur (Lamers, Junginger, Hamelinck, &Faaij, 2012). All these factors may lead to additional costs for inter-national or inter-region trades, thus should be taken into account inthe economic evaluation of the biofuel supply chain project. Addingto this point, carbon trading platforms would also contribute to theeconomic performance and sustainability of international biofuelsupply chains. Since countries may have deficits/surplus in theirrespective carbon emission quota, incorporating importing scenar-ios into the optimization model can help to achieve efficient use ofresources while to meet the environmental restrictions. Anotheralternative is to trade renewable fuel credits on the internationalmarket and enable local optimization of supply chains. This methodcould provide sufficient incentive to increase biofuel productionabove local mandates.

5.6. Integration with other supply chains

5.6.1. Integration with petroleum supply chainsConventionally, biorefineries are considered independent of

petroleum supply chains. However, with the development of

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

rces in

abotsc2

•

•

•

muuutbrpbtiYbts

tclfbat

Fig. 4. Entry points for biomass resou

dvanced hydrocarbon biofuels, great opportunities emerge foriofuel industry to take advantage of the existing infrastructuref petroleum supply chains (mainly refinery units and pipelines),hus saving considerable amounts of capital and operation costs. Ashown in Fig. 4, three possible insertion points for biofuel supplyhain to entry the petroleum infrastructure include (Bunting et al.,010; DOE/EERE, 2013d),

Insertion point (a): A bio-crude that can be co-processed withconventional crude oilInsertion point (b): Refinery-ready intermediates that are com-patible with specific refinery streams for further processing at therefineryInsertion point (c): A near-finished fuel or blend stock that willbe minimally processed at the refinery.

For insertion point (a), bio-crude (e.g., pyrolysis oil) can beixed with petroleum crude oil, then sent to crude distillation

nits, and finally converted to fuel products via a series of upgradingnits. For insertion point (b), bio-crude is directly sent to upgradingnits after certain treatment to remove oxygen and other con-aminants. For insertion point (c), there are two options, namelylending biofuels with conventional fossil-fuels in the petroleumefinery and then shipping them to the customers using existingipelines, or delivery of biofuels directly to distribution centers andlending there. Note that, minimum capital costs are still requiredo retrofit petroleum refinery units to be compatible with thesensertion points (Tong, Gleeson, Rong, & You, 2013; Tong, Gong,ue, & You, 2013). Depending on regional air emissions standards,lending of a finished biofuel might require specific adjustments ofhe petroleum fuel to meet requirements for Reformulated Blend-tock for Oxygenate Blending (RBOB).

From the modeling perspective, there are abundant existingools for the design, planning and scheduling of petroleum supplyhains due to its maturity, whereas there is currently a significantack of supporting tools for the decision-making in emerging bio-

uel supply chains. Therefore, with the resemblance and differenceetween the two types of supply chains in mind, we should takedvantage of the existing tools and exploit the synergies amonghe various activities in the integrated system. A superstructure

to existing petroleum infrastructure.

can be constructed to incorporate the biofuel supply chain intoexisting petroleum supply chain systems with binary variables rep-resenting the selection of biomass/biofuel insertion points, choiceof conversion technologies, existence and capacity level of biomasspretreatment units, as well as other retrofitting decisions associ-ated with catalysts, processing methods, process heat integration,hydrogen demand, etc. Furthermore, as additional material streamsintroduced from the biofuel supply chain are involved in the inte-grated production system, more efforts on planning are required forthe purchase, sales, transportation, and storage, but more impor-tantly, on the scheduling in raw material processing and fuelproduction to coordinate the additional material streams in orderto achieve maximum utilization of refinery units and maximumoverall profit from both biofuel and petroleum products.

5.6.2. Integration with carbon mitigationCO2 capture and sequestration (CCS) is a major means to reduce

greenhouse gas emissions, especially from centralized energyconversion sites (e.g., power stations, petroleum refineries). Con-ventional practice for carbon sequestration is to remove CO2from the atmosphere and deposit in a reservoir (e.g., soil, ocean)(DOE/NETL, 2010). However, the recent development of algae-based biofuel technology creates new options for CO2 sequestrationas well as the opportunities to take advantages of the carbon con-tent for the growth of algae biomass, thus increasing the atomefficiency and leading to CO2 capture and utilization (CCU).

There is a large body of literature investigating the potentialof using algae for carbon mitigation, but research on how to con-nect the carbon-sources with algae-based carbon-sinks is relativelylimited. Some researchers suggest building the algae cultivationfacilities close to the carbon-sources (Wilcox, 2012). However, thisoption requires considerable amounts of capital costs and may notbe optimal with a large number of carbon-sources that are locatedgeographically apart. In this scenario, optimization on sitting ofalgae cultivation sites and design of CO2 pipeline transport networkmay come into play. When the total amount of CO2 needed to be

transported is rather small, simple source-sink matching approachmay be applicable at the early stage of CCU development. How-ever, in scenarios with a large amount of CO2 to be transportedand a great number of carbon-sources to be matched, source-sink

emica

mon

taspwocrbtCsaopdvaaraB

6s

tIoeaowlc

6

ttdmi(Lu2aboeadecH

6

a

D. Yue et al. / Computers and Ch

atching approach appears less efficient and superstructure-basedptimization approaches for the design of CO2 pipeline transportetwork may become more appealing.

The major costs of the CO2 pipeline transport network involvehe capital costs for installing the pipelines and pump stations,s well as the costs associated with operation and energy con-umption (Zhou, Liu, & Li 2013). It is the pressure provided at theump stations that drives the CO2 to flow through the pipelines,hich has to be above a certain value to maintain the supercritical

r dense-phase liquid states of CO2 to ensure transportation effi-iency, meanwhile cannot exceed certain limits in order to avoidisks of explosion and other safety issues. Therefore, the tradeoffetween the number of pump stations and the distance of CO2ransportation needs to be established. An insight on the design ofO2 pipeline network, is that one needs to avoid valleys and depres-ions because a pipeline failure could starve a region of breathableir. Binary variables can be introduced to represent the existencef a pump station at a certain candidate node, the existence of aipeline along a certain segment, as well as the selection of theiameters of the pipelines with discrete standard sizes. Continuousariables can be used to represent the pressure rise, flow rate of CO2,nd the amount of CO2 available at carbon-sources and the capacityt algae cultivation sites. Existing works on optimization of natu-al gas pipeline networks may be good references for the designnd operation of CO2 pipeline transport networks (Baumrucker &iegler, 2010; Gopalakrishnan & Biegler, 2013).

. Modeling sustainability issues in bioenergy and biofuelupply chains

As a renewable alternative to conventional fossil-fuels, the sus-ainability issues associated with biofuels deserve our attention.n order to reduce adverse impacts on ecosystem and improve theverall economic profitability and social benefits, systematic mod-ling and optimization frameworks are required to simultaneouslyssess and identify the sustainable solutions for the design andperation of biofuel supply chains (You et al., 2012). In this section,e will introduce how modeling and optimization tools can be uti-

ized to address the sustainability issues in biofuel supply chains,oncerning environment, society, and economy, respectively.

.1. Environmental impacts

Accompanied by the rise in environmental awareness andightening of environmental policies, the study of environmen-al sustainability has received increasing attention in the pastecades. Among the various approaches, life cycle assessment (LCA)ethodology stands out to be the most successful tool for evaluat-

ng and analyzing the environmental impacts of product systemsAzapagic, 1999; Azapagic & Clift, 1999; Fava, 1994). The classicalCA framework is based on stages and processes, which is well reg-lated by ISO standards and specific methodology guidelines (ISO,006). In addition, input–output models and hybrid methods arelso promising tools for LCA, of which Eco-LCA as an example, isased on an integrated ecological-economic model of the U.S. econ-my and accounts for various ecosystem goods and services (Bakshit al., 2013). We note that the uncertainties in environmental evalu-tion can have significant influences on the reliability of LCA-basedecisions. These uncertainties may stem from the limited knowl-dge about the physical processes under study and the normativehoices regarding scenarios and mathematical models (Huijbregts,eijungs, & Hellweg, 2004).

.1.1. Four-phase life cycle assessmentHere in this paper, considering its relative maturity and wider

cceptance, we would like to briefly introduce the classical

l Engineering 66 (2014) 36–56 49

process-based LCA methodology, following the four-phase proce-dure suggested by ISO.

6.1.1.1. Goal and scope definition. In this phase, project description,goal of study, boundary of system and definition of functional unitare stated. Depending on the goal of study, the system boundarycan be “cradle-to-grave”, “cradle-to-gate” or “gate-to-gate”. In thestudy of biofuel supply chains, the “cradle-to-gate” boundary isoften adopted, including the life cycle stages from biomass culti-vation, harvesting, pretreatment, through transportation, storageand conversion, to the gate of finished-product distribution centers.Moreover, the definition of functional unit is also critical, based onwhich, all the calculations will be performed and normalized. Forinstance, for a biofuel supply chain that produces biomass-derivedgasoline, diesel and jet fuel, the “gasoline gallon equivalent (GGE)”– a unit based on energy content – can be selected as the functionalunit (Wikipedia, 2013b).

6.1.1.2. Inventory analysis. This second stage of LCA involves thecompilation and quantification of life cycle inventory (LCI) associ-ated with each process/stage within the life cycle boundary. LCI issimply a list of material inputs and outputs for a given product sys-tem throughout its life cycle, including the cumulative extractionof resources from the environment (e.g., primary energy, mineralresources) as well as the cumulative emissions to the environment(e.g., gas emissions, liquid and solid wastes). LCI for a lot of pro-cesses can be found in commercial LCA databases such as Ecoinvent(2008), GaBi (2011), GREET (2012), and openLCA (2012). However,for new processes that are not documented, cumulative LCI shouldbe calculated according to energy and mass balance, using the unitprocess raw data and LCI of background processes. Sometimes, on-site measurements may be necessary.

6.1.1.3. Impact assessment. In this phase, the LCI obtained from theprevious phase is translated according to certain damage assess-ment models. These models can generate a unified or a series ofenvironmental performance indicators, which are easily under-standable and user-friendly numbers. The most widely-used LCAmetrics include GWP (IPCC, 2007), Eco-indicator 99 (Goedkoop& Spriensma, 2001), IMPACT 2002+ (Humbert, Schryver, Bengoa,Margni, & Jolliet, 2012), etc. GWP focuses on greenhouse gas (GHG)emissions causing the global warming effects, while Eco-indicator99 and IMPACT 2002+ evaluate the environmental impacts in morecomprehensive categories (e.g., human health, ecosystem qualityand resources). We note one critical problem that has been recentlyhighlighted is about the water footprint of biofuel production sys-tems (Chiu, Suh, Pfister, Hellweg, & Koehler, 2012; Gerbens-Leenes& Hoekstra, 2011), which should also be explicitly dealt with in bio-fuel supply chain design (Bernardi, Giarola, & Bezzo, 2012; Bernardi,Giarola, & Bezzo, 2013; Cucek, Klemes, & Kravanja, 2012a,b).

6.1.1.4. Interpretation. In the final phase, the LCA results are ana-lyzed to provide a set of conclusions and recommendations, usuallyin the form of written reports. In this regard, when we considerthe biofuel supply chain, the goal of LCA is to provide criteriaand quantitative measurements for comparing different networklayouts and operation alternatives. However, one of the criticaldrawbacks of classical LCA framework is that it lacks a system-atic approach for generating such alternatives and identifying the

best one in terms of environmental performance (Grossmann &Guillén-Gosálbez, 2010). Therefore, most researchers merely useLCA as a post-optimization tool for the evaluation of environmentalsustainability.

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

ization

6

mS2TcetctoawOpc

6

qefbotTueoni&Gs(

6

6

dmloP2duma

Fig. 5. Life cycle optim

.1.2. Life cycle optimizationTo circumvent the aforementioned limitation, the life cycle opti-

ization framework was proposed in recent years (Gebreslassie,livinsky, Wang, & You, 2013a; Gebreslassie, Waymire, & You,013b; Wang, Gebreslassie, & You, 2013; Yue & You, 2013b).his general modeling framework couples optimization tools withlassical LCA methodology, allowing us not only to evaluate thenvironmental impacts of diverse design and operation alterna-ives but also to identify the optimal solution (Fig. 5). When otherriteria (e.g., economic and social) are considered at the same time,he life cycle optimization framework is also able to perform multi-bjective optimization. In this case, the results are often presenteds a set of Pareto-optimal solutions, consisting of a Pareto frontierhich reveals the tradeoffs in decision under multiple objectives.ne can choose among these Pareto-optimal solutions based on thereference for the design and operation of potential biofuel supplyhains.

.1.3. Functional-unit-based evaluationFor most economic objectives, we often try to optimize the total

uantity, i.e., maximizing total profit or minimizing total cost. How-ver, for environmental criteria, the “environmental impact perunctional unit” is more critical (Yue, Kim, & You, 2013b). This isecause, for example, one may care more about the carbon footprintf producing one gallon of biomass-derived gasoline, rather thanhe annual total CO2 emission of the entire biofuel supply chain.herefore, the environmental impact evaluated based on functionalnit is, in fact, the indicator of how “green” the supply chain is. How-ver, one arising issue is that functional-unit-based environmentalbjective function can involve fractional terms, thus introducingonlinearity into the model. Advanced solution methods includ-

ng parametric algorithm (You, Castro, & Grossmann, 2009; Zhong You, 2014), reformulation–linearization method (Yue, Guillén-osálbez, & You, 2013a), as well as several general-purpose MINLPolvers and global optimizers can be utilized to resolve the issueFloudas, 1999; Grossmann, 2002).

.2. Social factor

.2.1. Job creationThe most important perspective in the social dimension for the

esign and operation of biofuel supply chains may be the employ-ent effect, which can be measured by the number of accrued

ocal jobs (full-time equivalent for a year) in a regional econ-my (Bamufleh, Ponce-Ortega, & El-Halwagi, 2013; Lira-Barragán,once-Ortega, Serna-González, & El-Halwagi, 2013; You et al.,012). The Jobs and Economic Development Impact (JEDI) model

eveloped at National Renewable Energy Laboratory (NREL) can besed and integrated into the aforementioned multi-objective opti-ization framework, to systematically evaluate the social impacts

ssociated with the construction and operations of biofuel supply

modeling framework.

chains in the United States (NREL, 2012). To evaluate the numberof jobs that will accrue to the state or local region from a project,JEDI performs an input–output multiplier analysis, where a mul-tiplier is a simple ratio of total systemic change over the initialchange (e.g., employment) resulting from a given economic activity(e.g., startup or termination of a project). The multipliers are esti-mated through economic input–output (EIO) models, which wereoriginally developed to trace supply linkages in the economy, andto quantify the effects of change of expenditure within a regionaleconomy in multiple industry sectors. In the input-out analysis,the employment effects from the construction and operation of abiofuel supply chain can be classified into three categories: direct,indirect and induced.

• Direct effect: The immediate or on-site effect created by expen-diture. For example, in constructing a bio-refinery plant, directeffects may include the on-site contractors and hired crews inthe civil engineering team. In addition, direct effects may includethe jobs for building the process equipment units at bio-refineryplants.

• Indirect effect: The increase in economic activity that occurswhen contractors, vendors, or manufacturers receive paymentfor goods or services and in turn are able to pay others who sup-port their business. For instance, indirect effects may include thebanker who finances the contractor, the accountant who keepsthe contractor’s books, and the steel mills, electrical manufactur-ers and other suppliers that provide the necessary materials.

• Induced effect: This refers to the change in wealth that occurs oris induced by the expenditure of those persons who are directlyor indirectly employed by the project.

The total employment effect from a single expenditure canbe calculated by summing up the three kinds of effect, usingregional-specific multipliers and personal expenditure patterns.State-by-state multipliers for employment, economic activity, andpersonal expenditure patterns can be derived from the IMPLANProfessional model (IMPLAN, 2013). In this way, the changesin employment brought about by investment in the design andoperation of a biofuel supply chain can be matched with thecorresponding multipliers for each industry sector. Both the one-time impacts resulting from the construction phase and theyearly impacts resulting from the annual operations along theproject lifetime should be considered in the measure of socialimpacts.

6.2.2. Other perspectivesBesides job creation, the social dimension still involves many

other perspectives such as human rights, labor practices, decentwork conditions, society wellness, product responsibility, etc.Despite of the existence of several conceptual frameworks, thequantitative evaluation methodology for such social perspectives

emical Engineering 66 (2014) 36–56 51

iuttptSWLie

6

aafcsror

oocobaucoeasrprtsbtitt2mt

7s

dfmthlsfiip

D. Yue et al. / Computers and Ch

s still immature (UNEP, 2009). One of the promising approachesnder research right now is to broaden the scope of LCA frameworko include social inventory results. However, by the time of writinghis review, the choice and formulation of indicators as well as theosition towards the use of generic data are still under debate inhe social LCA community (Benoît et al., 2010; Dreyer, Hauschild, &chierbeck, 2006; Jørgensen, Bocq, Nazarkina, & Hauschild, 2008;eidema, 2006). Although no appropriate software tools for social

CA are currently available, there are several pilot developmentsncluding openLCA (openLCA, 2012), Earthster (GreenDelta, 2013),tc.

.3. Economic factor

The idea of economic sustainability is to promote the use ofvailable resources in a way that is both efficient and responsible,nd likely to provide long-term benefits. In addition, an alternativeeedstock can reduce risk from disruption of a highly-concentratedorrelated market. In the case of design and operation of a biofuelupply chain, it calls for using resources (e.g., capital, land, labor,esearch) wisely so that the business continues to healthily functionver a number of years, while consistently returning a risk-adjustedeturn.

Several metrics can be used as the economic objectives in theptimization model for biofuel supply chains. The most commonnes are net present value (Wikipedia, 2013c) and annualized totalost (Wikipedia, 2013a). Both metrics account for the time valuef money over the lifetime of the project (usually 15–20 years foriofuel supply chains) by using discounted cash flow analysis. Inddition, it is believed that minimizing the cost associated with pernit of biofuel products would lead to a more cost-effective supplyhain, thus promoting the competitiveness and market acceptancef the biofuel products. Besides, as a renewable alternative, thefficiency of energy and resource use in biofuel supply chains islso an important indicator. We can minimize the cumulative fos-il energy demand or the consumption of certain type of criticalesources (e.g., labor, materials) associated with per unit of biofuelroducts, in order to achieve the maximum utilization of availableesources (Althaus et al., 2010). Responsiveness is another impor-ant indicator affecting the economic performance of the biofuelupply chain, because to what extent the order of customers cane fulfilled on time is a touchstone for determining the quality ofhe design and operation of a supply chain. By using the stochasticnventory theory, we can employ the guaranteed service time fromhe gate of distribution centers to biofuel retailers or individual cus-omers as one of the optimization objectives (You & Grossmann,008a, 2011a). Again, since multiple objectives may be involved,ulti-objective optimization techniques can be utilized to simul-

aneously optimize the objectives of interest.

. Uncertainty analysis and risk management for biofuelupply chains

Uncertainties are ubiquitous, of various types and varyingegrees, and emerging from all stages and activities in the bio-uel supply chain (Awudu & Zhang, 2012). Corresponding to the

ulti-scale decisions, we can also broadly categorize the uncer-ainties into strategic and operational levels, which are usuallyandled with by optimization models of different scales and formu-

ations (Lee, 2012). The most widely employed approaches includecenario analysis, stochastic programming, robust optimization,

uzzing programming, chance-constraint optimization, stochasticnventory theory, etc (Sahinidis, 2004). A summary is providedn Fig. 6, including a selected number of uncertainty sources andromising approaches for optimization under uncertainty.

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

7.1. Strategic uncertainty

The most significant type of strategic uncertainty for biofuelsupply chains might associate with the government incentives andpolicies (MASBI, 2013). Due to the relatively high production cost ofbiofuel products and their limited market acceptance at the currentstage of development, the incentives and supporting policies serveas a main driver of the biofuel industry, making the biofuel productscompetitive with their petroleum counterpart both in terms of eco-nomic cost and sustainable performance. However, the incentivesand policies may be changed after the biofuel supply chain is built,due to the changes in economic, political or technical environments,thus causing deviation between the realized profit return and theoriginal anticipation. Furthermore, political uncertainty increasesperceived market risk, increasing financing costs to account forrisk-adjusted returns. Scenario analysis would be straightforwardto address the uncertainty in incentives and policies by inves-tigating the performances in optimistic, neutral and pessimisticscenarios. Alternatively, stochastic programming can be employedto maximize the profit expectation among a large number ofpossible scenarios. Facility disruption under catastrophic events(e.g., earthquake, hurricane, plant explosion, or terrorist action) isanother type of strategic uncertainty that might cause the paral-ysis or malfunctioning of the biofuel supply chain system. Metricsregarding the resilience under destructive disruptions can be incor-porated into the optimization model as one of the objectives at thesupply chain design phase to enhance the robustness of the supplychain design, though usually with additional costs. Alternatively, incases that probabilities of disruption can be estimated from histor-ical data or operating experiences, stochastic programming wouldbe a great tool to handle facility disruption. Besides, technologyevolution is an important type of strategic uncertainty. Consider-ing that many biofuel technologies are relatively young, the processconversion rate, energy efficiency, and material handling costsare expected to improve in the future. Therefore, one may need

to determine whether to employ mature technologies with lessimproving space or nascent technologies with considerable poten-tial, which makes stochastic programming (Birge and Louveaux,1997) a proper tool for solving this issue. The last type of strategic

5 emica

umbccRaa

7

tftoetas(sbewceemfpsotrtTtM2wGra(bmdvG

8d

bibf

8

obT

2 D. Yue et al. / Computers and Ch

ncertainty we would like to mention corresponds to the cli-ate and weathers. Different from petroleum resources, the annual

iomass yield is largely dependent on the sunlight, precipitation,limate, as well as pest threats. For biomass, there are potentialatastrophic risks at both the seasonal and multi-seasonal scales.obust optimization might be the appropriate tool for the designnd operation of biofuel supply chains to guarantee the quantitynd quality of biomass supplies.

.2. Operational uncertainty

Compared to the strategic uncertainties, the operational uncer-ainties are encountered thus need to be taken care of on a morerequent basis. The most recognized type of operational uncer-ainty is the volatility in prices and costs, namely purchasing pricesf biomass resources at farms, bills for utilities (e.g., natural gas,lectricity, stream, etc.) and costs associated with various activi-ies (e.g., labor, transport fuel, etc.). Note that pricing uncertaintiesre different between biomass and petroleum. Petroleum feed-tock and product prices are correlated. However, biomass productsbiofuels) are coordinated to petroleum fuels, while biomass feed-tock costs are more correlated to climate and weather. In addition,ecause of the variability in biomass feedstocks and the lack ofxperienced workers, the uncertainty in process yields associatedith disturbances in operating conditions, errors in operation, etc.

an lead to unqualified biofuel products and delay in order deliv-ry. We suggest probability theory and statistical tools can help toxplore the uncertainty dynamics. Whereas, robust optimizationay provide assistance to ensure satisfactory performances of bio-

uel supply chains under volatile market environment and unstablerocess outcomes. Another two types of operational uncertainty –upply delays and demand fluctuation – are common topics in theperations research community. Supply delays, also known as leadime variability, often occur during batch production and mate-ial transportation, while demand fluctuation is closely related tohe local consumption behavior and regional economy structure.here is no doubt that accurate forecasting models are criticalo make biofuel companies well-informed of the future market.

etrics such as conditional value-at-risk (Rockafellar & Uryasev,002) and downside risk (Eppen, Martin, & Schrage, 1989), coupledith the corresponding models and algorithms (You, Wassick, &rossmann, 2009), can also be employed to control and manage the

isks at the optimal design and operation of biofuel supply chains. Inddition, optimization models based on stochastic inventory theoryGraves & Willems, 2000, 2005; You & Grossmann, 2008b) mighte the most promising approach, because they can explicitly opti-ize the working and safety inventory to hedge against the supply

elay and demand fluctuation, whilst guarantee the required ser-ice level (You & Grossmann, 2011b; Yue & You, 2013a; You, Pinto,rossmann, & Megan, 2011).

. Supply chains of non-fuel end-uses for biomasserivatives

Besides the production of biofuels for the transportation sector,iomass sources also provide a sustainable alternative for produc-

ng electric power and heat. Furthermore, valuable co-products andyproducts can be obtained accompany with the production of bio-uels, which leads to additional profit for the biorefinery industry.

.1. Biopower and biomass-to-heat supply chain

Among the renewable power generation options, biopowerffers the advantage of being dispatchable, that is, the biomass cane stored for the use of power plant operation to serve the load.his avoids many of the grid integration challenges associated with

l Engineering 66 (2014) 36–56

intermittent sources like wind and solar energy (EPRI, 2010). Inaddition, there has been a long history of the use of biomass forheat supply. In modern practices, electric power and heat are oftengenerated simultaneously, namely cogeneration or combined heatand power (CHP) (Shipley, Hampson, Hedman, Garland, & Bautista,2008).

Compared to the biofuel supply chain system, the upstreamof the biomass-to-power-and-heat supply chain is almost thesame, including biomass harvesting, logistics, and pretreatmentinto pellet, chip, or pyrolysis oil. Therefore, the aforementionedoptimization tools for the upstream of the biofuel supply chain alsoapply in this case. Furthermore, the downstream of the biomass-to-power-and-heat supply chain is identical to the current distributionsystem for power and heat. Biopower can be easily integrated withthe grid and serve the loads in remote areas, while heat can be car-ried by hydrosystems to serve nearby facilities. Hence, the majormodeling challenge is at the production phase.

According to the technology characterizations developedby NREL and EPRI (Electric Power Research Institute), thereare three primary conversion routes for CHP applications ofbiomass—direct combustion, gasification, and cofiring. Moreover,there are numerous types of boiler technologies, e.g., spreaderstocker, fluidized-bed boiler, cyclone boiler, etc., which vary in cost,efficiency, technology maturity, as well as sustainability perfor-mances (Bain, Amos, Downing, & Perlack, 2003; Wiltsee, 2000).Binary variables can be used to facilitate the optimal selectionof these routes and technologies in the supply chain system, interms of economic, environmental, and/or social metrics. In sce-narios where biomass is co-fired with coal, scheduling tools canhelp to coordinate the two material streams. Additionally, in sce-narios where existing fossil-based power plants are to be retrofittedand renovated for CHP from biomass, multi-period planning toolswould contribute to the optimization of project evaluation, con-struction planning, resource allocation, etc. Superstructure-basedoptimization always plays an important role in the process design,no matter for biorefineries or biomass-based CHP facilities.

8.2. Biomass-to-chemicals supply chain

The petrochemical industry makes a wide range of productsfrom fossil fuels. These products include plastics, fibers, solvents,and other products that are ubiquitous in our daily life. Actually,the same or similar products can, for the most part, be potentiallymade from biomass as well (NREL, 2013a; PNNL, 2013). The U.S.Department of Energy have identified a list of “Top 10” buildingblock or platform chemicals in 2004 from more than 300 candidates(Werpy & Petersen, 2004), which is recently revisited by Bozelland Petersen (2010). At the time when the report was produced,most platform chemicals are sugar-based. However, with the devel-opment of advanced biofuel technologies, it becomes appealingto produce high-value-added low-volume bioproducts accompanywith the high-volume but low-value hydrocarbon biofuels. Differ-ent from petroleum feedstock, biomass has much higher oxygencontent than hydrocarbons fuels. Therefore, its conversion to fuelsrequires release of the oxygen content either as CO2 or other oxy-genated species, which might lead to distinct synergeis to producehighly-oxygenated biobased chemicals along with drop-in biofuels.From fossil sources, these compounds would have been producedby oxydizing reduced hydrocarbons (Petersen, Bozell, & White,2013).

To enhance the overall profitability and productivity andstimulate expansion of the biorefining industry, multi-scale mod-

eling and optimization can play an important role in integratingthe production of biofuel and advanced bioproducts. Systematicapproaches can be employed to make the assessment, screeningand identification of promising platform chemicals and reaction

emica

pgomstSucitosia

9

rccpmaeipmafistssirstbefpa

A

I(AsOUurcppe

R

A

D. Yue et al. / Computers and Ch

athways more efficient. Furthermore, process synthesis and inte-ration and multi-physics process simulation can help to lower theverall energy intensity of unit operations in biorefineries, maxi-ize the use of all feedstock components, byproducts and waste

treams, and use economies of scale, common processing opera-ions, materials, and equipment to drive down all production costs.ince the additional material streams related to advanced bioprod-cts may be similar to those in the petroleumchemical supplyhains, we can also leverage the tools on planning and schedul-ng to coordinate the use of resources in production and optimizehe logistic activities in the downstream distribution system. More-ver, the upstream of the biomass-to-chemicals supply chain is theame as the biofuel supply chain, thus the aforementioned model-ng and optimization methods for biomass harvesting and logisticslso apply in this case.

. Conclusion

Although ethanol and bio-diesel are the most worldwideecognized biomass-derived liquid transportation fuel productsurrently, it is foreseeable that in the near future advanced hydro-arbon biofuels would experience a rapid growth due to theolitical expectation, economic promise, environmental require-ent, and social benefits. As reviewed in this paper, cellulosic and

lgae-based hydrocarbon biofuels have numerous advantages overthanol and bio-diesel in compatibility with existing distributionnfrastructure, implication on food price and land competition, andotential for substituting various high-value petrochemicals. Theulti-scale modeling and optimization framework can preserve

holistic view and facilitate informed systematical researches inelds of molecular engineering, unit operation, process design,upply chain management, and sustainability assessment. Oppor-unities and challenges emerge from multiple spatial and temporalcales in the biofuel supply chain systems are discussed. Per-pectives on centralized-distributed network design, multi-playernteraction, and integration with other supply chain systems areevealed in this paper. Additionally, we classify decisions intotrategic and operational levels according to their influences onhe temporal scale. We also investigate issues associated withiofuel supply chains in the threefold sustainability—economy,nvironment, and society, as well as the uncertainty and risks. Non-ood end-uses of biomass derivatives are also introduced in thisaper, identifying the opportunities for power and heat supply anddvanced bioproducts.

cknowledgement

The authors gratefully acknowledge financial support from thenitiative for Sustainability and Energy at Northwestern UniversityISEN) and Argonne National Laboratory via a Northwestern-rgonne Early Career Investigator Award for Energy Research. Theubmitted manuscript has been created by UChicago Argonne, LLC,perator of Argonne National Laboratory (“Argonne”). Argonne, a.S. Department of Energy Office of Science laboratory, is operatednder Contract No. DE-AC02-06CH11357. The U.S. Governmentetains for itself, and others acting on its behalf, a paid-up nonex-lusive, irrevocable worldwide license in said article to reproduce,repare derivative works, distribute copies to the public, anderform publicly and display publicly, by or on behalf of the Gov-rnment.

eferences

kgul, O., Shah, N., & Papageorgiou, L. G. (2012). Economic optimisation of a UKadvanced biofuel supply chain. Biomass & Bioenergy, 41, 57–72.

l Engineering 66 (2014) 36–56 53

Akgul, O., Zamboni, A., Bezzo, F., Shah, N., & Papageorgiou, L. G. (2011). Optimization-based approaches for bioethanol supply chains. Industrial & EngineeringChemistry Research, 50, 4927–4938.

Aksoy, B., Cullinan, H., Webster, D., Gue, K., Sukumaran, S., Eden, M., & Sammons,N. (2011). Woody biomass and mill waste utilization opportunities in Alabama:Transportation cost minimization, optimum facility location, economic feasibil-ity and impact. Environmental Progress & Sustainable Energy, 30, 720–732.

Althaus, H.-J., Bauer, C., Doka, G., Dones, R., Frischknecht, R., Hellweg, S., Humbert,S., Jungbluth, N., Köllner, T., Loerincik, Y., Margni, M., & Nemecek, T. (2010). InR. Hischier, & B. Weidema (Eds.), Implementation of life cycle impact assessmentmethods. EcoInvent Centre, Swiss Centre for Life Cycle Invetories.

An, H. J., Wilhelm, W. E., & Searcy, S. W. (2011a). Biofuel and petroleum-based fuelsupply 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 alignocellulosic biofuel supply chain system with a case study based on a regionin Central Texas. Bioresource Technology, 102, 7860–7870.

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.,& Dutta, A. (2010). Techno-economic comparison of biomass-to-transportationfuels via pyrolysis, gasification, and biochemical pathways. Fuel, 89(Supplement1), S29–S35.

Avami, A. (2012). A model for biodiesel supply chain: A case study in Iran. Renewable& Sustainable Energy Reviews, 16, 4196–4203.

Awudu, I., & Zhang, J. (2012). Uncertainties and sustainability concepts in biofuelsupply chain management: A review. Renewable & Sustainable Energy Reviews,16, 1359–1368.

Azapagic, A. (1999). Life cycle assessment and its application to process selectiondesign and optimisation. Chemical Engineering Journal, 73, 1–21.

Azapagic, A., & Clift, R. (1999). The application of life cycle assessment to processoptimisation. Computers & Chemical Engineering, 23, 1509–1526.

Bai, Y., Hwang, T. S., Kang, S., & Ouyang, Y. F. (2011). Biofuel refinery location andsupply chain planning under traffic congestion. Transportation Research Part B:Methodological, 45, 162–175.

Bai, Y., Ouyang, Y. F., & Pang, J. S. (2012). Biofuel supply chain design under compet-itive agricultural land use and feedstock market equilibrium. Energy Economics,34, 1623–1633.

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.National Renewable Energy Laboratory.

Bakshi, B. R., Fiksel, J., Baral, A., Bolinger, J., Mishchenko, O., Singh, S., Ukidwe,N., Urban, B., & Zhang, Y. (2013). Eco-LCA: Ecologically-based life cycle assess-ment. OH: Process Systems Engineering Group, and Center of Resilience, TheOhio State University., http://resilience.eng.ohio-state.edu/eco-lca/. Accessedon: 7/17/2013.

Bals, B., Rogers, C., Jin, M., Balan, V., & Dale, B. (2010). Evaluation of ammonia fibreexpansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass har-vested in different seasons and locations. Biotechnology for Biofuels, 3, 1.

Bamufleh, H., Ponce-Ortega, J., & El-Halwagi, M. (2013). Multi-objective opti-mization of process cogeneration systems with economic, environmentaland social tradeoffs. Clean Technologies and Environmental Policy, 15,185–197.

Baumrucker, B. T., & Biegler, L. T. (2010). MPEC strategies for cost optimization ofpipeline operations. Computers & Chemical Engineering, 34, 900–913.

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 ofproducts: Just in time!. The International Journal of Life Cycle Assessment, 15,156–163.

Bernardi, A., Giarola, S., & Bezzo, F. (2012). Optimizing the economics and the car-bon and water footprints of bioethanol supply chains. Biofuels. Bioproducts andBiorefining, 6, 656–672.

Bernardi, A., Giarola, S., & Bezzo, F. (2013). Spatially explicit multiobjective optimiza-tion for the strategic design of first and second generation biorefineries includingcarbon and water footprints. Industrial & Engineering Chemistry Research, 52,7170–7180.

Biddy, M., & Jones, S. (2013). Catalytic upgrading of sugars to hydrocarbons technol-ogy pathway. In NREL/TP-5100-58055, PNNL-22319. National Renewable EnergyLaboratory, Pacific Northwest National Laboratory.

Birge, J. R., & Louveaux, F. V. (1997). Introduction to stochastic programming. Springer.Bowling, I. M., Ponce-Ortega, J. M., & El-Halwagi, M. M. (2011). Facility location and

supply chain optimization for a biorefinery. Industrial & Engineering ChemistryResearch, 50, 6276–6286.

Bozell, J. J., & Petersen, G. R. (2010). Technology development for the productionof biobased products from biorefinery carbohydrates—The US Department ofEnergy’s ‘Top 10’ revisited. Green Chemistry, 12, 539–554.

Bunting, B., Bunce, M., Barone, T., & Storey, J. (2010). Fungible and compatible biofuels:Literature search, summary, and recommendations. Oak Ridge: Oak Ridge NationalLaboratory, Fuels, Engines, and Emissions Research Center.

Chen, C. W., & Fan, Y. Y. (2012). Bioethanol supply chain system planning undersupply and demand uncertainties. Transportation Research Part E: Logistics andTransportation 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:http://dx.doi.org/10.2139/ssrn.2084313.

Chiu, Y.-W., Suh, S., Pfister, S., Hellweg, S., & Koehler, A. (2012). Measuring ecologicalimpact of water consumption by bioethanol using life cycle impact assessment.The International Journal of Life Cycle Assessment, 17, 16–24.

5 emica

C

C

C

C

C

C

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

D

E

E

E

E

E

E

4 D. Yue et al. / Computers and Ch

hu, Y., & You, F. (2012). Integration of scheduling and control with onlineclosed-loop implementation: Fast computational strategy and large-scale globaloptimization algorithm. Computers & Chemical Engineering, 47, 248–268.

orsano, G., Vecchietti, A. R., & Montagna, J. M. (2011). Optimal design for sustain-able bioethanol supply chain considering detailed plant performance model.Computers & Chemical Engineering, 35, 1384–1398.

ˇucek, L., Klemes, J. J., & Kravanja, Z. (2012). A review of footprint analysis tools formonitoring impacts on sustainability. Journal of Cleaner Production, 34, 9–20.

ucek, L., Lam, H. L., Klemes, J. J., Varbanov, P. S., & Kravanja, Z. (2010). Synthesis ofregional networks for the supply of energy and bioproducts. Clean Technologiesand Environmental Policy, 12, 635–645.

ˇucek, L., Varbanov, P. S., Klemes, J. J., & Kravanja, Z. (2012). Total footprints-basedmulti-criteria optimisation of regional biomass energy supply chains. Energy,44, 135–145.

undiff, J. S., Dias, N., & Sherali, H. D. (1997). A linear programming approach fordesigning a herbaceous biomass delivery system. Bioresource Technology, 59,47–55.

al-Mas, M., Giarola, S., Zamboni, A., & Bezzo, F. (2011). Strategic design and invest-ment capacity planning of the ethanol supply chain under price uncertainty.Biomass & Bioenergy, 35, 2059–2071.

aoutidis, P., Marvin, W. A., Rangarajan, S., & Torres, A. I. (2013). Engineering biomassconversion processes: A systems perspective. AIChE Journal, 59, 3–18.

avis, R., Aden, A., & Pienkos, P. T. (2011). Techno-economic analysis of autotrophicmicroalgae for fuel production. Applied Energy, 88, 3524–3531.

avis, R., Biddy, M., & Jones, S. (2013). Algal lipid extraction and upgrading to hydro-carbons technology pathway. In NREL/TP-5100-58049, PNNL-22315. NREL.

avis, R., Biddy, M., Tan, E., Tao, L., & Jones, S. (2013). Biological conversion of sug-ars to hydrocarbons technology pathway. In NREL/TP-5100-58054, PNNL-22318.National Renewable Energy Laboratory, Pacific Northwest National Laboratory.

OE. (2010). Report to congress, dedicated ethanol pipeline feasibility study. U.S.Department of Energy.

OE/EERE. (2010a). Biopower technical strategy workshop summary report. Denver,CO (2009): U.S. Department of Energy, Office of Energy Efficiency & RenewableEnergy.

OE/EERE. (2010b). National algal biofuels technology roadmap. U.S. Department ofEnergy, Office of Energy Efficiency & Renewable Energy.

OE/EERE. (2011a). Biochemical conversion: Using hydrolysis, fermentation, and catal-ysis to make fuels and chemicals. U.S. Department of Energy, Office of EnergyEfficiency & Renewable Energy.

OE/EERE. (2011b). Thermochemical conversion: Using heat and chemistry to makefuels and bioproducts. U.S. Department of Energy, Office of Energy Efficiency &Renewable Energy.

OE/EERE. (2011c). U.S. billion-ton update: Biomass supply for a bioenergy andbioproducts industry. In R. D. Perlack, & B. J. Stokes (Eds.), ORNL/TM-2011/224(p. 227). Oak Ridge, TN: Oak Ridge National Laboratory.

OE/EERE. (2013a). Advanced uniform-format feedstock supply system. U.S. Depart-ment of Energy, Office of Energy Efficiency & Renewable Energy. Available at:http://www1.eere.energy.gov/bioenergy/biomass feedstocks.html#AdvancedUniform Format Feedstock Supply (accessed on 7/9/2013).

OE/EERE. (2013b). Federal and state laws and incentives. Department ofEnergy, Office of Energy Efficiency & Renewable Energy. Available at:http://www.afdc.energy.gov/laws/ (accessed on: 7/19/2013).

OE/EERE. (2013c). Feedstock supply and logistics: Biomass as a commodity. U.S.Department of Energy, Office of Energy Efficiency & Renewable Energy.

OE/EERE. (2013d). Replacing the whole barrel: To reduce U.S. dependence on oil. U.S.Department of Energy, Office of Energy Efficiency & Renewable Energy.

OE/NETL. (2010). DOE/NETL carbon dioxide capture and storage RD&D roadmap. U.S.Department of Energy and National Energy Technology Laboratory.

reyer, L., Hauschild, M., & Schierbeck, J. (2006). A Framework for Social Life CycleImpact Assessment (10 pp). The International Journal of Life Cycle Assessment, 11,88–97.

unnett, A., Adjiman, C., & Shah, N. (2007). Biomass to heat supply chains applica-tions of process optimization. Process Safety and Environmental Protection, 85,419–429.

unnett, A., Adjiman, C., & Shah, N. (2008). A spatially explicit whole-system modelof the lignocellulosic bioethanol supply chain: An assessment of decentralisedprocessing potential. Biotechnology for Biofuels, 1.

utta, P. K. (1999). Strategies and games: Theory and practice. Massachusetts Instituteof Technology.

coinvent. (2008). The ecoinvent database V2.0. In Technical Report. Swiss Centre forLife Cycle Inventories., http://www.ecoinvent.ch/.

ISA. (2007). Energy Independence and Security Act of 2007. Public Law 110–140, 110thCongress. EISA.

ksioglu, S. D., Acharya, A., Leightley, L. E., & Arora, S. (2009). Analyzing the designand management of biomass-to-biorefinery supply chain. Computers & IndustrialEngineering, 57, 1342–1352.

ksioglu, S. D., Li, S., Zhang, S., Sokhansanj, S., & Petrolia, D. (2010). Analyzing impactof intermodal facilities on design and management of biofuel supply chain.Transportation Research Record, 144–151.

lia, J. A., Baliban, R. C., & Floudas, C. A. (2012). Nationwide energy supply chain anal-ysis for hybrid feedstock processes with significant CO2 emissions reduction.

AIChE Journal, 58, 2142–2154.

lia, J. A., Baliban, R. C., Xiao, X., & Floudas, C. A. (2011). Optimal energy supply net-work determination and life cycle analysis for hybrid coal, biomass and naturalgas to liquid (CBGTL) plants using carbon-based hydrogen production. Comput-ers & Chemical Engineering, 35, 1399–1430.

l Engineering 66 (2014) 36–56

Engell, S., & Harjunkoski, I. (2012). Optimal operation: Scheduling advanced controland their integration. Computers & Chemical Engineering, 47, 121–133.

Eppen, G. D., Martin, R. K., & Schrage, L. (1989). A scenario approach to capacityplanning. Operations Research, 37, 517–527.

EPRI. (2010). Biopower generation: Biomass issues, fuels, technologies, and research,development, demonstration, and deployment opportunities. Electric PowerResearch Institute.

Esri. (2012). ArcGIS—Mapping and spatial analysis for understanding our world. ,http://www.esri.com/software/arcgis.

Fahy, E., Subramaniam, S., Brown, H. A., Glass, C. K., Merrill, A. H., Murphy, R. C., Raetz,C. R. H., Russell, D. W., Seyama, Y., Shaw, W., Shimizu, T., Spener, F., van Meer,G., VanNieuwenhze, M. S., White, S. H., Witztum, J. L., & Dennis, E. A. (2005). Acomprehensive classification system for lipids. European Journal of Lipid Scienceand Technology, 107, 337–364.

FAO. (2009). Algae-based biofuels: A review of challenges and opportunities for devel-oping countries. Rome, Italy: Food and Agriculture Organization of the UnitedNations.

Fava, J. A. (1994). A techical framework for life-cycle assessment. Pensacola: SETACFoudation.

Ferreira, L., & Borenstein, D. (2011). Normative agent-based simulation for supplychain planning. Journal of the Operational Research Society, 62, 501–514.

Flisberg, P., Frisk, M., & Ronnqvist, M. (2012). FuelOpt: A decision support system forforest fuel logistics. Journal of the Operational Research Society, 63, 1600–1612.

Floudas, C. A. (1999). Deterministic global optimization: Theory, methods and applica-tions. Boston: Kluwer Academic Publishers.

Frombo, F., Mindardi, R., Robba, M., Rosso, F., & Sacile, R. (2009). Planning woodybiomass logistics for energy production: A strategic decision model. Biomass &Bioenergy, 33, 372–383.

GaBi. (2011). Software and database contents for life cycle engineering. PE Interna-tional., http://www.gabi-software.com.

Gan, J. B., & Smith, C. T. (2011). Optimal plant size and feedstock supply radius: Amodeling approach to minimize bioenergy production costs. Biomass & Bioen-ergy, 35, 3350–3359.

Gebreslassie, B. H., Slivinsky, M., Wang, B., & You, F. (2013). Life cycle optimization forsustainable design and operations of hydrocarbon biorefinery via fast pyrolysishydrotreating and hydrocracking. Computers & Chemical Engineering, 50, 71–91.

Gebreslassie, B. H., Waymire, R., & You, F. (2013). Sustainable design and synthesis ofalgae-based biorefinery for simultaneous hydrocarbon biofuel production andcarbon sequestration. AIChE JournalV 59, 1599–1621.

Gebreslassie, B. H., Yao, Y., & You, F. (2012). Design under uncertainty of hydrocar-bon biorefinery supply chains: Multiobjective stochastic programming models,decomposition algorithm and a Comparison between CVaR and downside risk.AIChE Journal, 58, 2155–2179.

Gerbens-Leenes, W., & Hoekstra, A. Y. (2011). The water footprint of biofuel-basedtransport. Energy & Environmental Science, 4, 2658–2668.

Giarola, S., Zamboni, A., & Bezzo, F. (2011). Spatially explicit multi-objectiveoptimisation for design and planning of hybrid first and second generationbiorefineries. Computers & Chemical Engineering, 35, 1782–1797.

Gjerdrum, J., Shah, N., & Papageorgiou, L. G. (2001). Transfer prices for multienter-prise supply chain optimization. Industrial & Engineering Chemistry Research, 40,1650–1660.

Goedkoop, M., & Spriensma, R. (2001). Eco-indicator 99: A damage oriented methodfor life cycle impact assessment; methodology report. PRé, Product Ecology Con-sultants.

Gold, S., & Seuring, S. (2011). Supply chain and logistics issues of bio-energy produc-tion. Journal of Cleaner Production, 19, 32–42.

Gopalakrishnan, A., & Biegler, L. T. (2013). Economic nonlinear model predictivecontrol for periodic optimal operation of gas pipeline networks. Computers &Chemical Engineering, 52, 90–99.

Graves, S. C., & Willems, S. P. (2000). Optimizing strategic safety stock place-ment in supply chains. Manufacturing & Service Operations Management, 2,68–83.

Graves, S. C., & Willems, S. P. (2005). Optimizing the supply chain configuration fornew products. Management Science, 51, 1165–1180.

GreenDelta. (2013). Earthster 2 turbo. In Tools & consulting for sustain-ability. GreenDelta. Available at: http://www.greendeltatc.com/Earthster-2-Turbo.200.0.html (accessed on: 7/16/2013).

GREET. (2012). The greenhouse gases, regulated emissions, and energy use in trans-portation model. Argonne National Laboratory., http://greet.es.anl.gov/.

Grossmann, I. (2002). Review of nonlinear mixed-integer and disjunctive program-ming techniques. Optimization and Engineering, 3, 227–252.

Grossmann, I. E., & Guillén-Gosálbez, G. (2010). Scope for the application of math-ematical programming techniques in the synthesis and planning of sustainableprocesses. Computers & Chemical Engineering, 34, 1365–1376.

Hasan, M. R., & Chakrabarti, R. (2009). Use of algae and aquatic macrophytes as feed insmall-scale aquaculture: A review. Rome, Italy: Food and Agriculture Organizationof the United Nations.

Hosseini, S. A., & Shah, N. (2011). Multi-scale process and supply chain mod-elling: From lignocellulosic feedstock to process and products. Interface Focus,1, 255–262.

Huang, Y. X., Chen, C. W., & Fan, Y. Y. (2010). Multistage optimization of the supply

chains of biofuels. Transportation Research Part E: Logistics and TransportationReview, 46, 820–830.

Huijbregts, M., Heijungs, R., & Hellweg, S. (2004). Complexity and integratedresource management: Uncertainty in LCA. The International Journal of Life CycleAssessment, 9, 341–342.

emica

H

I

I

I

I

J

J

J

K

K

K

K

L

L

L

L

L

L

L

M

M

M

M

M

M

M

M

M

D. Yue et al. / Computers and Ch

umbert, S., Schryver, A. D., Bengoa, X., Margni, M., & Jolliet, O. (2012). IMPACT2002+: User guide draft for version Q2.21. Quantis., http://www.quantis-intl.com/index.php.

akovou, E., Karagiannidis, A., Vlachos, D., Toka, A., & Malamakis, A. (2010). Wastebiomass-to-energy supply chain management: A critical synthesis. Waste Man-agement, 30, 1860–1870.

MPLAN. (2013). IMPLAN professional model. IMPLAN Group LLC. Available at:http://implan.com/V4/Index.php (accessed date: 7/16/2013).

PCC. (2007). Glossary in fourth assessment report: Climate change 2007. ,http://www.ipcc.ch/pdf/assessment-report/ar4/syr/ar4 syr appendix.pdf.

SO. (2006). ISO 14040: Environmental management-life cycle assessment-principlesand frame work. International Standard: ISO Standards.

onesF S.B., Holladay, J. E., Valkenburg, C., Stevens, D. J., Walton, C. W., Kinchin, C.,Elliott, D. C., & Czernik, S. (2009). Production of gasoline and diesel from biomass viafast pyrolysis, hydrotreating and hydrocracking: A design case. Pacific NorthwestNational Laboratory. PNNL-18284.

ørgensen, A., Bocq, A., Nazarkina, L., & Hauschild, M. (2008). Methodologies forsocial life cycle assessment. The International Journal of Life Cycle Assessment,13, 96–103.

unginger, M., Bolkesjø, T., Bradley, D., Dolzan, P., Faaij, A., Heinimö, J., Hektor,B., Leistad, Ø., Ling, E., Perry, M., Piacente, E., Rosillo-Calle, F., Ryckmans, Y.,Schouwenberg, P.-P., Solberg, B., Trømborg, E., Walter, A. d. S., & Wit, M. d.(2008). Developments in international bioenergy trade. Biomass and Bioenergy,32, 717–729.

aditi, E., & Swinnen, J. (2007). International biomass trade—What aresound policy approaches to guarantee a fair trade of biomass with-out further depletion of natural resources in exporting countries? InAgrinergy Workshop CEPS, Brussels, November 20, 2007,. Available at:http://agrinergy.ecologic.eu/download/kaditi agrinergy trade.pdf.

eirstead, J., Samsatli, N., Pantaleo, A. M., & Shah, N. (2012). Evaluating biomassenergy strategies for a UK eco-town with an MILP optimization model. Biomass& Bioenergy, 39, 306–316.

im, J., Realff, M. J., & Lee, J. H. (2011). Optimal design and global sensitivity analysisof biomass supply chain networks for biofuels under uncertainty. Computers &Chemical Engineering, 35, 1738–1751.

im, J., Realff, M. J., Lee, J. H., Whittaker, C., & Furtner, L. (2011). Design of biomassprocessing network for biofuel production using an MILP model. Biomass &Bioenergy, 35, 853–871.

amers, P., Junginger, M., Hamelinck, C., & Faaij, A. (2012). Developments in interna-tional solid biofuel trade—An analysis of volumes, policies, and market factors.Renewable and Sustainable Energy Reviews, 16, 3176–3199.

educ, S., Lundgren, J., Franklin, O., & Dotzauer, E. (2010). Location of a biomass basedmethanol production plant: A dynamic problem in northern Sweden. AppliedEnergy, 87, 68–75.

educ, S., Natarajan, K., Dotzauer, E., McCallum, I., & Obersteiner, M. (2009). Opti-mizing biodiesel production in India. Applied Energy, 86, S125–S131.

educ, S., Starfelt, F., Dotzauer, E., Kindermann, G., McCallum, I., Obersteiner, M., &Lundgren, J. (2010). Optimal location of lignocellulosic ethanol refineries withpolygeneration in Sweden. Energy, 35, 2709–2716.

ee, J. H. (2012). Energy supply chain optimization: A challenge for control engineers.In V. Kariwala, L. Samavedham, & R. D. Braatz (Eds.), Eighth IFAC internationalsymposium on advanced control of chemical processes (pp. 361–370). FuramaRiverfront, Singapore: International Federation of Automatic Control.

ira-Barragán, L. F., Ponce-Ortega, J. M., Serna-González, M., & El-Halwagi, M. M.(2013). Synthesis of integrated absorption refrigeration systems involving eco-nomic and environmental objectives and quantifying social benefits. AppliedThermal Engineering, 52, 402–419.

undquist, T. J., Woertz, I. C., Quinn, N. W. T., & Benemann, J. R. (2010). A realistictechnology and engineering assessment of algae biofuel production. Berkeley, CA:Energy Biosciences Institute, University of California.

arquardt, W., Harwardt, A., Hechinger, M., Kraemer, K., Viell, J., & Voll, A. (2010).The biorenewables opportunity—Toward next generation process and productsystems. AIChE Journal, 56, 2228–2235.

arvin, W. A., Schmidt, L. D., Benjaafar, S., Tiffany, D. G., & Daoutidis, P. (2012).Economic optimization of a lignocellulosic biomass-to-ethanol supply chain.Chemical Engineering Science, 67, 68–79.

ASBI. (2013). Fueling a sustainable future for aviation. Midwest Aviation SustainableBiofuels Initiative., www.masbi.org.

cNew, K., & Griffith, D. (2005). Measuring the impact of ethanol plants on localgrain prices. Review of Agricultural Economics, 27, 164–180.

ele, F. D., Guillen-Gosalbez, G., & Jimenez, L. (2009). Optimal planning of supplychains for bioethanol and sugar production with economic and environmentaloncerns. In J. Jezowski, & T. Thullie (Eds.), 19th European symposium on computeraided process engineering (26) (pp. 997–1002). Amsterdam: Elsevier ScienceBv.

ele, F. D., Kostin, A. M., Guillen-Gosalbez, G., & Jimenez, L. (2011). Multiobjec-tive model for more sustainable fuel supply chains. A case study of the sugarcane industry in Argentina. Industrial & Engineering Chemistry Research, 50,4939–4958.

elero, J. A., Iglesias, J., & Garcia, A. (2012). Biomass as renewable feedstock instandard refinery units. Feasibility opportunities and challenges. Energy & Envi-

ronmental Science, 5, 7393–7420.

obini, M., Sowlati, T., & Sokhansanj, S. (2011). Forest biomass supply logistics fora power plant using the discrete-event simulation approach. Applied Energy, 88,1241–1250.

yerson, R. B. (1997). Game theory: Analysis of conflict. Harvard University Press.

l Engineering 66 (2014) 36–56 55

NABC. (2013). National advanced biofuels consortium process strategies. NationalAdvanced Biofuels Consortium. Available at: http://www.nabcprojects.org/process strategies.html (accessed on: 7/19/2013).

Nash, J. F., Jr. (1950). The bargaining problem. Econometrica, 18, 155–162.NCEP. (2009). Task force on biofuels infrastructure. National Commission on Energy

Policy.NREL. (2012). JEDI models: Jobs and economic development impact models. NREL.NREL. (2013a). Bioproducts. National Renewable Energy Laboratory. Available at:

http://www.nrel.gov/learning/re bioproducts.html (accessed on 7/15/2013).NREL. (2013b). Energy analysis—International trade of wood pellets. In NREL/BR-

6A20-56791. National Renewable Energy Laboratory.openLCA. (2012). A new open source software for life cycle assessment. GreenDeltaTC

GmbH., http://www.openlca.org.Papapostolou, C., Kondili, E., & Kaldellis, J. K. (2011). Development and implementa-

tion of an optimisation model for biofuels supply chain. Energy, 36, 6019–6026.Parker, N., Tittmann, P., Hart, Q., Nelson, R., Skog, K., Schmidt, A., Gray, E., & Jenkins,

B. (2010). Development of a biorefinery optimized biofuel supply curve for theWestern United States. Biomass & Bioenergy, 34, 1597–1607.

Petersen, G., Bozell, J., & White, J. (2013). Selecting targets. In Catalytic process devel-opment for renewable materials. Wiley-VCH Verlag GmbH & Co. KGaA.

PNNL. (2013). Bio-based product research. Pacific Northwest National Laboratory.Available at: http://www.pnnl.gov/biobased/ (accessed on 7/15/2013).

Rajagapol, D., Sexton, S., Hochman, G., Roland-Holst, D., & Zilberman, D. D. (2009).Model estimates food-versus-biofuel trade-off. California Agriculture, 63.

Rentizelas, A. A., Tatsiopoulos, I. P., & Tolis, A. (2009). An optimization model formulti-biomass tri-generation energy supply. Biomass & Bioenergy, 33, 223–233.

Rentizelas, A. A., Tolis, A. J., & Tatsiopoulos, I. P. (2009). Logistics issues of biomass:The storage problem and the multi-biomass supply chain. Renewable & Sustain-able Energy Reviews, 13, 887–894.

Ringer, M., Putsche, V., & Scahill, J. (2006). Large-scale pyrolysis oil production: Atechnology assessment and economic analysis. In NREL/TP-510-37779. NationalRenewable Energy Laboratory.

Rockafellar, R. T., & Uryasev, S. (2002). Conditional value-at-risk for general lossdistributions. Journal of Banking & Finance, 26, 1443–1471.

Sahinidis, N. V. (2004). Optimization under uncertainty: State-of-the-art and oppor-tunities. Computers & Chemical Engineering, 28, 971–983.

Santibanez-Aguilar, J. E., Gonzalez-Campos, J. B., Ponce-Ortega, J. M., Serna-Gonzalez,M., & El-Halwagi, M. M. (2011). Optimal planning of a biomass conversion sys-tem considering economic and environmental aspects. Industrial & EngineeringChemistry Research, 50, 8558–8570.

Sharma, B., Ingalls, R. G., Jones, C. L., & Khanchi, A. (2013). Biomass supply chaindesign and analysis: Basis, overview, modeling, challenges, and future. Renew-able and Sustainable Energy Reviews, 24, 608–627.

Sharma, P., Sarker, B. R., & Romagnoli, J. A. (2011). A decision support tool for strate-gic planning of sustainable biorefineries. Computers & Chemical Engineering, 35,1767–1781.

Shastri, Y. N., Rodriguez, L. F., Hansen, A. C., & Ting, K. C. (2012). Impact of distributedstorage and pre-processing on Miscanthus production and provision systems.Biofuels Bioproducts & Biorefining-Biofpr, 6, 21–31.

Shipley, A., Hampson, A., Hedman, B., Garland, P., & Bautista, P. (2008). Combinedheat and power—Effective energy solutions for a sustainable future. In ORNL/TM-2008/224. Oak Ridge National Laboratory.

Sokhansanj, S., Mani, S., Turhollow, A., Kumar, A., Bransby, D., Lynd, L., & Laser, M.(2009). Large-scale production, harvest and logistics of switchgrass (Panicumvirgatum L.)—Current technology and envisioning a mature technology. Biofuels,Bioproducts and Biorefining, 3, 124–141.

Swanson, R. M., Satrio, J. A., Brown, R. C., Platon, A., & Hsu, D. D. (2010). Techno-economic analysis of biofuels production based on gasification. In NREL/TP-6A20-46587. National Renewable Energy Laboratory.

Talmadge, M., Biddy, M., Dutta, A., Jones, S., & Meyer, A. (2013). Syngas upgradingto hydrocarbon fuels technology pathway. In NREL/TP-5100-58052, PNNL-22323.National Renewable Energy Laboratory, Pacific Northwest National Laboratory.

Tatsiopoulos, I. P., & Tolis, A. J. (2003). Economic aspects of the cotton-stalk biomasslogistics and comparison of supply chain methods. Biomass & Bioenergy, 24,199–214.

Tittmann, P. W., Parker, N. C., Hart, Q. J., & Jenkins, B. M. (2010). A spatially explicittechno-economic model of bioenergy and biofuels production in California. Jour-nal of Transport Geography, 18, 715–728.

Tong, K., Gleeson, M. J., Rong, G., & You, F. (2013). Optimal design ofadvanced ‘drop-in’ hydrocarbon biofuel supply chain integrating withexisting petroleum refineries under uncertainty. Biomass & Bioenergy,http://dx.doi.org/10.1016/j.biombioe.2013.10.023, accepted.

Tong, K., Gong, J., Yue, D., & You, F. (2013). Stochastic programmingapproach to optimal design and operations of integrated hydrocarbon bio-fuel and petroleum supply chains. ACS Sustainable Chemistry & Engineering,http://dx.doi.org/10.1021/sc400267t, accepted.

UNEP. (2009). Guidelines for social life cycle assessment of products. In UnitedNations Environment Programme. United Nations.

USDA. (2007). Ethanol transportation backgrounder. U.S. Department of Agriculture.USDA. (2010). Field crops usual planting and harvesting dates. United States

Department of Agriculture, National Agricultural Statistics Service. Agricultural

Handbook Number 628.

Uslu, A., Faaij, A. P. C., & Bergman, P. C. A. (2008). Pre-treatment technologies,and their effect on international bioenergy supply chain logistics. Techno-economic evaluation of torrefaction, fast pyrolysis and pelletisation. Energy, 33,1206–1223.

5 emica

V

v

W

W

W

W

W

W

W

W

W

W

W

Y

Y

Y

Y

Y

cations in process systems engineering. Computers & Chemical Engineering, 61,

6 D. Yue et al. / Computers and Ch

elazquez-Marti, B., & Fernandez-Gonzalez, E. (2010). Mathematical algorithms tolocate factories to transform biomass in bioenergy focused on logistic networkconstruction. Renewable Energy, 35, 2136–2142.

on Stackelberg, H. (1934). Marktform und gleichgewicht. Wien & Berlin: Springer.VI, 138 S. 8◦: J. Springer.

alther, G., Schatka, A., & Spengler, T. S. (2012). Design of regional productionnetworks for second generation synthetic bio-fuel—A case study in NorthernGermany. European Journal of Operational Research, 218, 280–292.

ang, B., Gebreslassie, B. H., & You, F. (2013). Sustainable design and synthe-sis of hydrocarbon biorefinery via gasification pathway: Integrated life cycleassessment and technoeconomic analysis with multiobjective superstructureoptimization. Computers & Chemical Engineering, 52, 55–76.

assick, J. M., Agarwal, A., Akiya, N., Ferrio, J., Bury, S., & You, F. (2012). Address-ing the operational challenges in the development, manufacture, and supply ofadvanced materials and performance products. Computers & Chemical Engineer-ing, 47, 157–169.

eidema, B. (2006). The integration of economic and social aspects in lifecycle impact assessment. The International Journal of Life Cycle Assessment, 11,89–96.

erpy, T., & Petersen, G. (2004). Top value added chemicals from biomass volumeI—Results of screening for potential candidates from sugars and synthesis gas. U.S.Department of Energy.

ikipedia. (2013a). Equivalent annual cost. In Wikipedia, the free encyclopedia.Available at: http://en.wikipedia.org/wiki/Equivalent annual cost (accessed on:7/16/2013).

ikipedia. (2013b). Gasoline gallon equivalent. In Wikipedia, the free encyclopedia.Available at: http://en.wikipedia.org/wiki/Gasoline gallon equivalent (accessedon: 7/16/2013).

ikipedia. (2013c). Net present value. In Wikipedia, the free encyclopedia. Availableat: http://en.wikipedia.org/wiki/Net present value (accessed on: 7/16/2013).

ilcox, J. (2012). The role of algae in carbon capture. In Carbon capture sci-ence + business media LLC. Springer (Chapter 7).

iltsee, G. (2000). Lessons learned from existing biomass power plants. In NREL/SR-570-26946. National Renewable Energy Laboratory.

right, M. M., Satrio, J. A., Brown, R. C., Daugaard, D. E., & Hsu, D. D. (2010). Techno-economic analysis of biomass fast pyrolysis to transportation fuels. In NREL/TP-6A20-46586. National Renewable Energy Labratory.

eomans, H., & Grossmann, I. E. (1999). A systematic modeling framework of super-structure optimization in process synthesis. Computers & Chemical Engineering,23(6), 709–731.

ou, F., Castro, P. M., & Grossmann, I. E. (2009). Dinkelbach’s algorithm as an effi-cient method to solve a class of MINLP models for large-scale cyclic schedulingproblems. Computers & Chemical Engineering, 33, 1879–1889.

ou, F., & Grossmann, I. E. (2008a). Design of responsive supply chains under demanduncertainty. Computers & Chemical Engineering, 32, 3090–3111.

ou, F., & Grossmann, I. E. (2008b). Mixed-integer nonlinear programming models

and algorithms for large-scale supply chain design with stochastic inventorymanagement. Industrial & Engineering Chemistry Research, 47, 7802–7817.

ou, F., & Grossmann, I. E. (2011a). Balancing responsiveness and economics in pro-cess supply chain design with multi-echelon stochastic inventory. AIChE Journal,57, 178–192.

l Engineering 66 (2014) 36–56

You, F., & Grossmann, I. E. (2011b). Stochastic inventory management for tacti-cal process planning under uncertainties: MINLP models and algorithms. AIChEJournal, 57, 1250–1277.

You, F., Grossmann, I. E., & Wassick, J. M. (2011). Multisite capacity, production anddistributionplanning with reactor modifications: MILP model, bilevel decompo-sition algorithm vs. lagrangean decomposition scheme. Industrial & EngineeringChemistry Research, 50, 4831–4849.

You, F., Tao, L., Graziano, D. J., & Snyder, S. W. (2012). Optimal design ofsustainable cellulosic biofuel supply chains: Multiobjective optimization cou-pled with life cycle assessment and input–output analysis. AIChE Journal, 58,1157–1180.

You, F., & Wang, B. (2011). Life cycle optimization of biomass-to-liquid supplychains with distributed—Centralized processing networks. Industrial & Engineer-ing Chemistry Research, 50, 10102–10127.

Yue, D., Guillén-Gosálbez, G., & You, F. (2013). Global optimization oflarge-scale mixed-integer linear fractional programming problems: Areformulation–linearization method and process scheduling applications.AIChE Journal, 59, 4255–4272.

Yue, D., Kim, M. A., & You, F. (2013). Design of sustainable product systems andsupply chains with life cycle optimization based on functional unit: Generalmodeling framework, mixed-integer nonlinear programming algorithms andcase study on hydrocarbon biofuels. ACS Sustainable Chemistry & Engineering, 1,1003–1014.

You, F., Pinto, J. M., Grossmann, I. E., & Megan, L. (2011). Optimal distribution-inventory planning of industrial gases: ii. minlp models and algorithmsfor stochastic cases. Industrial & Engineering Chemistry Research, 50,2928–2945.

You, F. Q., Wassick, J. M., & Grossmann, I. E. (2009). Risk management for a globalsupply chain planning under uncertainty: Models and algorithms. AIChE Journal,55, 931–946.

Yue, D., & You, F. (2013a). Planning and scheduling of flexible process networks underuncertainty with stochastic inventory: MINLP models and algorithm. AIChE Jour-nal, 59, 1511–1532.

Yue, D., & You, F. (2013b). Sustainable scheduling of batch processes under economicand environmental criteria with minlp models and algorithms. Computers &Chemical Engineering, 54, 44–59.

Zamboni, A., Bezzo, F., & Shah, N. (2009). Spatially explicit static modelfor the strategic design of future bioethanol production systems.2. Multi-objective environmental optimization. Energy & Fuels, 23,5134–5143.

Zamboni, A., Shah, N., & Bezzo, F. (2009). Spatially explicit static model for the strate-gic design of future bioethanol production systems. 1. Cost minimization. Energy& Fuels, 23, 5121–5133.

Zhong, Z., & You, F. (2014). Globally convergent exact and inexact parametricalgorithms for solving large-scale mixed-integer fractional programs and appli-

90–101.Zhou, C., Liu, P., & Li, Z. (2013). A superstructure based mixed-integer programing

approach. In A. Kraslawski, & I. Turunen (Eds.), 23rd European symposium oncomputer aided process engineering Lappeenranta, Finland.