Feedstock and Technology Options for Bioethanol ... - ACS Publications

Aug 3, 2012 - South Africa: Technoeconomic Prefeasibility Study ... production costs and that the economic feasibility of such technologies relies hea...
1 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/EF

Feedstock and Technology Options for Bioethanol Production in South Africa: Technoeconomic Prefeasibility Study Bamikole Amigun,†,‡ Daniel Petrie,‡ and Johann Görgens*,‡ †

Sustainable Energy Futures, Natural Resources and the Environment, Council for Scientific and Industrial Research (CSIR), Stellenbosch, South Africa ‡ Department of Process Engineering, Stellenbosch University, Private Bag X1, Stellenbosch, 7602, South Africa ABSTRACT: The production of fuel grade ethanol from nonfood grade crops in South Africa has the potential to reduce reliance on imported oil and minimize the negative environmental consequences of fossil fuels. This article presents a preliminary assessment of the technical and economic feasibility of producing fuel-grade ethanol from nonfood, small grain crops cultivated on marginal lands in the Western Cape of South Africa, namely triticale, low-grade wheat, feed barley, and malt barley. It also explores the potential of grain fiber fractionation (separation of starch from bran) on process economics for a dry-mill process. A conservative “base-case” economic model was developed for a processing capacity of 200 000 metric tonnes per annum (tpa) at maximum expected feedstock prices. The overall average unit costs of production for each grain were compared for three possible process configurations, i.e. a conventional dry-mill starch-to-ethanol plant, an advanced starch-to-ethanol plant with fractionation and energy recovery from bran, and a hybrid integrated cellulosic plant with fractionation, hydrolysis, and fermentation of bran. Triticale demonstrated the greatest economic potential of all grains, regardless of technology type, while the advanced process technology optionfiber fractionation and steam generation from bran combustionachieved the lowest overall costs of production, across all grain types. For the conservative base-case model, it was found that the government subsidy of ZAR 1.50/L proposed under the South African Biofuels Strategy is insufficient to ensure economic feasibility for any of the grains or technology options under scenarios with high feedstock costs. The government subsidy would need to be increased by at least 124% before profitable operation during times with high feedstock prices would be possible. A sensitivity analysis of the economic assumptions of the base-case model demonstrated that feedstock price is the most important determinant of production costs and that the economic feasibility of such technologies relies heavily on a favorable ethanol selling price. When assuming maximum theoretical starch−ethanol yields, and considering fluctuation in feedstock prices in the range from ZAR 1589/t to ZAR 3680/t (as observed in South Africa during the period between 2005 and 2012), the best economic results were observed for an advanced starch-to-ethanol plant processing triticale grain at a rate of 200 000 tpa (production cost between ZAR 4.25/L and ZAR 8.48/L). Due to uncertainty of commodity price fluctuations, it is recommended that static, deterministic models for economic feasibility should therefore be enhanced to incorporate uncertainty and quantitative risk assessment in process and economic parameters, by adopting stochastic simulation methods to account for price volatility.

1. INTRODUCTION In their commitment to sustainable development, South African policy makers are concerned with devising strategies to address the country’s dependence on imported oil, escalating greenhouse gas emissions, and an economically unstable agricultural sector. In the National Industrial Biofuels Strategy,1 the government established a safe biofuels production target of 400 MLpa that can be produced without risk of fuel/food competition. To encourage investment in the industry, the government has proposed a 100% exemption from the fuel levy, equivalent to ZAR 1.50/L in 2010.1 In the Western Cape region of South Africa, small grains such as wheat, barley (both feed and malt), and triticale are farmed extensively and are being considered as feedstocks for bioethanol production. It is estimated that there is sufficient areas of marginal land available to produce approximately 200 000 metric tonnes per annum (tpa) of grain without affecting the supply of food grain (wheat and barley), brewing grain (malt barley), or animal feed grain (triticale).2 There are currently no starch-to-ethanol plants operating in South Africa. As such, there is little knowledge regarding the economic © 2012 American Chemical Society

feasibility of such operations. Richardson et al. conducted a prefeasibility study for dry-mill bioethanol production from wheat in the Western Cape and found that government subsidy of ZAR 1.03/L was required to ensure economically feasible production.3 That study used capital cost estimates based on USA literature sources4 and USA feedstock price.5 There is an opportunity to design advanced starch-to-ethanol plants that will fractionate the grain and remove bran/fiber from the rest of the carbohydrates by abrasive mechanical means prior to hydrolysis and fermentation of the starch. It is expected that this will reduce energy consumption (electricity for milling, typically from fossil sources) as well as capital and operating expenditure (smaller fermentation and distillation equipment and lower heating demand for distillation boilers), while at the same time reducing the enzyme load required for hydrolysis.6 Furthermore, the isolated bran can be used for steam generation or as a feedstock for further ethanol Received: May 14, 2012 Revised: August 3, 2012 Published: August 3, 2012 5887

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

Figure 1. Conventional dry-mill ethanol plant (adapted from ICM13).

Figure 2. Advanced ethanol plant with bran fractionation and steam generation (adapted from ICM13).

production.6 It is argued that these improvements will improve the overall energy balance challenges of starch-based ethanol production.7,8 In practice, bran might also serve as a feedstock for the production of other high-value organic chemicals, which is not considered within the scope of this study. The enhancement also improves the quality of the ethanol production process byproduct, dried distillers grain and solubles (DDGS). By removing the bran, the relative protein content is increased by as much as 40%.9 The relatively higher protein content DDGS can substitute for other high-valued animal feeds such as soyabean oilcake, which is currently imported at over a million tonnes per year.10,11 Such technology has been widely investigated for corn-fed ethanol plants6,12 and is in commercial development by firms in Canada and the USA,8,13−16 but there is limited knowledge on

the economic potential for small-grain fractionation in ethanol production. Kaylen et al.17 reported that the cost of the feedstock represents a large portion of the total cost of most biomassrelated industries. For instance, with conventional dry-mill ethanol production, feedstock can account for over 50% of the total cost of production.18 Yoosin and Sorapipatana19 also concluded that the competitiveness of ethanol production is highly dependent on the feedstock cost and the volatility of the feedstock prices in the market. They further reported that feedstock contributes more that 50% of the total cost of production and proposed that plants be designed to treat multiple feedstocks as a strategy to remain resilient against volatile market prices. 5888

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

Figure 3. Integrated ethanol plant with bran fractionation, hydrolysis and fermentation (adapted from ICM13). Scenario III. Scenario III describes a hybrid ethanol production process where a cellulosic-ethanol plant is incorporated into a traditional/conventional dry-mill ethanol plant, by integrating the cellulosic front-end process with a traditional starch-based ethanol plant (Figure 3). In such a plant the fiber is removed from grain using the same fractionation technology as discussed under Scenario II. However, rather than using the fiber for energy (steam/electricity) generation, the fiber is pretreated by steam-explosion, followed by enzymatic hydrolysis to liberate simple sugars for subsequent fermentation to ethanol. The fermentation beer is purified to fuel ethanol by distillation. By integrating cellulosic ethanol technology within a starch-based ethanol production process, there is scope to exploit other lignocellulosic feedstocks in the future and mitigate the risk of feedstock failure (as a result of supply side issues), without the need for excessive capital outlay.20 For the purpose of this study it is assumed that dry grains are processed by the proposed ethanol plant at a process capacity of 200 000 metric tonnes per annum (tpa). This “base-case” was determined on the basis of the estimated crop yield from land that can be made available in the Western Cape without affecting food supply.2 The ethanol plant is assumed to operate for 330 days a year, continuously. The yield of ethanol from conversion of starch in different grains is illustrated in Table 1. For the purposes of comparing grains and technology options, the yield was assumed to be the “maximum theoretical yield”, or the ethanol yield per metric tonne feedstock based on the starch content, using the maximum theoretical yield of

This paper aims to assess the economic feasibility of ethanol production from different small grains by comparing the total costs of production with the estimated market value of fuel ethanol, on an energy-equivalent volumetric basis, and, in so doing, validate the need for government subsidy to encourage growth in the South African bioethanol industry. It also evaluates the potential of three process configuration options to determine the economic impact of grain fiber (bran) processing and fermentation.

2. METHODOLOGY 2.1. Technical Prefeasibility Assessment. Three process scenarios were investigated as part of this study: Scenario I. Scenario I represents a typical dry-mill ethanol process. Milled grain is cooked and hydrolyzed by enzymes to release sugars from starch. The sugar solution (or “hydrolysate”) is fed to the fermenters, where yeast and nutrients are added. During fermentation, CO2 is evolved and ethanol is produced to form a beer. The noncarbohydrate constituents and nonutilized starch of the grain remain as DDGS. The fermentation beer is then fed to the distillation train where ethanol is purified through distillation and dehydration to produce 200 proof (99.5%) ethanol. The addition of five percent petroleum denaturant is required by law to prepare the ethanol for fuel use. This concept is illustrated in Figure 1. Scenario II. Scenario II considers an advanced ethanol plant and incorporates a dry fractionation step to pretreat the grain prior to conversion to ethanol (Figure 2). Fractionation separates the grain into its three constituents: germ, endosperm, and bran. The bran is isolated and combusted as an auxiliary fuel for supplementing process heat.6 The remaining carbohydrates (the endosperm) are hydrolyzed, fermented and purified as in Scenario I. The resulting DDGS, though reduced in volume, is of a higher quality, with protein content improved from 35% to 65%.6 This allows it to compete directly against soybean meal, a high valued feed product. The upfront investment in this type of plant is estimated to be 30− 50% percent higher than a conventional plant.6 However, it is argued that fractionation improves production efficiency both operationally and economically.7

Table 1. Starch Content and Potential Ethanol Yield for Different Grains grain wheat triticale feed barley malt barley a

5889

maximum theoretical ethanol yield (L/t)

starch contenta (%)

nondetergent fiberb (NDF) %

450 470 400

60−63 60−65 55−60

16 18 18

396

50−58

13

Brand, 2003.23 bPatzek, 2006.22 dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

0.5 kg ethanol per kilogram starch.21 Newer types of starch-to-ethanol plants are able to achieve 95−98% of the maximum theoretical yield.22 For scenarios II and III, conservative estimates of the fiber content of grain, in the form of reported values of the nondetergent fiber (NDF) content, was used for the analysis (Table 1). Starch content and NDF values are taken from Patzek,22 which compare well with local values reported by Brand et al.23 The conversion of grain fiber to ethanol in Scenario III was assumed to be the same as values reported by Aden et al.24 at 340 L/t. Meredith25 indicated that dry-mill plants have similar energy demands, regardless whether they are fed wheat or corn and typically draw 3.55 kWh/L. By isolating bran by fractionation, the energy demands of the plant are estimated to be reduced by 20% and where the bran is burned for energy recovery, the net energy load is reduced by 40%.9 2.3. Economic Prefeasibility Assessment. The economic feasibility of ethanol plants is dependent on a number of key parameters. For the purposes of this prefeasibility study, different grains and process scenarios are assessed in terms of the break-even ethanol selling price. This is a simplified economic assessment indicator that suits the purposes of first-estimate prefeasibility studies26,27 and is similar to the “minimum ethanol selling price” employed by Eggeman et al.28 The per-unit break-even selling price of ethanol is calculated as follows:

PETOH = C F + CO&M + C I + C R − C B

feedstock is one of the most influential factors, often accounting for up to 50% or more of the per unit production cost of bioethanol.29 Grain production costs in the Western Cape have varied significantly in recent years due to drought and fuel price fluctuations.30 Nonfood grade small grains are typically sold at the market price of B3 wheat, which is fixed at an 8% discount of the SAFEX wheat price.2 For the purpose of this study, the base-case feedstock prices were fixed according to the 2008 grain price of B3 wheat at ZAR 3450/t.2 This is close to the maximum price observed over the last five years (Figure 4) and thus ensures a conservative assessment of the economic feasibility of ethanol production. Operation and maintenance costs include processing demands (chemicals, energy, etc.) as well as management expenses such as labor and administrative expenses. Chemicals (including enzyme, yeast, antibiotics) are priced according to data presented by Tiffany and Eidman4 as are maintenance and repair costs, which are estimated as 1% of the total capital cost. Labor, management, insurance, and licensing fees are also estimated according to data presented by Tiffany and Eidman.4 Administration costs are estimated as a percentage of total labor costs. Estimating the capital cost of a process plant can vary from a rapid ballpark estimate to a carefully prepared, detailed calculation, depending on how much information is available, level of accuracy required, how much time and effort is available to do the estimate.31,32 The capital investment cost represents the costs of all necessary equipment, including installation, and indirect expenses associated with the plant’s construction. Land value in South Africa varies substantially by location, and the possibility that bioethanol projects might be undertaken by government agencies makes the prediction of land costs difficult and costs of land have been excluded from the assessment. At the level of prefeasibility, a range of uncertainty of −15/+30% is acceptable. Capital costs were determined according to estimates from South African technology suppliers16 and scaled for different production capacities using a 0.6 scaling factor as illustrated in eq 2.33−37

(1)

where: PETOH is the break-even selling price of ethanol including government subsidy (ZAR/L), CF is the per unit cost of feedstock (ZAR/L), CO&M is the per unit cost of operation and maintenance (ZAR/L), CI is the per unit cost of initial capital investment (ZAR/L), CR is the per unit opportunity cost: expected return on capital (ZAR/ L), and CB is the per unit earnings from any byproduct (ZAR/L). The economic feasibility is first determined for the base case parameters and financial assumptions (Tables 2 and 3). The cost of

⎡ Q ⎤n C1 = ⎢ 1⎥ ⎢⎣ Q 2 ⎥⎦ C2

Table 2. Total CAPEX for Different Grains and Technology Options

or

⎡ Q ⎤n C1 = C2· ⎢ 1 ⎥ = k · Q 1n ⎢⎣ Q 2 ⎥⎦

(2)

CAPEX (in ZAR millions) grain

wheat/triticale

barley (feed/malt)

scenario I scenario II scenario III

383 455 741

422 494 782

where C1 = cost of the item at size or scale Q1, C2 = cost of the reference item at the size or scale Q2, n = scale exponent or cost capacity factor, k = correlation constant (normal cost of the item at unit size or scale). These capital cost estimates are considerably larger than those used by Richardson et al.,3 possibly due to increases in the market price of steel and concrete.16

Table 3. Base Case Operating Parameters and Assumptions unit description grain throughput electricity pricea feedstock priceb DDGS product price DDGS product yield CO2 price CO2 yield water price interest rate on capital loanc expected return on investmentd debt:equity a

tpa R/kWh R/tgrain R/tDDGS t/tgrain R/tCO2 t/tgrain R/ kLwater

scenario I

scenario II

scenario III

conventional dry-mill starch plant 200000 0.19 3450 2500 0.33 132 0.33 6.58

advanced plant: fiber fractionation and energy recovery 200000 0.19 3450 5000 0.25 132 0.33 6.58

integrated hybrid plant: fiber fractionation and fermentation 200000 0.19 3450 5000 0.25 132 0.37 6.58

14.1%

14.1%

14.1%

25%

25%

25%

70:30

70:30

70:30

ESKOM, 2008.49 bKleynhans, 2009.2 cPrime interest rate +2% (ten Cate, 2009).50 dIndustry standard (Richardson et al, 2007).3 5890

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

Figure 4. Fluctuations in feedstock prices: 2005−2012.30 Assumed price is shown by solid horizontal line at R3450/t with an error band of ±50%.

Figure 5. Average costs of production for all scenarios at base-case operating conditions. which it is most likely to be used as a substitute.11,41 CO2 prices are taken from AFROX,42 a local industrial gas supplier. The quality of CO2 is calculated based on the assumption that 90% of the producer gas is captured. All cost and revenue values are presented in Table 2, normalized to a per-unit ethanol basis (ZAR/L). The petroleum pricing mechanism applied in South Africa is known as the basic fuel price (BFP). The formula ties the domestic retail price of fuel ethanol to the international price for crude oil. Because of the lower volumetric energy content of ethanol compared to petroleum, fuel ethanol is priced at 66% of the retail price for petroleum. It is a legal requirement that fuel ethanol be denatured to prevent oral consumption by the public.43 As such it is assumed that bioethanol produced is denatured with 5 vol % petroleum. This study assumes a petroleum price of ZAR 6.879/L, based on the May 2008 figure for 95 octane unleaded petrol.43 The market price-at-the-pump of denatured ethanol is thus ZAR 4.66/L. The overall break-even selling price (average costs and expected return less average revenue from byproduct) is compared with the existing market price for fuel ethanol to determine the economic feasibility of the project. This gives an indication of the subsidy required from government in order to achieve economically viable operation.

South Africa is a Type A countrya technologically advanced developing country with well-diversified and comprehensive industrial, energy, and R&D infrastructure. It is thus expected that operating cost estimates from developed nations will be transferable to local feasibility studies.38 The different technology scenarios demand different capital expenditure based on costs of fractionation, boilers, pretreatment (steam explosion), and hydrolysis. For Scenario I, barley-fed plants are assumed to demand higher capital expenditure (10% more) than wheat/triticale fed plants due to the expected wear-and-tear on equipment for milling barley, which has a very abrasive husk.39 It is assumed that the initial capital outlay is financed as 70% debt, and amortized to an annualized cost of capital, recovered over the lifetime of the plant, with a fixed average interest rate over an expected plant life of 15 years. A common approach is to annualize the capital cost using the equation developed by Blank and Tarquin:40 ⎡ i(1 + i)n ⎤ annualized capital cost(A) = P ⎢ ⎥ ⎣ (1 + i)n − 1 ⎦

(3)

where A = annual payments (ZAR/year), P = present worth of the first investment cost (ZAR), i = annual interest rate in percent, n = project life in years. The investors’ opportunity cost is accounted for as an expected return on investment (ROI), based on industrial benchmarks.3 Earnings from byproduct include sales of DDGS and CO2. DDGS prices are estimated relative to the price of soybean meal in 2008, for

3. RESULTS AND DISCUSSION The technical and economic feasibility of producing fuel-grade ethanol from nonfood small grains in the Western Cape province of South Africa has been investigated. At a process 5891

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

Figure 6. Average revenue for different scenarios at base-case operating conditions.

Figure 7. Break-even ethanol selling price for all scenarios at base-case operating conditions.

higher yields per tonne of grain thus result in a lower per-liter cost of feedstock and are shown (see Figure 7) to be economically more attractive (advanced/hybrid technology options, and wheat/triticale over barley). Scenario III results in a reduction in per-liter cost of feedstock of over 10% but at higher capital and chemical costs. The contribution of capital expenditure to the average cost of production (as shown in Figure 5) is significant (ranging between 6 and 10% for different grains and technology options). This share is more significant for Scenario III than the other process options, due to the additional cost for the pretreatment (steam explosion), scale up of the fermentation units and additional auxiliaries, which could not be offset by the increase in the production rate of ethanol associated with fermentation of fiber. It is expected that technological advances in the fields of pretreatment and enzymatic hydrolysis may reduce the capital requirement for second-generation technologies in the future, making technology options such as Scenario III more economically attractive. Energy costs also affected the overall costs of production, but remained similar across all grains and scenarios (variation of less than ZAR/L 0.27 for all grains between scenarios). While there are significant savings in energy with the incorporation of

capacity of 200 000 tpa of grain, an ethanol production capacity from 83 to 111 MLpa is achievable, depending on the grain type and technology option considered (Table 1). Average costs of production for the base case operating and financial assumptions, including feedstock, amortized capital, and operating costs, are presented for each grain and scenario on a per-liter basis in Figure 5. The average revenue earned from the sale of byproduct (DDGS and CO2) is illustrated in Figure 6 while Figure 7 presents the break-even ethanol selling price. 3.1. Economic Analysis of Process Scenarios and Grains under Base-case Conditions. From analysis of the average costs of production under base-case operating (Figure 5), it is apparent that triticale and wheat exhibit slightly lower production costs per liter of bioethanol, when compared with feed and malt barley. The variation in average production cost between different process scenarios (Figure 5) is less significant (less than ZAR/L 0.20 for all grain types). The most critical economic factor to consider for input costs of ethanol production is noticeably the price of feedstock, accounting for over 50% of the average costs of production across all grains and scenarios, in accordance with values reported by Kaylen et al.,17 Hamelinck and Faaij,29 and Yoosin and Sorapipatana.19 The scenarios and grains that achieve 5892

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

grain fractionation, the economic benefit of this remains relatively low, with energy costs making up just 3−7% of the average costs of production. The advantage of grain fractionation is thus realized rather in added value to the DDGS product. This observation is subject to current energy prices, which are expected to increase,44 suggesting the savings in energy by fractionation may have a more significant impact in the future. Chemical, enzyme, and water costs make a significant contribution to the average production cost, most noticeably for Scenario III, where enzymatic hydrolysis of cellulose has high enzyme demand. This clearly demonstrated that, following feedstock, enzyme cost is one of the major obstacles to commercialization of cellulosic ethanol production, which may be reduced in future by further commercial development of enzyme as well as consolidated bioprocessing (CBP) strategies.45 Revenue from byproduct is comparable between all grain types, at approximately ZAR/L 2.00−3.00). Variation between process scenarios is higher than for average costs for different grain types (ZAR/L 0.86−1.02) (Figure 6). This suggests that revenue of byproduct (specifically sales of DDGS) has a stronger impact on assessing economic feasibility than average costs of production. The proportion of revenue earned from DDGS sales is presented in Figure 6 and demonstrates that advanced plants, which add value to the DDGS product by increasing relative protein, achieve significantly higher earnings. Even without bran combustion for energy recovery or fermentation for extra ethanol product, the value added to the DDGS product outweighs the extra cost of capital (Figures 5 and 6). The break-even ethanol selling price is shown to vary between grains and scenarios (Figure 7), but none of the grains achieves economic feasibility under any of the processing scenarios, even with the proposed government subsidy of ZAR/ L 1.50. This suggests that bioethanol production from small grains in South Africa is not economically viable during times of high feedstock prices, even when revenue streams (ethanol and DDGS) are well-priced. Triticale and wheat exhibit lower overall break-even prices to barley under all processing scenarios, with triticale demonstrating lowest break-even selling prices by a margin of just over 3% for all scenarios. Barley grains require higher selling prices than triticale to break even, by a margin of 13−17%. Since there is no significant benefit of processing one over the other, producers may choose to process both wheat and triticale to mitigate feedstock supply and price risk. Scenario II (dry fractionation and energy recovery) shows the most promising economic performance across all grains under base case operating parameters and financial assumptions (Figure 7). The next favored technology option is a dry-mill scenario (Scenario I), which is 10% higher, and then Scenario III (12% higher). This suggests that advanced ethanol process technology involving grain fractionation and energy recovery should be considered over dry-milling and hybrid-cellulosic technologies as illustrated in Figure 7. Acknowledging, however, the risk associated with novel technologies such as dry fractionation, certain investors may opt for the more established dry-mill technology despite its inferior economic performance. Table 4 reports the percentage increase required in the proposed subsidy, under base case operating parameters and financial assumptions, for each grain being treated by the best

Table 4. Required Subsidy Augmentation for Scenario II wheat

barley malt

unit

total production cost expected profit total revenue from byproduct overall production cost market price ethanol required subsidy for break-even proposed subsidy (100% fuel levy exemption) required subsidy increase required subsidy increase (percentage)

ZAR/L

9.83

9.48

10.97

11.06

ZAR/L ZAR/L

1.20 2.72

1.15 2.61

1.47 3.06

1.48 3.09

ZAR/L

8.31

8.02

9.37

9.45

ZAR/L

4.66

4.66

4.66

4.66

ZAR/L

3.65

3.36

4.71

4.79

ZAR/L

1.50

1.50

1.50

1.50

ZAR/L

2.72

2.61

3.06

3.09

143%

triticale

barley feed

grain type

124%

215%

219%

technology (Scenario II), in order for the plant to be economically feasible. This gives a clear indication that the proposed subsidy is inadequate to ensure economic feasibility under all possible scenarios of feedstock costs for any of the grains. Typical variations in feedstock prices would require that the government subsidy has to be increased by at least 124% before profitable operation would be possible during times with high feedstock prices. There are also options for the government to consider other forms of subsidy, such as capital funding, consumer incentives, funding for feedstock (feedstock establishment to reduce the cost), or funding for supply chain (infrastructure use and installation for handling, processing and storage).46 3.2. Sensitivity to Capacity and Feedstock Price. The economic feasibility of such an enterprise is dependent on a number of influences, some of which can be controlled by the investor (such as grain type, production capacity, acceptable ROI, debt:equity financing mix), as well as numerous factors beyond the control of the investor (such as fluctuating feedstock price, ethanol price, and DDGS price). It is therefore useful to expand the base-case model to evaluate the potential impact of such changing influences on the break-even ethanol selling price for the best grain-technology combination observed under the base-case model (Scenario II; triticale), keeping all other parameters equal. The results of this analysis are plotted in Figure 8. Substantial volatility in feedstock prices has been observed in South Africa, with B3 wheat prices in the range from ZAR 1589/t to 3680/t during May 2005 to March 2010 (Figure 4). Within this range, significant changes to the overall production cost are observed, with the plant becoming economically viable at 25% reduction in feedstock price to ZAR 2588/t (accounting for the proposed subsidy of ZAR 1.50/L). The production of bioethanol from small grains in the Western Cape of South Africa is likely to be constrained by the availability of agricultural land that can be utilized without competing with food production. From Figure 8, apparent economies of scale are observed; however, the limitation on arable land and available feedstock (estimated 200 000 tpa) will dictate the eventual rated capacity of such a plant. This result explains the economically feasible results observed by Richardson et al.,3 where a larger plant capacity (approximately 5893

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

Figure 8. Spider diagram of sensitivity in production cost of triticale-derived ethanol to capacity and feedstock price for Scenario II technology.

feedstock at R 3450/t) the project is not feasible with the existing ethanol market price and subsidies. This is mostly due to the high costs of feedstock, making up more than 50% of the production cost of ethanol from grain. Under this scenario, the producer support mechanism subsidy of ZAR 1.50/L proposed by government will need to be increased by at least 124% to encourage investors expecting a return on investment (ROI) of 25%. This highlights the importance of political engagement and the need to include political risks in economic feasibility assessments where subsidies and policy incentives are proposed by government as ways to develop and sustain biofuel markets. Fiber fractionation and energy recovery (Scenario II) demonstrates the best economic potential compared with traditional dry-mill processes and hybrid-cellulosic technologies for all grain types. This is largely attributed to the value-added byproduct DDGS, without the need for excessive capital expenditure. Until such dry-fractionation technologies reach commercial maturity, however, their implicit risk may deter investors. Of all grains considered, triticale and wheat show the greatest economic potential, followed by feed barley and malt barley. The project is shown to be most sensitive to changes in feedstock price, becoming economically viable at a triticale grain price of ZAR 2588/t, representing a 25% reduction compared to the conservative base-case grain price estimate (R 3450/t). There is potential for improved economic performance at higher production capacity, but this will be limited by the amount of land available for nonfood crop cultivation and proximity to pre-existing infrastructure. Further economic modeling of South African bioethanol production feasibility should be based on stochastic methods that account for possible simultaneous variations in input parameters and assess economic viability over the duration of the plant lifespan.

350 000 tpa) was used as a base-case. The CAPEX estimates used by Richardson et al.3 and feedstock prices were also considerably lower. Of all factors influencing the success of the proposed investment, the price of oil, and by association fuel ethanol is perhaps the least predictable. Fuel prices reported by SAPIA43 indicate ethanol price fluctuation between ZAR 3.00 and 7.00/ L. At this upper limit (50% price increase on the assumed basecase value), both triticale and wheat-fed plants will be economically viable with subsidy and Scenario II process technologies. The contribution of revenue from DDGS is crucial to the economic feasibility of he enterprise and is subject to the price such a product can fetch on the South African market. Over the past 10 years, soya oilcake imported for animal feed has sold at prices between ZAR 1500 and 4000/t.41 It is therefore reasonable to expect significant fluctuation to DDGS prices. However, even at a 50% price increase, revenue earned from DDGS is unable to guarantee economic feasibility with subsidy. The influence of fluctuation in feedstock cost on the proposed investment’s economic feasibility is shown to be more significant than changes in plant capacity (Figure 8), ethanol price, or DDGS price, but the “deterministic” approach to this assessment does not account for simultaneous variations in parameters and correlation between seemingly isolated variables (feedstock, energy, labor) and does not quantify the probability of economic success. To address uncertainty of this nature in economic feasibility studies, stochastic methods such as Monte Carlo simulation should be adopted.47 This accounts for long-term variation in process and economic parameters, quantifiably assessing a project’s economic risk (likelihood of economic success). Amigun et al.48 applied this approach to evaluate the economic potential of producing bioethanol from triticale in the Western Cape.

■

4. CONCLUSIONS The outcomes of this study show clearly the significant economic challenges that confront South African ethanol producers competing with conventional petroleum fuel in the market. Preliminary economic assessment of a small grain ethanol plant shows that at base-case operating conditions (200 000 tpa production capacity and conservatively priced

AUTHOR INFORMATION

Corresponding Author

*Tel.: +27 21 808 3503. E-mail address: [email protected]. Notes

The authors declare no competing financial interest. 5894

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

■

Article

(19) Yoosin, S.; Sorapipatana, C. A Study of Ethanol Production Cost for Gasoline Substitution in Thailand and Its Competitiveness. Thammasat Int. J. Sci. Technol. 2007, 12, 69−80. (20) Burke M. Vice president and MD of SunOpta. www. bioproductsalberta.com; 2007. (21) Novazymes. Mass balance and the quest for yield, Webinar Series. http://bioenergy.novozymes.com/en/learn-more/webinars/ Documents/Questforyield(webinarversion).pdf; 2009. (22) Patzek, T. W. A statistical analysis of the theoretical yield of ethanol from corn starch. Nat. Resourc. Res. 2006, 15 (3), 205−212. (23) Brand, T. S.; Cruywagen, C. W.; Brandt, D. A.; Viljoen, M.; Burger, W. W. South African J. Animal Sci. 2003, 33 (2), 117−125. (24) Aden, A.; Ruth, M.; Ibsen, K.; Jechura, J.; Neeves, K.; Sheehan, J.; Wallace, B. Lignocellulosic biomass to ethanol process design and economics utilizing Co-Current dilute acid prehydrolysis and enzymatic hydrolysis for corn stover; National Renewable Energy Laboratory, Golden, CO, 2002. (25) Meredith, J. Dryhouse design: Focusing on reliability and return on investment. In The Alcohol Textbook. A Reference for the Beverage, Fuel and Industrial Alcohol Industries, 4th ed.; 2004; pp 364−375. (26) Amigun, B.; Muller-Langer, F.; von Blottnitz, H. Predicting the costs of biodiesel production in Africa: learning from Germany. Energy for Sustainable Dev. 2008, 12 (1), 5−21. (27) Hettinga, W. G.; Junginer, H. M.; Dekker, S. C.; Hoogwijk, M.; McAloon, A. J.; Hicks, K. B. Understanding the reductions in US corn ethanol production costs: An experience curve approach. Energy Policy 2009, 37 (1), 190−203. (28) Eggeman, T.; Elander, R. T. Process and economic analysis of pre-treatment technologies. Bioresour. Technol. 2005, 96, 2019−2025. (29) Hamelinck, C.; Faaij, A. Production of advanced biofuels. Int. Sugar J. 2006, 108 (12387), 168−175. (30) GrainSA. Prince information (Domestic and international), Market Info. http://www.grainsa.co.za/body.php?mostvisited= 1&quicklinkgroupid=243&subcatid=91&pagetitle= Market%20Info&pagesubtitle= PRYS%20INLIGTING%20%28PLAASLIK%20EN%20INTERNASIO NAAL%29%20/ %20PRICE%20INFORMATION%20%28DOMESTIC%20AN D%20INTERNATIONAL%29; 2010. (31) Montaner, M. I.; Gadaleta, L.; Parin, M. A.; Zugarramurdi, A. Estimation of investment costs in fish processing plants. Int. J. Prod. Econ. 1995, 40, 153−161. (32) Amigun, B.; von Blottnitz, H. Cost analyses and predictions for a fuel ethanol plant in a rural and landlocked African country: Lang factor approach. Int. J. Prod. Econ. 2009, 119, 207−216. (33) Brennan, D. J. Process Industry Economics−An International perspective; Institution of Chemical Engineers, Warwickshire, UK, 1998; pp 47−59. (34) Zugarramurdi, A.; Parin, M. A.; Carrizo, G. A.; Gadaleta, L.; Lupini, H. M. Investment and production costs for fishmeal plants in developing and developed countries. Int. J. Prod. Econ. 2002, 76, 53− 59. (35) Remer, D. S.; Mattos, F. B. Cost and scale-up factors, international inflation indexes and location factor. Int. J. Prod. Econ. 2003, 84, 1−16. (36) Marouli, A. Z.; Maroulis, Z. B. Cost data analysis for food industry. J. Food Eng. 2005, 67, 289−299. (37) Remer, D. S.; Lin, S.; Yu, N.; Hsin, Y. An update on cost and scale-up factors, international inflation indexes and location factors. Int. J. Prod. Econ. 2008, 114 (2008), 333−346. (38) Amigun, B.; Sigamoney, R.; von Blottnitz, H. Commercialisation of biofuel industry in Africa: A review. Renewable Sustainable Energy Rev. 2008, 12, 690−711. (39) Hicks, K. B.; Flores, R. A.; Taylor, F.; McAloon, A. J.; Moreau, R. A.; Johnsmetric ton, D. B.; Senske, G. E.; Brooks, W. S.; Griffey, C. A. Current and potential use of barley in fuel ethanol. Presented at the 2005 EWW/SSGW Conference, May 9−12, Bowling Green, KY; 2005. (40) Blank, L.; Tarquin, A. Engineering Economy, 3rd ed.; McGrawHill Higher Education: New York, 1989.

DEDICATION The lead author of this paper, Bamikole Amigun, died tragically in a car accident in his home country of Nigeria as this article was being reviewed for publication. The final version has been edited in his absence and is the last of his work to be published. His gentle kindness and insightful approach to energy research in Africa will be greatly missed by friends and colleagues. Thank you, Kole.

■

REFERENCES

(1) Final Draft Bio-fuels Industrial Strategy of the Republic of South Africa. Department of Minerals and Energy, Pretoria; 2007; http:// www.dme.gov.za/pdfs/energy/renewable/biofuels_indus_strat. pdf%282%29.pdf. (2) Kleynhans, T. Personal Communication; Department of Agricultural Economics, University of Stellenbosch, February, 2009. (3) Richardson, J. W.; Outlaw, J. L.; Lemmer, W. J. Bio-ethanol Production from Wheat in the Winter Rainfall Region of South Africa: A Quantitative Risk Analysis. Int. Food Agribusiness Manage. Rev. 2007, 10 (2), 181−204. (4) Tiffany, D. G., Eidman., V. R. Factors Associated with Success of Fuel Ethanol Producers; Staff Paper Series, Department of Applied Economics, University of Minnesota, Minneapolis, MN, 2003. (5) FAPRI. Food and Agricultural Policy Research Institute. Iowa State University, January 2007 Baseline. http://www.fapri.iastate.edu/ brfbk07/. Data Table File: Wheat; 2007. (6) Best, M.; Shumaker, G.; Flanders, A. McKissick, J. The Economic Feasibility of Operating an Advanced Ethanol Production Facility in Georgia; Centre for Agribusiness and Economic Development College of Agricultural and Environmental Sciences; 2005; Pub. Number: FR05-09. (7) Elander, R. T.; Hsu, T. Processing and economic impacts of biomass delignification for ethanol production. Appl. Biochem. Biotechnol. 1995, 463 (45/46), 463−478. (8) POET. Poet Energy. http://www.poet.com/innovation/index.asp; 2009. (9) Robert, J. P. Enhanced profitability through technology, integration and diversification. Proceedings of the 2006 Bioenergy Conference, Tomar, Portugal, 15−18, 2006; http://services.bepress. com/cgi/viewcontent.cgi?article=1003&context=eci/bioenergy_i. (10) Coltrain, D. Economic Issues with Ethanol. Proceedings of the 2001 Risk and Profit Conference, Kansas, USA; 2001. (11) Strydom, D. B. The economic impact of maize-based ethanol production on the South African animal feed industry. Masters Thesis, Department of Agricultural Economics, Faculty of Natural and Agricultural Sciences, University of the Free State, Bloemfontein, 2009. (12) Rajogopalan, S.; Ponnampalam, E.; McCalla, D.; Stowers, M. Enhancing Profitability of Dry Mill Ethanol Plants: Process Modeling and Economics of Conversion of Degermed Defibred Corn to Ethanol. Appl. Biochem. Biotechnol. 2005, 120, 37−50. (13) ICM, Inc. Ethanol Production. http://www.icminc.com/ethanol/ ; 2008. (14) PureVision Technology Inc. Biomass Fractionation. http://www. purevisiontechnology.com/biomass-fractionation.html; 2008. (15) Permolex International, L.P. Company website. http://www. permolex.com/aboutus.asp; 2008. (16) Herbst, I. C. Logichem Process Engineering. Personal communication, March, 2008. (17) Kaylen, M.; Van Dyne, D. L.; Youn-Sang Choi, Y.-S.; Blase, M. Economic feasibility of producing ethanol from lignocellulosic feedstocks. Bioresour. Technol. 2000, 72, 19−32. (18) Shapouri, H.; Gallagher, P. USDA’s 2002 Ethanol Cost-ofproduction survey, Agricultural Economic Report Number 841, Office of Energy Policy and New Uses, United States Department of Agriculture, 2005; http://www.usda.gov/oce/reports/energy/USDA_ 2002_ETHANOL.pdf. 5895

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896

Energy & Fuels

Article

(41) BFAP. The South African Agricultural Baseline Report, Bureau for Food and Agricultural Policy. http://www.bfap.co.za/ BFAP%20Agricultural%20Outlook%202011.PDF; 2011. (42) AFROX. Personal Communication, February 2009. (43) SAPIA. Petrol and diesel in South Africa and the impact on air quality. www.sapia.co.za; 2009. (44) NERSA. Nersa’s decision on ESKOM’s application for interim price increase. Media Statement. www.nersa.org.za; 2009. (45) Lynd, L. R.; van Zyl, W. H.; McBride, J. E.; Laser, M. Consolidated bioprocessing of cellulosic biomass: an update. Curr. Opin. Biotechnol. 2005, 16, 577−583. (46) Shepherd, K. Funding & Support Mechanisms for Bioenergy. http://www.bioenergywm.org/documents/ Bioenergy%20Funding%20&%20Support%20Mechanisms.pdf; 2009. (47) Lefley, F. Approaches to risk and uncertainty in the appraisal of new technology capital projects. Int. J. Prod. Econ. 1987, 53, 21−33. (48) Amigun, B.; Daniel, P.; Görgens, J. Economic risk assessment of advanced process technologies for bio-ethanol production in South Africa: Monte Carlo analysis. Renewable Energy 2011, 36 (2011), 3178−3186. (49) ESKOM. Tariff History. http://www.eskom.co.za/live/content. php?Category_ID=285’; 2009. (50) ten Cate, S. Personal Communications, Western Cape Ethanol, November, 2009.

5896

dx.doi.org/10.1021/ef3008272 | Energy Fuels 2012, 26, 5887−5896