Economic Risk Assessment of Early Stage Designs ... - ACS Publications

Article pubs.acs.org/IECR

Economic Risk Assessment of Early Stage Designs for Glycerol Valorization in Biorefinery Concepts Carina L. Gargalo,† Peam Cheali,† John A. Posada,‡ Krist V. Gernaey,† and Gürkan Sin*,† †

Downloaded via UNIV OF SUSSEX on July 9, 2018 at 22:20:44 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

CAPEC-PROCESS Research Center, Department of Chemical and Biochemical Engineering, Technical University of Denmark, Building 229, DK-2800 Lyngby, Denmark ‡ Section Biotechnology & Society, Department of Biotechnology, Faculty of Applied Sciences, Delft University of Technology, Julianalaan 67, 2628BC Delft, The Netherlands S Supporting Information *

ABSTRACT: A systematic methodology to critically assess and screen among early stage design alternatives was developed for the use of glycerol. Through deterministic sensitivity analysis it was found that variations in the product and feedstock prices, total production cost, fixed capital investment, and discount rate, among others, have high impact on the project’s profitability analysis. Therefore, the profitability was tested under uncertainties by using NPV and MSP as economic metrics. The robust ranking of solutions is presented with respect to minimum economic risk of the project being nonprofitable (failure to achieve a positive NPV times the consequential profit loss). It was found that the best potential options for glycerol valorization is through the the production of either (i) lactic acid (9 MM$ economic risk with 63% probability of failure to achieve a positive NPV); (ii) succinic acid (14 MM$ economic risk with 76% probability of failure to achieve a positive NPV); or finally, (iii) 1,2-propanediol (16 MM$ economic risk with 68% probability of failure to achieve a positive NPV). As a risk reduction strategy, a multiproduct biorefinery is suggested which is capable of switching between the production of lactic acid and succinic acid. This solution comes with increased capital investment; however, it leads to more robust NPV and decreased economic risk by approximately 20%, therefore creating a production plant that can continuously adapt to market forces and thus optimize profitability. replacement of oil-based chemicals.1 Therefore, research into innovation on the topic of integrated biorefineries for the production of biofuels and chemicals, and its challenges, has been growing at a significant pace. These studies investigate not only

1. INTRODUCTION Concerns about the environment, dependency on fossil fuels and their persistent supply and security prices have motivated the development of a biobased economy. The latter aims at replacing fossil fuels on a large scale, not exclusively for energy purposes, but furthermore to develop/support the production of chemicals and materials. If biomass conversion is implemented, it would allow the decrease of nonrenewable fuel usage and encourage the establishment of a robust bioindustry, and potentially lead to the © 2016 American Chemical Society

Received: Revised: Accepted: Published: 6801

December 1, 2015 May 10, 2016 May 19, 2016 May 19, 2016 DOI: 10.1021/acs.iecr.5b04593 Ind. Eng. Chem. Res. 2016, 55, 6801−6814

Article

Industrial & Engineering Chemistry Research Table 1. Glycerol-Based Biorefinery Concepts Investigated products

conversion technologies and downstream processing

production rate (ton/y)

glycerol fermentation by an engineered strain of B. succinicis producens DD1 + reactive extraction w/ TOA and 1-octanol glycerol fermentation by an engineered strain of K. pneumoniae + reactive extraction w/ isobutyraldehyde glycerol fermentation by an engineered strain of P. acidopropionoci ACK-Tet + reactive extraction w/TO and ethyl acetate glycerol fermentation by an engineered strain of C. necator + blending w/surfactant solution + hypochlorite digestion centrifugation + spray drying sequential processes of dehydrogenation-hydrogenation via hydroxyacetone glycerol fermentation by an engineered strain of E.coli + + reactive extraction w/TOA and DCE

10200

7300

H2 n-butanol

chemical conversion of glycerol to propanal via acrolein + methanol conversion to methanal + methanal and propanal condensed to methacrolein + hydrogentation of methacrolein to isobutyl alcohol steam reforming of glycerol for hydrogen production over Ni/SiO2 catalyst glycerol anaerobic fermentation by an engineered strain of C. pasteurianum + vacuum stripping

ethanol

glycerol fermentation by an engineered strain of E.coli + dehydration

5211

acrolein

glycerol dehydration

4870

succinic acid 1,3-propanediol propionic acid PHB 1,2-propanediol lactic acid isobutyl-alcohol

how to transition toward a biobased economy, but also how to deal with its challenges, such as sustainability concerns of the biomass feedstock, use efficiency, and potential economy of scales in biomass utilization.2 However, challenges are still present such as how to compete with the efficiency of the current petrochemical industry to convert biomass into fuels, high-value chemicals, and other biobased building blocks. The production of building blocks, the processes to convert these building blocks to final products and, the capital investment necessary to take the technology to commercial scale, need a clear economic incentive in order to be pursued. The concept of an integrated biorefinery for biofuel production provides a large-volume product to achieve economies of scale, while lower-volume, biobased coproducts and power production can improve the overall economics of the facility, hence, providing the economic incentive for the industrial acceptance of renewable carbon as feedstock. Nevertheless, the choice of product portfolio faces significant challenges such as (i) the great range of alternative target products, at present produced by the oil-based industry, or yet to be manufactured from biorefinery building blocks, and (ii) difficult process screening due to the lack of established technologies and real plant data. Furthermore, the conversion of biomass-based carbon compounds to chemicals has been reported as the most complex technology and is still remains under development.1,3 This implies that during the first stages of biorefinery design and development, there are several alternative technologies, feedstocks, and products, generating a great number of potential processing pathways; therefore, there is a need for screening and gathering the most appropriate configurations with regards to economics, environmental constraints, and overall sustainability. However, since real data are often incomplete or not available, early stage design of biorefineries is marked by assumptions, hypotheses, and simplifications that need to be made in order to represent the complexity of the problem. Consequently, strategic uncertainty as well as operational uncertainty on the techno-economic parameters is expected,4 and needs to be appropriately handled.5

5500 6796 3600 7740 7931

659 3000

yield

data sources

1.02 g/g glycerol (0.80 mol/mol) 0.55 g/g glycerol (0.67 mol/mol) 0.68 g/g glycerol (0.844 mol/mol) 0.36 g/g glycerol

8, 14

0.774 g/g glycerol 0.79 g/g glycerol (0.81 mol/mol) 0.73 g/g glycerol

19, 20 21

0.66 g/g glycerol 0.30 g/g glycerol (0.37 mol/mol) 0.52 g/g glycerol (1.02 mol/mol) 0.486 g/g glycerol (0.80 mol/mol glycerol)

24, 25 8, 26, 27 27, 28

15, 16 17 18

22, 23

29

Therefore, since preliminary process engineering assessments can address “what if” questions, early stage analysis can reduce risks related to the decision-making process.1,6 Therefore, early studies in our group have developed a systematic and robust methodology that aims to assist the user to manage the complexity of input data and their intrinsic uncertainties.6 In this work, the methodology is extended by (i) identifying the sources of uncertainty which affect the economic performance the most using deterministic sensitivity analysis; (ii) proactively incorporating uncertainties through stochastic optimization and scenario-based analysis; (iii) quantitative and qualitative analysis of economic metrics; (iv) quantifying the risk of potential profit loss for the top-ranked alternatives taking into account time value of money; and, finally (v) exploring the implementation of flexible concepts as reduction strategy for the minimization of future risk. The work aims ultimately at providing a robust assessment and decision-making tool for the identification of optimal biorefinery concepts at the screening phase, under uncertainties with regards to technical and economic criteria. This systematic approach is applied to a significant case study focusing on value added products from glycerol. Today 2/3 of the world glycerol supply is produced alongside with biodiesel as a coproduct. The world biodiesel production has significantly increased with an average growth of 11% annually (growth rate end-2009 through 2014) and it reached a maximum of 125 billion liters in 2014.7 The increased biodiesel production has created a significant amount of surplus crude glycerol, and it has resulted in a 10-fold decrease in crude glycerol prices.1 Henceforth, glycerol valorization by chemical and biochemical conversion to biofuels, high value-added chemicals, and building blocks has been pointed out as a powerful alternative to add value to this side stream.1,8 By transforming the conventional biodiesel industry into an integrated biorefinery, it would not only increase its economic viability, but also provide a potential replacement platform for fossil-based products, instigating a probable overall improvement of environmental performance. Thus, in this study we aim at identifying and ranking possible glycerol valorization concepts to 6802

DOI: 10.1021/acs.iecr.5b04593 Ind. Eng. Chem. Res. 2016, 55, 6801−6814

Article

Industrial & Engineering Chemistry Research

value than an equal amount received or paid today. Because the goal of investing is to increase the shareholder’s wealth, an investment is worth undertaking if it creates value, and the investment creates value if it crates more value than it costs, considering the time value of money (discounted cash in-flows and out-flows). Therefore, a process to be economically viable/ acceptable has to present a NPV higher than zero (at which point it breaks-even, value created is equal to costs), which means that it is expected to generate more revenue than could be obtained by gaining the discount rate, representing the time value of money which is the amount that could be obtained by investing in other alternatives.30,31 Capturing the time dependency of cashflows during the project is very important to investors first because not all the money has to be financed immediately, and second because the sooner the capital is repaid to the investors, the sooner it can be used to fund another investment.31 This model is defined to be the interest rate at which all cash flows must be discounted in order for the net present value of the project to be equal to zero (breakeven). Therefore, it is based on the calculation of the net present value by setting a discount rate r, which is the interest rate that represents the minimum rate of return that the company is willing to accept for a new investment. The model also enables the calculation of minimum selling price to comprehensively assess and compare the process alternatives (eqs 2 to 9). Noteworthy is that this method is recommended when significant uncertainties are present and therefore risk is a challenge.32 A brief description of the assumptions for the economic model is presented in Supporting Information (section B) and a summary is presented in Table 2.

be implemented into the biodiesel plants as an add-on. A database containing relevant processing technologies for production of a range of high value chemicals and fuels are compiled from literature. Economic metrics such as net present value (NPV) and minimum selling price (MSP) are used to perform a comparative analysis of the suggested processing pathways under uncertainties such as market price fluctuations, capital investment, sales volume, and total product production cost variability. The manuscript is organized as follows: the superstructure formulation and data collection (database) for glycerol biorefinery economic model for the assessment are presented in Materials and Methods; followed by the Results, Discussion, and Conclusions sections.

2. MATERIALS AND METHODS 2.1. Case-Study: Glycerol-Based Biorefinery Concepts. The superstructure representing the glycerol-based biorefinery concepts was formulated (please see Supporting Information, Figure A1). The current product portfolio composing the database included in the superstructure was selected based on the following criteria: (i) important reports on biobased chemicals, such as the references9−12 in which the main biobased chemicals (value-added products) that could be coproduced with bioenergy/biofuel in integrated biorefineries are reported, in this case specifically checking for the ones that can be potentially obtained from glycerol, a biodiesel coproduct; and crosschecking this information with the (ii) availability and completeness of information for techno-economic assessment (given by previous studies and process simulations with sufficient detail). As having a comprehensive database is by itself very valuable, an extensive literature review has been done, gathering, to the best of our knowledge, the most significant contributions for the current set of products. Several building blocks for a wide range of applications (various bulk and niche chemicals) are included in the product portfolio, such as (i) poly-3hydroxybutyrate (PHB), lactic acid, and 1,3-propanediol, which are polymer building blocks for biopolyesters, and polylactic acid (PLA), polyethers, polyurethanes, and plastics; and (ii) succinic acid, which can be used as food additive (natural flavor), excipient in pharmaceutical products (to control acidity) and also as a precursor of polyesters and thermoplastics (polybutylene terephthalate, PBT). Table 1 presents a summary of the biorefinery concepts analyzed in this work; the full database and the superstructure of the glycerol conversion to value-added chemicals are presented in the Supporting Information, Table A1. The glycerol-based biorefinery concepts investigated are based on a medium-sized biodiesel plant in Europe,13 providing approximately 10 kton glycerol/year, corresponding to 10% (w/w) of the total annual biodiesel production rate (∼100 kton/year). 2.2. Techno-Economic Analysis: Methods and Assumptions. To make an investment decision, the foreseen profit from an investment must be evaluated relative to some profitability standards, a quantitative measurement of profit with respect to the investment necessary to generate that profit. In this work, for the project evaluation, the economic model used is the discounted cash-flow rate of return (DCFR). This model is used to estimate the attractiveness of an investment. It uses future cash flow predictions and discounts them to get a present value estimate, which is then used to analyze its potential for investment. Therefore, it considers the time value of money, that is, the DCFR takes into consideration that money to be received or paid at some time in the future is viewed as having less

T

NPV =

∑ t=1

Ct − C0 (1 + r )t

(2)

Table 2. Summary of the Assumptions Used for the Discounted Cash-Flow Rate of Return30,33,34 parameter

assumption

plant life (years) discount rate (mar) depreciation period (years) equity loan interest loan term (years) construction period (years) % spent in year −1 % spent in year 0 start-up time (years) product production/feedstock use (% of normal) variable costs (% of normal) fixed cost (% of normal) income tax rate cost year for analysis

20 10% 10 (MACRS system) 40% 5% 10 2 60% 40% 0.50 50% 75% 100% 35% 2014

where t = 1, ..., T, Ct = net cash inflow during the lifetime of the plant, C0 = total initial investment costs NPV = annual present value T

−

∑ t =−2

6803

(annual fixed investment cost) (3) DOI: 10.1021/acs.iecr.5b04593 Ind. Eng. Chem. Res. 2016, 55, 6801−6814

Article

Industrial & Engineering Chemistry Research Table 3. Algorithm for Economic Assessment under Uncertainty step

description

1

identify economic metrics (Yk) and input model variables (Xi)

2

initialize model → deterministic solution (base case)

output

Xi = {x1, x 2 , ..., xM} Yk = {y1 , y2 , ..., yK } = {MSP, NPV, ...}

Y0 = {NPV0 , MSP0 , ...}

sensitivity analysis of deterministic model: identify model variables subject to uncertainty that have the highest impact on the model (θj)

3

⎧ dy dy dy ⎫ ⎬ SA = ⎨ , , ..., d x d x d xM ⎭ ⎩ 1 2 the higher is

4

model simulation and propagation of uncertainties to the outcomes

4A

characterize θj by assigning probability distribution functions + generation of N samples with LHS from the uncertainty domain of θj,...,M

dy → the higher is the impact to fx → θ ∈ X dx

⎡ ssi , j ⋯ ssi , M ⎤ ⎢ ⎥ SN × M = ⎢ ⋮ ⋱ ⋮ ⎥ ⎢⎣ ssN , j ⋯ ssN , M ⎥⎦

max f (x , y) = max NPV = min MSP ⎡ yi , k ⋯ yi , K ⎤ ⎢ ⎥ =⎢ ⋮ ⋱ ⋮ ⎥ ⎢ ss ⎥ ⎣ N , k ⋯ yN , K ⎦

4B

deterministic optimization problem solved for each scenario ssi,j → mapping and analysis of solutions (see Supporting Information, section C)

5

economic risk analysis and mitigation

5A

risk-based economic analysis

risk =

5B

risk mitigation strategies

flexible multiproduct network

⎞ 1 t⎟ ⎝ (1 + r ) ⎠

∑ annual cash inflow × ⎜ t=1

i

⎛ $ ⎞ ⎛ ton ⎞ variable operating costs ⎜ ⎟ = raw materialsl ⎜ ⎟ ⎝ year ⎠ ⎝ year ⎠

⎛

T

∑ Pi × Mi = ∑ P(NPV ≤ 0) × (NPVi ) i

annual present value(net present revenue) =

YN × K

⎛ ton ⎞ ⎛ $ ⎞ ⎟ + utilities ⎜ × market pricel ⎜ ⎟ k ⎝ ton ⎠ ⎝ year ⎠

(4)

⎛ ton ⎞ ⎛ $ ⎞ ⎟ + waste disposal ⎜ × market pricek ⎜ ⎟ ⎝ ton ⎠ ⎝ year ⎠

where

⎛ $ ⎞ ⎛ $ ⎞ ⎟ + repairs ⎜ ⎟ × price ⎜ ⎝ ton ⎠ ⎝ ton ⎠ ⎛ $ ⎞ ⎛ $ ⎞ + operating supplies ⎜ ⎟ + royalities ⎜ ⎟ ⎝ year ⎠ ⎝ year ⎠

⎛ 1 ⎞ ⎜ t ⎟ = discount factor ⎝ (1 + r ) ⎠

where k = 1, ..., K (total number of utilities), l = 1, ..., L (total number of raw materials) (for the factorial methodology see Supporting Information, Table B2)

r = discount rate = annual cash inflow (net revenue) = total annual sales − total production cost − income tax − loan payment

(8)

fixed operating costs

(5)

= labor (operators × salaries) + maintenance + property insurance and tax

⎛ ton ⎞ total annual sales = production ratei ⎜ ⎟ ⎝ year ⎠ ⎛ $ ⎞ ⎟ × market pricei ⎜ ⎝ ton ⎠

3. METHODOLOGY A comprehensive economic assessment is performed to characterize the uncertainty and incorporate it in the economic metrics for project evaluation. The economic assessment follows the algorithm presented in Table 3 which briefly includes the following: (step 1) relevant economic metrics for the analysis is defined as well as the required input for the economic model; (step 2) the output from step 1 is used to perform a deterministic analysis of economic metrics for each alternative; (step 3) sensitivity analysis is applied to the results obtained in step 2 which provides a ranking of input data regarding their impact on the model; (step 4) using the most important parameters

(6)

where i = 1, ..., I total annual production cost = variable operating costs + fixed operating costs

(9)

(7) 6804


Article


(see Table 4). A satisfactory design represents a plant that can produce a product which generates a profit. Therefore, by eliminating the alternatives whose objective function values are negative, economically infeasible options are removed. For the calculation of the fixed and variable operating costs, the capital investment (FCI) needs to be identified prior to applying the factorial methodology (Supporting Information, Table B2). The capital investment was calculated based on the factorial methodology (Supporting Information, Table B1) and by dividing the processing routes/pathways into three sections: (1) glycerol separation and purification; (2) conversion step; and (3) downstream processing (DSP). The capital investment for section (1) was calculated, as above-mentioned, based on a medium-sized biodiesel plant in Europe,13 providing approximately 10 kton glycerol/year, corresponding to 10% (w/w) of the total annual biodiesel production rate (∼100 kton/year). The detailed equipment and their cost in $2003 and the updated costs for $2014, as well as the purchased, installed capital investment and total fixed capital investment can be found in the Supporting Information, section B. The total purchased capital investment for section (1) is 1.61 M$, including the utilities system (Supporting Information, section B). For sections (2) and (3), recovery and purification of products, the capital investment was calculated by a careful and independent calculation for all the products being produced based on a literature review, scaled to the production rate set by the verified stoichiometry. Table 4 presents the purchased equipment costs, the data sources and the remaining parts (calculated based on the factorial methodology). The purchased equipment costs were calculated based on the product specific references (base) and adapted to appropriate capacities (plant X), the purchase costs were updated to year 2014, as shown in eq 10.30,46 The seven-tenths rule46 was used as exponent (n in eq 10) for scale up/down for all equipment except for the fermenters, for which an exponent of 0.75 was used.47

identified in step 3, an uncertainty analysis is performed on the economic metrics by first performing a Latin Hypercube Sampling (LHS) of input uncertainty domain and then solving an optimization problem for each LH sample (so-called uncertainty mapping); (step 5) based on the uncertainty mapping, the top three product candidates are selected for further economic risk analysis, and further risk mitigation strategies are proposed and evaluated. The application of these steps are further detailed and analyzed in the results section below.

4. RESULTS Step 1: Identify Economic Model and Input Model Variables. As presented by the algorithm in Table 3, Step 1, instructs to select the economic model and metrics on which the projects will be evaluated, they are referred to as Yk, where we can therefore identify the input variables to the model (Xi). As abovementioned, the economic model used in this work is the discounted cash-flow rate of return (DCFR). For the calculation of NPV, the question is which interest rate (represented by r in eq 2) is to be used to discount the cash-flows. This interest rate is usually set by the investors or management department and it represents the minimum rate of return that the company is willing to accept for a new investment (company’s adjustment to risk). On the other hand, the user can also estimate the internal rate of return (IRR, which corresponds to the discount rate when NPV is set to zero) or minimum selling price (MSP) for which the project would break-even (rearranging eq 2) by using the goal-seek function and setting the NPV to zero, while maintaining all the other variables constant. This analysis is performed by using the NREL excel file (http://www.nrel.gov/ extranet/biorefinery/aspen_models) for the discounted cashflow rate of return calculation after adapting it with the appropriate model assumptions for the present case study (please see Table 2). To this end, the assumptions described in Table 3 together with the mean value of the remaining input variables are used to establish the base case (for the mean raw material cost and products sell price, see Table 4).

costX

products

mean 2014 ($/kg)

std.

data points ref

0.368 2.010 0.707 1.55836 0.536 1.66236 1.59036 4.510,36 2.042 1.52436 2.02 2.029

0.048 0.23 0.137 0.421 0.138 0.28 0.131 0.197 0.041 0.12 0.35 0.27

35 36 23, 33 37 38 39 40 41 43 36 36, 44 45

⎡ ⎛ capacity ⎞n⎤ CEPCI 2014 X ⎟ ⎥ = ⎢costbase × ⎜⎜ ⎟ ⎥ × CEPCI ⎢⎣ capacity ⎝ base base ⎠ ⎦ (10)

costX2014

where represents the purchased cost for plant X calculated from the costbase of a similar plant with similar functionality; CEPCI2014 and CEPCIbase represent the Chemical Engineering’s Plant Cost Index (see Supporting Information, Figure B1) for 2014 and for the year base (reference year for each case), respectively. It is important to stress that, as this work focused on systematically screening alternatives for glycerolbased bioproducts and their comprehensive uncertainty analysis at an early stage design, rigorous simulations of shortlisted process candidates are considered to be out of scope. It would be natural, however, at a later stage of the process development life cycle to perform more detailed process simulations for process design optimization and refinement of the cost estimation. The NPV and MSP estimates based on the nominal values are shown in Table 5. A first ranking of alternatives can be obtained; however, as only the production of lactic acid presents a positive NPV, further analysis is needed especially on the high impact variables that might be subject to high degrees of uncertainty. It is important to mention that, like the production of ethanol, isobutyl alcohol, and n-butanol, the production of H2 in the current project’s setting is not profitable. Its market price should increase at least to its MSP level (please see Table 5) in order for the project to break-even. However, special attention should be

Table 4. Product Prices and Respective Standard Deviations crude glycerol (60% w/w) succinic acid (HSuc) ethanol (EtOH) n-butanol H2 1,2-propanediol (1,2PDO) propionic acid polyhydroxybutyrate (PHB) lactic acid (HLac) isobutyl alcohol 1,3-propanediol (1,3-PDO) acrolein

2014

Step 2: Deterministic Solution. In this step, the model is initialized to generate the base case and to estimate the first ranking of solutions, therefore the economic metrics are calculated for the base case scenario. The problem is formulated and solved by estimating the Net Present Value (eq 2 to 9) for each of the processing networks (alternatives) using the nominal/average market prices of product and raw material 6805


Article


Table 5. Estimation of the Purchased Capital Cost to Be Used for the Factorial Methodology (Updated to 2014 Prices by Using the Appropriate CEPCI) and Remaining Needed Parameters. NPV and MSP Are Estimated at the Conditions Reported in Table 2

purchased capital cost, E′ (MM$)

fixed capital investment, FCIa (MM$)

total product cost without depreciation, TPCa ($/kg)

utilities (MM$/y)

11.21748

35.221

1.625

0.956

22.278 24.817

2.045 1.369

45.903 16.280 26.161 21.107 21.627 22.615 30.022 27.314

a

succinic acid 1,3 PDO propionic acid PHB 1,2 PDO lactic acid isobutyl alcohol H2 n-butanol ethanol acrolein a

[ref]

5.34744 4.74740,48 15.02049 4.71320 4.92948 6.70522 5.55125 5.20250 8.22723,33 4.92729,51

sales (MM$/y)

average product price ($/kg)

NPV @ 10 (MM $)

MSP @10 ($/kg)

19.4

2.00

−2.72

2.18

1.971 0.474

10.6 10.3

2.02 1.59

−28.1 −16.4

2.71 1.94

3.643 1.421 1.218 1.471

1.948 2.848 0.529 0.921

15.4 12.2 15.1 11.5

4.50 1.66 2.00 1.524

−26.0 −6.0 11.5 −22.6

5.59 1.79 1.74 1.91

1.445 3.172 1.708 3.133

1.666 1.434 0.551 0.761

3.4 4.4 3.5 9.3

0.536 1.558 0.707 2.00

−75.9 −66.4 −76.6 −79.1

1.96 4.37 2.59 4.07

a

It includes the separation and purification of crude glycerol.

illustrative example, the sensitivity analysis is performed for the 1,2-PDO production from glycerol and it presented in Table 6 and illustrated in Figure 2. The results of the economic indicator NPV, when subjected to the variation of the input parameters for the DCFR model are presented in Figure 1. The data categories are ordered so that the level of effect on the economic indicators decreases from top to bottom; that is, the economic indicators on top of these figures correspond to the parameters that have relatively higher importance on the model. Product price, feedstock price, and capital investment have been previously identified as important and common sources of uncertainty on the economic performance across different biorefinery types.58−60 The effect of economies of scale can be perceived by the fact that increasing the plant capacity 20% has less impact on the NPV than the opposite decrease of 20%. It means that the fixed capital investment has a nonlinear relationship with the NPV; the higher is the plant capacity the less impact has the fixed capital investment on the NPV. The same effect is observed with the relationship for the total annual production cost versus the NPV. As a complementary scenario analysis in order to investigate further the yearly impact of the discount rate on the economic model, Figure 2 shows the cumulative discounted cash-flows using the deterministic model varying only the discount rate. It shows the discounted cash flow diagram generated using the deterministic model based on the 10 year Modified Accelerated Cost Recovery System (MACRS) depreciation method (see Table 2, and more details in Supporting Information section B). The blue line represents the DCFR for a discount rate of 10% corresponding to a low risk project; and the red line represents the DCFR for a 24% discount rate, corresponding to a medium/ high risk project (see Table 7). Analyzing Figure 2, one can see that at the 10 years’ mark (depreciation period) the project at both discount rates starts to be profitable (NPV > 0); however, at 24% the profit is barely visible since the investors set such a high rate of return to adjust for the possible risk the investment faces. The impact of the selected discount rate on the economic performance can be further analyzed by estimating the product minimum selling price (MSP) obtained by setting NPV to zero, for a constant discount rate. The MSP for the 1,2-propanediol production at the conservative discount rate of 24% is 2.05 $/kg, whereas for a discount rate of 10%, the MSP of 1,2-PDO is 1.79

payed to H2 due to the fact that it is not straightforwardly transported or stored. Therefore, if along with the manufacture of high value-added product(s), synergies were to be explored and these fuels were to be produced and consumed inside the running plant (fulfilling part or all of its energetic needs), it may potentially result in a profitable project. Hence, in future work, different business plans together with supply chain models will be analyzed including the assessment of potential synergies through product portfolio combinations. Step 3A: Deterministic Sensitivity Analysis. The investment and cash-flows are first calculated for the baseline conditions, and then the NPV is re-estimated by changing one variable at a time over the expected range of variability (presented in Table 6). This will express how sensitive the DCFR model is to variations in the input information. Therefore, this sensitivity analysis provides an idea about the degree of risk involved in foreseeing the economic performance of the project and also indicates which input data have the highest impact. As an Table 6. Key Economic Factors under Variability for Sensitivity Analysis of the DCFR Model data sources product price feedstock price fixed capital investment discount rate plant lifetime loan interest construction time income tax rate total annual production cost (TPC) plant capacity

lower limit (% of the baseline)

base case (1,2-PDO)

upper limit (% of the baseline)

52 52 53,54

−20% −10% −20%

1.662 $/kg 0.368 16.280 MM$

+20% +30% +50%

52,55, this work this work this work this work 56 57, this work

8%

10%

24%

10 years

20 years

30 years

4%

5%

8%

1 year

2 years

3 years

−20% −20%

35% 1.421 $/kg

+20% +20%

this work

−20%

7740

+20%

6806


Article


Figure 1. Sensitivity analysis of NPV to variations in key economic parameters for the production of 1,2-PDO. Dark gray represents the negative impact on the NPV and the light gray represents the positive impact on the NPV.

Figure 2. Cumulative discounted cash flow for 1,2-PDO production. Baseline assumptions + varying the internal rate of return for the investment: the red line represents 10% discount rate and the blue line represents 24% discount rate.

There are several methods to get price forecasts, such as the ones discussed in (Towler and Sinnott, 2013);52 however, the consistent price information is a continuous challenge in this field/community, since companies cannot provide prices. Therefore, in this work we have chosen to use historical price data and use this to define a scenario in which future price forecasts are sampled from historical price distributions. The historical trends of the products under consideration have been surveyed in the period of past 10 years and used to construct the historical price distributions for each product (please see Table 4). An example of the historical data price collection and their fitting through distribution functions, which will be used as input for the following steps, is presented in Figure 3, for the crude glycerol and 1,2-PDO. The same procedure was performed for the remaining 10 products being considered here. Through LHS with correlation structure/control61 (see Supporting Information, Table D1) 500 future scenarios were generated for each of the input parameters. The remaining relevant sources of uncertainty were analyzed through scenariobased analysis in Step 4B. Step 4B: Uncertainty Mapping and Analysis of Solutions. In this step, a deterministic optimization problem is solved for each of the scenarios generated by Monte Carlo sampling performed in the previous step, where the con-

Table 7. Suggested Discount Rate According to Levels of Risk Per Type of Investment30 type of investment basissafe corporate investment new capacity for established corporate market position new product entering into established market, or new process technology new product or process in a new application Everything new, high R&D and market effort

risk level

r(%/year)

safe low

4−8 8−16

medium

16−24

high very high

24−32 32−48+

$/kg (for the MSP of the remaining products estimated at discount rates of 10% and 24% please see Table 5). Step 4A: Uncertainty characterization and sampling from input uncertainty domain. Several sources of uncertainty on the economic model predictions were pointed out through deterministic sensitivity analysis, performed in Step 3A. Therefore, the base case scenario considered for further analysis was built on (i) historical data on the raw materials and products market prices described through appropriate probability distribution functions, and (ii) considering variability of the fixed capital investment over its typical range of variation (represented by uniform distribution varying between −20 to +50%). 6807


Article


Figure 3. Historical price data collection (left) and their fitting through normal probability distribution functions (right). The 1st and 2nd row correspond to 1,2-PDO,36,39 and crude glycerol,35 respectively.

data corresponding to those uncertainty realizations. It displays high-dimensional data sets, where each y axis (“columns”) represents one variable; there are 500 values for each y, corresponding to the 500 LH samples. These values are then joined, creating multiple polylines that represent the scenarios across variables. Variables (“columns”) were normalized to a fixed range (0−100) which is equivalent to working with standardized variables (to avoid the influence of one variable onto the others due to scaling). Furthermore, to take the uncertainties in a proactive way and further understand its impact on the ranking/mapping of solutions, three extra scenarios were built based on different

sequences of the data uncertainty on the decision-making problem are studied and mapped. The result comprises a distribution of 500 optimal processing networks that are mapped and statically analyzed. The frequency of selection for each scenario is presented in Figure 4. The top-3 ranking of alternatives are lactic acid, succinic acid, and 1,2-PDO. Figure 5 presents the parallel coordinate plot of the system under uncertainty. It allows a quick quantitative and qualitative visualization of results obtained from the optimization problem for each scenario generated by the Monte Carlo simulations, in this way depicting how input uncertainties affect the optimal decision. Hence, it presents the optimization results and the raw 6808


Article


Further, if future developments take the prices of the top selected products identified in the base case scenario (scenario (ii), to decrease to 80% of its average price, then the 1,2-PDO is selected by 34% of the samples, followed by lactic acid which is selected with 26%, 1,3-PDO, and succinic acid with 16 and 13%, respectively. Finally, scenario (iii) is built on the sales volume decreasing to 80% of its average value. In this case, lactic acid is selected by 56% of the samples, followed by succinic acid and 1,2PDO which is selected with 18% and 12%, respectively (please find the histograms corresponding to scenarios (i) to (iii) in the Supporting Information, section D, Figures D1 to D3). Therefore, as expected from the deterministic sensitivity analysis performed in Step 3A, the results from scenario (ii) show that the ranking of solutions is highly dependent on the products’ selling price. However, within the price scenario established (for the current and referenced price data, Table 4), the ranking of solutions is robust, which can be interpreted from the comparison of the rankings obtained in scenarios (i) and (iii) to the ranking obtained on base case scenario. Step 5A: Economic Risk Quantification. As mentioned earlier, risk is defined as the probability of occurrence times the consequence of that same event to occur. In this work, the NPV is used as economic metric/indicator for project evaluation in order to support a risk-aware decision making. To this end, the risk is given by the probability of failing to achieve the targeted NPV (“being lower or equal to”) times the magnitude of the consequence of that event occurring (“loss of profit”). The economic risk is defined by the probability of the project being nonprofitable (NPV < 0) times the consequence (loss of profit); its respective mathematical description is as follows.

Figure 4. Network frequency of selection, where the top-3 selected are lactic acid, succinic acid, and 1,2-PDO.

assumptions of future uncertainties through the realization of Monte Carlo simulations. It is important to note that these scenarios were built on the base case in which uniform distribution of capital investment and historical price data were used to describe preidentified uncertainties. The scenarios are as follows: (i) the total production costs varies normally with 20% standard deviation, (ii) prices of top products selected crash to 80% of its average price, and, (iii) total sales volume decreases to 80% of its normal volume. The total production cost, essential for the estimation of the economic metrics, can vary for several reasons, such as raw materials, chemicals, utilities, and solvents prices variation, as well labor and maintenance, among others. Therefore, scenario (i) considers the total production cost of each alternative in the base case to vary with normal distribution within its probable range of variation. The realization of uncertainties through the optimization problem is quantified and, in this scenario, lactic acid and succinic acid lead the top with 38% and 30%, followed by PHB and 1,2-PDO, with 16% and 10%, respectively.

risk =

∑ Pi × Mi = ∑ P(NPV ≤ 0) × (NPV)i i

i

(11)

where i is the occurrence of the undesirable event, Pi is the probability of that event to occur, and Mi is the magnitude of the consequence (in MM$) of the undesirable event. It is important to note that, in this study, the overall project to be undertaken represents the implementation of new technologies, and therefore the project discount rate is adjusted to

Figure 5. Parallel coordinate plot of the system under uncertainty. Orange, blue, and light blue represent the top-4 best solutions: lactic acid, succinic acid, 1,2-PDO and PHB. The yellow lines represent the remaining selected concepts. 6809


Article


Figure 6. Cumulative distribution function for the lactic acid, succinic acid, and 1,2-PDO production from glycerol. The highlighted area represents the risk of the project being nonprofitable. Blue represents NPV obtained for a discount rate of 10%. Orange represents NPV obtained for a discount rate of 24%.

investment, and the low variable operating costs when comparing to the remaining concepts. Various studies on glycerol conversion have been performed along the years;24,29,62,63 a fewer number have considered techno-economic indicators, but noteworthy is that, to the best of our knowledge, up to now there are no studies on glycerol valorization concepts that are subjected to uncertainties on the input data. Vlysidis et al.63 analyzed the coproduction of succinic acid to add value to the glycerol side stream of the biodiesel industry, where the coproduction of succinic acid is forecasted to be a positive and potentially profitable solution to valorize glycerol. Posada (2011)24 also performed a deterministic analysis of possible schemes for the conversion of glycerol to value-added products, assessed only based on the production cost. Both studies, being deterministic assessments for early stage screening and design of biorefinery concepts, present some shortcomings related to the possible lack of awareness of market fluctuations and technical variabilities. They embody a substantial risk concerning the selection of optimal process concepts, and so, long-term competitiveness and robustness cannot be guaranteed due to the variability, scarcity, and uncertainty of input information. Therefore, in this work, a systematic study is performed by incorporating stochastic uncertainty in the technoeconomic analysis, according to which the most suitable set of process concepts are identified. The input uncertainty on the feedstock and product prices was depicted based on the empirical historical data using appropriate probability distribution functions. The fixed capital investment variability was described by a uniform distribution of values around their typical predicted range of variation. Then, the input uncertainty was propagated to the objective function though Monte Carlo simulation enhanced with LHS. The obtained uncertainty mapping of solutions pointed to first lactic acid and second succinic acid as the best potential alternatives regarding economic criteria (positive and high NPV), followed by 1,2PDO. To further analyze the robustness of the ranking of alternatives under uncertainties, three additional scenarios were tested. As expected from the deterministic sensitivity analysis performed in Step 3A, the ranking of solutions is robust. This can

offset risk and attract investors, therefore considered to be somewhere between medium to high, and so the minimum acceptable rate of return is set to be 24% (see Table 7). However, to also depict the effect of the company choices with respect to the level of discount rate used, the economic evaluation is here performed considering both discount rates of 10% and 24%. The cumulative distribution function of NPV is presented in Figure 6. The importance of incorporating cost uncertainties in the NPV calculations through stochastic modeling has also been explored/demonstrated in ref 12 for analysis of biorefinery concepts. According to eq 11, the calculation of risk is equal to the shaded area in the cumulative distribution function for NPV shown in Figure 6, where two curves are depicted in red and blue, representing NPV obtained at 10% and 24% internal rate of return, respectively. A summary of results is presented in Table 8. Table 8. Summary of Results for the Calculation of Economic Risk for the Top-3: Lactic Acid, Succinic Acid, and 1,2-PDO

frequency of selection Pr (NPV ≤ 0) @10% Pr (NPV ≤ 0) @24% risk @10% (MM$) risk @24% (MM$) net present revenue @10% risk @10%/net present revenue @10%

lactic acid

succinic acid

1,2-PDO

216/500 0.632 0.99 8.69 19.21 26.4 0.3291

138/500 0.764 0.962 13.74 23.86 17.4 0.787

60/500 0.68 0.89 15.45 16.43 6.43 2.403

5. DISCUSSION A first ranking of solutions is given by solving the deterministic problem by maximizing the potential profit of the biorefinery concepts considering no variability on the input data (“single point” analysis). It was found that the lactic acid, succinic acid, and 1,2-PDO are the top-3 ranked solutions when only using NPV nominal data. This is justified based on the balance between the products selling price and production capacity, fixed capital 6810


Article


On the other hand, succinic acid has a potential loss of 14 MM $, which represents around 80% of the potential net present revenue. Finally, based on the current data, 1,2-PDO has a high potential risk of profit loss of 16 MM$, representing a risk which is 2.4 times higher than the potential present revenue. To decrease the risk associated with the mentioned concepts, risk mitigation strategies can be applied. In this work, the flexible multiproduct biorefinery is tested as a risk reduction strategy, and its discussion follows in the next section. Step 5B: Flexible Multiproduct Glycerol-Based Biorefinery Concepts. In this step, the aim is to identify the optimal trade-off between the operating flexibility and capital investment according to different uncertainty realizations of product prices. Therefore, as a risk mitigation strategy, a possible multiproduct network (comb) was tested where lactic acid and succinic acid are produced depending on the market environment. Moreover, as previously mentioned, based on the large ratio risk/net present revenue for the production of 1,2-PDO, its coallocation combination with the production of lactic acid is economically infeasible. Following the same principles as the economic risk estimated in step 5A, the risk was estimated for comb and it is reported in Table 9. There is a reduction in the

be concluded from the comparison of the rankings obtained in scenarios (i) and (iii) to the ranking obtained on the base case scenario which remains the same. The effect of the fixed capital investment variability was verified to have a significant impact on the project profitability and, consequentially, it substantially influences the optimal solution (decision-making) at the concept screening phase where a broad range of uncertainty is expected. Furthermore, the findings of this study pointing to lactic acid as the most suitable and robust candidate for the valorization of glycerol is corroborated by the existing commercial plants such as Corbion Purac, Natureworks, and Galactic, among others, which are manufacturing lactic acid by fermentation of carbohydrates.64,65 Lactic acid is a bulk chemical and it has been reported to have an average annual growth rate of 10%,64 with and a long history of uses in the food, cosmetic, and pharmaceutical sectors, and as a monomer for the production of biodegradable PLA (polylactic acid).64 The high growth rate of lactic acid lies fundamentally on the demand of PLA, the biodegradable polymer. It has approximately a demand of 25 000 ton/year (2013) and can reach up to 650 000 ton/year in 2025.64 On the other hand, succinic acid was also identified as a potential candidate for the glycerol valorization. This is also supported by commercial production plants such as Reverdia producing 10 kton/y of biosuccinium from renewable carbon sources (fermentable sugars) in Italy, since 2012;66 Succinity, a joint venture between Corbion Purac and BASF established in 2013, also operating a 10kton/y plant in Spain since 2013; and Myriant and Bioamber, with the same range of installed capacities, operating in Canada and North America, respectively.65 Succinic acid (and its salts) has been predicted to be one of the future platform chemicals that can be derived from renewable resources and has significant potential as a building block; for the production of biopolymers,67 to be used directly in food, pharmaceutical, and cosmetic industries as well as its conversion into solvents and other current petro-based chemicals.14 Finally, 1,2-PDO, having several applications in industries, food, and pharmaceuticals, is produced in industrial and USP grades, respectively. The world leader of biobased 1,2PDO production is ADM, followed by Oleon, manufacturing it from alternative carbohydrate sources such as glycerol, sorbitol, and dextrose. Moreover, as mentioned in above sections, several reports and review articles and reports such as refs 65, 64, and 68 have continuously identified lactic acid, succinic acid, and 1,2-PDO, among others, reporting them as having strong market growth and highlighting their potential to be used as platform chemicals and their suitability as building blocks for the synthesis of a variety of chemicals through chemical conversion. Therefore, in light of this, a deeper analysis was undertaken into the top-3 concepts through economic risk analysis, that is, the probability of NPV being lower than zero times the consequence of that event occurring. In other words, the economic risk will reflect the probability of the biorefinery concept under study to be nonprofitable which corresponds to a NPV lower than zero. To this end, the glycerol valorization through the production of lactic acid, succinic acid, and 1,2propanediol were compared based on the quantified economic risk of being nonprofitable. It was found that the lactic acid production from glycerol has a potential risk of profit loss of approximately 9 MM$ (estimated at discount rate of 10%), over a net present revenue of 26 MM$ (20 years of plant lifetime). In this case, the net present revenue, over 20 years of plant lifetime, is approximately 3 times the potential profit loss risk.

Table 9. Quantified Risk for Single-Product and Multiproduct Combinations (Estimated at Discount Rate of 10%) combination

risk (MM$) @ 10%

risk @10%/net revenue

lactic acid succinic acid comb → lactic acid + succinic acid

8.69 13.74 7.03

0.33 0.79 0.28

quantified risk by about 20%, and consequentially the net present revenue to risk ratio has improved to be approximately 3.6 times the potential risk. Therefore, it indicates that a flexible multiproduct network for the valorization of glycerol through the coallocation of lactic acid and succinic acid is a promising solution to mitigate the effect of uncertainties in the technoeconomic metrics. The multiproduct network allows the products to be, in turn manufactured depending on favorable market demands and prices. A plant capable of switching between final products has advantages since it is then able to guarantee that it is continuously using the pathway that optimizes the profitability, considering the market demands, market prices, and their corresponding volatilities. By mimicking the economic success story of the oil-based industry, a multiproduct plant would be able to reduce the risk and therefore yield a more sustainable and competitive biobased industry. Finally, the results support the argument that early stage design of processes (concept screening phase) are significantly affected by uncertainties due to the scarcity and variability of the input data needed. Nevertheless, if reasonable approximations are used to describe the techno-economic reality and if carefully considering the potential (and probable) uncertainties, the methodology is still a powerful tool for screening out unfeasible process-product pathways and, in this way, improve the effectiveness of decision making for screening of process concepts.

6. CONCLUSIONS The growing market for biodiesel is significantly changing the cost and availability of glycerol. Biodiesel manufacturers having little economic encouragement to purify the crude glycerol, it is 6811



■

critical to develop processes to explore its potential and make use of it in an efficient way. In the future, if supplies of biodiesel continue to expand, glycerol will develop into a flexible building block for the production of high-value-added products as an addon to the biodiesel plant within an integrated biorefinery concept. This research studied possible pathways for the valueadded conversion of crude glycerol. Therefore, by following a systematic methodology for identification and screening of early stage concepts for glycerol valorization, 11 possible routes/ technologies for the production of value-added chemicals are assessed. The input uncertainties, such as product and glycerol market prices and fixed capital investment, were propagated to the model outcomes (NPV and MSP). Product historical market prices were taken to depict possible future scenarios; however, as the methodology used is systematic, it can be iterated and solved for different price scenarios such as one where the market prices are affected by increasing the availability of these products in the market, or any other price forecasting approach as the ones proposed in (Towler and Sinnott, 2013).52 Finally, the results support the development of an integrated multiproduct concept comb through the production coallocation of lactic acid and succinic acid according to the market environment. This seems to be a promising solution for the production of chemicals from the biodiesel byproduct, potentially adding value and enhancing the bioindustry’s overall robustness and sustainability. Furthermore, the methodology used in this work for technoeconomic assessment under uncertainty is systematic and flexible, since the database generation step allows it to be further expanded and updated appropriately. As this work focuses on technology assessment using a fixed design capacity corresponding to the glycerol feedstock, in future work, the methodology in the future will be applied to supply chain studies which will enable the analysis of (i) different business plans (e.g., shipment of glycerol to where it is needed in order to manufacture a valuable product on site instead of building a new plant); and, (ii) potential synergies through diverse product portfolio combinations. As such, the methodology can be used in wider range of applications as guiding tool for informed decision making and robust evaluation of process concepts.

■

REFERENCES

(1) Bozell, J. J.; Petersen, G. R. Cutting-edge research for a greener sustainable future Technology development for the production of biobased products from biorefinery carbohydrates−the US Department of Energy ’ s “ Top 10 ” revisited. Green Chem. 2010, 12 (4), No. b922014c. (2) Scarlat, N.; Dallemand, J. F.; Monforti-Ferrario, F.; Nita, V. The role of biomass and bioenergy in a future bioeconomy: Policies and facts. Environ. Dev. 2015, 15, 3−34. (3) Weissermel, K, Arpe, H. J.. Industrial Organic Chemistry, 4th ed.; Wiley-VCH: Weinheim, Germany; 2003. (4) Yue, D.; You, F.; Snyder, S. W. Biomass-to-bioenergy and biofuel supply chain optimization: Overview, key issues and challenges. Comput. Chem. Eng. 2014, 66, 36−56. (5) Turton, R. Analysis, Synthesis and Design of Chemical Processes; Prentice Hall: 2009. (6) Cheali, P.; Quaglia, A.; Gernaey, K. V.; Sin, G. Effect of Market Price Uncertainties on the Design of Optimal Biore fi nery Systems - A Systematic Approach. Ind. Eng. Chem. Res. 2014, 53 (14), 6021−6032. (7) Foley, T., Thornton, K., Hinrichs-rahlwes, R., et al. Renewables 2015 Global Status Report; REN21: 2015; http://www.ren21.net/wpcontent/uploads/2015/07/REN12-GSR2015_Onlinebook_low_ nolinks.pdf. (8) Yang, F.; Hanna, M a; Sun, R. Value-added uses for crude glycerol– a byproduct of biodiesel production. Biotechnol. Biofuels 2012, 5 (1), 13. (9) Jong, E D. Bio-based Chemicals Value Added Products from Biorefineries; IEA Bioenergy-Task: 2011; http://www.qibebt.ac.cn/ xwzx/kydt/201202/P020120223409482956847.pdf. (10) Koutinas, A a; Vlysidis, A.; Pleissner, D.; et al. Valorization of industrial waste and by-product streams via fermentation for the production of chemicals and biopolymers. Chem. Soc. Rev. 2014, 43 (8), 2587−2627. (11) Reinhardt, G., Gärtner, S., Häfele, S., et al. Sustainable and Integrated Production of Liquid Biofuels, Green Chemicals and Bioenergy from Glycerol in Biorefineries: Environmental Assessment; IFEU, Heidelberg, Germany; 2012. (12) EuroBioref. EuroBioRef Public Final Booklet; 2014. DOI:10.13140/RG.2.1.2450.5687. (13) Green Business. Factbox: Biodiesel plants in the EU. http://www. reuters.com/article/us-biofuels-europe-biodieselidUSTRE6172JY20100208, 2010 (accessed January 1, 2015). (14) Scholten, E.; Renz, T.; Thomas, J. Continuous cultivation approach for fermentative succinic acid production from crude glycerol by Basfia succiniciproducens DD1. Biotechnol. Lett. 2009, 31 (12), 1947−1951. (15) Zeng, A.-P.; Biebl, H. Bulk chemicals from biotechnology: the case of 1,3-propanediol production and the new trends. Adv. Biochem Eng. Biotechnol. 2002, 74, 239−259. (16) Nakamura, C. E.; Whited, G. M. Metabolic engineering for the microbial production of 1,3-propanediol. Curr. Opin. Biotechnol. 2003, 14 (5), 454−459. (17) Barbirato, F.; Chedaille, D.; Bories, a. Propionic acid fermentation from glycerol: Comparison with conventional substrates. Appl. Microbiol. Biotechnol. 1997, 47 (4), 441−446. (18) Cavalheiro, J. M. B. T.; de Almeida, M. C.; Grandfils, C.; da Fonseca, M. M. R. Poly(3-hydroxybutyrate) production by Cupriavidus necator using waste glycerol. Process Biochem. 2009, 44 (5), 509−515. (19) Bricker, M. L., Le, L. Methods for Converting Glycerol to Propanol. US Patent US8101807 B22012; http://www.google.com/patents/. (20) Cabaniss, S., Park, D., Silvinsky, M., Wagoner, J., Chen, D., You, F. Design G2. Northwestern University Chemical Process Design Open Textbook; Northwestern University: 2014. https://processdesign. mccormick.northwestern.edu/index.php/Design_G2. (21) Posada, J.; Cardona, C.; Gonzalez, R. Analysis of the production process of optically pure D-lactic acid from raw glycerol using engineered Escherichia coli strains. Appl. Biochem. Biotechnol. 2012, 166 (3), 680−699.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.iecr.5b04593. Description of process alternatives; economic model and assumptions, optimization formulation for the deterministic and stochastic problem; uncertainty characterization (PDF)

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank the Reneseng Project, Marie Curie Initial Networks (ITN), Grant Agreement 607415 for the financial support. 6812


Article

Industrial & Engineering Chemistry Research (22) Bauer, F.; Hulteberg, C. Isobutyl alcohol from glycerine − A techno-economic evaluation of a new biofuel production process. Appl. Energy 2014, 122, 261−268. (23) Tao, L.; Tan, E. C. D.; McCormick, R.; et al. Techno-economic analysis and life-cycle assessment of cellulosic isobutyl alcohol and comparison with cellulosic ethanol and n-butanol. Biofuels, Bioprod. Biorefin. 2014, 6 (3), 246−256. (24) Posada Duque, J. A. Design and Analysis of Technological Schemes for Glycerol Conversion to Added Value Products. Ph.D. Thesis, Universidad Nacional de Colombia, 2011 (25) Zhang, Y.; Brown, T. R.; Hu, G.; Brown, R. C. Comparative techno-economic analysis of biohydrogen production via bio-oil gasification and bio-oil reforming. Biomass Bioenergy 2013, 51, 99−108. (26) Taconi, K. A.; Venkataramanan, K. P.; Johnson, D. T. Growth and solvent production by Clostridium pasteurianum ATCC® 6013 utilizing biodiesel-derived crude glycerol as the sole carbon source. Environ. Prog. Sustainable Energy 2009, 28 (1), 100−110. (27) Schindler, B. D.Understanding and Improving Respiratory Succinate Production. Ph.D. Dissertation, Michigan State University, 2011 (28) Shams Yazdani, S.; Gonzalez, R. Engineering Escherichia coli for the efficient conversion of glycerol to ethanol and co-products. Metab. Eng. 2008, 10 (6), 340−351. (29) Liu, L.; Ye, X. P.; Bozell, J. J. A comparative review of petroleumbased and bio-based acrolein production. ChemSusChem 2012, 5 (7), 1162−1180. (30) Peters, M. S., Timmerhaus, K. D., West, R. E. Plant Design and Economics for Chemical Engineers; McGraw-Hill; 2003. (31) Towler, G, Sinnott, R. Economic Evaluation of Projects. In Chemical Engineering Design, 2nd ed.; Elsevier Ltd: 2013;389−429. DOI:http://dx.doi.org/10.1016/B978-0-08-096659-5.00009-2. (32) Short, W, Packey, D. J., Holt, T. A Manual for the Economic Evaluation of Energy Efficiency and Renewable Energy Technologies; No. DE-AC36-83CH10093; National Renewable Energy Laboratory, US Department of Energy: CO, 1995. (33) Humbird, D., Davis, R., Tao, L., et al. Process Design and Economics for Biochemical Conversion of Lignocellulosic Biomass to Ethanol; NREL/ TP-5100-47764; National Renewable Energy Laboratory: Golden, Colorado, 2011. (34) Waldron, K. W. Advances in Biorefineries: Biomass and Waste Supply Chain Exploitation; Woodhead Publishing: 2014. (35) ICIS. Price HistoryCrude Glycerol, 25 August 2010 to 25 August 2015. ICS Dashboard; http://www.icis.com/about/dashboard/. (36) Davies, P. Biomaterials and Intermediates Including Biobased Biopolymers and Their Petrochemical Equivalent. Chemical Business Focus 2014; 8. (37) n-Butanol; Techno-orbichem., 2013; http://www.orbichem.com/ userfiles/CNF%20Samples/nbt_13_11.pdf. (38) Natural Gas price−EIA. http://www.eia.gov/dnav/ng/hist/ n3035us3m.htm (accessed August 5, 2015). (39) Techno-orbichem. Techno-orbichem: monopropylene glycol. http://www.orbichem.com/userfiles/CNF%20Samples/mpg_13_11. pdf. (40) Tufvesson, P.; Ekman, A.; Sardari, R. R. R.; Engdahl, K.; Tufvesson, L. Economic and environmental assessment of propionic acid production by fermentation using different renewable raw materials. Bioresour. Technol. 2013, 149, 556−564. (41) Roland-Holst, D., Triolo, R., Heft-Neal, S., Bayrami, B. Bioplastics in California: Economic assessment of market conditions for PHA/PHB bioplastics produced from waste methane; CalRecycle: 2013; http://www. calrecycle.ca.gov/publications/Documents/1469\20131469.pdf. (42) Liew, T.; King, L.; Haur, C. J.; Yu, L. Overview of Poly(Lactic Acid) production with oil palm biomass as potential feedstock. Int. J. Eng. Appl. Sci. 2014, 5 (04), 1. (43) Investegate: PLA. http://www.investegate.co.uk/Compdata. aspx?PLA (accessed October 20, 2015). (44) Molel, E., Phillips, H., Smith, A. 1,3-Propanediol from Crude Glycerol. University of Pennsylvania: 2015. (45) Supplementary Charts. Chemical Business Focus; Technoorbichem.: http://www.orbichem.com/userfiles/

CBF%20Supplementary%20Tables%20&%20Charts/ma_sup_c.pdf (Accessed August 2015). (46) Ereev, S. Y., Patel, M. K. Standardized cost estimation for new technologies (SCENT) - methodology and tool. J. Bus Chem. 2012;9 (1). (47) Stanbury, P. F. Principles of Fermentation Technology; Pergamon: 1995. (48) Vaswani, S. Process Economics Program; IHS: California; 2014. (49) Choi, J.; Lee, S. Y. Bioprocess Eng. 1997, 17, 335−342. (50) Nilsson, R.; Bauer, F.; Mesfun, S.; et al. Techno-economics of carbon preserving butanol production using a combined fermentative and catalytic approach. Bioresour. Technol. 2014, 161, 263−269. (51) Banu, I.; Guta, G.; Bildea, C. S.; Bozga, G. Design and performance of a plant for glycerol conversion to acrolein. Environ. Eng. Manage. J. 2015, 14 (3), 509−517. (52) Towler, G, Sinnott, R. Chemical Engineering Design: Principles, Practice and Economics of Plant and Process Design, 2nd ed.; Elsevier: 2013. (53) Christensen, P., Dysert, L. R. Cost Estimate Classification System As Applied in Engineering, Procurement, and Construction for the Process Industries; AACE: 2005. (54) Cheali, P., Gernaey, K., Sin, G. Uncertainties in early-stage capital cost estimation of process designA case study on biorefinery design. Front. Energy Res. 2015;3(3). DOI:10.3389/fenrg.2015.00003. (55) El-Temtamy, S a.; Gendy, T. S. Economic evaluation and sensitivity analysis of some fuel oil upgrading processes. Egypt. J. Pet. 2014, 23 (4), 397−407. (56) Nguyen, N.; Demirel, Y. Economic analysis of biodiesel and glycerol carbonate production plant by glycerolysis. J. Sustainable Bioenergy Syst. 2013, 3, 209−216. (57) Åkerberg, C.; Zacchi, G. An economic evaluation of the fermentative production of lactic acid from wheat flour. Bioresour. Technol. 2000, 75 (2), 119−126. (58) Tay, D. H. S.; Ng, D. K. S.; Tan, R. R. Robust Optimization Approach for Synthesis of Integrated Biorefineries with Supply and Demand Uncertainties. Environ. Prog. Sustainable Energy 2013, 32 (2), 404−409. (59) Hytö nen, E, Stuart, P. R., M, E-HM. Technoeconomic Assessment and Risk Analysis of Biorefinery Processes. In Integrated Biorefineries: Design, Analysis and Optimization; CRC Press; 2013; Vol I; Http://www.energy.gov/Eere/Bioenergy/Integrated-Biorefineries. (60) Wang, K.; Ou, L.; Brown, R. C. Beyond ethanol: a technoeconomic analysis of an integrated corn biorefinery for the production of hydrocarbon fuels and chemicals. Biofuels, Bioprod. Biorefin. 2015, 9, 190−200. (61) Helton, J. C.; Davis, F. J. Latin hypercube sampling and the propagation of uncertainty in analyses of complex systems. Reliab Eng. Syst. Saf. 2003, 81 (1), 23−69. (62) Talebian-Kiakalaieh, A.; Amin, N. A. S.; Hezaveh, H. Glycerol for renewable acrolein production by catalytic dehydration. Renewable Sustainable Energy Rev. 2014, 40, 28−59. (63) Vlysidis, A.; Binns, M.; Webb, C.; Theodoropoulos, C. A technoeconomic analysis of biodiesel biorefineries: Assessment of integrated designs for the co-production of fuels and chemicals. Energy 2011, 36 (8), 4671−4683. (64) Jong, E.; de Higson, A.; Walsh, P.; Wellisch, M. Bio-Based Chemicals: Value Added Products from Biorefineries; IEA Bioenergy: 2012; http://www.ieabioenergy.com/publications/bio-based-chemicalsvalue-added-products-from-biorefineries/. (65) From the Sugar Platform to Biofuels and Biochemicals - Final Report for the European Comission Directorate-General Energy; Contract No. ENER/C2/423-2012/SI2.673791; E4tech: London; 2015. (66) Biosuccinium Sustainable Succinic Acid: A Green Building Block to Enable Production of Bio-based Specialty Chemicals. http://www. reverdia.com/products/biosuccinium/ (2015). (67) Bechthold, I.; Bretz, K.; Kabasci, S.; Kopitzky, R.; Springer, A. Succinic acid: A new platform chemical for biobased polymers from renewable resources. Chem. Eng. Technol. 2008, 31 (5), 647−654. 6813


Article

Industrial & Engineering Chemistry Research (68) Werpy, T.; Petersen, G. Top Value Added Chemicals from Biomass; Volume IResults of Screening for Potential Candidates from Sugars and Synthesis Gas; Pacific Northwest National Laboratory: 2004.

6814


Economic Risk Assessment of Early Stage Designs ... - ACS Publications

Recommend Documents