Introduction
General
This project explores how different parametric uncertainties affect the material requirements for ethanol production, focusing on feedstock and land usage By analyzing the entire process from harvesting switchgrass to conversion, nutrient inputs, and land area needs, the study identifies which steps contribute most to outcome uncertainty Key processes include pretreatment, enzymatic hydrolysis, and fermentation, as illustrated in Figure 1.
Figure 1 Summary of processes for conversion of switchgrass to ethanol
Objectives
This project aims to assess how uncertainty affects the material input needs for producing a specific level of ethanol To accomplish this goal, several key steps must be undertaken.
1 Determine parameters with the most significant uncertainty in the process of producing ethanol from switchgrass.
2 Collect parameter data and characterize the nature of the uncertainty for each parameter determined in step 1, in the form of range and distribution type.
3 Develop a model for the ethanol production process and method to incorporate uncertainty into this model
4 Conduct sensitivity analysis of uncertainty parameters on land area and harvest switchgrass.
5 Determine range of land area and harvested switchgrass based on desired amount of annual ethanol production.
Literature Review
Switchgrass
Switchgrass (Panicum virgatum) is a native North American perennial grass recognized for its potential as a dedicated bioenergy feedstock Its high yield, extensive cultivation range, and minimal agronomic input needs have positioned switchgrass as a leading choice in bioenergy production Additionally, its resilience to heat, cold, and drought, along with its resistance to pests and diseases, allows various ecotypes of switchgrass to thrive in diverse climates and soil conditions across North America.
Switchgrass is categorized into two main ecotypes: lowland and upland Lowland ecotypes, such as Alamo and Kanlow, thrive in wet conditions and are characterized by their tall, vigorous, and thick-stemmed growth, making them suitable for the Deep South and mid-latitudes In contrast, upland ecotypes, including Cave-In-Rock, Blackwell, and Trailblazer, are shorter and better adapted to drier climates, making them ideal for central and northern states.
Switchgrass is generally harvested once in the fall or winter, following a killing freeze, which allows nutrients to move into the root system This process not only reduces the need for nutrient replacement but also enhances the yield of switchgrass Consequently, a single, late-season harvest is considered essential for establishing switchgrass production as a sustainable, low-input system The optimal harvest period for switchgrass typically occurs in November.
1 and March 1 (Larson et al., 2010).
Ethanol Production
To ensure a continuous supply of feedstock for ethanol production, biorefineries rely on stored switchgrass during non-harvest periods, as there is only one annual harvest after senescence This storage process is crucial within the switchgrass supply chain Research by Larson et al (2010) indicates that one-third of the harvested switchgrass is delivered to the biorefinery immediately during the harvest season, while the remaining two-thirds is stored and systematically supplied to the facility from March to October.
The U.S Department of Energy recognizes switchgrass as a leading herbaceous energy crop due to its high yield and minimal water and nutrient requirements This versatile plant offers significant environmental advantages and can flourish on marginal lands, making it an ideal choice for sustainable agriculture Additionally, switchgrass can be seamlessly integrated into existing farming operations since conventional farming equipment can be utilized for its cultivation and harvesting According to estimates from the Oak Ridge National Laboratory, the United States has the potential to produce 171 million tons of switchgrass annually, highlighting its economic viability as a renewable energy source.
Switchgrass primarily consists of lignocellulose, which is made up of cellulose, hemicellulose, and lignin in a complex crystalline structure To convert lignocellulosic material into ethanol, two key processes are involved: the hydrolysis of cellulose into fermentable reducing sugars and the fermentation of these sugars into ethanol Due to the limited accessibility of cellulose and hemicellulose for enzymatic hydrolysis, an initial pretreatment step is essential to enhance enzyme accessibility to the structural carbohydrates Various methods, including physical, chemical, and biological processes, are employed for biomass pretreatment.
Ammonia fiber explosion (AFEX) is an effective physiochemical pretreatment method that solubilizes and removes lignin and hemicellulose from biomass cellulose In the AFEX process, biomass is subjected to high-pressure liquid ammonia (100 to 400 psi) and moderate temperatures (70 to 200°C) for under 30 minutes, followed by a rapid pressure release that "explodes" the fibrous mass This treatment decrystallizes cellulose, hydrolyzes hemicellulose, and depolymerizes lignin, significantly enhancing the size of micropores on the cellulose surface As a result, the treated biomass exhibits increased susceptibility to enzymatic hydrolysis, allowing it to achieve sugar yields close to theoretical maximums.
After pretreatment, cellulose and hemicellulose undergo enzymatic hydrolysis, resulting in fermentable sugars like glucose and xylose This process, utilizing carbohydrate-degrading enzymes such as cellulases and hemicellulases, offers an eco-friendly alternative to traditional methods that rely on concentrated acids or alkaline reagents (Keshwani, 2009).
The maximum theoretical yield of reducing sugars from switchgrass is 800 mg/g of dry biomass, as determined by the complete hydrolysis of cellulose and hemicellulose (Dale et al 1996) Research indicates that optimal sugar yields occur under AFEX treatment conditions of 90 degrees Celsius, with an ammonia loading of 1 g per g of biomass and a moisture content of 15% Under these conditions, AFEX-treated samples produce 4 to 5 times more sugar than untreated controls when subjected to the same enzyme loading.
The primary benefit of separate hydrolysis and fermentation (SHF) over simultaneous saccharification and fermentation (SSF) is the ability to optimize each process individually for better efficiency (Taherzadeh and Karimi, 2007) This allows for the effective conversion of sugars into ethanol, which is then refined from the fermented mash into pure ethanol.
In order to assess the impact of various stages of this process, a method for modeling the overall system is required.
System Modeling
The input-output model detailed in Appendix A outlines the processes involved in producing ethanol from switchgrass, illustrating the interdependencies between various stages of the system (Miller and Blair, 2009) Each node within this model comprises multiple equations that collectively form a comprehensive system of equations This approach allows for a breakdown of each ethanol production process by analyzing inputs and outputs, where each output serves as the input for the next process, highlighting the interconnected nature of the production stages.
The diagram illustrates a detailed snapshot of the switchgrass to ethanol input-output model, highlighting the interconnections between individual processes Each process depicted corresponds to a specific series of equations, which are summarized in Table 1 for the four processes shown Furthermore, the processes are linked at nodes, where the output from one process serves as the input for the next Table 2 provides an overview of the connectivity between these processes by presenting the nodal equations.
Figure 2 Several processes of the switchgrass to ethanol input-output model
Table 1 Select process equations for the switchgrass to ethanol input-output model
Table 2 Nodal equations for process connectivity for select process of the switchgrass to ethanol input-output model
After developing a base-case system model, it is essential to incorporate uncertainty, and one effective method for doing so is Monte Carlo Simulation (MCS) This stochastic technique utilizes random sampling from probability distributions to integrate uncertainty into the model (Manly, 2007) By applying Monte Carlo simulation to the input-output model, it becomes possible to assess a range of outcomes for land area and the necessary harvested switchgrass requirements.
Monte Carlo simulation provides estimates with inherent uncertainty, necessitating consideration of multiple simulations to achieve results with higher certainty (Seppale, 2008) While the primary drawback of this method is the need for extensive simulations, tools like Matlab facilitate the rapid execution of thousands of simulations The outcomes can be visualized as empirical probability distributions, such as histograms, which illustrate the range and most likely output values This technique is invaluable as it captures the variability of a system, offering a comprehensive view of potential outcomes.
Monte Carlo simulation employs random selection to conduct multiple trials, producing various results for each output These outcomes can be visualized as empirical probability distributions, such as histograms, which illustrate the range and most likely values This technique is invaluable as it captures the variability of a system, providing insights into the range of possible outcomes.
Model Development
Assumptions
The results outlined in this report are based on a number of assumptions, outlined as follows:
1 Single, late-season switchgrass harvest.
2 Switchgrass is stored throughout the year to provide continuous supply of feedstock to plant for ethanol production.
3 Annual 95% ethanol production (stimulus variable) is set at 95,000,000 liters of ethanol produced per year.
4 Method of pretreatment is ammonia fiber explosion (AFEX).
5 Separate hydrolysis and fermentation (SHF).
6 Assume average time in storage (t) of 200 days.
7 Parameter values described in Table 3.
8 Process 5 and process 15 intentionally left out to remain consistent with original model.
Parameters
The parameters for each process are summarized in Table 3 and discussed in greater detail below Note that single values represent assumed constant values
In a 2008 report by T A Groode, it is noted that crop yield follows a normal distribution with a mean of 12.5 mt/ha and a standard deviation of 2.8 mt/ha, which is crucial for calculating the land needed to produce a specific amount of ethanol annually The rates of nitrogen, phosphate, potassium, and pesticide applications are influenced by crop yield, which in turn affects the required land area The application rates (kg/mt) are calculated by multiplying the assumed application per area (kg/ha) by the crop yield (ha/mt), as illustrated in Table 3.
Table 3 Input parameters, including constant and variable
Biomass storage results in losses due to the breakage of fragile components and the fermentation and degradation of carbohydrates (Sokhansanj et al., 2006) The extent of these storage losses is influenced by moisture content, as illustrated in Equation 3.1 According to Bals et al (2010), moisture content is randomly selected from a continuous uniform distribution, ranging from a minimum of 0.1 to a maximum of 0.25 The relationship is defined by the equation: k 2,3max = 0.3792 * Moisture Content + 0.0368 (3.1).
Equation 3.1 calculates the maximum dry matter loss, k2,3max The following equation assumes dry matter loss approaches a maximum value asymptotically, where t is time in storage (days) k 2 , 3 1 [k 2 , 3 max *(1 e ( t / 180 ) )] (3.2)
Equation 3.2 is used in the model to determine the overall switchgrass lost in storage (k2,3) The original model neglects storage loss, setting k2,3 at a value of 1 mt/mt Table 4 below compares the results for the original switchgrass storage loss coefficient (k2,3) to 5,000 simulations varying the moisture content and determining k2,3 as seen above in Equations 3.1 and 3.2 The results show approximately 7% change in both land area and harvested switchgrass requirements when accounting for switchgrass lost in storage, therefore, the model will account for switchgrass storage loss as determined by Equations 3.1 and 3.2 Additionally, the coefficient for switchgrass lost in storage calculated in Process 3, encompasses all storage loss Therefore, we neglect storage loss in Process 6 (k2,6 = 1) as seen in Table 3.
Table 4 Effect of switchgrass storage loss on land area and harvested switchgrass
Dale et al (1996) indicate that AFEX treated switchgrass can produce a sugar yield of 550-600 mg sugar/g biomass, represented by a continuous uniform distribution in Process 10, where each yield within this range has an equal probability of occurrence The transformation process's water and excess water requirements are influenced by the sugar yield, with coefficient values for both determined by multiplying the assumed rates, as outlined in Table 3, by the randomly selected sugar yield.
The fermentation process, represented by Process 11, yields approximately 0.51 g of ethanol per gram of sugars from pretreated switchgrass (Krishnan et al., 1999) Under fermentation conditions of 60 g Bactopeptone/L and 10 g yeast extract/L, typical ethanol yields range from 0.46 to 0.48 g ethanol/g glucose, achieving 90-94% of the theoretical yield (Krishnan et al., 1999) The yield is derived from a continuous uniform distribution, with k1,11 calculated based on a dilute ethanol concentration of 12.5 g/L Additionally, the fermentation process generates 0.489 kg of carbon dioxide per kg of sugar solution (Xu et al., 2010), and the coefficient values for Bactopeptone, yeast, and carbon dioxide are determined by multiplying the assumed rates, as detailed in Table 3.
3, by the randomly selected fermentation yield.
Procedure
A Matlab program was developed based on the model in Appendix A to calculate the necessary variables for producing 95 million liters of ethanol annually The program outputs the response variables into an Excel TM spreadsheet, with Table 5 detailing the variables, symbols, and units used in the model Key variables include land area (Y5,1) and harvested switchgrass (Y0,1) Additionally, the program was adjusted to incorporate uncertainty parameters, as outlined in the appendix.
In this study, we simultaneously varied four key parameters (k5,1, k2,3, k1,10, and k1,11) in Matlab code to explore all potential outcomes for land area and harvested switchgrass needed to achieve specific annual ethanol production targets By conducting multiple simulations, we were able to generate a comprehensive range of requirements for both land area and switchgrass harvest.
In their 2010 report, Larson et al estimated that a commercial cellulosic ethanol biorefinery could achieve an annual capacity of 25 million gallons (95 million liters) of 95% ethanol This projection was informed by discussions with executives from Genera Energy LLC and DuPont Danisco Cellulosic Ethanol LLC, highlighting the potential scale for first-generation biorefineries (Larson et al., 2010).
In this study, each parameter (k5,1, k2,3, k1,10, and k1,11) was varied independently while keeping the others constant to assess their direct impact on land area and the requirements for harvested switchgrass Linear regression analysis was employed to explore the relationship between land area, harvested switchgrass, and the parameters of interest Additionally, a sensitivity analysis was conducted by adjusting each parameter to its high and low values, resulting in tornado diagrams for both land area and harvested switchgrass.
Table 5 Name and symbol of variables
Results and Discussion
Simulation Results
The study varied four key parameters (k5,1, k2,3, k1,10, and k1,11) simultaneously to explore the potential outcomes for land area and harvested switchgrass needed to produce 95 million liters of ethanol annually The results of 5,000 simulations for land area (Y5,1) and harvested switchgrass (Y0,2) are illustrated in Figures 3 and 4 Additionally, Table 6 presents a comparison of the means and standard deviations for land area, harvested switchgrass, and the four parameters of interest, which are highlighted in gray and derived from the distributions outlined in Table 3.
Figure 3 Land Area (ha/yr) for varying parameters simultaneously
Figure 4 Harvested Switchgrass (mt/yr) for varying parameters simultaneously
Table 6 Results of varying all parameters simultaneously
In the study, each parameter (k5,1, k2,3, k1,10, and k1,11) was independently varied while keeping the other parameters constant Table 7 presents the land area requirements (ha/yr) and the corresponding harvest of switchgrass (mt/yr) for each parameter variation, based on 5,000 simulations The table includes both the mean values and standard deviations Notably, when varying the coefficient k5,1, the standard deviation for harvested switchgrass (Y0,1) remains at 0 mt/year, indicating that changes in land area requirements do not influence the switchgrass harvest requirements.
Table 7 Results of varying each parameter independently
Sensitivity Analysis
To evaluate the effects of parametric uncertainties on the material needs for ethanol production, particularly regarding feedstock requirements and the land area needed for its cultivation, a sensitivity analysis was conducted using linear regression and tornado diagrams.
Simultaneous variation of parameters yielded results illustrated in Figures 3 and 4 To analyze the direct impact of each parameter on land area and harvested switchgrass requirements, linear regression was conducted The findings are summarized in Tables 8 and 9, which detail the relationships between land area and harvested switchgrass requirements with respect to each parameter Notably, Table 8 indicates that crop yield, represented by coefficient k5,1, exhibits the highest slope and coefficient of determination, signifying its significant influence on land area requirements Meanwhile, Table 9 reveals that sugar yield, denoted by coefficient k1,10, has the most substantial effect on harvested switchgrass requirements, as evidenced by its greatest slope and coefficient of determination.
Table 8 Land area linear regression results
Table 9 Harvested switchgrass linear regression results
Figure 5 presents a tornado plot illustrating the effects of uncertain parameters on land area (ha/year), highlighting the variations in switchgrass yield, storage loss, sugar yield, and fermentation yield Additionally, Figure 6 depicts the range of harvested switchgrass by examining the impacts of storage loss, sugar yield, and fermentation yield.
Figure 5 Range of land area (ha/yr) when varying each parameter independently
Figure 6 Range of harvested switchgrass (mt/yr) when varying each parameter independently
Conclusion
Switchgrass yield significantly influences land area, with a standard deviation of approximately 110,000 ha/year when varying this parameter independently, as indicated in Table 7 In comparison, the variation of other parameters results in much smaller deviations This dependence is further supported by linear regression analysis, which reveals that switchgrass yield (k5,1) has the highest slope and coefficient of determination, as shown in Table 8 Additionally, the tornado plot in Figure 5 illustrates that switchgrass yield exerts the most substantial impact on land area requirements.
The amount of harvested switchgrass is influenced by land area and crop yield, but it exhibits the most significant variation when the sugar yield coefficient (k1,10) is altered This is illustrated in Table 7, where the standard deviation for harvested switchgrass requirements is highest when adjusting sugar yield while keeping other parameters constant Sensitivity analyses confirm that changes in sugar yield lead to greater variations in harvested switchgrass compared to other factors Furthermore, Table 9 indicates that sugar yield (k1,10) has the steepest slope and highest coefficient of determination in linear regression analysis Additionally, results in Figure 6 reveal that the range of harvested switchgrass is primarily dependent on sugar yield rather than fermentation or storage yield.
Recent research from Oak Ridge National Laboratory indicates that the United States has the potential to produce around 171 million tons of switchgrass each year This significant amount of switchgrass could yield approximately 3.75 billion liters of 95% ethanol annually, highlighting its viability as a renewable energy source.
Future research should expand on the impact of uncertainties in economic and energy flows related to lignocellulosic feedstock conversion to ethanol, an emerging technology with many unknowns The model may be enhanced to evaluate different ecotypes of switchgrass, nutrient inputs, and treatment methods Additionally, as new studies on value-added byproducts emerge, the model could incorporate methods for extracting proteins while producing fermentable sugars from AFEX pretreated switchgrass and utilizing hemicellulose, which constitutes 20-25% of switchgrass, to enhance the economics of ethanol production.
The sustainability of converting switchgrass to ethanol relies heavily on the chemical and energy inputs involved in the process Future research must prioritize these factors to create accurate models By integrating energy consumption, greenhouse gas emissions, and petroleum displacement throughout the life cycle of switchgrass-based ethanol, we can better assess the viability of large-scale, long-term ethanol production from this lignocellulosic biomass (Groode, 2008).
Alizadeh, H et al., 2005 Pretreatment of Switchgrass by Ammonia Fiber Explosion (AFEX)
Applied Biochemistry And Biotechnology, 121-124, pp.1133-1141.
Bals et al (2010) conducted a study evaluating the ammonia fibre expansion (AFEX) pretreatment method for enhancing the enzymatic hydrolysis of switchgrass The research focused on switchgrass harvested across various seasons and locations, highlighting the effectiveness of AFEX in improving biomass conversion to biofuels Their findings, published in Biotechnology for Biofuels, emphasize the potential of this pretreatment technique in optimizing the use of switchgrass as a renewable energy source The full article is accessible through PubMed Central.
Casler, M.D., 2005 Ecotypic Variation among Switchgrass Populations from the Northern USA
Chang, V.S et al., 2001 Oxidative Lime Pretreatment of High-Lignin Biomass Applied
Cundiff, J.S et al., 2009 Logistic Constraints in Developing Dedicated Large-Scale Bioenergy
Systems in the Southeastern United States Journal of Environmental Engineering, 135(11), pp.1086-1096 Available at: http://link.aip.org/link/JOEEDU/v135/i11/p1086/s1&Agg=doi.
Dale, B et al., 1996 Hydrolysis of lignocellulosics at low enzyme levels: Application of the
AFEX process Bioresource Technology, 56(1), pp.111-116 Available at: http://linkinghub.elsevier.com/retrieve/pii/0960852495001832.
Duffy, M & Nanhou, V.Y., 2001 Costs of Producing Switchgrass for Biomass in Southern Iowa.
Iowa State University: University Extension, pp.1-12.
Ferrer, A et al., 2002 Optimizing ammonia processing conditions to enhance susceptibility of legumes to fiber hydrolysis: Florigraze rhizoma peanut Applied Biochemistry and
Biotechnology, 98-100, pp.135-46 Available at: http://www.ncbi.nlm.nih.gov/pubmed/12018243.
Groode, T.A., 2008 Biomass to Ethanol: Potential Production and Environmental Impacts
Massachusetts Institute of Technology, Department of Mechanical Engineering, (2002), p.185.
Gunter, L E., Tuskan, G A., Wullschleger, S D., 1996 Diversity among populations of switchgrass based on RAPD markers Crop Sci 36 (4), 1017–1022.
Holtzapple, M.T et al., 1991 The Ammonia Freeze Explosion (AFEX) Process: A Practical
Lignocellulose Pretreatment Applied Biochemistry And Biotechnology, 28/29, pp.59-74.
Keshwani, D.R & Cheng, J.J., 2009 Switchgrass for Bioethanol and Other Value-Added
Applications: A Review Bioresource Technology, 100, pp.1515-1523 Available at: http://www.ncbi.nlm.nih.gov/pubmed/18976902.
Krishnan, M.S et al., 1997 Fuel Ethanol Production from Lignocellulosic Sugars: Studies Using a Genetically Engineered Saccharomyces Yeast In ACS Symposium Series pp 74-92.
Krishnan, M.S., Ho, N.W.Y & Tsao, G.T., 1999 Fermentation Kinetics of Ethanol Production from Glucose and Xylose by Recombinant Saccharomyces Applied Biochemistry And
Manly, Bryan F J., 2007 Randomization, Bootstrap and Monte Carlo Methods in Biology 3 rd ed Chapman & Hall/CRC Boca Raton, Florida.
Miller, Ronald E and Peter D Blair, 2009 Input-Output Analysis: Foundations and
Extensions, 2nd edition Cambridge University Press.
Mosier, N et al., 2005 Features of promising technologies for pretreatment of lignocellulosic biomass Bioresource technology, 96(6), pp.673-86 Available at: http://www.ncbi.nlm.nih.gov/pubmed/15588770.
Popp, M.P., 2007 Assessment of Alternative Fuel Production from Swithgrass: An Example from Arkansas Journal of Agricultural and Applied Economics, 39(2), pp.373-380.
Sanderson, M.A., Egg, R.P & Wiselogel, A.E., 1997 Biomass Losses During Harvest and
Storage of Switchgrass Biomass and Bioenergy, 12(2), pp.107-114.
Schmer, M.R et al., 2008 Net energy of cellulosic ethanol from switchgrass Proceedings of the
National Academy of Sciences of the United States of America, 105(2), pp.464-9
Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi? artid"06559&tool=pmcentrez&rendertypestract.
Seader, J.D and Ernest J Henley, 1990 Separation Process Principles.
Seppale, Tomi, 2008 Introduction to Monte Carlo Simulation and Modeling Department of
Business Technology Helsinki School of Economics Available at: http://www.evira.fi/attachments/elaintauti_ja_elintarviketutkimus/riskinarviointi/food_saf ety_simulation_evira.pdf
Shuler, M.L and F Kargi, 2002 Bioprocess Engineering: Basic Concepts 2 nd ed Prentice Hall.
Sokhansanj, S., Kumar, A & Turhollow, A.F., 2006 Development and Implementation of
Integrated Biomass Supply Analysis and Logistsics Model (IBSAL) Biomass and
Bioenergy, 30(10), pp.838-847 Available at: http://linkinghub.elsevier.com/retrieve/pii/S0961953406000912 [Accessed May 9, 2011].
Taherzadeh, M.J & Karimi, K., 2007 Enzyme-Based Hydrolysis Processes for Ethanol from
Lignocellulosic Materials: A Review BioResources, 2(4), pp.707-738.
Thomason, W.E et al., 2004 Switchgrass Response to Harvest Frequency and Time and Rate of
Applied Nitrogen Journal of Plant Nutrition, 27(7), pp.1199-1226 Available at: http://www.informaworld.com/openurl?genre=article&doi.1081/PLN-
120038544&magic=crossref||D404A21C5BB053405B1A640AFFD44AE3 [Accessed February 13, 2011].
Xu, Y., Isom, L & Hanna, M., 2010 Adding Value to Carbon Dioxide from Ethanol
Fermentations Bioresource technology, 101(10), pp.3311-9 Available at: http://www.ncbi.nlm.nih.gov/pubmed/20110166.
Appendix B: Switchgrass to Ethanol Library Process 1 – Switchgrass Cultivation
Y5,1 = Land Under Cultivation, ha / yr
Technology Coefficients k 1,1 = kg Nitrogen / mt Biomass = (112)(k 5,1 )* [1, 4, 5] k 2,1 = kg Phosphate / mt Biomass = (50)(k 5,1 ) † [3] k 3,1 = kg Potassium / mt Biomass = (112)(k 5,1 ) ‡ [3] k 4,1 = kg Pesticides / mt Biomass = (1.75)(k 5,1 ) § [1] k 5,1 = Hectare / mt Biomass = Normal Distribution (μ.5, σ=2.8) [2]
* k 1,1 = (112 kg Nitrogen/ha)(k 5,1 ha/mt) = kg Nitrogen/mt Biomass
† k 2,1 = (50 kg Phosphate/ha)(k 5,1 ha/mt) = kg Phosphate/mt Biomass
‡ k 3,1 = (112 kg Potassium/ha)(k 5,1 ha/mt) = kg Potassium/mt Biomass § k 4,1 = (1.75 kg pesticides/ha)(k 5,1 ha/mt) = kg Pesticides/mt Biomass
1 Casler, M.D., 2005 Ecotypic Variation among Switchgrass Populations from the Northern USA
2 Groode, T.A., 2008 Biomass to Ethanol: Potential Production and Environmental Impacts
Massachusetts Institute of Technology, Department of Mechanical Engineering, (2002), p.185.
3 Popp, M.P., 2007 Assessment of Alternative Fuel Production from Swithgrass: An Example from Arkansas Journal of Agricultural and Applied Economics, 39(2), pp.373-380.
4 Schmer, M.R et al., 2008 Net energy of cellulosic ethanol from switchgrass Proceedings of the
National Academy of Sciences of the United States of America, 105(2), pp.464-9 Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi? artid"06559&tool=pmcentrez&rendertypestract.
Thomason et al (2004) investigated the effects of harvest frequency, timing, and nitrogen application rates on switchgrass Their study, published in the Journal of Plant Nutrition, provides insights into optimizing switchgrass yield and nutrient management The findings highlight the importance of strategic harvesting and nitrogen use to enhance plant growth and sustainability.
Y1,2 – Hammer Milled Switchgrass, mt/yr
Assume 10% of switchgrass is lost during grinding.
Technology Coefficients k1,2 = mt harvested switchgrass = 0.9 mt harvested switchgrass
Bals et al (2010) conducted a study evaluating the effectiveness of ammonia fibre expansion (AFEX) pretreatment for enhancing enzymatic hydrolysis of switchgrass, focusing on samples harvested across various seasons and locations Their findings, published in Biotechnology for Biofuels, highlight the potential of AFEX in improving biofuel production efficiency from switchgrass The research underscores the significance of seasonal and geographical factors in optimizing biomass pretreatment processes For more details, the full article is available at PubMed Central.
2 Sanderson, M.A., Egg, R.P & Wiselogel, A.E., 1997 Biomass Losses During Harvest and Storage of Switchgrass Biomass and Bioenergy, 12(2), pp.107-114.
3 Sokhansanj, S., Kumar, A & Turhollow, A.F., 2006 Development and Implementation of Integrated Biomass Supply Analysis and Logistsics Model (IBSAL) Biomass and Bioenergy, 30(10), pp.838-847 Available at: http://linkinghub.elsevier.com/retrieve/pii/S0961953406000912.
Y1,3 – Hammer Milled Switchgrass, mt/yr
Y2,3 – Stored Hammer Milled Switchgrass, mt/yr
Technology Coefficients k2,3 = mt stored hammermilled switchgrass per yr = see below * mt hammermilled switchgrass per yr
* K 2,3 = Moisture Content = ~U(0.1,0.25) [1, 2] k 2,3max = (0.3793)*( K 2,3 )+0.0368 [3] k = mt stored hammermilled SG/mt hammermilled SG = 1-[k *(1-e (-t/180) )] [3]
Assume no switchgrass is lost during this stage of storage.
Y2,6 – Stored Delivered switchgrass, mt/yr
Technology Coefficients k1,6 = mt stored delivered switchgrass per yr = 1 mt delivered switchgrass per yr
Assume 5% of switchgrass is lost during size reduction.
Technology Coefficients k1,7 = mt reduced switchgrass per yr = 0.95 mt delivered switchgrass per yr
Y0,8 – pretreated biomass (dry weight), mt/yr
The technology coefficients for biomass processing include k1,8, which denotes that 1 metric ton of pretreated switchgrass is produced annually; k2,8, indicating the use of 1000 kilograms of liquid ammonia per year for each metric ton of pretreated biomass; k3,8, which specifies that 990 liters of recycled ammonia are generated annually per metric ton of pretreated biomass; and k4,8, reflecting that 0.11 metric tons of water are utilized per year for each metric ton of pretreated biomass.
1 Alizadeh, H et al., 2005 Pretreatment of Switchgrass by Ammonia Fiber Explosion (AFEX) Applied Biochemistry And Biotechnology, 121-124, pp.1133-1141.
2 Bals, B et al., 2010 Evaluation of ammonia fibre expansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass harvested in different seasons and locations
3 Dale, B et al., 1996 Hydrolysis of lignocellulosics at low enzyme levels:
Application of the AFEX process Bioresource Technology, 56(1), pp.111-116 Available at: http://linkinghub.elsevier.com/retrieve/pii/0960852495001832.
4 Ferrer, A et al., 2002 Optimizing ammonia processing conditions to enhance susceptibility of legumes to fiber hydrolysis: Florigraze rhizoma peanut Applied
Biochemistry and Biotechnology, 98-100, pp.135-46 Available at: http://www.ncbi.nlm.nih.gov/pubmed/12018243.
5 Holtzapple, M.T et al., 1991 The Ammonia Freeze Explosion (AFEX) Process: A Practical Lignocellulose Pretreatment Applied Biochemistry And Biotechnology, 28/29,
1 Bals, B et al., 2010 Evaluation of ammonia fibre expansion (AFEX) pretreatment for enzymatic hydrolysis of switchgrass harvested in different seasons and locations
Biotechnology for biofuels, 3(1), pp.1-11 Available at: http://www.pubmedcentral.nih.gov/articlerender.fcgi? artid(23726&tool=pmcentrez&rendertypestract.
2 Ferrer, A et al., 2002 Optimizing ammonia processing conditions to enhance susceptibility of legumes to fiber hydrolysis: Florigraze rhizoma peanut Applied
Biochemistry and Biotechnology, 98-100, pp.135-46 Available at: http://www.ncbi.nlm.nih.gov/pubmed/12018243.
3 Holtzapple, M.T et al., 1991 The Ammonia Freeze Explosion (AFEX) Process: A Practical Lignocellulose Pretreatment Applied Biochemistry And Biotechnology, 28/29, pp.59-74.
4 Mosier, N et al., 2005 Features of promising technologies for pretreatment of lignocellulosic biomass Bioresource technology, 96(6), pp.673-86 Available at: http://www.ncbi.nlm.nih.gov/pubmed/15588770.
The technology coefficients for biomass processing include k1,9, which represents 1 metric ton of hydrolyzed biomass per year from pretreated biomass; k2,9, indicating the enzyme production capacity of 5 million FPU per year for hydrolyzed biomass; and k3,9, denoting the annual requirement of 20,000 liters of citrate buffer for the hydrolysis process.
Process 10 – Lignin Recovery/Filtration of Sugar Stream
1 Chang, S.V., W.E Kaar, B Burr, and M.T Holtzapple 2001 Simultaneous saccharification and fermentation of lime-treated biomass Biotechnology Letters, 23: 1327-1333
2 Dale, B.E., C.K Long, T.K Pham, V.M Esquivel, I Rios, and V.M Latimer 1996 Hydrolysis of Lignocellulosics at Low Enzyme Levels: Appication of the AFEX Process Bioresource Technology, 56: 111-116.
3 Ferrer, A., F.M Byers, B Sulbaran-de-Ferrer, B.E Dale, and C Aiello 2002 Optimizing Ammonia Processing Conditions to Enhance Susceptibility of Legumes to Fiber Hydrolysis Applied Biochemistry and Biotechnology, 98-100: 123-134.
The technology coefficients for biomass processing reveal critical metrics: k1,10 indicates the annual production of sugar solution, estimated at approximately U(0.550, 0.600) million tons from hydrolyzed biomass Meanwhile, k2,10 represents the annual output of spent solids, fixed at 0.629 million tons from the same biomass source Additionally, k3,10 calculates the yearly water consumption at 4,900 times the value of k1,10, while k4,10 measures the excess water produced, which is 7.258 times k1,10 These coefficients are essential for optimizing biomass conversion efficiency.
* k 1,10 = Continuous Uniform Distribution from 550-600 mg sugar/g BM
† k 3,1 = (4900 liters H 2 0/mt sugar solution)(k 1,10 ) = liters H 2 0/mt hydrolyzed biomass
‡ k 4,1 = ( 7.258 mt excess H20/mt sugar solution)(k 1,10) = mt excess H20/ mt hydrolyzed biomass
Process 11 – Fermentation of Glucose & Xylose to Ethanol
1 Krishnan, M.S et al., 1997 Fuel Ethanol Production from Lignocellulosic Sugars: Studies Using a Genetically Engineered Saccharomyces Yeast In ACS Symposium
2 Krishnan, M.S., Ho, N.W.Y & Tsao, G.T., 1999 Fermentation Kinetics of Ethanol Production from Glucose and Xylose by Recombinant Saccharomyces Applied
3 Xu, Y., Isom, L & Hanna, M a, 2010 Adding Value to Carbon Dioxide from Ethanol Fermentations Bioresource technology, 101(10), pp.3311-9 Available at: http://www.ncbi.nlm.nih.gov/pubmed/20110166.
Y0,11 = Broth (3g/L cell mass & 12.5g/L ethanol), L/yr
Y1,11 = Sugar Soln.(10g/L xylose, 20g/L glucose), mt/yr
The technology coefficients for the fermentation process are defined as follows: k1,11 represents the quantity of L Ferm Broth, while k2,11 denotes the amount of Bactopeptone, calculated as (0.02)(k1,11) Additionally, k3,11 indicates the quantity of Yeast Extract at (0.01)(k1,11), and k4,11 specifies the amount of Mutant Yeast, which is (0.0006)(k1,11) Lastly, k5,11 corresponds to the production of Carbon Dioxide, calculated at (0.489)(k1,11).
* k 1,11 = [(12.5 g/L ethanol)/((90-94%)(0.51 g ethanol/g glucose))] -1 *10 6 = liter/mt
† k 2,11 = (0.02 kg Bactopepetone/liters Broth)(k 1,11 ) = kg Bactopepetone/mt
‡ k 3,11 = (0.01 kg Yeast/liters Broth)(k 1,11 ) = kg Yeast/mt § k 4,11 = (0.0006 kg Yeast/liters Broth)(k 1,11 ) = mg Mutant Yeast/mt
** k5 ,11 = (0.489 kg CO 2 /kg sugar)(k 1,11 ) = kg CO 2 /mt
Process 12 – Separation of single-cell protein from broth
1 Shuler, M.L and F Kargi, 2002 Bioprocess Engineering: Basic Concepts 2 nd ed Prentice Hall.
Y1,12 = Broth (3g/L cell mass & 12.5g/L ethanol), L/yr
Y2,12 = Yeast Cell Mass, kg/yr
L Broth k2,12 = kg Yeast Cell Mass = 0.003 *
Process 13 – Ethanol Recovery and Purification
1 Seader, J.D and Ernest J Henley, 1990 Separation Process Principles.
Y1,13 = Dilute Ethanol (12.5g/L or 1.25% mass ethanol), L/yr
* k 2,13 = (800 l waste water/l conc eth)*(0.00125 l conc eth) = 1 l waste water/l dilute eth
Assume no switchgrass is lost during this stage of storage.
Technology Coefficients k2,14 = liter stored 95% ethanol per yr = 1 liter 95% ethanol per yr
1 Introduction to Chemical Engineering Thermodynamics, 6 th ed., J.M Smith, H.C Van Ness and M.M Abbott, McGraw-Hill, 2001.
2 Elementary Principles of Chemical Processes, 3rd ed., R Felder and R Rousseau, J Wiley, 2000.
3 McLaughlin, S., J Bouton, D Bransby, B Conger, W Ocumpaugh, D Parrish, C Taliaferro, K Vogel, and S Wullschleger 1999 Developing switchgrass as a bioenergy crop p 282-299 In: J Janick (ed.), Perspectives on new crops and new uses ASHS Press, Alexandria, VA.
4 Chang, S.V., W.E Kaar, B Burr, and M.T Holtzapple 2001 Simultaneous saccharification and fermentation of lime-treated biomass Biotechnology Letters, 23: 1327-1333.
Y0,16 = Spent (dry) switchgrass, mt/yr
Technology Coefficients k1,16 = mt ash = 0.05 [1-3] mt switchgrass k1,16 = mt ash = 0.273 [4] mt switchgrass k2,16 = mt H2O = 5.64 [1-3] mt switchgrass k3,16 = mt steam = 5.64 [1-3] mt switchgrass k3,16 = mt steam = 11.3 [4] mt switchgrass k4,16 = mt CO2 = 11.3 [1-3] mt switchgrass
Assume steam entering this process is equal to spent steam.
Technology Coefficients k1,17 = mt spent steam = 1 mt steam
% Department of Biological and Environmental Engineering
% PURPOSE: Function in which the main parameters of interest
% in Processes 1, 3, 10 and 11 are all varied to
% calculate all material flows for switchgrass conversion
% REFERENCE: Masters of Engineering Project: "Assessing the Impact of
% Uncertainty on Ethanol Production Outcomes"
% P2 - Switchgrass Grinding k1_2 = 9; %mt harvested SG/mt harvested SG
%k2_3 = 1; %mt stored hammermilled SG/mt hamermilled SG
% P6 - Swithgrass Storage k2_6 = 1; %mt stored delivered SG/mt delivered SG
% P7 - Switchgrass Grinding k1_7 = 95; %mt reduced SG/mt delivered SG
% P8 - Switchgrass Pretreatment k1_8 = 1; %mt pretreated SG/mt reduced BM k2_8 = 1000; %kg liquid ammonia/mt reduced BM k3_8 = 990; %l recycled ammonia/mt reduced BM k4_8 = 0.110; %mt H20/mt reduced BM
% P9 - Cellulose Hydrolysis k1_9 = 1.0; %mt hydrolyzed BM/mt pretreated BM k2_9 = 5.0*10^6; %FPU enzymes/mt pretreated BM k3_9 = 2*10^4; %l buffer (citrate)/mt pretreated BM k2_10 = 0.629; %mt spent solids/mt hydrolyzed BM
% P12 - Cellulose Hydrolysis k1_12 = 1.0; %l dilute ethanol/l borth k2_12 = 003; %kg/l borth
% P13 - Separation of single-cell protein from broth k1_13 = 0.00125; %l concentrated ethanol/l dilute ethanol k2_13 = 1.0; %l waste water/l dilute ethanol
% P14 - Ethanol Storage k2_14 = 1; %l ethanol efflux/l ethanol influx
% P16 - Switchgrass Boiler k1_16 = 05; %mt ash/mt switchgrass k2_16 = 5.64; %mt water/mt switchgrass k3_16 = 5.64; %mt steam/mt switchgrass k4_16 = 3.80; %mt carbon dioxide/mt switchgrass
% P17 - Steam Condensation k1_17 = 1.0; %mt spent steam/mt steam
% Y2_14 = input('Enter Value for 95% Ethanol Produced liters/yr, Y2,14: ');
% Mapping of Solution Vector to Material Flows
% Check the size of the matrix
% fprintf(' 1 Number of row = %3.0f and Number of Column = %3.0f\r',n,m);
% % Check Whether the System of Equations has a Determinant
% fprintf(' 2 The Determinant of A is %8.1f 1\r',d);
% fprintf(' Press any key to continue!\r');
% Solve Linear System of Equations
Y1_2 = y(8); % Hammer Milled Switchgrass mt/yr
Y1_3 = y(9); % Hammer Milled Switchgrass mt/yr
Y0_4 = y(11); % Switchgrass transport to plan mt/yr
Y0_7 = y(14); % Size Reduced Switchgrass mt/yr
Y0_11 = y(30); % Ferm Broth (3g/L cell mass & 12.5 g/l ethanol) l/yr Y1_11 = y(31); % Sugar Solution (10g/L xylose & 20.0 g/l glucose) mt/yr Y2_11 = y(32); % Bactopepton kg/yr
Y1_12 = y(37); % Ferm Broth (3g/L cell mass & 12.5 g/l ethanol) l/yr Y2_12 = y(38); % High Protein Bioprod kg/yr
% Department of Biological and Environmental Engineering
% PURPOSE: To call function Switch_ModelV5 in which the main
% parameters of interest in Processes 1, 3, 10 and 11
% are all varied to calculate all material flows for
% REFERENCE: Masters of Engineering Project: "Assessing the Impact of
% Uncertainty on Ethanol Production Outcomes"
% clear all; %Clear All clc; %Clear Command Window Display s = input('Enter the number of simulations, s: ');
Y2_14 = input('Enter Value for 95% Ethanol Produced liters/yr, Y2,14: '); for n=1:s
%Select Crop Yield from Normal Distribution ~N(12.5, 2.8) (mt Biomass/ha) k5_1(n) = (normrnd(12.5,2.8))^-1; %ha/mt Biomass
%Fixed application rate of Nitrogen
%Fixed application rate of Phosphate
K2_1 = 50; % kg Phosphate/ha k2_1(n)= (K2_1)*k5_1(n); %kg P/mt
%Fixed application rate of Potassium
K3_1 = 112; % kg Potassium/ha k3_1(n)= (K3_1)*k5_1(n); %kg K/mt
%Fixed application rate of Pesticides
K4_1 = 1.75; % kg Pesticides/ha k4_1(n)= (K4_1)*k5_1(n); %kg Pesticides/mt
K2_3(n) = unifrnd(0.1,0.25); %(continuous uniform distribution) %moisture content k2_3_max(n) = 0.3793*K2_3(n)+0.0368; %max dry matter loss t_stor= 200; %time in days k2_3(n)= 1-((k2_3_max(n))*(1-exp(-t_stor/180))); %mt stored hammermilled SG/mt hamermilled SG at time of storage
% P10 - Lignin recovery/filtration of sugar stream k1_10(n) = unifrnd(0.550,0.600); %(continuous uniform distribution) mt sugar solution/mt hydrolyzed BM
K3_10 = 4900; %l water/mt sugar solution k3_10(n) = (K3_10)*(k1_10(n)); %mt spent solids/mt hydrolyzed BM
K4_10 = 7.258; %mt excess water/mt sugar solution k4_10(n) = (K4_10)*(k1_10(n)); %mt excess water/mt hydrolyzed BM
K1_11(n) = unifrnd(0.9,0.94); %(continuous uniform distribution), theoretical yield = 0.51 g ethanol/g glucose) k1_11(n) = (((12.5)/(0.51*K1_11(n)))^-1)*(10^6); %liters Broth/ mt sugar solution
K2_11 = 0.02; %kg Bactopepetone/liters Broth k2_11(n) = (K2_11)*(k1_11(n)); %kg Bactopepetone/mt sugar solution
K3_11 = 0.01; %kg Yeast/liters Broth k3_11(n) = (K3_11)*(k1_11(n)); %kg Yeast/mt sugar solution
K4_11 = 0.0006; %kg Yeast/liters Broth k4_11(n) = (K4_11)*(k1_11(n)); %kg Yeast/mt sugar solution
K5_11 = 0.489; %kg CO2/kg sugar solution k5_11(n) = (K5_11)*(k1_11(n)); %kg CO2/mt sugar solution
[y(:,n)] = Switch_ModelV5(k1_1(n), k2_1(n), k3_1(n), k4_1(n), k5_1(n), k2_3(n), k1_10(n), k3_10(n), k4_10(n),k1_11(n), k2_11(n), k3_11(n), k4_11(n), k5_11(n), Y2_14); M=[transpose(k5_1) transpose(k2_3) transpose(k1_10) transpose(k1_11) transpose(y(1,:))]; %creates matrix for linear regression end
% %Send Results to Excel Spreadsheet titled 'SGResults'
The analysis of the results reveals key statistics regarding land area and harvested switchgrass The average land area is calculated to be approximately 8.2 hectares per year, with a standard deviation of the same value Additionally, the average harvested switchgrass is reported at around 8.2 metric tons per year, accompanied by a standard deviation that reflects the variability in the data These metrics provide valuable insights into agricultural productivity and land use efficiency.
The analysis reveals that the average crop yield for k5,1 is 8.2 ha/mt, with a standard deviation of 8.2 ha/mt Additionally, for k2,3, the average switchgrass storage loss is measured at 8.2 mt/mt, accompanied by a standard deviation of 8.2 mt/mt.
\n',std(k2_3)); fprintf('k1,10 - Average Sugar Yield = %8.2f mt/mt \n',mean(k1_10)); fprintf('k1,10 - Standard Deviation Sugar Yield = %8.2f mt/mt \n',std(k1_10)); fprintf('k1,11 - Average Fermentation Yield = %8.2f l/mt \n',mean(k1_11)); fprintf('k1,11 - Standard Deviation Fermentation Yield = %8.2f l/mt \n',std(k1_11));
% %Dispay results from single simulation
The results from the single run indicate the following metrics: harvested switchgrass is recorded at Y0,1 with a yield of %8.1f mt/yr, while nitrogen (Y1,1) is at %8.1f kg/yr, and phosphate (P2O5, Y2,1) stands at %8.1f kg/yr Potassium (K2O, Y3,1) is noted at %8.1f kg/yr, and pesticide usage (Y4,1) is measured at %8.1f kg/yr Land utilization (Y5,1) is documented at %8.1f ha/yr Additional metrics include hammer milled switchgrass yields (Y1,2 and Y1,3) at %8.1f mt/yr and %8.1f mt respectively, along with switchgrass transport to the plant (Y0,4) at %8.1f mt/yr Delivered switchgrass metrics (Y1,6 and Y2,6) are at %8.1f mt/yr, while size-reduced switchgrass (Y0,7) is %8.1f mt/yr Pretreated biomass (Y0,8) is recorded at %8.1f mt/yr, with liquid ammonia (Y2,8) at %8.1f kg/yr Hydrolyzed biomass (Y0,9) yields %8.1f mt/yr, and sugar solution (Y0,10) is at %8.1f mt/yr The fermentation process results in various outputs, including diluted ethanol (Y0,12) and high protein bioproducts (Y2,12) Finally, the analysis concludes with metrics for steam and CO2 production, with significant values recorded for both A histogram of harvested switchgrass (Y0,1) is presented to visualize the distribution of the data.
% figure(6) hist(y(6,:)); title('Land Area Y_5_,_1 (ha/yr)');
% title('Switchgrass Storage Loss k_2_,_3 (mt/mt)');
% title('Hammermilled Switchgrass Y_1_,_2 (mt/yr)');
% title('Reduced Switchgrass Y_0_,_7 (mt/yr)');
% title('Hydrolyzed Biomass Y_0_,_9 (mt/yr)');
% title('Sugar Solution Y_0_,_10 (mt/yr)');