Optimal design of microalgae-based biorefinery: Economics, opportunities and challenges

Applied Energy 150 (2015) 69–79

Contents lists available at ScienceDirect

Applied Energy journal homepage: www.elsevier.com/locate/apenergy

Optimal design of microalgae-based biorefinery: Economics, opportunities and challenges Muhammad Rizwan a, Jay H. Lee a,⇑, Rafiqul Gani b a Department of Chemical and Biomolecular Engineering, Korea Advanced Institute of Science and Technology (KAIST), 291 Daehak-ro, Yuseong-gu, Daejeon 305-701, Republic of Korea b CAPEC, Department of Chemical & Biochemical Engineering, Technical University of Denmark, DK-2800 Kgs Lyngby, Denmark

h i g h l i g h t s  A superstructure based optimization

model is developed.  Optimal structures of microalgal biorefinery are determined.  The optimization problem is formulated as an MINLP model.  Two different optimization scenarios are investigated.  Sensitivity analysis elaborates potential improvements.

g r a p h i c a l a b s t r a c t Feed

Cultivation

Harvesting

1,3

Pretreatment

Lipid extraction

1,4

1,5

2,4

2,5

3,4

3,5

4,4

4,5

Transesterification

Post transesterification purification

Pretreatment of residue

Conversion of residue

Products

1,8

1,9

Biodiesel

2,8

2,9

Glycerol

3,8

3,9

Bio-oil

4,8

4,9

Bioethanol

5,8

5,9

Biogas

1,6 2,3

2,6 3,3

1,7

3,6 1,2

4,3

5,5

Feed 2,2

6,5

5,3

6,3

5,4

7,5 4,6

Main stream

8,5 7,3 9,5

5,6

Microalgae residue Recycle stream

8,3 10,5

6,6 7,6

a r t i c l e

i n f o

Article history: Received 31 December 2014 Received in revised form 1 April 2015 Accepted 6 April 2015

Keywords: Microalgal biorefinery Biofuels Superstructure optimization Mixed integer nonlinear programming Lipid-extracted microalgae

2,7

a b s t r a c t Microalgae have great potential as a feedstock for the production of a wide range of end-products under the broad concept of biorefinery. In an earlier work, we proposed a superstructure based optimization model to find the optimal processing pathway for the production of biodiesel from microalgal biomass, and identified several challenges with the focus being on utilizing lipids extracted microalgal biomass for economic and environmentally friendly production of useful energy products. In this paper, we expand the previous optimization framework by considering the processing of microalgae residue previously treated as wastes. We develop an expanded biorefinery superstructure model, based on which a mixed integer nonlinear programming (MINLP) model is proposed to determine the optimal/promising biorefinery configurations with different choices of objective functions. The MINLP model is solved in GAMS using a database built in Excel. Economic sensitivity analysis is performed to elaborate the potential improvements in the overall economics, and set the targets that must be achieved in the future in order for microalgal biofuels to become economically viable. Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction ⇑ Corresponding author. Tel.: +82 42 350 3926. E-mail address: [email protected] (J.H. Lee). http://dx.doi.org/10.1016/j.apenergy.2015.04.018 0306-2619/Ó 2015 Elsevier Ltd. All rights reserved.

Microalgae have received significant attention lately as a promising renewable feedstock for producing a wide range of

70

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

products, including biofuels and a variety of value-added chemicals, owing to its high quantity of lipids, proteins, carbohydrates, vitamins, pigments, and enzymes [1]. These biofuels also contribute to the quality of the environment by producing less net amounts of carbon dioxide than fossil fuels [2], and may reduce dependency on fossil fuels. Microalgae also offer many potential advantages over terrestrial food crops, including their extremely rapid growth (with the typical doubling time of less than 24 h) and ability to be grown in non-arable land [3]. They can also utilize CO2 as a carbon source, and hence their large-scale cultivation can contribute to restoration of the carbon balance in the atmosphere [4]. Microalgae are composed of three main components; lipids, proteins, and carbohydrates. After extraction, lipids are mainly used for the production of biodiesel through the transesterification process (a well-known process for biodiesel production). One of the major challenges in the production of biodiesel from microalgae is its cost-effectiveness. Microalgal biodiesel is not costeffective at current time, as indicated in our previous study [5]. This challenge can be addressed by utilizing all the components of microalgae, not just the lipids, under the broad concept of biorefinery where the lipid contents can be utilized for the production of biodiesel, and (lipid-extracted) residue can be used for the production of value-added products including various biofuels, valuable chemicals, and other useful industrial intermediates [3,6]. ‘Spent microalgal biomass’ or ‘microalgae residue’, in this work, mean the residuals left over after the extraction of the lipids; they mainly contain proteins and carbohydrates. These microalgae residues account for approximately 70% of the whole microalgal biomass (on dry basis) and thus, have potential to serve for a number of applications [7]. Therefore, proper exploitation and utilization of microalgae residue can enhance the overall economics of biofuels production from microalgae. Unlike the lipid contents of microalgal biomass, the spent microalgal biomass has not been studied extensively and its applications are not well-documented [8]. According to a baseline study [8], spent microalgal biomass can be used to produce many products such as bio-ethanol, biohydrogen, bio-oil, bio-methane, fertilizer, plastics, nutrients, animal feed, sorbents, etc. Therefore, their processing can be considered under the overall frame of biorefinery. Despite the various benefits associated with the concomitant utilization of microalgal biomass and its residue, an economic feasibility of microalgae-based biorefinery has not yet been evaluated fully [9]. A systems engineering approach can be applied to address the challenges associated with the microalgae-based biorefinery, for example, by developing a systematic methodological framework to locate the promising biorefinery configurations in terms of cost-effectiveness, robustness, and environmental sustainability [10,11]. These challenges include: (1) existence of a huge number of processing pathways for the production of biofuels and chemicals from microalgae due to many technological alternatives available at each processing step and (2) inconsistent, uncertain, and preliminary nature of technological and economic data. The research in the field of microalgae-based biorefinery is in an early phase of development and yet to mature enough to produce a consistent and reliable dataset regarding the various technological alternatives for each step of microalgae processing. Therefore, the developed framework should have the capability to consider and address the inconsistent nature of the data, and give a set of promising production routes with high potential to become sustainable from both economic and environmental viewpoints. Systems approaches to process synthesis and optimization of processing networks have been proposed and described in many studies [12–14]. Based on these studies, many researchers developed systemic methodologies for a number of applications including the optimal synthesis of biogas production from organic and animal waste [15], processing of soybean oil [16], processing of

waste palm oil for biodiesel and fatty alcohol production [17], processing of lignocellulousic biomass [18–20], and processing of microalgal biomass [5,21–25]. Davis et al. [26] investigated and established the baseline economics for microalgae cultivation via open ponds and photobioreactors, and then further conversion to green diesel via hydrotreating. Delrue et al. [27] developed an economic, sustainability and energetic model for the production of biodiesel from microalgae. In their follow-up study [28], the techno-economic viabilities of hydrothermal liquefaction, oil secretion and alkane secretion have been evaluated. Slegers et al. [29] proposed a model-based combinatorial optimization approach for energy efficient conversion of microalgae into biodiesel. The performance is expressed in terms of net energy ratios. In a recent review on the application of process synthesis to biorefinery processes [30], process synthesis is recognized as a powerful tool to generate a cost-effective process for the production of bio-based products from biomass derived feedstocks, with high conversion efficiency. It also provides a detailed review of the studies on the process synthesis of biorefineries. Martin and Grossman [22] evaluated a superstructure for the production of biodiesel from cooking oil and algae by formulating an MINLP problem considering heat and water integration. Gebreslassie et al. [23] proposed a superstructure based optimization of algae based biorefinery for simultaneous production of hydrocarbon biofuels and carbon sequestration. A multi-objective MINLP model is developed that simultaneously maximizes the net present value (NPV) and minimizes the global warming potential (GWP). Gong and You [24] formulated an MINLP model to minimize the unit carbon sequestration and utilization cost for algae-based biorefinery processes. In their follow-up study [25], a generic modeling framework for global superstructure optimization is developed for the synthesis and sustainable design of algae processing networks for CO2 mitigation and biofuels production, considering the entire lifecycle. A multi-objective MINLP is developed to simultaneously optimize the unit annualized cost and GWP. Despite several contributions focusing on optimal design of biorefiney configurations: (1) no modeling framework addresses the integration of microalgal biodiesel production with the processing of microalgae residue in a comprehensive manner, i.e., by incorporating a number of potential technological alternatives available for the pre-treatment and further conversion of microalgae residue into biofuels and other useful products; (2) no biorefinery superstructure is developed based on the process data solely applicable to a specific microalgae species by considering the fact that dataset is strain-specific, and every strain has its own dataset starting from the cultivation to the final products. The coupling of residue processing with biodiesel production will enhance the overall economics of biofuels production from microalgae. It is thus the objective of this paper to address these research gaps by developing a systemic methodology to handle simultaneous production of biodiesel and other useful energy products from Chlorella vulgaris using a dataset of preliminary nature. In our previous work [5], we proposed a superstructure based modeling framework for the production of biodiesel from the lipid contents of microalgal biomass. In this work, we extend the optimization framework by adding models for the processing of microalgae residue into useful products so that the overall economics of the biofuels production from microalgae could be improved. The extended framework also accommodates the recycling of water and solvents. The biorefinery superstructure model is developed for the particular species of C. vulgaris but it can easily be extended to consider other species given that data of similar nature can be collected. In addition, the optimization formulation is not strain specific. The optimization results give insight about the promising configurations and cost-effectiveness of microalgae-based biorefinery. Due to the preliminary and inconsistent

71

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

nature of the input data, a set of ordered promising pathways is generated by adding integer cut constraints in the optimization model. The results obtained can guide the researchers to shortlist the candidates and focus on those production routes with high economic potential given the uncertain nature of data. Economic sensitivity study is performed to identify some of the influential model parameters with biggest impact on the overall economics, thus suggesting where to focus for further improvement. 2. Methodology 2.1. Problem definition Given a superstructure encompassing potential technological alternatives/options available at various processing stages, the problem is defined as: Determine an optimal processing route for the simultaneous production of biodiesel from the lipid contents of microalgae and processing of microalgae residue into useful products. Various objective functions for the optimization can be chosen, such as the maximization of yield of the desired products and/or the maximization of gross operating margin (GOM). 2.2. Superstructure development and process description The biorefinery superstructure is developed for the simultaneous production of biodiesel from the lipid contents of microalgae and processing of microalgae residue into a number of useful products. This superstructure (shown in Fig. 1) is based on the process data of C. vulgaris taken from the published literature. Besides its potential advantages for the production of biofuels [31], it also offers the opportunity of developing comprehensive dataset for the optimization formulation. Thus, it is considered as a model species for this study. The developed superstructure consists of eight major processing steps/stages: (1) cultivation of microalgae, (2) harvesting of microalgal biomass, (3) the pre-treatment step including drying and cell disruption of harvested biomass, (4) lipid extraction, (5) transesterification, (6) post-transesterfication purification, (7)

Feed

Cultivation

Harvesting

1,3

Pretreatment

Lipid extraction

1,4

1,5

2,4

2,5

3,4

3,5

4,4

4,5

Transesterification

pretreatment of microalgae residue, and (8) conversion of the residue into useful products. At each processing stage, a number of technological alternatives/options are modeled to perform the respective tasks. As shown in Fig. 1, each option in the superstructure is represented by two indices; the first index represents the option number and the subsequent second index represents the processing stage. The physical descriptions of all the options included in the biorefinery superstructure model are given in Table 1. When only one option is to be selected from each stage, the developed superstructure encompasses 1440 possible combinations of processing pathways for the production of biofuels from microalgae. As the technology for biofuels production from microalgae evolves, more options will be generated that can be incorporated into the superstructure, and this number may become much larger. Cultivation of microalgae: The proposed microalgae based processing network starts with the growth of microalgae. The amounts of carbon dioxide, water and nutrients required for the cultivation of microalgae are calculated on the basis of elemental composition of microalgae [32] described by Dassey et al. [33]. Two technological alternatives are included for the cultivation of microalgae; open pond system and photobioreactor. For the case of open pond system, water is supplied by considering the evaporation losses. 14% of water is assumed to be lost due to evaporation for the case of open ponds [34,35]. As light is essential for microalgae growth, only day time operation is considered. Harvesting of microalgal biomass: The microalgal biomass is concentrated at the harvesting stage. The concentration of biomass at the inlet of the harvesting stage depends on the specific alternative selected at the cultivation stage. In this work, 1.5 g/L and 4 g/L are assumed for the open pond system and photobioreactor, respectively [3,36]. Eight alternatives are included to carry out the harvesting of microalgal biomass. 90% of the harvest water is considered to be recycled and reused for the growth of microalgae at the cultivation stage [36]. Drying and cell disruption: Four technological alternatives are included at the pretreatment stage by considering various

Post transesterification purification

Pretreatment of residue

Conversion of residue

Products

1,8

1,9

Biodiesel

2,8

2,9

Glycerol

3,8

3,9

Bio-oil

4,8

4,9

Bioethanol

5,8

5,9

Biogas

1,6 2,3

2,6 3,3

1,7

3,6 1,2

4,3

5,5

Feed 2,2

6,5

5,3

6,3

5,4

7,5 4,6

Main stream

8,5 7,3 9,5

5,6

Microalgae residue Recycle stream

8,3 10,5

6,6 7,6

2,7

Fig. 1. Biorefinery superstructure for the production of biofuels from C. vulgaris.

72

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

combinations of drying, grinding, microwave, and ultra-sonication. Thus, the pretreatment can be performed in a single step or multiple sub-steps or the pretreatment step can be bypassed entirely which is modeled by the use of empty box, option ‘4,4’. Freeze drying [37,38] is considered for this study, and is modeled based on the flow of utilities required for the drying operation, which is taken from the published literature as cited in Table 1. Both drying and cell disruption steps consume significant energy. That is why the bypass of this stage is considered which naturally accommodates the wet lipid extraction method in the proposed biorefinery superstructure model. Lipid extraction: For lipid extraction stage, nine technological alternatives are included in the superstructure depending upon the options/conditions of the preceding stage. Empty box, ‘10,5’, is also added in this stage to bypass the extraction of lipids; this option can be coupled with ‘in-situ transesterification’ included in the subsequent stage of the superstructure model, where both

Table 1 List of processing alternatives/options and their references. Box No.

Technological alternative/option

Reference

1,1 1,2 2,2 1,3 2,3 3,3 4,3 5,3 6,3 7,3 8,3 1,4 2,4 3,4 4,4 5,4 1,5 2,5 3,5 4,5 5,5 6,5 7,5 8,5 9,5 10,5 1,6 2,6 3,6 4,6 5,6 6,6 7,6 1,7

Feed Open pond system Photobioreactor Flocculation with poly electrolyte Flocculation with NaOH Flocculation with PGA Flocculation with chitosan acid solution Bioflocculation + centrifugation Centrifugation Auto flocculation (induced by high pH) Microfiltration + centrifugation Grinding in liquid nitrogen Drying + ultrasound Drying + grinding + microwave + ultrasound Empty Drying Grinding-assisted lipid extraction Ultrasound assisted extraction by [Bmim][MeSO4] Ultrasound and microwave assisted lipid extraction Wet lipid extraction Solvent extraction (Bligh and Dyer’s Method) Solvent extraction (Modified Bligh and Dyer’s Method) Supercritical fluid extraction Extraction by ionic liquids mixture Extraction by [Bmim][MeSO4] Empty Base catalyzed transesterification Acid catalyzed transesterification Enzymatic transesterification Alkaline in-situ transesterification Acidic in-situ transesterification Enzymatic in-situ transesterification Empty Methanol recovery + gravity separation + washing of biodiesel layer + purification of biodiesel layer + washing of glycerol layer + flash separation Empty Empty Empty Empty Enzymatic hydrolysis Empty Empty Empty Fast pyrolysis Fermentation Anaerobic digestion Biodiesel Glycerol Bio-oil Bioethanol Biogas

[31] [26,33] [26,33] [39] [40] [41] [42] [43] [43] [44] [45] [46] [47] [37]

2,7 1,8 2,8 3,8 4,8 5,8 1,9 2,9 3,9 4,9 5,9 1,10 2,10 3,10 4,10 5,10

[37,38] [46] [47] [37] [48] [49] [49] [50,51] [52] [47] [53] [54] [55] [56] [56] [55] [57–59]

[60]

[61] [60] [62] [3,63,64] [65–66] [61] [60] [62,67]

lipid extraction and transesterification occur in a single step. In case if lipid extraction is not bypassed, the outflow from this stage is split into two streams: (1) microalgal lipids, the main process stream, and (2) microalgae residue which is left over after the extraction of lipids, with the help of a split factor. The lipid stream is sent to the transesterification stage for conversion into biodiesel whereas the non-lipid contents of microalgae are collected in the empty box ‘7,6’ to be sent to next stages for further downstream processing and conversion into the useful products. Transesterification: Six technological alternatives are considered for this stage; three for transesterifcation, and the rest three for insitu transesterification. The technological alternatives to carry out in-situ transesterifcation are selected only if the preceding lipid extraction stage is bypassed. If in-situ transesterifcation is selected, then again the outflow is split into two streams; (1) biodiesel stream which is composed of biodiesel, glycerol as a byproduct, catalyst, and unreacted methanol, and (2) microalgae residue. The biodiesel stream is sent to the next stage for purification whereas the residue is collected in the empty box ‘2,7’ from which it is sent for pretreatment and conversion into various products. The splitting of stream is not required if transesterifcation is carried out through one of the first three alternatives of this stage. Post-transesterifcation purification: For post-transesterification, no technological alternatives are considered, and thus no decision regarding the technology selection is required to carry out the purification step. Here, the empty box ‘2,7’ just allows the microalgae residue to bypass this stage and pass on to next processing stages for pretreatment and conversion to the final products. Pretreatment and conversion of microalgae residue: Pretreatment may be required prior to conversion of microalgae residue into products, depending upon microalgae species and target products. Due to the lack of process data about the processing of microalgae residue, only three technological alternatives are included in the superstructure for pretreatment and conversion of microalgae residue, resulting in the production of bio-oil, bioethanol and biogas from the residue. 2.3. Modeling In our earlier work [5], a superstructure based optimization model was proposed for the production of biodiesel from microalgal biomass. In this section, we extend the framework to accommodate the processing of microalgae residue and recycling of the water and solvents. The extended framework includes mass balance constraints, energy balance constraints, and objective function. 2.3.1. Mass balance constraints Mass balances at each processing stage must be satisfied. In the generic form of the model, the mass balance at option k of stage j with recycle is modeled through Eqs. (1)–(15). The modified general flow diagram for a processing stage and that for each technological alternative/option within a stage are given by the illustrations in Fig. 2(a) and (b), respectively. The nomenclature is given in Table 2. The detailed explanation regarding how indices are arranged can be found in Rizwan et al. [5]. As illustrated in Fig. 2(a), there are three kinds of incoming streams to stage j for each component i (which is an index used to keep track of all the components involved including those in the feed, product, and additive streams); (1) process stream F i0 ðiÞ;j1 continuing from stage j  1 onto stage j, (2) recycle stream Ri,j+1 coming from stage j + 1 to stage j and (3) externally added/makeup stream Qi,j fed to stage j. Here to keep the model form simple, it is assumed that the recycling is considered from stage j + 1 to stage j, as it is the case for the recycling of water from the

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

73

Binary variable yk,j is used for the selection of option k from stage j (if the corresponding option is selected, yk,j equals to 1; otherwise yk,j equals to 0). These binary variables determine the optimal processing pathways and thus are the main decision variables. In the proposed biorefinery superstructure, it is assumed that only one option can be selected at each processing stage. Therefore, the following constraint is enforced:

X yk;j 6 1

ð1Þ

k

Fig. 2. Representation of (a) stage j and (b) option k in stage j, considering recycling.

Table 2 Nomenclature. Indices i, i0 l j k r m

Component (i = lipids, carbohydrate, CO2, water, etc.) Utility Processing stage Option/technological alternative Reaction Key reactant, a subset of i

Parameters Fraction of chemical i added with respect to the incoming flow of component i0 in option k of processing stage j bl,i,k,j Fraction of utility l added with respect to the incoming flow of component i in option k of processing stage j ei,k,j Split factor of component i in option k of processing stage j ci,r,k,j Stoichiometric ratio coefficient of product component i during reaction r in option k of processing stage j hm,r,k,j Fractional conversion of reactant m during reaction r in option k of processing stage j MWi Molecular weight of component i ki;i0 ;k;j Yield coefficient of product i with respect to the incoming flow of component i0 in option k of processing stage j lrecycle Recycle fraction of component i in option k of processing stage j

ai;i0 ;k;j

i;k;j

lwaste i;k;j

Waste fraction of component i in option k of processing stage j

P 1i P 2i P 3i P 4l

Sale price of products Cost of raw material (feed) Cost of chemicals/solvents Cost of utilities

Binary variable yk,j Binary variable; 1 if option k from stage j is selected and 0 if otherwise Continuous variables Fi,j1 Flow of component i in the process stream coming from stage j  1 Fi,j Flow of component i in the process stream leaving stage j Ri,j Flow of component i in the recycle stream leaving stage j Qi,j Flow of component i in the chemical/solvent stream added to stage j Wi,j Flow of component i in the waste stream leaving stage j Ul,j Flow of utility l added to stage j Flow of component i in the process stream leaving option k of stage j F^i;k;j ^ i;k;j R ^ Q

i;k;j

^ i;k;j W ^ l;k;j U

Flow of component i in the recycle stream leaving option k of stage j Flow of component i in the chemical/solvent stream added to option k of stage j Flow of component i in the waste stream leaving option k of stage j Flow of utility l added to option k of stage j

harvesting stage to the preceding one, i.e., the cultivation stage, and it does hold true for the recovery and recycling of methanol as well. However, the index j + 1 can be generalized to incorporate recycling from any later stage. In the similar fashion to incoming streams, there are also three kinds of outgoing streams; (1) process stream Fi,j leaving stage j and continuing onto stage j + 1, (2) recycle stream leaving stage j and will be added in stage j  1 and (3) waste stream Wi,j leaving stage j for disposal.

However, for the processing stages coming after the lipid extraction, an empty box must be selected to model the processing of microalgae residue. In this superstructure, empty boxes can be categorized into two groups. The first type serves the purpose to bypass a processing stage, e.g., ‘4,4’, ‘10,5’, ‘3,8’, ‘5,8’ and these empty boxes are modeled by Eq. (1) as binary decisions are involved in this case and only one option will be selected considering these empty boxes as technological alternatives. The second type of empty boxes is not used to bypass a processing stage, rather to provide a path for a process stream composed of certain components. These includes: ‘7,6’ and ‘2,7’ which are used to send the microalgae residue to next processing stages; ‘1,8’ and ‘1,9’ which give the path to biodiesel to reach product stage; ‘2,8 and ‘2,9’ serving the same purpose for glycerol. These empty boxes are always selected and no binary decisions are involved for their selection. Therefore, the processing stages containing the empty boxes of second type are modeled by Eq. (1) excluding these empty boxes. Given the constraint as in Eq. (1), F i;j , the flow of process stream leaving stage j is given by:

F i;j ¼

X ðyk;j  F^i;k;j Þ

ð2Þ

k

where F^i;k;j is the flow of component i in process stream leaving option k of stage j. The flow of component i in waste stream leaving the stage j ^ i;k , the flow of stream without continuing on to the next stage W Qi,j, the flow of recycle stream Ri,j, and the flow of utility stream Ul,j are modeled in a similar fashion as below:

W i;j ¼

X ^ i;k;j Þ ðyk;j  W

ð3Þ

k

Q i;j ¼

X ^ i;k;j Þ ðyk;j  Q

ð4Þ

k

Ri;j ¼

X ^ i;k;j Þ ðyk;j  R

ð5Þ

k

U ‘;j ¼

X ^ ‘;k;j Þ ðyk;j  U

ð6Þ

k

As shown in Fig. 2(b), it is assumed that a sequence of tasks is occurring for each option/technological alternative box. These tasks include (1) mixing (2) reaction and (3) separation. Mixing occurs at the inlet of the box where three incoming streams ^ i;k;j are being mixed to define the inlet flow of comFi,j1, Ri,j+1 and Q ponent i fed to option k of stage j, as modeled by Eq. (7).

^ F^in i;k;j ¼ ei;k;j  F i;j1 þ Ri;jþ1 þ Q i;k;j

ð7Þ

where Fi,j1 is the flow of process stream of component i going from stage j  1 to option k of stage j, ei,k,j is a split factor of component i entering to option k of stage j and is used for the separation of microalgae residue from the lipid stream, Ri,j+1 is the flow of the recycle stream of component i coming from stage j + 1 to option k

74

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

^ i;k;j is the flow of the externally added/makeup of stage j and Q stream of component i fed to option k of stage j which is given by:

^ i;k;j ¼ Q

X ðai;i0 ;k;j  F i0 ;j1 Þ

ð8Þ

i0

where ai;i0 ;k;j is the fraction of chemicals/solvents (component i) added with respect to incoming component i0 . Reactions are assumed to occur inside the box which is modeled by adding the consumption and generation terms with the help of parameters such as the stoichiometric coefficients and fractional conversions of key reactants. After the reaction, the flow of the process stream at the outlet of box F^out is given by: i;k;j

^in F^out i;k;j ¼ F i;k;j þ

X

ci;r;k;j  hm;r;k;j 

r;m

F^in m;k;j

!!  MW i

MW m

ð9Þ

where ci,r,k,j is the stoichiometric coefficient of product i in reaction r, hm,r,k,j is the fractional conversion of reactant m for reaction r, F^in m;k;j is the flow of reactant m at the inlet of option k of stage j, MWi is the molecular weight of components. The conversion of microalgae residue is modeled with the help of yield coefficient, ki;i0 ;k;j . However, it can also be modeled stoichiometrically but due to the lack of stoichiometry data, the residue processing is modeled with the help of yield parameter which is defined with respect to the incoming flows, as given by Eq. (10):

^in F^out i;k;j ¼ F i;k;j þ

 X ki;i0 ;k;j  F^in i0 ;k;j

ð10Þ

i0

superstructure based on the mass flows. These energy requirements are met by the use of utilities such as steam, electricity, ^ l;k;j fed to option k of stage j is given etc. The flow of utility stream U by:

^ ‘;k;j ¼ U

 X b‘;i;k;j  F^in i;k;j

ð16Þ

i

where bl,i,k,j is the fraction of utilities added and is calculated with respect to each component i of the incoming process streams. 2.3.3. Objective function Two different objective functions tried for the optimization are the maximization of yield of biofuels and the maximization of GOM. Yield: The yield of biofuels is proportional to the flow of biofuels out of the final stage (product stage i.e., stage 10), and is defined on the basis of gasoline gallon equivalent (GGE). The yield can be expressed by:

Yield ¼

X F i;10  GGEi i

ð17Þ

qi

where GGEi is the gasoline gallon equivalent conversion coefficient of component (product) i and qi is the density of component (product) i. GOM: GOM is defined as the difference between total sales and the operating cost. Product sales are given by:

Sales ¼

X ðP1i  F i;10 Þ

ð18Þ

i

where ki;i0 ;k;j is the yield coefficient of product (obtained from residue) i with respect to the incoming flow of component (components in the residue) i0 in option k of stage j, F^in0 is the flow of component i ;k;j

(components in the residue) i0 in option k of stage j. Separation occurs at the outlet where process stream is separated into recycle and waste streams as given below:

F^i;k;j ¼

F^out i;k;j

^ i;k;j  W ^ i;k;j R

Operating Cost ¼ ð11Þ

where F^i;k;j is the flow of process stream of component i leaving ^ i;k;j is the flow of recycle stream leaving option option k of stage j. R k of stage j which is given by: recycle i;k;j

^ i;k;j ¼ l R

 F^out i;k;j

where

ð13Þ

lwaste is the split factor for the waste stream. i;k;j X W i;j

ð14Þ

i;j

Raw material/Feed assignment Option 1 of stage 1 describes the feed quantity in terms of carbon dioxide, water and nutrients required for the cultivation of microalgae which is represented by:

F^i;1;1 ¼ /i

i

i;j

ð19Þ

l;j

where P 2i , P3i and P4l are the cost factors for raw material (feed), chemicals/solvents and utilities, respectively. GOM is given by:

GOM ¼ Sales  Operating Cost

ð20Þ

2.4. Model solution and computational issues

Total waste production is modeled as

waste ¼

X X X ðP2i  F i;1 Þ þ ðP3i  Q i;j Þ þ ðP4l  U l;j Þ

ð12Þ

where lrecycle is the split factor for the recycle stream. i;k;j ^ W i;k;j is the flow of waste stream leaving option k of stage j which is given by:

^ i;k;j ¼ lwaste  F^out W i;k;j i;k;j

where P 1i is the sale price of product i. In Eqs. (17) and (18), the component index i covers over the set of products only, which includes biodiesel, glycerol, bio-oil, bioethanol and biogas. In this work, the operating cost is considered as the sum of raw material cost, chemicals/solvents cost and utilities cost which is given by:

ð15Þ

The MINLP model is formulated and solved in the software package GAMS with the DICOPT solver using a database built in Excel. The problem database contains the input values of all model parameters. These values are taken from the published literature and are given in the Supplementary Data along with the source of information. In our earlier work [5], the MINLP model was linearized to an equivalent MILP form in order to investigate the ease of solution procedure as well as to see the effect of linearization on the optimization results. The results underlined that there is no effect of linearization on the obtained results. However by linearizing the model, the size of the problem gets bigger while the ease of solution remains the same. Therefore in this work, only MINLP form is used in the solution. 3. Results and discussion

where /i is the composition of feed/raw material. 2.3.2. Energy balance constraints Energy balances involve the amount of energy required to operate all the technologies included in the proposed biorefinery

The proposed modeling framework is implemented to determine the optimal biorefinery configurations for the production of biofuels from microalgae. It is assumed that the microalgal biomass (C. vulgaris) is composed of 31.5% lipids, 54.5% proteins and

75

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

14% carbohydrates [68,69]. To gain more insight into microalgaebased biorefinery, two different optimization scenarios are formulated and investigated by employing different objective functions: Scenario-1: Maximization of products yield. Scenario-2: Maximization of GOM. 3.1. Scenario-1: Maximization of product yield A summary of the model and solution statistics for scenario-1 is given in Table 3, and the optimization results are given in Table 4. The optimal processing pathway (Fig. 3) corresponding to the maximum products yield is composed of cultivation of microalgae in photobioreactor, harvesting of microalgal biomass by microfiltration followed by centrifugation, drying of the harvested biomass, acidic in-situ transesterification, post-transesterification purification, and anaerobic digestion for conversion of microalgae residue into biogas.

Table 3 Model summary and solution statistics for scenario-1 and scenario-2.

Description Objective function Constraints Number of equations Number of binary variables Number of continuous variables Number of iterations Optimality gap CPU time (s)

Scenario-1

Scenario-2

Maximization of products yield Given by Eq. (17) Given in Sections 2.3.1 and 2.3.2 23,212 90

Maximization of GOM Given by Eq. (20) Given in Sections 2.3.1 and 2.3.2 23,212 90

29,608

29,608

157 0 0.126

266 0 0.156

3.2. Scenario-2: Maximization of GOM

Table 4 Optimization results (Yield is based on 100 kg of dry biomass). Yield biodiesel Yield glycerol Yield biogas Waste (GGEs) (GGEs) (GGEs) (kg) Scenario-1 Scenario-2

9.189 8.814

0.488 0.468

5.121 4.912

Chlorella vulgaris (Basis: 100 kg dry Photobioreactor (2,2) biomass)

The maximum yield of biodiesel is found to be 9.189 GGEs/ 100 kg of dry biomass. Glycerol is obtained as a byproduct of transesterification process and its yield is found to be 0.488 GGEs/100 kg of dry biomass. In this scenario, microalgae residue is determined to be utilized for the production of biogas with yield of 5.121 GGEs/100 kg of dry biomass. The amount of waste produced with this pathway is found to be 74.741 kg/100 kg of dry biomass. It includes chemicals/solvents utilized at various processing stages. Scenario-1 describes the production of biodiesel from lipid contents of microalgae via in-situ transesterification along with the production of biogas from their residue via anaerobic digestion. In-situ transesterification and anaerobic digestion are selected mainly due to their high conversion of microalgal lipids and residue, respectively. In-situ transesterification also comes with the advantage of directly converting microalgal lipids into biodiesel by avoiding/bypassing the lipid extraction stage, thus reducing the number of processing steps required for biodiesel production. Whereas all the components of residue including the lipids (present in small fraction), proteins and carbohydrates contribute toward production of biogas during anaerobic digestion [70] such is not the case with the production of bioethanol from residue where only carbohydrates play a major role [60]. The fact that the yield of biogas production from microalgae residue is higher than that of bioethanol production or even bio-oil production is the main reason why it is selected as the option for residue processing in the optimal processing pathway. However, very little is known about the processing of microalgae residue from a biological perspective.

GOM ($)

74.741 59.626 75.972 46.493

Microfiltration + Centrifugation (8,3)

Scenario-2 deals with the maximization of GOM. A summary of the model along with solution statistics is given in Table 3, and the optimization results are given in Table 4. The optimal processing pathway (Fig. 4) determined for this scenario is slightly different as it consists of open pond cultivation, harvesting of microalgal biomass by flocculation using poly electrolyte as a flocculent, drying of harvested biomass, acidic in-situ transesterification, post-transesterification purification, and anaerobic digestion for conversion of microalgae residue into biogas. In contrast to scenario-1, here open pond cultivation and flocculation are selected for the growth of microalgae and harvesting, respectively, while all the other technological alternatives

Drying (5,4)

Acidic in-situ transesterification (5,6)

Empty (10,5)

Biodiesel (9.189 GGEs)

Empty (1,9)

Empty (1,8)

Purification of biodiesel

Washing of biodiesel layer

Glycerol (0.488 GGEs)

Empty (2,9)

Empty (2,8)

Flash separation

Washing of glycerol layer

Gravity separation

Methanol recovery

(1,7)

Biogas (5.121 GGEs)

Anaerobic digestion (5,9)

Empty (5,8)

Empty (2,7) Main stream Microalgae residue

Fig. 3. Optimal processing pathway for scenario-1 (maximization of products yield).

76

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

Chlorella vulgaris (Basis: 100 kg dry biomass)

Open pond cultivation (1,2)

Flocculation by poly electrolyte (1,3)

Drying (5,4)

Acidic in-situ transesterification (5,6)

Empty (10,5)

Biodiesel (8.814 GGEs)

Empty (1,9)

Empty (1,8)

Purification of biodiesel

Washing of biodiesel layer

Glycerol (0.468 GGEs)

Empty (2,9)

Empty (2,8)

Flash separation

Washing of glycerol layer

Gravity separation

Methanol recovery

(1,7)

Biogas (4.912 GGEs)

Anaerobic digestion (5,9)

Empty (5,8)

Empty (2,7) Main stream Microalgae residue

Fig. 4. Optimal processing pathway for scenario-2 (maximization of GOM).

Fig. 5b. Breakdown of operating cost.

Fig. 5a. Stage-wise cost distribution and sales of products for scenario-1 and scenario-2.

remain the same. These two are selected because of their lower operating costs compared to other technological alternatives in their respective processing stages. In this scenario the product yield is decreased but the GOM is improved, though it is not improved enough to enforce the biofuels production from microalgae to become economically feasible. It is found to be a loss of $46.493 which implies that the operating cost is higher than the revenue. The amount of waste produced with this pathway is found to be 75.972 kg/100 kg of dry biomass. The major share in the high operating cost is the cost of drying for both the scenarios as shown in Fig. 5a which shows the breakdown of the operating cost for the two scenarios along with the revenue generated by selling the products. Drying of microalgal biomass is generally required prior to the lipid extraction process to effectively and efficiently extract lipids but it requires an input of significant energy. It is also required even for the case of in-situ transesterification. Wet lipid extraction can be a good option to alleviate these problems associated with lipid extraction but it has other drawbacks like low lipid yield and the requirement of large volume of solvent which can be costly. However, yield of lipid extraction process is strain-specific and is more related to the biology of microalgae rather engineering aspects. For the case of C. vulgaris (as chosen in this study), in-situ transesterification of dried microalgal biomass is determined to be the promising one. Therefore, there is a need to evolve some strategies to reduce the

cost of drying or to improve the wet lipid extraction method. Operating cost of transesterification is made higher also by the higher methanol/lipids ratio required for the in-situ transesterification. As given by Eq. (19), operating cost includes cost of feed, cost of chemicals/solvents and cost of utilities, and their contributions are given in Fig. 5b. Cost of utilities takes the largest share for both the scenarios; it implies that a lot of input energy is needed for the production of biofuels. Therefore, at this stage many technical breakthroughs are needed to develop less energy-intensive alternatives which can lead to significant improvements in the overall economics of biofuels production from microalgae. On the contrary, cost of feed is a bit higher for scenario-2 because in open pond cultivation, there are water losses as well due to evaporation, and thus more water intake is required for the growth of microalgae than in the photobioreactor-based cultivation. 3.3. Ordered promising processing pathways The systematic generation of an ordered list of top processing routes is done via generation of ordered solutions by adding integer cut constraints (given by Eq. (21)). This is done to investigate the economic potential of other (promising) pathways apart from the optimal one. More detail on integer cut constraints can be found in Grossmann [71] and Fazlollahi et al. [72]. According to optimization formulation, these processing pathways are suboptimal but still have a lot of potential for researchers to consider for possible improvements even if they are not the optimal choices with respect to the selected objective function.

77

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

GOM ($)

Promising processing pathways

Pathway-1 (Base case) Pathway-2

46.493

Pathway-3

48.894

Pathway-4

49.301

Pathway-5

50.070

Pathway-6

51.612

1,1 1,2 1,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 5,8 1,9 2,9 5,9 1,10 2,10 5,10 1,1 2,2 1,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 5,8 1,9 2,9 5,9 1,10 2,10 5,10 1,1 2,2 3,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 5,8 1,9 2,9 5,9 1,10 2,10 5,10 1,1 1,2 1,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 4,8 1,9 2,9 4,9 1,10 2,10 4,10 1,1 2,2 1,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 4,8 1,9 2,9 4,9 1,10 2,10 4,10 1,1 2,2 3,3 5,4 10,5 5,6 1,7 2,7 1,8 2,8 4,8 1,9 2,9 4,9 1,10 2,10 4,10

47.262

Relative variation of GOM (%)

Table 5 A set of ordered potential processing pathways (scenario-2).

5

Cost of feed Lipid contents Fractional conversion of lipids Residue conversion Lipid yield

4

3

2

1

0 -10

-8

-6

-4

-2

0

2

4

6

8

10

Variation of model parameters (%) Numbers in bold represent the differences in selected alternatives with respect to the base case.

X p2A

n

ðyk;j Þp 

X p2B

n

ðyk;j Þp 6 jBn j  1 n ¼ 1; . . . ; N

ð21Þ

where An = {p|(yk,j)p = 1}, Bn = {p|(yk,j)p = 0}, n = 1, . . ., N. This systematic generation of ordered production routes allows us to evaluate other nearby potential alternatives in view of the preliminary and inconsistent nature of problem dataset. The results in Table 5 show the first five ordered pathways in addition to the optimal one which is also termed as the base case. These pathways are generated for the case of the maximization of GOM (scenario-2). The GOM differences among these processing pathways are not very large as only small variations were seen among their selected alternatives. As shown in Table 5, both the alternatives of the cultivation stage (open pond and photo-bioreactor) have the potential to serve the purpose of cultivation with only a small effect on GOM, and need more precise, accurate and reliable data for the final selection. Furthermore, the selection of photobioreactor ends up with more variations in the downstream processing than the selection of an open pond system. From the harvesting stage, flocculation with NaOH also shows the potential to carry out the harvesting task. From the perspective of residue processing and conversion, bioethanol is emerged as a potential candidate and spotted as a good candidate next to biogas. This option may evolve as even more attractive for the starch-enriched microalgae residue. The rest of alternatives remain the same throughout this analysis. However, no specific pattern is observed among the solutions. Also, as the processing technologies improve and additional technologies are developed, these results may change. Here, we have explored the economic potential of multiple processing routes other than the optimal one. 3.4. Sensitivity analysis To investigate and evaluate the effect of key model parameters on GOM, sensitivity analysis is performed. The parameters evaluated include cost of feed, lipid contents, fractional conversion of lipids into biodiesel, residue conversion (which is defined in term of product yield from residue), and lipid extraction yield. The results of sensitivity analysis are presented in Fig. 6, where the value of GOM obtained in scenario-2 is used as a reference. GOM is found to be most sensitive to the cost of feed which includes the cost of carbon dioxide, water and nutrients. With 10% decrease in feed cost, GOM is increased by 3.9%. Surprisingly, residue conversion is found to be the second most sensitive parameter followed by fractional conversion of lipids and lipid contents. With 10% increase in residue conversion, GOM is increased by 3.65%. These results underline that the processing of microalgae residue is an important and promising aspect of microalgal biorefinery, and technological improvements in this field can create a huge

Fig. 6. Sensitivity analysis performed for scenario-2.

impact on the economics of microalgal biorefinery. Therefore, there is a need to pay equal attention to develop promising technologies for converting residue into useful products or industrial intermediate, along with the production of biodiesel from lipid contents. Increase in lipid contents does not impact the bottom line as significantly as expected. Increase in lipid contents indirectly results in less residue and thus less revenue is generated from residue-based products. As a result, GOM is not increased significantly by increasing lipid contents and, inherently it also strengthens the fact about the potential of microalgal residue. The yield of lipid extraction process does not affect GOM at all because biodiesel is produced via insitu transesterification and even 10% increase in the lipid yield is not enough to cause any change in the selected alternatives. Therefore, lipid yield has no effect on overall results until and unless an alternative from the lipid extraction stage is selected. These results also highlight that by evaluating ±10% variations in these model parameters, there is no change in the optimal processing pathway because these variations are globally applied to the whole superstructure, and thus result in affecting the value of objective function only with no change in the optimal production route. These key model parameters are related to both process development/improvements and microalgal biology except cost of feed which is more related to engineering aspects. Therefore, these parameters can create a huge impact on overall economics of microalgal biorefinery through improvements from both engineering and biological standpoints. 3.5. Potential strategies for improvements in GOM Based on the results of sensitivity analysis, some new strategies are suggested for potential improvements in GOM. Furthermore, this analysis also suggests possible targets that must be achieved in order to make GOM to approach break-even point. The obtained results are shown in Table 6 and theses strategies are further explained below: Case-1: Cost of feed is the most sensitive parameter as evaluated by the sensitivity analysis. Thus, they can be first targeted for improvement. In particular, CO2 purity requirement for microalgae cultivation deserves a close look. For this, we analyzed the economics with an added assumption that CO2 is available at no cost. At this point, without the carbon tax, this implies that CO2 from Table 6 Effect of different strategies/parameters combinations on GOM. Base case

Case-1

Case-2

Case-3

Case-4

Case-5

GOM ($) 46.493 36.993 30.123 25.231 21.738 0 % Increase in GOM 0 20.4 35.2 45.7 53.2 100

78

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

fossil fuels-based combustion systems can be added directly to the cultivation system. As shown in Table 6 and 20.4% increase in GOM can be achieved with respect to the base case (scenario-2) by assuming that CO2 is available at no cost. Additionally, this will also simplify CO2 separation from flue gas significantly [73]. Case-2: Cost of feed also includes the cost of nutrients required for the growth of microalgae. In this case, the enhanced effect of case-1 is combined with 90% recycling of the nutrients. Water recycling is already considered in the base case. Overall 35.2% increase in GOM is observed by assuming 90% recycling of nutrients and then combining with case-1. Case-3: Residue conversion is also determined as one of the influential parameters. As the processing technologies continue to improve or additional technologies may develop, this aspect should see significant improvement. By assuming 30% increase in the residue conversion and then by integrating with case-2, overall 45.7% increase in GOM is observed. Case-4: Maximum achievable lipid contents of C. vulgaris is reported as 58% based on dry biomass basis [68]. By integrating the maximum achievable lipid contents with enhanced effect of case-3, GOM is improved by 53.2%. The impact of lipid contents on GOM is not as significant as expected, as discussed in Section 3.3. Case-5: After exploring all the possible potential combinations of these key model parameters, GOM is still found to be a loss of $21.7 and therefore, needs more reduction in operating cost for the operating margin to approach the break-even point. In this case, by considering improvements in process technologies, the energy requirements for the drying operation are assumed to be reduced by 45%. When this reduction is integrated with the results of case-4, GOM indeed was seen to approach the break-even point. These strategies set the potential targets that must be achieved in order to make biofuels production from microalgae economically viable. The key results of these strategies highlight that in general there is an essential need (1) to improve existing technologies and/or to develop new processes that can significantly reduce costs, (2) to develop innovative techniques based on the concept of genetic engineering and synthetic biology because most of the parameters are related with the biology of microalgae and can be improved drastically through biological improvements. Therefore, by the innovative combinations of both engineering and biological improvement opportunities, cost can be reduced significantly, and the production of biofuels from microalgae may become economically viable one day.

4. Conclusions In this contribution, an MINLP model is developed to determine optimal/promising biorefinery configurations from a large number of processing alternatives for the production of biofuels from microalgae. Optimization results showed that the GOM lies below the breakeven point, and hence the production of microalgal biofuels is not economically viable. Economic sensitivity analysis revealed the parameters that should be targeted for substantial improvements in the economics of microalgae-derived fuels. Therefore, few potential strategies are devised that elaborate the targets and/or the challenges that must be handled in the future by technical breakthroughs from both engineering and biological perspectives of microalgae. The proposed optimization framework is computationally very efficient and has the capability to quickly scan through all the potential processing alternatives to locate the optimal ones under various choices of objective functions. Due to uncertain nature of input data, there is a need to manage and address uncertainties in model parameters to ensure robust decision making. Our future

work, therefore, will include the development of systematic method to address the uncertainties present in the dataset. Acknowledgement This work was supported by the Advanced Biomass R&D Center (ABC) of Global Frontier Project funded by the Ministry of Education, Science and Technology, South Korea (ABC- 20110031354). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.apenergy.2015. 04.018. References [1] Harun R, Singh M, Forde GM, Danquah MK. Bioprocess engineering of microalgae to produce a variety of consumer products. Renew Sustain Energy Rev 2010;14:1037–47. [2] Nigam PS, Singh A. Production of liquid biofuels from renewable resources. Prog Energy Combust Sci 2011;37:52–68. [3] Chisti Y. Biodiesel from microalgae. Biotechnol Adv 2007;25:294–306. [4] Pires JCM, Alvim-Ferraz MCM, Martins FG, Simoes M. Carbon dioxide capture from flues gasses using microalgae: engineering aspects and biorefinery concept. Renew Sustain Energy Rev 2012;16:3043–53. [5] Rizwan M, Lee JH, Gani R. Optimal processing pathway for the production of biodiesel from microalgal biomass: a superstructure based approach. Comput Chem Eng 2013;58:305–14. [6] Rawat KRR, Mutanda T, Bux F. Biodiesel from microalgae: a critical evaluation from laboratory to large scale production. Appl Energy 2013;103:444–67. [7] Park JH, Yoon JJ, Park HD, Lim DJ, Kim SH. Anaerobic digestibility of algal bioethanol residue. Bioresour Technol 2012;113:78–82. [8] Rashid N, Rehman MSU, Han J. Recycling and reuse of spent microalgal biomass for sustainable biofuels. Biochem Eng J 2013;75:101–7. [9] Kim J, Yoo G, Lee H, Lim J, Kim K, Kim CW, et al. Methods of downstream processing for the production of biodiesel from microalgae. Biotechnol Adv 2013;31:862–76. [10] Liu P, Pistikopoulos EN, Li Z. An energy systems engineering approach to polygeneration and hydrogen infrastructure systems analysis and design. Chem Eng Trans 2009;18:373–8. [11] Yue D, You F, Snyder SW. Biomass-to-bioenergy and biofuel supply chain optimization: overview, key issues and challenges. Comput Chem Eng 2014;66:36–56. [12] Grossmann IE. Mixed-integer programming approach for the synthesis of integrated process flowsheets. Comput Chem Eng 1985;9:463–82. [13] Grossmann IE. Mixed-integer nonlinear programming techniques for the synthesis of engineering systems. Res Eng Des 1990;1:205–28. [14] Yeomans H, Grossmann IE. A systematic modeling framework of superstructure optimization in process synthesis. Comput Chem Eng 1999;23:709–31. [15] Drobez R, Pintaric ZN, Pahore B, Kravanja Z. MINLP synthesis of processes for the production of biogas from organic and animal waste. Chem Biochem Eng Q 2009;23:445–59. [16] Quaglia A, Sarup B, Sin G, Gani R. Integrated business and engineering framework for synthesis and design of enterprise-wide processing networks. Comput Chem Eng 2012;38:213–23. [17] Simasatitkul L, Arpornwichanop A, Gani R. Design methodology for bio-based processing: biodiesel and fatty alcohol production. Comput Chem Eng 2013;57:48–62. [18] Zondervan E, Nawaz M, de Haan AB, Woodley JM, Gani R. Optimal design of a multi-product biorefinery system. Comput Chem Eng 2011;35:1752–66. [19] Martin M, Grossmann IE. Process optimization of FT-Diesel production from lignocellulosic switchgrass. Ind Eng Chem Res 2011;50:13485–99. [20] Ng RTL, Tay DHS, Ng DKS. Simultaneous process synthesis, heat and power integration in a sustainable integrated biorefinery. Energy Fuels 2012;26:7316–30. [21] Rizwan M, Lee JH, Gani R. Superstructure optimization of biodiesel production form microalgal biomass. In: Proceedings of the 10th IFAC international symposium on dynamics and control of process systems (DYCOPS). Elsevier Science; 2013. p. 111–6. [22] Martin M, Grossmann IE. Simultaneous optimization of heat integration for biodiesel production from cooking oil and algae. Ind Eng Chem Res 2012;51:7998–8014. [23] Gebreslassie BH, Waymire R, You F. Sustainable design and synthesis of algaebased biorefinery for simultaneous hydrocarbon biofuels production and carbon sequestration. AIChE J 2013;59:1599–621. [24] Gong J, You F. Optimal design and synthesis of algal biorefinery processes for biological carbon sequestration and utilization with zero direct greenhouse

M. Rizwan et al. / Applied Energy 150 (2015) 69–79

[25]

[26] [27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41] [42] [43]

[44]

[45]

[46]

[47]

gas emissions: MINLP model and global optimization algorithm. Ind Eng Chem Res 2014;53:1563–79. Gong J, You F. Global optimization for sustainable design and synthesis of algae processing network for CO2 mitigation and biofuel production using life cycle optimization. AIChE J 2014;60:3195–210. Davis R, Aden A, Pienkos PT. Techno-economic analysis of autotrophic microalgae for fuel production. Appl Energy 2011;88:3524–31. Delrue F, Setier PA, Sahut C, Cournac L, Roubaud A, Peltier G, et al. An economic, sustainability, and energetic model of biodiesel production from microalgae. Bioresour Technol 2012;111:191–200. Delrue F, Beisseon YL, Setier PA, Sahut C, Roubaud A, Froment AK, et al. Comparison of various microalgae liquid biofuel production pathways based on energetic, economic and environmental criteria. Bioresour Technol 2013;136:205–12. Slegers PM, Koetzier BJ, Fasaei F, Wijffels RH, van Straten G, van Boxtel AJB. A model-based combinatorial approach for energy-efficient processing of microalgae. Algal Res 2014;5:140–57. Yuan Z, Chen B, Gani R. Application of process synthesis: moving from conventional chemical processes towards biorefinery processes. Comput Chem Eng 2013;49:217–29. Safi C, Zebib B, Merah O, Pontalier P, Vaca-Garcia C. Morphology, composition, production, processing and applications of Chlorella vulgaris: a review. Renew Sustain Energy Rev 2014;35:265–78. Williams PJB, Laurens LML. Microalgae as biodiesel and biomass feedstocks: review and analysis of the biochemistry, energetics and economics. Energy Environ Sci 2010;3:554–90. Dassey AJ, Hall SG, Theegala CS. An analysis of energy consumption for algal biodiesel production: comparing the literature with current estimates. Algal Res 2014;4:89–95. Guieysse B, Bechet Q, Shilton A. Variability and uncertainty in water demand and water footprint assessments of fresh algae cultivation based on case studies from five climatic regions. Bioresour Technol 2013;128:317–23. Weissman JC, Geobel RP. Design and analysis of microalgal open pond systems for the purpose of producing fuels, SERI/SP-231-2840. Golden, Colorado: Solar Energy Research Institute; 1978. Richardson JW, Johnson MD, Zhang X, Zemke P, Chen W, Hu Q. A financial assessment of two alternative cultivation systems and their contributions to algae 50 biofuel economic viability. Algal Res 2014;4:96–104. Sostaric M, Klinar D, Bricelj M, Golob J, Berovic M, Likozar B. Growth, lipid extraction and thermal degradation of the microalgal Chlorella vulgaris. New Biotechnol 2012;29:325–31. Lee AK, Lewis DM, Ashman PJ. Disruption of microalgal cells for the extraction of lipids for biofuels: processes and specific energy requirements. Biomass Bioenergy 2012;46:89–101. Granados MR, Acien FG, Gomez C, Fernandez-Sevilla JM, Grima EM. Evaluation of flocculants for the recovery of freshwater microalgae. Bioresour Technol 2012;118:102–10. Araujo GS, Matos LJBL, Goncalves LRB, Fernandes FAN, Farias WRL. Bioprospecting for oil producing microalgal strains: evaluation of oil and biomass production for ten microalgal strains. Bioresour Technol 2011;102: 5248–50. Zheng H, Gao Z, Yin J, Tang X, Ji X, Huang H. Harvesting of microalgae by flocculation with poly (c-glutamic acid). Bioresour Technol 2012;112:212–20. Rashid N, Rehman SU, Han J. Rapid harvesting of freshwater microalgae using chitosan. Process Biochem 2013;2013(48):1107–10. Salim S, Vermue MH, Wijffels RH. Ratio between autoflocculating and target microalgae affects the energy-efficient harvesting by bio-flocculation. Bioresour Technol 2012;118:49–55. Vandamme D, Foubert I, Fraeye I, Meesschaert B, Muylaert K. Flocculation of Chlorella vulgaris induced by high pH: role of magnesium and calcium and practical implications. Bioresour Technol 2012;105:114–9. Bilad MR, Vandamme D, Foubert I, Muylaert K, Vankelecom IFJ. Harvesting microalgal biomass using submerged microfiltration membranes. Bioresour Technol 2012;111:343–52. Zheng H, Yin J, Gao Z, Huang H, Ji X, Dou G. Disruption of chlorella vulgaris cells for the release of biodiesel-producing lipids: a comparison of grinding, ultrasonication, bead milling, enzymatic lysis, and microwaves. Appl Biochem Biotechnol 2011;164:1215–24. Kim Y, Park S, Kim MH, Choi Y, Yang Y, Kim HJ, et al. Ultrasound-assisted extraction of lipids from Chlorella vulgaris using [Bmim][MeSO4. Biomass Bioenergy 2013;56:99–103.

79

[48] Park JY, Oh YK, Lee JS, Lee K, Jeong MJ, Choi SA. Acid-catalyzed hot-water extraction of lipids from Chlorella vulgaris. Bioresour Technol 2014;153: 408–12. [49] Jeon J, Choi H, Yoo G, Choi Y, Choi K, Park H, et al. New mixture composition of organic solvents for efficient extraction of lipids from Chlorella vulgairs. Biomass Bioenergy 2013;59:279–84. [50] Mendes RL, Coelho JP, Fernandes HL, Marrucho IJ, Cabral JMS, Novais JM. Application of supercritical CO2 extraction to microalgae and plants. J Chem Technol Biotechnol 1995;62:53–9. [51] Brentner LB, Eckelman MJ, Zimmerman JB. Combinatorial life cycle assessment to inform process design of industrial production of algal biodiesel. Environ Sci Technol 2011;45:7060–7. [52] Choi S, Oh Y, Jeong M, Kim SW, Lee J, Park J. Effects of ionic liquid mixtures on l ipid extraction from Chlorella vulgaris. Renew Energy 2014;65:169–74. [53] Plata V, Kafarov V, Moreno N. Optimization of third generation biofuels production: biodiesel from microalgae oil by homogenous transesterification. Chem Eng Trans 2010;21:1201–6. [54] Lam MK, Lee KT. Catalytic transesterification of high viscosity crude microalgae lipid to biodiesel: effect of co-solvent. Fuel Process Technol 2013;110:242–8. [55] Tran DT, Yeh KL, Chen CL, Chang JS. Enzymatic transesterification of microalgal oil from Chlorella vulgaris ESP-31 for biodiesel synthesis using immobilized Burkholderia lipase. Bioresour Technol 2012;108:119–27. [56] Orta SBV, Lee JGM, Harvey A. Alkaline in situ transesterification of Chlorella vulgaris. Fuel 2012;94:544–50. [57] Dhar BR, Kirtania K. Excess methanol recovery in biodiesel production process using a distillation column: a simulation study. Chem Eng Res Bull 2009;13:55–60. [58] Glisic SB, Skala DU. Design and optimization of purification procedure for biodiesel washing. Chem Ind Chem Eng Q 2009;15:159–68. [59] Zhang Y, Dube MA, McLean DD, Kates M. Biodiesel production from waste cooking oil: 1. Process design and technological assessment. Bioresour Technol 2003;89:1–16. [60] Moncada J, Jaramillo JJ, Higuita JC, Younes C, Cardona CA. Production of bioethanol using Chlorella vulgaris Cake: a technoeconomic and environmental assessment in the Colombian context. Ind Eng Chem Res 2013;52:16786–94. [61] Wang K, Brown RC, Homsy S, Martinez L, Sidhu SS. Fast pyrolysis of microalgae remnants in a fluidized bed reactor for bio-oil and biochar production. Bioresour Technol 2013;27:494–9. [62] Ehimen EA, Connaughton S, Sun Z, Carrington GC. Energy recovery from lipid extracted, transesterified and glycerol codigested microalgae biomass. GCB Bioenergy 2009;1:371–81. [63] Demirbas A. Competitive liquid biofuels from biomass. Appl Energy 2011;88:17–28. [64] Amaro HM, Guedes AC, Malcata FX. Advances and perspectives in using microalgae to produce biodiesel. Appl Energy 2011;88:3402–10. [65] Moon C, Ahn JH, Kim SW, Sang BI, Um Y. Effect of biodiesel-derived raw glycerol on 1,3-propanediol production by different microorganisms. Appl Biochem Biotechnol 2010;161:502–10. [66] Mu Y, Teng H, Zhang DJ, Wang W, Xiu ZL. Microbial production of 1,3propanediol by Klebsiella pneumoniae using crude glycerol from biodiesel preparations. Biotechnol Lett 2006;28:1755–9. [67] Holm-Nielsen JB, Al Seadi T, Oleskowicz-Popiel P. The future of anaerobic digestion and biogas utilization. Bioresour Technol 2009;100:5478–84. [68] Mata TM, Martins AA, Caetano NS. Microalgae for biodiesel production and other applications: a review. Renew Sustain Energy Rev 2010;14:217–32. [69] Becker EW. Microalgae in human and animal nutrition. In: Richmond A, editor. Handbook of microalgal culture. Oxford: Blackwell Publishing; 2004. p. 312–51. [70] Sialve B, Bernet N, Bernard O. Anaerobic digestion of microalgae as a necessary step to make microalgal biodiesel sustainable. Biotechnol Adv 2009;27:409–16. [71] Grossmann IE. Mixed-integer nonlinear programming techniques for process systems engineering. Comput Chem Eng 1995;19:189–204. [72] Fazlollahi S, Mandel P, Becker G, Marechal F. Methods for multi-objective investment and operating optimization of complex energy systems. Energy 2012;45:12–22. [73] Olaizola M. Microalgal removal of CO2 from flue gases: changes in medium pH and flue gas composition do not appear to affect the photochemical yield of microalgal cultures. Biotechnol. Bioprocess Eng 2003;8:360–7.