DEVELOPMENT AND APPLICATION OF A DECISION SUPPORT TOOL FOR BIOMASS CO - FIRING IN EXISTING COAL FIRED POWER PLANTS By Jason S. Smith A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Biosystems Engineering Master of Science 201 5 ABSTRACT DEVELOPMENT AND APPLICATION OF A DECISION SUPPORT TOOL FOR BIOMASS CO - FIRING IN EXISTING COAL FIRED POWER PLANTS By Jason S. Smith Biomass co - firing h as the potential to be a low cost source of renewable energy that can utilize the existing infrastructure of existing coal fired power plants, while reducing the overall environmental impact of the existing systems. Thoug h there are technical barriers to the development of co - firing systems, there is a growing need to utilize biomass resources in renewable energy production and several systems have shown the ability to do so successfully. In moving forward, project develo pers need tools to identify potential technical, economic or logistical issues when planning the development of such systems. The purpose of this study is to use the aggregated information regarding various combustion technologies, pre - treatment technolog ies and available biomass feedstocks to generate a decision support tool for energy providers that will help identify economic, environmental and social impacts of developing site specific biomass co - firing projects at existing coal fired power plants. Th is decision support tool will then be utilized to examine existing power plant and biomass data to generate a site specific case study in the state of Michigan. iii Dedicated to my loving wife Emily and to my awesome family. I love you all! iv A CKNOWLEDGEMENTS There are many people without whom this thesis would not be possible. Firstly, I would like to thank my advisor Dr. Steven Safferman for his guidance in every aspect of the project and for his kindness in I would also like to thank my committee, Dr. Chris Saffron and Dr. Tim Harrigan. Their support throughout the proj ect and valuable insight made this project the success that it is. I would like to thank members of the development team Dr. Mike Thomas and Mr. David Binkley, without whom this project would not have been possible. I would also like to offer a special tha belief in me to carry out this work. Last but not least, I would like to thank again my wife and family for their patience and their support throughout this journey. v TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ................................ ............ vii LIST OF FIGURES ................................ ................................ ................................ ................................ ........... ix 1 INTRODUCTION ................................ ................................ ................................ ................................ ..... 1 2 OBJECTIVES ................................ ................................ ................................ ................................ ........... 3 3 LITERATURE REVIEW ................................ ................................ ................................ ............................. 4 3.1 Combustion Boiler Types and Methodologies ................................ ................................ .............. 4 3.1.1 Pulverized Fuel Combustion ................................ ................................ ................................ . 4 3.1.2 Fixed Bed ................................ ................................ ................................ ............................... 5 3.1.3 Fluidized Bed ................................ ................................ ................................ ......................... 6 3.2 Combustion Plant Configurations ................................ ................................ ................................ . 7 3.2.1 Direct Co - Firing ................................ ................................ ................................ ..................... 7 3.2.2 Parallel Co - Firing ................................ ................................ ................................ ................... 9 3.3 Biomass Pre - Treatments ................................ ................................ ................................ ............. 10 3.3.1 Torrefaction ................................ ................................ ................................ ........................ 10 3.3.2 Pelleting ................................ ................................ ................................ .............................. 11 3.3.3 Drying ................................ ................................ ................................ ................................ .. 13 3.4 Biomass Feedstocks ................................ ................................ ................................ .................... 14 3.5 Previous Modeling Efforts ................................ ................................ ................................ ........... 15 3.5.1 Kinetic & Mechanistic ................................ ................................ ................................ ......... 16 3.5.2 Techno - Economic ................................ ................................ ................................ ................ 17 3.6 Co - Firing Case Studies ................................ ................................ ................................ ................. 18 3.6.1 Field Testing & Full Time Units ................................ ................................ ............................ 18 3.6.2 Feasibility Studies ................................ ................................ ................................ ................ 19 4 METHODS ................................ ................................ ................................ ................................ ............ 21 4.1 Process Outline & Model Scope ................................ ................................ ................................ .. 21 4.2 Model Outline ................................ ................................ ................................ ............................. 23 4.3 Proces s Modeling ................................ ................................ ................................ ........................ 27 4.3.1 Combustion Modeling ................................ ................................ ................................ ......... 27 4.3.2 Biomass Depot Modeling ................................ ................................ ................................ .... 33 4.3.3 Torrefaction ................................ ................................ ................................ ........................ 34 4.4 Financ ial Modeling ................................ ................................ ................................ ...................... 37 4.4.1 Power Plant Capital Costs ................................ ................................ ................................ ... 37 4.4.2 Power Plant Operation and Maintenance Costs ................................ ................................ . 39 4.4.3 Biomass Depot Capital and Maintenance Costs ................................ ................................ . 40 4.4.4 Biomass Depot Operation and Maintenance Costs ................................ ............................ 40 4.4.5 Net Present Value (NPV) ................................ ................................ ................................ ..... 41 4.4.6 Levelized Co st of Electricity (LCOE) ................................ ................................ ..................... 42 4.4.7 Energy Returned on Energy Invested ................................ ................................ .................. 42 4.5 Statistical Methods ................................ ................................ ................................ ..................... 43 vi 4.5. 1 Fuel Variation ................................ ................................ ................................ ...................... 44 4.5.2 Biomass Yield Variation ................................ ................................ ................................ ....... 45 4.5.3 Monte Carlo Analysis ................................ ................................ ................................ .......... 46 5 RESULTS AND DISCUSSION ................................ ................................ ................................ .................. 48 5. 1 Model Verification and Validation ................................ ................................ .............................. 48 5.1.1 Comparison to known scenarios ................................ ................................ ......................... 50 5.1.2 Validation Results and Limitations ................................ ................................ ...................... 54 5.2 Scenario Setup ................................ ................................ ................................ ............................ 55 5.2.1 J.H. Campbell Power Plant ................................ ................................ ................................ .. 55 5.2.2 Feedstock Collection ................................ ................................ ................................ ........... 56 5.2.3 Biomass Production Capability: ................................ ................................ .......................... 58 5.3 Scenario Analysis ................................ ................................ ................................ ......................... 59 5.4 Scenario 1 Pelleted Poplar ................................ ................................ ................................ ....... 60 5.4.1 Energy Assessment ................................ ................................ ................................ ............. 61 5.4.2 Economic Assessment ................................ ................................ ................................ ......... 63 5. 4.3 Emissions Analysis ................................ ................................ ................................ ............... 65 5.5 Scenario 2 TOP Poplar ................................ ................................ ................................ .............. 66 5.5.1 Energy Assessment ................................ ................................ ................................ ............. 67 5.5.2 Economic Assessment ................................ ................................ ................................ ......... 69 5. 5.3 Emissions Analysis ................................ ................................ ................................ ............... 71 5.6 Scenario 3 Pelleted Switchgrass ................................ ................................ ............................... 72 5.6.1 Energy Assessment ................................ ................................ ................................ ............. 73 5.6.2 Economic Assessment ................................ ................................ ................................ ......... 74 5. 6.3 Emissions Analysis ................................ ................................ ................................ ............... 77 5.7 Scenario 4 TOP Switchgrass ................................ ................................ ................................ ..... 77 5.7.1 Energy Assessment ................................ ................................ ................................ ............. 78 5.7.2 Economic Assessment ................................ ................................ ................................ ......... 79 5.7.3 Emissions Analysis ................................ ................................ ................................ ............... 82 5.8 Comparison of Results ................................ ................................ ................................ ................ 83 6 CONCLUSIONS ................................ ................................ ................................ ................................ ..... 85 APPENDICES ................................ ................................ ................................ ................................ ................ 89 Appendix A Sup plemental Results Charts ................................ ................................ ............................ 90 Appendix B Economic Assumptions ................................ ................................ ................................ ..... 96 Appendix C Parameter Look - Up Table Values ................................ ................................ ..................... 99 Appendix D - Screenshots ................................ ................................ ................................ ..................... 107 BIBLIOGRAPHY ................................ ................................ ................................ ................................ .......... 113 vii LIST OF TABLES Table 1: Biomass Sizing Requirements for Various Boiler Types (FEMP, 2004) ................................ .......... 10 Table 2: Typical properties of different solid fuels - CEN - 335 - Solid biofuels, Fuel specifications and classes, March 2003 ................................ ................................ ................................ ................................ .... 15 Table 3: Ultimate and Proximate Analysis Values Used for Modeling as determined by ECN's PHYLLIS Database (ECN, 2015) ................................ ................................ ................................ ................................ . 28 Table 4: Torrefaction Yield and Efficiency Relation to Moisture Content (Batidzirai et al., 2013) ............. 35 Table 5: EROEI for Common Fuels (Murphy & Hall, 2010) ................................ ................................ ......... 43 Table 6: Sample Fuel Properties Variation - Switchgrass ................................ ................................ ............ 44 Table 7: 2014 MSU Field Plot Data for Isabella County - Yield and Dry Matter ................................ .......... 45 Table 8: Sample of Dis tribution Parameters for Monte Carlo Analysis ................................ ...................... 47 Table 9: EPRI Base Cases: Pretax Cost of Power from Cofiring In PC Boilers (Biom ass Cofiring Guidelines - 2015 Dollars) ................................ ................................ ................................ ................................ ............... 51 Table 10: CREDIT Model Outputs for Selected EPRI Cases ................................ ................................ ......... 52 Table 11: 2012 NASS Field Data Derived from (NASS, 2012) ................................ ................................ ... 57 Table 12: EROI Accounting for Scenario 1 ................................ ................................ ................................ ... 62 Table 13: Scenario 1 Economic Summary ................................ ................................ ................................ ... 64 Table 14: Scenario 1 Emissions Analysis ................................ ................................ ................................ ..... 66 Table 15: Scenario 2 Energy Assessment ................................ ................................ ................................ .... 68 Table 16: Scenario 2 Economic Summary ................................ ................................ ................................ ... 70 Table 17: Scenario 2 Emissions Analysis ................................ ................................ ................................ ..... 72 Table 18: Scenario 3 Energy Accounting ................................ ................................ ................................ ..... 7 4 Table 19: Scenario 3 Economic Assessment ................................ ................................ ............................... 75 Table 20: Scenario 3 Emissions Report ................................ ................................ ................................ ....... 77 viii Table 21: Scenario 4 Energy Accounting ................................ ................................ ................................ ..... 79 Table 22: Scenario 4 Financial Summary ................................ ................................ ................................ .... 81 Table 23: Scenario 4 Emissions Report ................................ ................................ ................................ ....... 82 Table 24: EIA LCOE Projections for 2020 (EIA, 2015)In combination with study ouputs. .......................... 84 Table 25: Scenario 1 Monte Carlo Statistics ................................ ................................ ............................... 92 Table 26: Scenario 2 Monte Carlo Statistics ................................ ................................ ............................... 93 Table 27: Scenario 3 Monte Carlo Statistics ................................ ................................ ............................... 94 Table 28: Scenario 4 Monte Carlo Statistics ................................ ................................ ............................... 95 Table 29: Variable Cost Assumptions ................................ ................................ ................................ .......... 96 Table 30: Capital Cost Assumptions ................................ ................................ ................................ ............ 97 Table 31: Depreciation Schedule for Equipment and Buildings ................................ ................................ . 98 Table 32: Miscanthus Fuel Properties (PHYLLIS2, ECN) ................................ ................................ .............. 99 Table 33: Wheat Straw Fuel Properties (PHYLLIS2, ECN) ................................ ................................ .......... 100 Table 34: Willow Fuel Properties (PHYLLIS2, ECN) ................................ ................................ ................... 101 Table 35: Poplar Fuel Properties (PHYLLIS2, ECN) ................................ ................................ .................... 102 Table 36: Switchgrass Fuel Properties (PHYLLIS2, ECN) ................................ ................................ ............ 103 Table 37: Coal Proximate and Ultimate Analysis (PHYLLIS2, ECN) ................................ ............................ 104 Table 38: Transportation Conditions ................................ ................................ ................................ ........ 104 Table 39: Coal Fired Powerplant Data - 1 ................................ ................................ ................................ . 105 Table 40: Coal Fired Power Plant Data - 2 ................................ ................................ ................................ 106 ix LIST OF FIGURES Figure 1: Cross Section o f Overfeed, Water - Cooled, Vibrating Grate Boiler (EPA, 2007) ............................ 5 Figure 2: Direct Fire Biomass Pathways ................................ ................................ ................................ ........ 8 Figure 3: Generalized Parallel Co - Firing Configuration ................................ ................................ ................. 9 Figure 4: Process Configuration ................................ ................................ ................................ .................. 22 Figure 5: Biomass Incorporation Routes at Coal Fired Power Plants ................................ .......................... 23 Figure 6: CREDIT data flow chart (Blue denotes user definable areas, purple denotes databases, and green denotes calculation sheets) ................................ ................................ ................................ .............. 24 Figure 7: CREDIT Screenshot, Biomass Blend Calculation ................................ ................................ ........... 25 Figure 8: CREDIT Screenshot, Biomass Collection Data ................................ ................................ .............. 25 Figure 9: CREDIT Screenshot, Biomass Depot Quest ionnaire ................................ ................................ ..... 26 Figure 10: CREDIT Screenshot, Electricity and Biomass Price Optimization Macros ................................ .. 26 Figure 11: Generalized Mass Flow Diagram of a Coal Fired Power Plant Utilizing Co - Firing ...................... 30 Figure 12: Biomass Depot Process Configuration ................................ ................................ ....................... 33 Figure 13: Pulverized Coal Equipment Decision Tree ................................ ................................ ................. 38 Figure 14: Isabella County Switchgrass Annual Yield ................................ ................................ .................. 46 Figure 15: Pulverized Coal Co - Firing Equipment Configuration Logic Tree ................................ ................ 49 Figure 16: Normalized Capital Cost Comparison Between EPRI and CREDIT Cases. Error Bars Reflect One Standard Deviation as Determined by Monte Carlo Analysis ................................ ................................ ..... 53 Figure 17: Levelized Cost of Electricity Comparison Between EPRI and CREDIT Cases. Error Bars Reflect One Standard Deviation as Determined by Monte Carlo Analysis ................................ ............................. 54 Figure 18: J.H. Campbell Power Plant (Courtesy of Consumers Energy) ................................ .................... 56 Figure 19: J.H. Campbell Power Plant Generalized Process Flow Diagram (courtesy of Consumers Energy) ................................ ................................ ................................ ................................ ................................ .... 56 x Figure 20: Study Collection Area ................................ ................................ ................................ ................. 58 Figure 21: Scenario Outline ................................ ................................ ................................ ......................... 60 Figure 22: Pelleted Poplar Monte Carlo Analysis ................................ ................................ ........................ 63 Figure 23: Cost distribution of biomass processing for scenario 1 ................................ ............................. 65 Figure 24: Monte Carlo Histogram - Scenario 2 EROI ................................ ................................ ................. 68 Figure 25: Monte Carlo Histogram - Scenario 2 LCOE ................................ ................................ ................ 69 Figure 26: Scenario 2 Processes Biomass Cost Distribution ................................ ................................ ....... 71 Figure 27: Monte Carlo Histogram for Scenario 3 EROI ................................ ................................ ............. 73 Figure 28: Scenario 3 Biomass Depot Proces sing Cost Distribution ................................ ........................... 76 Figure 29: LCOE of TOP Switchgrass Monte Carlo Histogram ................................ ................................ ..... 80 Figure 30: Scenario 4 Processed Biomass Cost Distribution ................................ ................................ ....... 82 Figure 31: Delivered Feedstock Cost Distribution ................................ ................................ ...................... 90 Figure 32: Poplar LCOE Sensitivity to Depot Distance from Power Plant ................................ ................... 91 Figure 33: User Input Screen Capture ................................ ................................ ................................ ....... 107 Figure 34: Investment Cost Screen Capture ................................ ................................ ............................. 108 Figure 35: Transportation Mass and Energy Balance Screenshot ................................ ............................ 109 Figure 36: Biomass Depot Mass and Energy Balance Screenshot ................................ ............................ 110 Figure 37: Co - Firing Mass and Energy Balance Screenshot ................................ ................................ ...... 111 Figure 38: Pro - Forma Screenshot ................................ ................................ ................................ ............. 112 1 1 INTRODUCTION With renewable energy standards becoming more prevalent in the U.S. and environmental regulations resulting in the potential closure of existing coal fired power plants, energy companies are increasingly interested in finding new ways to reduce carbon emissions. In this effort, biomass co - firing is being considered as a transition option toward a low carbon or carbon free power sector (Lempp, 2013) . Stand - alone biomass plants have the proven ability to produce reliable energy using a car bon neutral fuel source , but can be cost prohibitive when compared to other renewable energy options and may compete with other industry for biomass feedstock supplies (Kinney, 2012; Lempp, 2013) , resulting in energy insecurity and a high Le velized Cost of Electricity (LCOE) . According to the U.S Energy $100/MWh, as compared to $72.6/MWh for wind and $42.8/MWh for geothermal (EIA, 2015) . This is due to the seasonal nature of some forms of biomass and is exacerbated by the fact that biomass resources may be dispersed without a supply chain as well established as that of coal or other competing fossil fuels. Additionally, transport efficiencies for biomass tend to be much lower than conventional fossil fuels due to poor bulk dens ity, energy density, and high moisture content (Demirbas, 2005) . Co - firing biomass and coal at exiting coal fired power plants can negate some of the aforemention ed issues . An existing coal supply line and combustion system can help to buffer the system against fluxuations in biomass availability. (Williams, Pourkashanian, & Jones, 2001) . When compared to dedicated biomass systems, co - fired fired operations can be quit e large and take advantage of the improved efficiencies of such systems (Williams et al., 2001) . Finally, modifying existing systems is much more cost effective than building new, as capital costs can be kept low through the co - opting of existing infrastructure (Centre, 2005) . This is also true of operation and maintenance costs, as existing labor and equipment can be utilized in most cases to carry out daily activities. All of 2 these benefits can be realized with relative certainty because co - firing is a commercial ready technology as a r ecent review of co - firing exp erience identified over 100 successful co - firing field demonstrations (Baxter, 2005) . However, co - firing is not without challenges or barriers to ove rcome. Previous studies have noted that biomass fuel cannot be conditioned in the same manner as coal. The fibrous nature of biomass precludes it from being directly injected to size reduction technologies such as coal pulverizers. In fact, in pulverize d coal systems, biomass has been shown to be problematic at blending rates greater than 3% by heat without specific pretreatment and conditioning. Additionally, high alkali contents of some biomass fuels can lead to untimely boiler corr osion and increase maintenance requirements . Coal fired power plants that sell fly ash for beneficial use also need to ensure that the addition of biomass ash does not compromise the chemical composition of saleable byproducts. Though there are technical barriers to the de velopment of co - firing systems, there is a growing need to utilize biomass resources in renewable energy production and several systems have shown the ability to do so successfully. In moving forward, project developers need tools to identify potential te chnical, economic or logistical issues when planning the development of such systems. The purpose of this study is to use the aggregated information regarding various combustion technologies, pre - treatment technologies and available biomass feedstocks to generate a decision support tool for energy providers that will help id entify economic, environmental and social impacts of developing site specific biomass co - firing projects at existing coal fired power plants. This decision support tool will then be ut ilized to examine existing power plant and biomass data to generate a site specific case study in the state of Michigan. 3 2 O BJECTIVES T he objective of th is thesis is t o develop a decision support tool to evaluate the strategic use of biomass co - firing at ex isting power plants to include biomass aggregation, processing, and integrati on into existing infrastructure, to calibrate this tool using literature values, to verify the tool through matching logic structures to existing literature, and validate the tool through comparison with existing case studies. The tool will then be used to investigate scenarios surrounding a real coal fired power plant. Outputs to be estimated by the tool shall include : Cost of capital investment normalized to renewable electrici ty production, levelized cost of electricity production , associated reduction in net CO 2 emissions and energy returned on energy invested to produce biomass energy. 4 3 LITERATURE REVIEW A review of previously published literature was performed in order to establish the baseline assumptions and governing equations contained within the decision support tool. First, a review of combustion technologies applied at coal fired power plants was performed in order to understand the application and limitation of biomass firing in such systems. Secondly, a review of previously established biomass co - firing methodologies was performed to determine how biomass is commonly used within the context of e xisting coal fired power plants. As the model is also intended to ascertain the effect of biomass pre - treatment on project feasibility, the literature review was extended to include the technical nature and use of biomass pre - treatment techniques such as drying, densification, and torrefaction. Lastly a review of existing models and case studies was performed in order to better understand the current state of the technology, and the application of technical and economic models to biomass co - firing. 3.1 Combus tion Boiler Types and Methodologies The methods of combustion most commonly used in existing coal fired power plants. Combustion methods can be divided into 3 main categories: fixed bed, pulverized fuel, and fluidized bed combustion systems. There are se veral variations of each and an overview of each to understand the benefits and limitations associated with using biomass in such systems are discussed in the sequential sections. 3.1.1 Pulverized Fuel Combustion Pulverized f uel or d ust c ombustion systems (PCs) use compressed air to feed fuel into a combustion cyclone (Van Loo & Koppejan, 2008). Because the feedstock needs to be reliably moved by compressed air, f uel quality in dust combustion needs to be constant with a maximum fuel particle size of 10 - 20 mm an d a moisture content of no more than 20 wt% ( wet basis ) (Van Loo & Koppejan, 2008). Common in existing coal - fired power plants, the feedstock is pulverized prior to combustion to increase 5 completeness of reaction before feeding into a whirling cyclone of combustion air. Feedstock in this process have the most stringent size and moisture specifications, as feedstock needs to be injected in the same manner as pulverized coal but cannot mat or bind during the feeding process. This is by far the most common form of coal boiler used in the state of Michigan. According to an analysis performed of US Energy Information Administration (EIA) databases, 99% of coal fired capacity in the state of Michigan comes from pulverized coal boilers (EIA, 2012) 3.1.2 Fixed Bed In fixed bed system biomass is loaded onto a metal mesh or grate upon which the combustion reaction is performed. H ere, air passes through a fixed bed (grate) inside of a combustion chamber. Inside where in the primary combustion chamber the drying , gasification and ignition phases of combustion are performed in the presence of excess oxygen. Heavy particles remain securely on the grate while fines and gasses are lifted and combusted in a separate zone (Van Loo & Kopp ejan, 2008) . An annotated depiction of one such system is shown in Figure 1. Figure 1 : Cross Section of Overfeed, Water - Cooled, Vibrating Grate Boiler (EPA, 2007) Such furnaces are designed to accommodate fuels that have a high moisture content, varying parti cle sizes, and high ash content (Van Loo & Koppejan, 2008). G rate furnace s come in various types which are 6 commonly defined by grate config uration (f ixe d, moving, travelling, rotating and vibrating ) as well as air movement patterns (underfeed, counter curr ent, co - current and cross flow) (Van Loo & Koppejan, 2008). F ixed bed boilers are not commonly associated with direct co - firing as most la rge scale coal fired power plants do not currently utilize this boiler type. According to an analysis of EIA form 860 data for 2012 (EIA, 2012) less than 1% of coal fired generation capacity in Michigan comes from stoker boilers. However, these systems can be a part of more complex co - firing strategies such as in - direct co - firing . (Van Loo & Koppejan, 2008) . In these systems, biomass is combusted separately but in the same plant as coal and is used to drive th e same generators. This is useful where biomass cannot be directly mixed with coal due to particle size, matting, or other concerns. There is, however, a significant and often prohibitive added cost to adding a new combustion unit to an existing plant (Maciejewska, Veringa, Sanders, & Peteves, 2006) . 3.1.3 Fluidized Bed Fluidized b ed c ombustion uses a mixture of inert material , such as sand , in suspension with the fuel particles . Compressed air is continuously fed through the bottom of the combustion bed to force the bed to fluidize, thus improving available surface are and particle interaction. (Van Loo & Koppejan, 2008). In these beds, inert material makes up rough ly 90 - 98% of the total mixture by mass, with the fuel constituting the remainder of the mixture. The bed material provides high thermal inertia , allows for greater particle interaction (i.e. improved heat transfer) and stabilizes the combustion process . However, like pulverized fuel boilers, the fuel particle size must be consistently less than 80mm (Van Loo & Koppejan, 2008). The combustion temperature has to be kept low (800 - 900ºC) to prevent ash from coalescing with the inert particles and de - fluidizing the bed in a process called sintering (Van Loo & Koppejan, 2008). 7 3.2 Combustion Plant Configurations Another key consideration in biomass co - firing is determining where in the coal fired power plant process biomass will be utilized point of insertion relative to the existing process flow is called the combustion configuration. The two most commonly used combustion configurations are direct co - firing (also known as direct mixing) and parallel co - firing. 3.2.1 Direct Co - Firi ng As the name s uggests, direct co - firing involves mixing the biomass and coal feedstocks at a pre - determined ratio prior to insertion into the power plant boiler. At low biomass blending rates, t his configuration is the most economical and thus the most widely used . De pending on the biomass fuel characteristics , biomass specific mills may be required to ensure particle sizing and flow . Tillman et al. describes three primary methods by which direct co - firing can take place at coal fired power plant (Tillman, 2000) , which are depicted in Figure 1. The first option Tillman presents is to blend the fuel directly at the fuel pile, pr ior to size reduction or firing of the coal. This is logistically the simplest and least expensive option as mostly existing infrastructure is used and very little new capital investment is required . H owever this option can only be accomplished at low bi omass blending rations, usually <5% (Tillman, 2000) due to inconsistencies between biomass and coal physical properti es . The viable blending ratio is even lower for pulverized fuel boilers utilizing herbaceous or fibrous biomass; usually <3% (FEMP, 2004; Maciejewska et al., 2006) . This is due to the inabilit y of coal pulverizers to crush large quantities of biomass without biomass matting, resulting in the inhibiting of fuel flow. The second option partially addresses this issue by utilizing separate milling of biomass. After milling, compressed air is use d to insert biomass into the existing coal fired system. This can be accomplished by tapping in to the existing coal pipeline entering the burner, or through a separate injection port at 8 the burner itself. This approach involves a higher investment but c o - firing at higher ratios ( up to 15% biomass by heating value ) can be achieved (FEMP, 2004) . The final option is related to option 2. Here again, biomass is milled separately and through a separate injection port infused directly into the boiler. However, in this option the injection port is strategically placed so that biomass is used as a re - burn fuel to control the NO x emissions. NOx re - burn fuel systems take advantage of multi stage co mbustion zones inside of the boiler to limit NOx production. In this scenario, the first stage or main combustion zone would be where coal is burned. The second stage is a re - burning zone. Here excess fuels with a high volatile carbon and low nitrogen c ontent are added to create a reducing zone. As will be discussed in later sections, biomass has the required properties for a re - burn fuel, and can be utilized in this manner Figure 2 : Direct Fire Biomass Pathways 9 3.2.2 Parallel Co - Firing In parallel co - firing operations , biomass is in a parallel boiler in order to produce low grade steam. This steam is then upgraded to match the specifications of the coal boiler steam and injected into the steam cycle either at or ahead of the g enerator , as shown in Figure 3 . Capital costs for parallel co - firing installations are significantly higher than with direct mixing due to the need to install a new boiler system and steam upgrading equipment. However, there are a number of advantages th at can be realized as well according to (Van Loo & Koppejan, 2008) . Figure 3 : Generalized Parallel Co - Firing Configuration First, different combustion methodologies can be employed for the biomass, reducing the need for biomass conditioning prior to combustion. Most coal fired power plants in the United States utilize either pulverized coal, fluidized bed or cyclone boilers (EIA, 2012) . These systems have strict requirements for part icle sizing prior to combustion, as seen in Table 1 . Biomass boilers are generally fixed/stoker boilers that can accommodate larger particle sizes. Having a separate biomass boiler also allows for the combustion conditions to be tailored for a lower heat ing value fuel, improving overall system efficiency and reducing the need for pre - combustion conditioning activities such as drying. 10 Table 1 : Biomass Sizing Requirements for Various Boiler Types (FEMP, 2004) Existing Boiler Type Particle Size Required ( mm ) Pulverized Coal <0. 0002 Fluidized Bed < 50 Stoker 20 - 90 (Van Loo & Kopp ejan, 2008) Having a separate biomass boiler system can also be useful for fuels with high alkali and/or chlorine contents (Van Loo & Koppejan, 2008) . High alkali and chlorine content biomass sources such as wheat straw can be problematic to boiler ope rators, as these traits can cause corrosion of conventional coal boilers, as well as harmful byproducts (such as dioxin) under conditions of incomplete combustion. As boiler conditions and materials can be tailored to the biomass in parallel co - firing, th ese concerns are greatly reduced. Lastly, it is critical to note that coal and biomass ash are kept separate in parallel co - firing operations. This may be an important and deciding factor for operations that need to keep ash chemical composition within sp ecific limits for the purpose of beneficial re - sale. 3.3 Biomass Pre - Treatments As observed from the previous section s , the type of combustion system depends on the plant layout, and the blending ratio of biomass greatly depends on the chemical and physical conditions of the biomass delivered to the power plant. Chemical and physical conditions of the biomass depend on the inherent properties of the biomass and pre - treatment. The following sub section outlines key biomass pre - treatments that can be employed to condition biomass prior to delivery. 3.3.1 Torrefaction Torrefaction is a thermo - chemical process conducted in an anaerobic environment , at a temperature of 200 - 300 o C . Residence times vary depending on biomass moisture content but generally are 1 hour. Because of the lack of oxygen, the biomass does not combust at these temperatures but rather 11 decomposes d uring which volatiles such as carbon monoxide are given off and the solid portion of the biomass that remains as the final product. This is sometimes called char, bio - char or bio - coal. (Bergman, 2005) . Combustible torrefaction gas (torgas) is also produced during the reaction, which is often combusted to help heat the torrefaction reaction. Torrefaction is a promising pre - treatment option, improving the properties of biomass , includ ing the following (Bergman, 2005) . 1. Increased specific heating value . 2. I mproved water resistance (hydrophobicity) . 3. Increased friability (ease of grinding) . 4. Greater Uniformity . Torrefied b iomass will act more like coal as it is more brittle and is eas ily pulverized to a fine dust without binding. This property is known as friability and has further implications in energy consumption. The improved friability of biomass through torrefaction can reduce particle size reduction energy inputs by 70 - 90% , c ompared to non - torrefied biomass (Bergman, 2005) . Furthermore, the improved hydrophobicity of the feedstock allows for easier storage, as torrefied biomass will not spoil, ferment or otherwise degrade as regular biomass will if not kept in a controlled environment. Torrefaction is still an emerging technology, with several unit s in the pre - commercial development stages but few in full scale industrial operation. The resultant model makes the assumption that full scale production of a viable torrefaction product will be available in the near term. 3.3.2 Pelleting Pelletizing is a comp acting process that changes the biomass into a homogenous , high energy dens e cylindrical shape with dimension of 6 - 8 mm diameter (Van Loo & Koppejan, 2008) . At present , pellets are the form of biomass most often u sed in coal - fired plants for co - combustion (Bergman, 2005) . 12 The most important role of pelleting is that it improves both the bulk density and energy density of biomass feedstocks . Low bulk density can cause a number of problems along the supply chain and within combustion operations . Such feeds tocks are orders of magnitude more expensive to transport to the combustion facility. Once at the facility, feedstocks with a low energy density need to be treated differently in the combustion process. If the energy density is low, adding too much of th e low energy density feedstock to the combustion reaction will lower the combustion temperature and reduce overall combustion efficiency while also releasing harmful byproducts of incomplete combustion. For this reason, non - densified feedstocks may need t o be fed a lower ratio to coal, thus limiting their potential impact. (Bologa, Paur, Seifert, Woletz, & Ulbricht, 2012; Dunajski, Kruk, & Nowak, 2013; Gil et al., 2010; Svanberg, Olofsson, Flodén, & Nordin, 2013) . Pellets also have the added benefit of being more uniform in size and shape than u nprocessed biomass, reducing concerns of binding during feedstock handling. The process of pelleting is defined by 5 steps; drying, milling, steam - conditioning, densification, and cooling. Drying is necessary if and only if biomass does not meet the prere quisite conditions for pelleting. Depending on the type of biomass being pelleted, initial moisture content needs to be between 8 and 12% by weight (w.b.). This is because pellet stability is a function of the friction between the pelleting apparatus and the raw material. If the material is too dry, the friction from pressing will cause material carbonization. If the biomass is too moist, residual moisture in the formed pellet will cause expansion over time. In either case, pellet stability is lost. A fter drying biomass is hammer milled to reduce particle size and sometimes steam conditioned to improve cohesion. The biomass is then densified and quickly cooled to increase durability . (Van Loo & Koppejan, 2008) . The primary drawback of pelleting is that it does not completely address the issue of moisture management. Although pellets are dry, they readily absorb moisture and will swell or complete 13 disintegrate under wet conditions (Gil, 2010) . Consequently, pelleted biomass still need s to be st ored in a controlled environment before use in combustion. 3.3.3 Drying Drying is the process of vaporizing moisture found in biomass. T his is often performed in a controlled chamber with the aid of an external combustion reactor; however air drying can also be effective. (Van Loo & Koppejan, 2008) . This process is necessary to improve overall handling and combustion efficiency. Moisture content in raw biomass can be quite high. For raw woody biomass , for instance (sometimes called greenwood) , the moisture c ontents often exceed 50 % by weight (Svoboda K., 2005) . Combustion can be difficult to sustain if biomass moisture exceeds 60% wt . (Van Loo & Koppejan, 200 8) . Above this moisture content , support fuel is needed to sustain the reaction (Svoboda K., 2005) . Thus drying plays an integral role in the combustion process. H owever , the need to dry needs to be balanced against the financial cost of drying. Reducing moisture content from 50% to levels lower than 10 - 15% has been shown to be cost prohibitive even in large installations . (Svanberg, 2013). In addition to reduced boile r efficiency, high moisture content is also associated with biomass decomposition and associated energy loss during storage. The following drying options are available (Gebreegzia bher, Oyedun, & Hui, 2013; Svanberg et al., 2013; Werther & Ogada, 1999) . Open - air drying : harvested biomass is left in the open air. The targeted resultant moisture content for woody biomass in this situation is around 30% (Van Loo & Koppejan, 2008) . However, precipitation events can negate this effect. With f ew input costs, open air drying is the cheapest, yet least reliable drying option. Mechanical d ryers : These include belt, drum, tube and fluidized bed style boilers. All driers are capable of reducing biomass to user defined moisture contents, however th ey require an energy 14 input (fossil fuel, electricity, or excess biomass). If the raw biomass source is in log or branch form, size reduction will be required ahead of the drying process. 3.4 Biomass Feedstocks Biomass feedstocks can generally be divided into three categories; dedicated energy crops, crop residues, and feedstocks of opportunity. Dedicated energy crops, are those purposefully grown to be utilized as energy crops. These include energy grasses, short rotation poplar, and willow. Bio mass residue s are those feedstocks which are a byproduct of other agricultural processes such as wheat straw, corn stover and forest slash. Feedstocks of opportunity are generally defined as those which can be obtained as a waste product of a residential or industria l process such as sawdust, biosolids, and construction debris. As shown in Table 2 , the biomass and coal differ in a few key composition characteristics. Notably, as previously discussed biomass naturally has higher moisture content than most coal species, thus lowering its net calorific value. Further, biomass is generally lower in sulfur content, and has a lower C/O ratio than coal. Certain biomass residues such as straw are also known to have higher chlorine con tent than coal. If not properly managed, this can be corrosive to boilers. Similarly, high alkali and high ash content fuels such as switchgrass and miscanthus can cause issues with slagging if not properly managed. Finally, the high volatile matter cont ent when combined with ambient moisture can result in mass loss during storage if not properly managed. This mass loss can be as high as 1% per month {Srivastava, 2011 #231}. Torrefaction, and to a degree pelleting can manage this loss. For the purposes of this study, processing is assumed to be expedient and mass losses are thus considered to be negligible. 15 Table 2 : Typical properties of different solid fuels - CEN - 335 - Solid biofuels, Fuel specific ations and classes, March 2003 Coal Peat Wood without bark Bark Forest residues (conifer ou s tree with needles) Willow Straw Reed canary grass (spring harvest) As h content ( db. ) 8 . 5 - 10. 9 4 - 7 0 . 4 - 0 . 5 2 - 3 1 - 3 1 . 1 - 4 . 0 5 6. 2 - 7 . 5 Moisture content (w t %) 6 - 10 40 - 55 5 - 60 45 - 65 50 - 60 5 0 - 60 1 7 - 25 15 - 20 NCV (MJ/kg) 26 - 28. 3 20. 9 - 21 . 3 18. 5 - 20 18. 5 - 23 18. 5 - 20 18. 4 - 19 . 2 17. 4 17. 1 - 17 . 5 C, % db. 76 - 87 52 - 56 48 - 52 48 - 52 48 - 52 4 7 - 51 4 5 - 47 45,5 - 46,1 H, % db. 3 . 5 - 5 5 - 6 . 5 6. 2 - 6. 4 5 . 7 - 6 . 8 6 - 6 . 2 5. 8 - 6 . 7 5. 8 - 6 . 0 5 . 7 - 5 . 8 N, % db. 0. 8 - 1. 5 1 - 3 0. 1 - 0. 5 0. 3 - 0. 8 0. 3 - 0. 5 0. 2 - 0. 8 0. 4 - 0. 6 0. 65 - 1 . 04 O, % db. 2 . 8 - 11 . 3 30 - 40 38 - 42 24 . 3 - 40 . 2 40 - 44 4 0 - 46 4 0 - 46 44 S , % db. 0 . 5 - 3. 1 <0. 05 - 0 . 3 <0. 05 <0. 05 <0. 05 0. 02 - 0 . 10 0. 05 - 0. 2 0. 08 - 0 . 13 Cl , % db. <0. 1 0. 0 2 - 0. 06 0. 01 - 0 . 03 0. 01 - 0. 03 0. 01 - 0. 04 0. 01 - 0. 05 0. 14 - 0. 97 0. 09 3.5 Previous Modeling Efforts Several models have been developed previously to describe individual components of the biomass co - firing process chain that CREDIT is attempting to describe. These studies should be divided into two distinct categories; mechanistic and techno - economic. K inetic models attempt to describe one or a series of physical/ chemical relationships between project parameters. An example of this would be a model that describes the relationship between fuel properties and energy generation or a model that describes the relationship between fuel volatile matter and NOx emissions. Techno - economic models attempt to describe the relationship between required process technologies, process mechanics, and cost . 16 3.5.1 Kinetic & Mechanistic The following kinetic or mechanistic models were investigated for use or incorporation into CREDIT. (Basu, 2013) and (Van Loo & Koppejan, 2008) derive stoichiometric relationships for the combustion of biomass based upon the carbon, hydrogen, nitrogen, and sulfur mass fractions of biomass and coal, the moisture content of biomass and coal, as wel l as the excess air ratio used by the boiler. These equations predict the rate of formation of carbon dioxide, sulfur dioxide and nitrogen compounds based upon these parameters . Both texts also offer equations for gross calorific content and net calorifi c content of feedstocks based on these characteristics, and methodologies to apply these values to boiler outputs. Van Loo & Koppejan 2008 also offers stoichiometric adjustments to these equations based upon the formation of incomplete combustion products . Based on live studies conducted Allen Fossil Plant, (Tillman, 2000) studied the relationship between co - firing rat es and NOx emission reduction. In these tes ts, sawdus t was blended into a cyclone boiler firing Utah coal as a base feed at a rate of 15% biomass by mass (7% by energy). Tillman found that NOx emissions were strongly correlated to the increase in volatil e carbon content in biomass . The relationship was described as NOx = 1.554(FN ) + .021(EO2) + 0.0013 (FR) + 1.46(V/FC) 1.75. Where FN was the fuel nitrogen percentage, EO2 was the excess oxygen in flue gas, FR was the firing rate and V/FC was the volati le to fixed carbon ratio. This relationship was observed to have an R 2 value of 0.87 . At the pre - described blending rates, NOx emissions were found to be reduced by 15% (mass) while the boiler system experienced no capacity loss and only minor reductions in efficiency. 17 3.5.2 Techno - Economic The following techno - economic models were investigated for use or incorporation into CREDIT, and also investigated to ensure CREDIT fits a niche not previously occupied by existing models. Caputo et. al developed a series of equations the cost of installation of new equipment associated with biomass gasification and direct combustion based on the study of several case studies. Select values and equations from this study were used to calibrate CRE DIT economic assumptions about the relative cost of processing equipment installations. (Caputo, Palumbo, Pelagagge, & Scacchia, 2005) COFIRE is a spreadsheet based techno - economic analysis developed in 1990 by the Energy Production Research Institute (EPRI) to assess the cost of modifications needed fo r various fuel blending ratios and generator sizes. Though the model itself and the calculations used in the model are not publicly available, reported model results have been used to calibrate the power plant assessment portion of CREDIT after adjusting the results to 2015 dollars. Batidzirai et al developed a techno - economic analysis of a biomass torrefaction reactor depot. For this model, Batizirai utilized a mass and energy balance analysis of a torrefaction reactor that utilized direct combustion of excess biomass as an energy source to size a system for economic assessment. Analysis was performed for systems ranging from 50 - 500 kilo tonnes per year. Elements of the energy balance from this study, as well as the method of associating reactor yield with the moisture content observed in previous studies were adopted for CREDIT analysis of torrefaction reactions . (Batidzirai, Mignot, Schakel, Junginger, & Faaij, 2013) As part of its expanding effort to provide decision support tools for renewable energy generation the National Renewable Energy Laboratory developed an execu t ab le computer program called the System Advisory Model (SAM) . SAM is capable of integrating online weather data, feedstock availability (per the Billion Ton Study), and other regional data to produce a 18 region specific techno - economic assessment for several renewable energy technologies. Installation of new dedicated biomass only combustion systems is covered by this model, but retrofitting of existing coal fired power plants is not. Functional elements of the model such as integration with existing databas es have been adopted by CREDIT to improve site specific accuracy. 3.6 Co - Firing Case Studies In addition to understanding the models developed around co - firing, it is crucial to look at the existing instances of full scale biomass co - firing technology applica tion in order to understand how the models compare, and to understand some of the practical limitations of technology deployment. The literature provides two such kinds of study. The first is the presentation of results from full scale field operations. Though several plants have reportedly tested biomass co - firing at various levels, not all have reported their findings. This is reflected in the literature review. Additionally, other entities have conducted and reported full scale engineering feasibili ty studies regarding biomass co - firing. Though these may not provide field level data, they do provide a useful look at other forms of situational analysis that can be used to develop a decision support model. 3.6.1 Field Testing & Full Time Units Tillman , 2000 reports the results of several co - firing case studies performed between the years of 1990 and 1999 at several facilities, all conducted under the auspices of the EPRI . These facilities include the Bailly, Seward, Shawville, Allen, and Michigan City Gener ating stations. The results from these studies are aggregated in (Tillman, 2000) . As this dataset includes several s tudies performed by the same author with the same methodology, they represent a good source of data for comparison. Economic and e mission data from these case studies were used to validate CREDIT. In these studies, Tillman investigates the blending effec ts of varying biomass 19 injection rates on NOx reduction, boiler efficiency, pulverizer energy requirements, and capital investment requirements. In 2003, co - firing of torrefied wood was tested in PC - plant in the Netherlands, where torrefied wood was mixed w ith coal at a ratio of up to 9% (energy basis). The conclusion was that a 9% torrefied biomass blend was non - problematic and that co - firing at higher ratios may even be possible (Weststeyn A., 2004) . 3.6.2 Feasibility Studies The Idaho National Laboratory conducted a feasibility study to determine the cost of woody biomass co - firing at a blending rate of 20% by mass. INL utilized combustion models developed in ASPEN to calculate mass and energy balances, in conjunction with a spreadsheet based economic assessment. The INL report also utilized a version of a Monte Carlo analysis to a scertain model result distribution based upon variation across multiple parameters . This method of analysis was duplicated within the CREDIT scenario analysis and data validation process as a method to deal with the uncertainty inherent in assumption base d feasibility studies. Srivastava et. al. investigated the cost effectiveness of biomass pre - treatment via torrefaction and pelleting in the state of Michigan. A scenario was developed investigating the production of farmed willow crops followed by torrefaction, and pelleting p rior to combustion. The effects of pre - treatment and depoting were evaluated at multiple distances in order to determine the distance at which torrefaction became cost effective through the realization of transport efficiencies. Values for Michigan speci fic biomass generation rates as well as machinery costs were used to calibrate CREDIT. (Srivastava, Abbas, Saffron, & Pan, 2011) 20 Hartmann and Kaltschmitt conducted a life cycle analysis (LCA) of a German coal fired power plant that co - fed a 10% residual wood and straw blend with coal. (Hartmann & Kaltschmitt, 1999) 21 4 M ETHODS A spreadsheet model was developed using literature derived relationships and equations and then us ed to evaluate the long term cost, GHG emissions, and energy generation potential associated with co - firing biomass at exiting coal fired power plants based on analysis of publicly available data and data provided by MSU extension agents . Microsoft Excel was utilized to perform the necessary calculations and conditional relationship statements in order to ensure that the tool could be used by plant operators, policy makers, and members of the public at large without the need for specialized software. Two b iomass pre - processing routines were investigated in order to determine the logistic and economic viability of using biomass depots as a portion of the biomass co - firing lifecycle process. The calculations, methods, and assumptions contained within the spr eadsheet model are outlined in the following sections. 4.1 Process Outline & Model Scope Figure 4 illustrates the scope of the analysis. The model begins by assuming feedstock is purchased at - and raw physical and chemical characteristics are assumed at this point from feedstock specific literature derived values. Biomass is then assum ed to be transported via truck from the farm gate to an aggregation point called a biomass depot. Once at the biomass depot, biomass may be processed through milling, drying, densification or torrefaction. Figure 4 displays the two processing schematics used in later scenario analysis. After depoting, biomass is transported from the depot to the power plant either by rail or by truck. 22 Figure 4 : Process Configuration At the power plant, biomass will be offloaded, stored, and where necessary processed to meet boiler fuel specifications. Processing activities at the power plant may include drying, milling, or incorporation into a separate boiler. Figure 5 offers a more detailed description of the incorporation routes that can be utilized at the power plant. 23 Figure 5 : Biomass Incorporation Routes at Coal Fired Power Plants 4.2 Model Outline The model that was derived from this effort was designate d the Combustion of Renewable Energy Development Iterative Tool (CREDIT) . CREDIT is capable of drawing upon existing databases as well as regional data relating to biomass production and land availability to generate energy production cost estimates, proj ected CO 2 mitigation data, as well as scenario analysis of various biomass pre - processing techniques. The logical structure derived for this task is shown in Figure 6 . CREDIT leverages site data collected and aggregated by the US EIA to populate several key parameters regarding the selected site, including but not limited to boiler type, boiler capacity, boiler feeding rate, fuel type, and emission data. The tool requests that the user select which power plant and boiler is under investigation through a dropdown menu. These parameters are used as the baseline for several calculations regarding biomass requirements and capacity. 24 Figure 6 : CREDIT d ata flow chart (Blue denotes user definable areas, purple denotes databases, and green denotes calculation sheets) After site identification, the user is prompted to identify the type of biomass that is to be co - fired from a pre - define d list of biomass sources. Table 2 lists the relevant feedstocks along with their associated chemical properties. Based on the answers given to the first 2 questions, an advised blending ratio is recommended which maximizes the amount of biomass which ca n be blended given the type of biomass, the boiler type and the combustion configuration ( Figure 7 ) . Users may accept this value or enter their own before calculating the amount of biomass required to meet this blending ratio. 25 Figure 7 : CREDIT Screenshot, Biomass Blend Calculation Users are then asked a series of questions regarding biomass collection, transportation and logistics in order to ascertain the collection area needed to achieve the desired blending ratio . Figure 8 is a screenshot from CREDIT depicting the information needed for this step. Green cells are user defined inputs, while blue cells are automated lookup values associated with user defined feedstock. Figure 8 : CREDIT Screenshot, Biomass Collection Data Additionally , if biomass is intended to be aggregated at a depot prior to use at the coal fired power plant, the user may define what depot operations will take place including drying, torrefaction, and 26 densification of the biomass . Figure 9 shows a screenshot of CREDIT showing the user inputs for this step. Figure 9 : CREDIT Screenshot, Biomass Depot Questionnaire C ollected data is then fed into several subroutines (each an individual excel worksheet) a s is outlined by the data flow diagram in Figure 6 . The calculations utilized in these sheets are detailed in the following sections. An additional critical point is the optimization loops associated with the economic analyses. Excel macros were develop ed to allow the user to determine the value of processed biomass feedstock exiting the biomass depot, as well as the break - even electricity sale pri ce for the power plant. These break - even prices are optimizations that set the net present value NPV either the depot or the power plant to zero, after a return on investment of 8% is reached. The resultant values are used to calculate the levelized cost of electricity production (LCOE) for the given scenario ( Figure 10 ) . Figure 10 : CREDIT Screenshot, Electricity and Biomass Price Optimization Macros 27 4.3 Process Modeling mass and energy balances associated with the individual processe s involved in the model. This section outlines the governing equations and assumptions made in producing these calculations at the power plant and at the biomass depot. 4.3.1 Combustion Modeling Combustion is a complex phenomenon involving simultaneous coupled heat and mass transfer with chemical reaction and fluid flow (Jenkins, Baxter, Miles Jr, & Miles, 1998) . In order to predict the mass and energy flows associated with these reactions, it is necessary to utilize knowledge of fuel properties and how those fuel properties effect the combustion reaction. In order to produce a reasonably accurate assumption of combustion conditions that are sufficient to meet the stated aims of the decision support tool , t he combustion reaction is presumed to proceed based on the basis of complete stoichiometric combustion as outline d in this section . Energy generation from the reaction is calculated based on the lower heating value of coal and biomass in accordance with the relationships developed by Basu 2013 as well as VanLoo and Koopejan 2008. (Basu, 2013; Van Loo & Koppejan, 2008) . The first step in modeling the combustion process of biomass fuel is to unders tand the chemical and physical composition of the fuel source . The primary methods are to perform an ultimate and proximate testing analysis of the feedstocks . Ultimate analysis is the laboratory defined elemental composition of the biomass which include s C, H, N, O, S and a sh content on a percent weight basis. Proximate analysis defines the combustion characteristics of the fuel by calculating the value of gross components such as moisture content, volatile matter, fixed carbon, and ash content on a per cent weight basis. V olatile matter is defined as the condensable and non - condensable vapor released when the fuel is heated, ash is defined as the inorganic solid residue left after the fuel is completely burned, 28 and fixed carbon is defined as the remaini ng portion of the biomass which cannot be defined as volatile matter, moisture or ash. For the purposes of this study, ultimate and proximate analyses were derived from the E nergy Centre of the Netherlands (ECN) Phyllis 2 database (ECN, 2015) , which aggregates ultimate and proxi mate analysis from internal laboratory work and literature . At present, the database has in excess of 3000 entries . The database further allows for the averaging and statistical analysis of multiple studies on similar type feedstocks. The values present ed in Table 3 represent the average values established by the ECN Phyllis2 database for the given biomass fuels. This t able is utilized by CREDIT as a VLOOKUP t able to determine feedstock chemical composition of user defined feedstocks. Table 3 : Ultimate and Proximate Analysis Values Used for Modeling as determined by ECN's PHYLLIS Database (ECN, 2015) Ultimate Analysis (%dm) Proximate Analysis (%dm) C H S O N Ash Fixed Carbon Volatile Matter Hybrid Poplar 49.4% 6.0% 0.1% 43.1% 0.2% 1.2% 13.7% 85.1% Willow Wood 49.9% 5.9% 0.1% 41.8% 0.6% 1.7% 16.1% 82.2% WWTP Biosolids 34.0% 4.9% 1.3% 20.0% 4.7% 35.0% 11.5% 53.5% Wheat Straw 46.0% 5.5% 0.1% 41.4% 1.7% 5.0% 0.0% 0.0% Switchgrass - Baled 47.8% 5.8% 0.1% 35.1% 1.2% 10.1% 0.0% 0.0% Miscanthus - Baled 44.9% 5.4% 0.1% 40.3% 0.5% 4.6% 19.5% 71.5% Hybrid Poplar (Torrefied) 53.0% 5.5% 0.0% 37.9% 0.5% 3.1% 24.7% 72.2% Willow Wood (Torrefied) 53.0% 5.5% 0.0% 37.9% 0.5% 3.1% 24.7% 72.2% Hybrid Poplar (T & P) 53.0% 5.5% 0.0% 37.9% 0.5% 3.1% 24.7% 72.2% Willow Wood (T & P) 53.0% 5.5% 0.0% 37.9% 0.5% 3.1% 24.7% 72.2% Switchgrass - Char 50.5% 2.8% 0.1% 19.7% 1.2% 28.2% 0.0% 0.0% Switchgrass - Pelleted 47.8% 5.8% 0.1% 35.1% 1.2% 10.1% 0.0% 0.0% The energy generated from the combustion of coal and biomass is predicated on the net calorific value or lower heating value (LHV) of the combined biomass and coal blend. Lower heating values of coal and biomass can be ascertained through proximate and ul timate analysis of the given feedstocks. To do this, the higher heating value of the fuel (HHV), also called the gross calorific value must be calculated. 29 Channiwala and Parikh (Channiwala & Parikh, 2002) d eveloped Equation 1 for the following unified correlation for HHV based on 15 existing correlations and 50 fuels, including biomass, gas, and coal . HHV = 349C + 1178.3H + 100.5S - 103.4O - 15.1N - 21.1ASH kJ/kg Equation 1 : Highe r Heating Value of Fuels (Channiwala and Parikh 2002) w here C, H, S, O, N, and ASH are percentages of carbon, hydrogen, sulfur, oxygen, nitrogen, and ash as determined by ultimate analysis on a dry basis. Using the as received analysis proximate analysis of the feedstock in conjunction with the calculated HHV , the LHV or net calorific value can be determined using Equation 2. kj/kg Equation 2 : Lower Heating Value of Fuels (Basu 2013) where LHV, HHV, H, and M are lower heating value, higher heating value, hydrogen percentage, and moisture percentage, respectivel y, on an as received basis. Here, h g is the latent heat of steam in the same units as HHV. The latent heat of vaporization when the reference temperature is 100 o C is 2260 kJ/kg (Basu, 2013) . Energy generation is calculated based on the LHV of constituent feedstocks because it accounts for the latent heat of vaporization of water. This is a critical factor to account for when considering biomass as a co - feedstock due to its relatively high moisture content in comparison with traditional fossil fuels. Figure 1 1 outlines the generalized mass balance for a pulverized coal fired power plant. The primary inputs are the biomass, coal, and atmospheric air . Water and st eam are used as a medium to deliver energy from the boiler to the turbine. In conventional operation though, water from this process is conserved and recycled in a continuous loop. For the purposes of this study, which is primarily focused on the utiliza tion of biomass and coal, it is assumed that 100% of process water is conserved. 30 Figure 11 : Generalized Mass Flow Diagram of a Coal Fired Power Plant Utilizing Co - Firing In order to perform the mass balance in an academically appropriate format that meets the tool goal of generating supported approximations of process parameters, the CREDIT model utilizes the stoichiometric relationship of complete combustion in pulverized fuel boilers as established by Basu, 2013 and Van Loo & Koopejan, 2008 . The first calculation to be performed is the determination of the mass of dry air that will be required to combust the given fuel under ideal conditions. Assuming that dry air cont ains 23.16% oxygen, 76.8% nitrogen and 0.04% inert gasses by mass, Basu states that the mass of dry air required can be found using the E quation 3. kg/kg of dry fuel Equation 3 : Mass of Dry Ai r (Basu, 2013) w here C, H, O, and S are the mass fractions of their respective elements on a dry matter basis. In CREDIT, these values are derived from the proximate analysis values previously defined for the chosen 31 feedstock . This value is, in turn, used to calculate the t otal mass of air needed to perform combustion using E quation 4 . Equation 4 : Mass of actual air required (Basu, 2013) w here EAC is the mass of excess air used in the combustion reaction and X m is the moisture cont ent of the boiler air. In C REDIT, EAC and X m are de termined using listed values for the boiler in question defined by the EIA boiler summary embedded in the model (EIA, 2012) . The total mass of flue gas (W c ) generated can then be calculated using Equation 5. kg/kg dry fuel Equation 5 : Mass o f Total Flue Gas (Basu, 2013) Using the same methodology, Basu further states that the mass of several key flue gas constituents can be calculated using E quations 6 - 10. kg/kg dry fuel Equation 6 : NOx Calculation (Basu, 2013) kg/kg fuel Equation 7 : Carbon Dioxide produced from fixed carbon in Flue Gas (Basu, 2013) kg/kg dry fuel Equation 8 : Water Vapor in Flue Gas (Basu, 2013) kg/kg fuel Equation 9 : Oxygen in Flue Gas (Basu, 2013) 32 kg/kg fuel Equation 10 : SO2 in Flue Gas (Basu, 2013) C o - blending of feedstocks has the potential to result in incomplete combustion of fuel due to the reduction in boiler op erating temperature, as well as the change in air injection ratios. As this model does not utilize computational fluid dynamic or advanced reaction kinetics and subsequently assumes complete combustion, it is necessary to account for this contingency in a nother way to accomplish the ultimate goal of CREDIT; which is to provide information to support decision making . This is accomplished through comparative examination of the process parameters required for complete combustion to the EIA defined boiler cap abilities. Key amongst these parameters are excess air injection rate capacity and fuel feed rate capacity. If the stated ratings for these parameters are sufficient to support complete combustion of the proposed feedstock blend, it is assumed that compl ete combustion will be feasible with the budgeted modifications to the boiler. If either required parameter exceeds the boiler rating provided by the EIA, a notification is generated for the user specifying that incomplete combustion may result from the p roposed project parameters. Ash is collected in 2 primary locations; the boiler bottom (bottom ash) and particle captured from the flue gas (fly ash). Because this model assumes complete combustion of fuel , ash is calculated using Equation 11. Ash out = M airin +Mass CoalIn + Mass BiomassIn Mass FlueGas Equation 11 : Generalized Mass Balance This model does not differentiate between bottom ash and fly ash generation, but , in general , bottom ash accounts for 90% of the ash total in pulverized fuel boiler systems. 33 4.3.2 Biomass Depot Modeling Figure 1 2 displays the generalized process flow diagram associated with the biomass depot investigated in this model. By adding or subtracting process elements or pieces of equipment, it is possible to arrive at several different configurations for a biomass depot pant. For the purposes of CREDIT and this study, four basic configurations are considered based on their p racticality and appearance in previous studies. These are : 1. Biomass dried, torrefied, and densified to produce a torrefied pellet (TOP) 2. Biomass dried and pelleted 3. Biomass pelleted as received 4. Biomass aggregated and stored onsite without modification (not p ictured) Figure 12 : Biomass Depot Process Configuration 34 Functionally, the modeling of these configurations is accomplished - conditionals within the excel spreadsheet. Thus, as the user defines the configuration, different process. 4.3.3 Torrefaction Torrefaction is a relatively complex process to model on the chemical level, as the condition of torref ied biomass is dependent upon the type of torrefier, temperature of torrefaction, duration of torrefaction, rate of oxygen infusion in to the process, and physical/chemical characteristics of the biomass being torrefied (Joshi, de Vries, Woudstra, & de Jong, 2014) (Park, Meng, Lim, Rojas, & Park, 2013) . It is unlikely that the user of this decision support tool inherently know s the desired values of these parameters . Even if they are known , combinations of these parameters are not studied thoroughly enough to produce an empirical model to predict the ultimate and proximate analysis of biomass based on their manipulation. As such, a more practical approach was employed by CREDIT whereby the entrance and exit conditions of the biomass into the torrefaction u nit are set as user defined givens and the physical and energy requirements for the process are back - calculated using the mass balance of the reaction as defined in Equation 12 . Eq uation 12 : Torrefaction Mass Balance w here m x is the mass of the respective mass flow component. The mass and moisture content of the biomass stream is defined by the previous steps, however , the torgas /torrefied biomass ratio i s relatively difficult to determine theoretically as it relies on variables such as torrefier type, operation temperature, residence time, and feedstock composition. These variables, in turn, are specifically chosen by plant operators in order to achieve specific properties of torrefied biomass. Thus, in order to 35 model this reaction, a number of assumptions are made. For the purposes of this study, it was assumed that the torrefaction plant would produce torrefied biomass at a yield rate consistent with previous studies defined by Table 4 as shown in Equations 14 and 15. Equation 13 : Torrefaction Mass Yield Equation 14 : Torrefaction Energy Yield Using the values listed in Table 4 , in combination with the yield rates defin e d in E quations 14 and 15 , it is possible to determine the mass yields of both torrefied bi omass and the torrefaction gas . Table 4 : Torrefaction Yield and Efficiency Relation to Moisture Content (Batidzirai et al., 2013) Moisture Content Model Values (torrefaction plant) Theoretical Values (Torrefaction Plant) Thermal Efficiency (%) Feedstock to Product Ratio Thermal Efficiency (%) Feedstock to Product Ratio 20% 96.4 1.59 98.0 1.35 30% 95.4 1.84 97.6 1.70 35% 94.8 1.99 97.3 1.88 40% 93.8 2.18 96.9 2.11 45% 92.7 2.41 96.3 2.35 50% 91.0 2.72 95.6 2.67 55% 88.6 3.11 95.0 3.08 Using t he process stream masses, along with associated LHVs of the constituent components, it is possible to calculate the thermal energy required to perform the torrefaction reaction using E quation 16. 36 Equation 15 : Torrefaction Energy Balance ( Batidzirai et al 2013) wh ere n torr is the thermal efficiency of the torrefaction reactor as de fined in Table 4 and E torr is the required energy input for torrefaction. In CREDIT, it is assumed that E torr will be supplied through the combustion of natural gas. 37 4.4 Financial Modeling One of the key aspects of the model was to determine the financial viability of proposed projects . In order to accomplish this, CREDIT estimates capital costs and operating costs for the proposed depot system and coal fired power plant. The methods and assumptions relating financial considerations are found in this sec tion. 4.4.1 Power Plant Capital Costs Power plant capital costs will vary greatly depending upon the selected biomass type, biomass blending rates, and energy incorporation strategy (direct vs. indirect combustion). This is due to the different types of infrastructure that will be required based upon these factors. In general, these infrastructure needs can be broken down into 5 categories ; capital costs of bo iler/generator modifications (CI mod ) biomass storage costs (CI BS ), biomass handling equipment (CI BH ) biomass conditioning equipment (CI CD ) , and boiler construction (CI BC ) . For example, if a power plant with a PC boiler is to co - fire chipped wood with a m oisture content of 30% at a blending ratio of 10% , by energy value , the biomass must first be dried and pulverized separately from coal because the PC system cannot incorporate more than 3% biomass directly into the coal pulverizer due to concerns of produ ct matting (FEMP, 2004) . Conversely, a parallel fired system would not need such conditioning equipment as the boiler would be built to handle the biomass. However, a separate cost would be asse ssed for the construction of the new boiler. Both systems would require separate biomass storage and handling systems. Thus, the capital costs for the direct fired system would be CI mod + CI BS + CI BH + CI BC , whereas the costs for the parallel fired syste m would be CI BC + CI BS + CI BH . This logic leads to several permutations of system capital cost requirements which are outlined by the logic diagram found in Figure 13 . 38 Figure 13 : Pulverized Coal Equipment Decision Tree - statements associated with user inputs regarding biomass type, pre - conditioning, and plant setup. The capital costs associated with the aforementioned infra structure categories were calculated by scaling the costs of known existing operations using a power scaling equation of the general form , as shown in Equation 20. Equation 16 : power scaling equation proposed system, and alpha is a scaling factor ranging from 0 - 1. Alpha values as well as referen ce pricing is readily available within existing literature. For power plant costs, CREDIT utilizes the reference costs and alpha values established by (De & Assadi, 2009) as well as (Caputo et al., 2005) . This is a 39 common methodology deployed in techno - economic ana lyses of complex systems where finding pricing of individual components for every conceivable size is impractical. Equations 17 - 21 illustrate these relationships in reference to the previously defined capital cost categories. Equation 17 : Boiler Modification Costs (Caputo et al., 2005; De & Assadi, 2009) Equation 18 : Biomass Storage (Caputo et al., 2005; De & Assadi, 2009) Equation 19 : Biomass Handling Equipment (Caputo et al., 2005; De & Assadi, 2009) Equation 20 : Biomass Conditioning Equipment (Caputo et al., 2005; De & Assadi, 2009) 4.4.2 Power Plant Operation and Maintenance Costs Maintenance and labor costs were estimated at eight percent of total capital costs annually (Batidzirai et al., 2013) . Additionally, replacement of the drier and biomass handling units was assumed to be required in year 10 . The cost of replacement was assumed to be equivalent to the purchase prices assessed in year 0 . 40 The summation of O&M costs were escalated by 1.5 percent each year through year 2 0 to account for inflation and the real value of money. This approach follows the practice of the American Society of Agricultural and Biological Engineers (Binkley, 2010) 4.4.3 Biomass Depot Capital and Maintenance Costs Depot capital costs were determined by applying the power scaling method defined in S ection 3.5.1. to individual components of the biomass depot process flow. As with the power plant, not all capital investments were necessary for each scenario. In this case, the required equipment is not determined by a logic flow structure but rather by user defined inputs . I f the user sp ecifies that the product reaching the power plant is torrefied and pelleted, a torrefier and a palletization system are added to the capital costs. The reference costs and capacities of individual components associated with this project are derived from t he literature (Srivastava et al., 2011) . 4.4.4 Biomass Depot Operation and Maintenance Costs Maintenance and labor costs were estimated at eight percent of total capital costs annually (Batidzirai et al., 2013) . Additionally , replacement equipment for the torrefaction reactor, drier, and pelletization unit was assumed to be required in year 10 and was v alued at the cost estimated by CREDIT in the investment cost module. Additionally, feedstock costs , electricity costs, and natural gas costs were calculated to scale with the plant. Electricity prices and natural gas prices are user defined values in the model . For the purposes of this study, these utility costs were set to May 2015 values as determined by the US EIA for the state of Michigan . The summation of O&M costs were escalated by 1.5 percent each year through year 20 to account for inflation and the real value of money . This ap proach follows the practice of the American Society of Agricul tural and Biological Engineers (Binkley, 2010) . 41 4.4.5 Net Present Value ( NPV ) A 20 year pro - forma analysis was produced based on the capital and operating costs for the economic analysis portion of CREDIT . The resulting NPV indicated whether the power plant investment generate s a pos itive or negative return on investment . In general, a positive NPV indicates that the project should be considered and a negative NPV indicated the opposite. The pro - forma and subsequent NPV calculation were based on projected electricity sale revenues and cost flows tracked within CRED IT . T he retail price of electricity was indexed to the real industrial price of electricity provided by the E IA . All scenarios analyzed in this study were assumed to be financed entirely by loans at an interest rate of 7 % over the 20 year useful life of the project . Grants funds considered within CREDIT as a revenue realized in year 0 . A salvage value was added to the cash flows in year 20 and any value that exceeded the remaining depreciation balance was valued as a capital gain. The net present value of the power - plant capital investment was calculated by Equation 21 derived by Binkley et al, 2015 , Equation 21 : Net Present Value Calculation (Binkley et al., 2015) where : = The net present value in year i = Electricity sales revenue = Taxable grant funding = Value of carbon credits = Operational expenditures including repairs, labor, and insurance. = Loan interest = Carryover losses 42 = Marginal tax rate = Grant funding = depreciation of equipment = loan principal = initial cash investment = = Opportunity cost of capital = Project useful life 4.4.6 Levelized Cost of Electricity (LCOE) In order to determine the levelized cost of electricity, a goal seek routine was added to the model to determine the saleable price of electricity such that the power plant NPV is equal to 0 after 20 years. The effective equation for this optimization takes the general form: Equation 22 : Generalized LCOE Equation Where N is the project lifetime in years, CAPEX is the capital expenditure made in year zero, OPEX is the sum of all operational expenditures in year i, r is the opportunity cost of capital investment, and e is the specific energy yield of the power plant in kWh for year i. 4.4.7 Energy Returned on Energy Invested Another key metric assessed by this model is energy returned on energy invested (EROEI) , Equation 22. Equation 23 : Generalized EROEI Equation Energy return on investment is a ratio of the useful energy obtained (electricity or combined heat and power) versus the energy invested in a system (i.e. transport fuel, feedstock processing, and feedstock 43 handling). A non - fractional value denotes a viab le project, whereas a fractional EROEI means the project is an energy sink . EROEI in this study is used to eliminate potential projects that may be financially viable, and theoretically meet renewable energy generation requirements, but in actuality requi re more energy to operate than they produce. It is also possible to compare the EROEI of this project with the EROEI of similar energy technologies such as those listed in Table 5 . Table 5 : EROEI for Common Fuels (Murphy & Hall, 2010) EROI Fuel 1.3 Biodiesel 3.0 Bitumen tar sands 80.0 Coal 1.3 Ethanol corn 5.0 Ethanol sugarcane 100.0 Hydro 35.0 Oil imports 1990 18.0 Oil imports 2005 12.0 Oil imports 2007 20.0 Oil production 10.0 Natural gas 2005 10.0 Nuclear (with diffusion enrichment ) 50 - 75 Nuclear (with centrifuge enrichment ) 6.8 Photovoltaic 5.0 Shale oil 1.6 Solar collector 1.9 Solar flat plate 18.0 Wind 4.5 Statistical Methods In biological systems there is inherently high variation and uncertainty with material properties and kinetics. For this model, these uncertainties are most clearly manifested in the variation of biomass chemical composition and produc tion rates (or yield). As outlined in the previous chapter, biomass physical and chemical composition plays a key role in determining the energy potential of the biomass, reaction kinetics of the combustion reaction , and combustion products. Similarly, w hen determining the amount of biomass available to a depot or a power plant, it is critical to understand the crop yield . 44 However, the yield can vary greatly from year to year and region to region. Consequently, it is important to understand the natural v ariation associated with these key biological components. Data gathering activities are outlined in this section. 4.5.1 Fuel Variation Fuel data was collected from both the NREL biomass feedstock database and the ECN Phyllis2 biomass feedstock database. Both a re web based databases that catalogue the results of internal and published analyses of biomass feedstocks. A simple statistical analysis was performed on this data to determine variation between samples. Table 5 provides a sample of this analysis for sw itchgrass. . A complete list of feedstock t able s is available in the appendix. Table 6 : Sample Fuel Properties Variation - Switchgrass Fuel Properties Unit Min Max Median Mean Std dev Samples Proximate Analysis Moisture content wt% (ar) 8.2 15.0 11.9 11.7 2.8 24% 5 Ash content wt% (dry) 1.9 10.1 6.3 6.3 1.4 22% 34 Volatile matter wt% (daf) 72.9 86.9 84.3 83.2 4.5 5% 8 Fixed carbon wt% (daf) 13.1 27.1 15.8 16.8 4.5 27% 8 Ultimate Analysis Carbon wt% (daf) 45.2 53.2 50.6 49.4 2.5 5% 13 Hydrogen wt% (daf) 5.6 6.5 6.1 6.1 0.4 6% 13 Nitrogen wt% (daf) 0.4 1.3 0.6 0.6 0.2 28% 30 Sulphur wt% (daf) 0.0 0.2 0.1 0.1 0.1 45% 13 Oxygen wt% (daf) 39.0 48.6 43.7 44.0 2.9 7% 13 Total (with halides) wt% (daf) 0.0 101.8 0.6 38.7 49.3 127% 34 Calorific Values Net calorific value (LHV) MJ/kg (daf) 16.9 18.9 17.7 17.8 0.7 4% 12 Gross calorific value (HHV) MJ/kg (daf) 18.3 20.2 18.9 19.2 0.7 4% 12 HHVMilne MJ/kg (daf) 16.9 21.6 19.5 19.5 1.2 6% 13 Halides Chlorine (Cl) mg/kg (daf) 370 5249 1062 1952 1943 1 5 Major elements Potassium (K) mg/kg (dry) 3400 3400 3400 3400 0 0 1 Sodium (Na) mg/kg (dry) 33 33 33 33 0 0 1 45 The results from these analyses were used to derive the anticipated distributions of biomass composition , which were in turn used for iterations of the Monte Carlo analysis discussed in the next section. 4.5.2 Biomass Yield Variation In order to account for the uncertainty associated with biomass yield, regional data was procured from active field studies in the investigated study area. MSU Extension Educators have managed 4 field plots within Isabella county since 2009. Table 7 disp lays the summary of field test results for the mature switchgrass crop harvested in 2014 (Pe nnington, 2015) . Though the sample size is not large, it is the best representative sample available for region specific production rates of switchgrass. As with the biomass chemical analysis, variation s noted in these studies can be incorporated into the Monte Carlo analysis described in the next section. Table 7 : 2014 MSU Field Plot Data for Isabella County - Yield and Dry Matter % Dry Matter Yield (dry lb/acre) Mean 59.5% 7.73 Standard Error 0.5% 0.21 Median 59.7% 7.64 Standard Deviation 1.0% 0.41 Sample Variance 0.0% 0.17 Range 2.3% 0.96 Minimum 58.1% 7.35 Maximum 60.4% 8.32 Count 4 4 Figure 1 4 displays the average annual yield for the 4 test plots represented in Table 7 . As can be seen, the plot increases yield over time. This is common among perennial grass energy crops, as full matur ity does not occur until year 4 or 5 of establishment. This natural variation will also need to be incorporated into any analysis of feedstock production potential. 46 Figure 14 : Isabella County Switchgrass Annual Yield 4.5.3 Monte Carlo Analysis For models with multiple parameters with ranges of variability it is helpful to visualize possible distribution s through a Monte Carlo simulation analysis. Monte Carlo analysis is used to simulate random variation in sets of related variables. In order to run a Monte Carlo analysis, a range of potential values must be established for each studied variable. In the case of this study, the range for the inve stigated input parameters was determined by values found in the literature. Where a sui t able range of values were not available, a range of ±20% from the default value was assumed. The Monte Carlo analysis then selects one value randomly from the given r anges for each parameter. The new set of input values generated from this activity is then used to calculate a result from the model under investigation. The process of randomizing variables over a range and re - running the model and recording the results is repeated until outcome distributions can be used to predict the result within a pre - determined error range (O'Donnel, Hickner, & Barna, 2002) . Monte Carlo methods can be 0 2 4 6 8 10 2008 2009 2010 2011 2012 2013 2014 2015 Yield (dry lb/acre) Year Isabella County Switchgrass Field Studies 47 particularly useful to help ascertain the risk associated with a model that relies on a large nu mber of independent variables such as CREDIT. For the purposes of this study, 16 input variables were defined for investigation with the range of values shown in Table 8 . A Monte Carlo simulation was performed for a 1000 iterations of the random variable assignment using an excel macro embed ded within CREDIT. The value of 1000 iterations was selected through operator attempts to reduce error while keeping model runtime low (<1 hr. ). Results from the analysis were subsequently tested for normality of distribution and used to calculate confid ence intervals for CREDIT results . Outputs investigated were NPV of the Power Plant, plant - gate feedstock pricing, and cost of CO 2 mitigation. Table 8 : Sample of Distribution Parameters for Monte Carlo Analysis Default Min Max Feedstock Production (kg/hectare/ yr. ) 1,120,622 896,498 1,344,746 Raw Biomass Moisture Content (% w.b.) 2 0% 10% 40% Biomass LHV (kJ/kg) 18,780 18,240 19,510 Market Electricity Price ($/kWh) $ 0.12 $ 0.08 $ 0.14 Farmgate Feedstock Price ($/dt) $ 0.05 $ 0.04 $ 0.05 Natural Gas Price ($/MJ) $ 0.003 $ 0.002 $ 0.003 Specific Transportation Cost ($/km) $ 2.25 $ 1.80 $ 2.70 Plant O&M Cost ($/kW installed capacity) $ 47.60 $ 38.08 $ 57.12 Boiler Modifications ($ @ ref case) $ 6,813 $ 5,450 $ 8,176 Plant Storage ($ @ ref case) $ 862,221 $ 689,777 $ 1,034,665 Plant Handling ($ @ ref case) $ 1,311,789 $ 1,049,4 31 $ 1,574,146 Plant Conditioning ($ @ ref case) $ 86,148 $ 68,918 $ 103,377 Depot O&M Costs (% of capital costs) 8% 6% 10% Depot Cap Costs $ 7,129,364 $ 5,703,491 $ 8,555,236 Boiler Efficiency (Thermal) 90% 80% 99% Torrefier Efficiency (Thermal) 90% 80% 99% Drier Efficiency (Thermal) 90% 80% 99% 48 5 RESULTS AND DISCUSSION The methods outlined in the previous section were applied to generate an excel based decision support tool for the co - firing of biomass in coal fired power plants. This section details the results of the verification and validation activities applied to the decision support tool, as well as the results of four scenario analyses that were conducted for a Michigan coal fired power plant using CREDIT. 5.1 Model Verification and Validation As with all decision support tools, CREDIT is an approximate imitation of a real world situation. As such, it is the intent of CREDIT to produce reasonable approximations of economic and environmental impacts that would be realized by implementing co - firi ng of biomass at existing coal fired power plants. The extent to which all models and decision support tools can accomplish this is governed by model verification and validation. verification of a model is the process of co nfirming it is correctly implemented with respect to the conceptual model ( i.e. it matches specifications and assumptions deemed accep t able for the given purpose of application) (Sargent, 2011) . Verification for CREDIT development has taken place iteratively throughout the development process by consulting with experts in the fields of economics, combustion, biomass harvesting, and biomass harvesting. It is further verified via the logical structure that was developed through the literature review and represented in the logic flow diagrams Figure 15 displays a sample logic diagram for depicting the required equipment for various biomass conditions at a pulverized coal plant . 49 Figure 15 : Pulverized Coa l Co - Firing Equipment Configuration Logic Tree The model has further been verified through cross comparing model parameter s with literature values such as feedstock composition, production rates, equipment efficiency, equipment cost and others to determine reasonable ranges of potential values . These values have in - turn been incorporated into the Monte Carlo simulations already outlined in order to determine a reasonable margin of error for predicted results. Validation of the model is of the real system (Sargent, 2011) . For many models, it is possible and highly preferable to accomplish this through direct experiment al testing of the model. However, it is difficult to fully validate CREDIT based on direct experimentation. Even one iteration of direct testing could cost several million dollars, many years, and would take the full cooperation of multiple entities. A feasible alternative is to run CREDIT with inputs from documented case studies and compare the predictive results against the 50 findings in the case study. This method has been used by other decision support tool developers in published literature (Binkley, 2010) . There are some drawbacks. First, it is improbable that published case studi es will supply all of the same required parameters and outputs by the model being tested. Thus it may be necessary to make standard assumptions for parameters not provided in the case study. Further, definitions of key terms and calculations may differ b etween case studies, thus making their results hard to compare. To combat this issue, case studies used for comparison were derived from a single entity (Tillman, 1997) in order to assure case study methods, calculations and terminology were standardized to the degree practicable. Finally, it can be difficult to compar e results across time and geographic location. To minimize this, studies for comparison were selected only from the United States and economic values were adjusted to reflect 2015 U.S. dollars. Despite these limitations it is important to note that Sargen t (Sargent, 2011) a model should be verified and validated to the degree needed for the model s intended purpose or application ate predictive power but rather to produce reasonable approximations of economic and environmental impacts that would be realized by implementing co - firing of biomass at existing coal fired power plants, sufficiently well to determine whether or not furthe r study and investment in the scenario is warranted . 5.1.1 Comparison to known scenarios For this this analysis, four case conducted by Tillman, T.A. (Tillman, 1997) were selected for comparison. The results are displayed in Table s 9 and 10 . Note that a few modifications were made to CREDIT in order to accurately track the EPRI cases. First, the EPRI studies were conducted in 1997 thus a cost conversion factor of 1.49 was applied to bring values to 2015 dollars. Secondly, the EPRI studies all assume a SO 2 credit of $80/ton . This was applied to CREDIT specifically for the purposes of this comparison, though it is largely irrelevant to current studies as 2015 SO 2 credits were traded at 51 $0.11/ton. Lastly, Tillman does not include the avoided cost of coal as a factor in his levelized cost of power calculation. CREDIT was adj usted for these studies to account for this in comparison, but it is not true of CREDIT as a whole. Results are presented in the Figures and Tables below. Table 9 : EPRI Base Cases: Pretax Cost of Power from Cofiring In PC Boilers ( Biomass Cofiring Guidelines - 2015 Dollars) Parameter EPRI Case 1 2 3 4 Technology PC PC PC PC Biomass Fuel Wood ( unspecified ) Wood ( unspecified ) Wood ( unspecified ) Wood ( unspecified ) Boiler Capacity (MW) 200 500 50 150 Cofiring %, Mass 5 5 20 15 Cofiring %, Heat 2.2 2.2 8.7 6.3 Coal Type Eastern Bituminous Western Bituminous Eastern Bituminous Eastern Bituminous Capacity Factor 0.75 0.75 0.75 0.75 Net Station Heat Rate on Coal (MJ/kWh) 11.08 10.02 11.61 10.55 Capital Cost for Cofiring System ($/kW) $ 49.00 $ 74.50 $ 342.70 $ 260.75 Biomass Cost ($/10 6 Btu) $ 1.27 $ 1.49 $ 1.27 $ 1.27 Biomass Cost ($/GJ) $ 1.20 $ 1.41 $ 1.20 $ 1.20 Biomass Cos ($/kg dry) $ 0.04 $ 0.05 $ 0.04 $ 0.04 Capacity Co - fired on Biomass (MW) 4.5 11.2 4.3 9.4 Pretax Levelized Cost of Power ($/kWh) $ 0.02 $ 0.02 $ 0.03 $ 0.03 52 Table 10 : CREDIT Model Outputs for Selected EPRI Cases Parameter CREDIT Case 1 2 3 4 Technology PC PC PC PC Biomass Fuel Poplar (40% moisture ) Poplar (40% moisture ) Poplar (40% moisture ) Poplar (40% moisture ) Boiler Capacity (MW) 200 500 50 150 Cofiring %, Mass 5 5 20 15 Cofiring %, Heat 2.0 2 8.9 6.5 Coal Type Eastern Bituminous Western Bituminous Eastern Bituminous Eastern Bituminous Capacity Factor 0.75 0.75 0.75 0.75 Net Station Heat Rate on Coal (MJ/kWh) 11.08 10.02 11.61 10.55 Capital Cost for Cofiring System ($/kW) $ 141.00 $ 112.00 $ 301.00 $ 235.00 Biomass Cost ($/10 6 Btu) $ 1.27 $ 1.49 $ 1.27 $ 1.27 Biomass Cost ($/GJ) $ 1.20 $ 1.41 $ 1.20 $ 1.20 Biomass Cos ($/kg dry) $ 0.04 $ 0.05 $ 0.04 $ 0.04 Capacity Co - fired on Biomass (MW) 4.0 10 4.45 9.75 Pretax Levelized Cost of Power ($/kWh) $ 0.02 $ 0.02 $ 0.04 $ 0.03 53 Figure 16 : Normalized Capital Cost Comparison Between EPRI and CREDIT Cases. Error Bars Reflect One Standard Deviation as Determined by Monte Carlo Analysis $- $50 $100 $150 $200 $250 $300 $350 $400 1 2 3 4 Capital Costs $/kW Installed Capacity Case Number Normalized Capital Cost Comparison EPRI Case CREDIT Case 54 Figure 17 : Levelized Cost of Electricity Comparison Between EPRI and CREDIT Cases. Error Bars Reflect One Standard Deviation as Determined by Monte Carlo Analysis 5.1.2 Validation Results and Limitations The model is moderately sensitive to capital cost input requirements at the power plant. As observed when compared with existing case studies, it can be insufficient to assume that any particular piece of the cost of individual equipment components based on throughput capacity, however, other site specific factors can have a significant impact on overall pricing of the equipment. These factors include, but are not limited to, space constraints, age of existing infrastructure, and plant layout. Any or all of these factors can increase the total project cost. For instance, in EPRI case 2, the estimated no rmalized capital cost for the plant redesign was $112/kWh whereas the observed capital cost was $75/kWh installed, case 3 that the predicted normalized ca pital cost of project implementation was $301 /kW for the $- $0.005 $0.010 $0.015 $0.020 $0.025 $0.030 $0.035 $0.040 1 2 3 4 Lelized Cost of Electricity $/kWh Case Number Pretax LCOE Comparison EPRI Case CREDIT Case 55 CREDIT case and $341 /kW for EPRI - conservative. This was due to case 2 having certain pre - processing infrastructure already on site (dryer, screeners) and case 3 ne eding additional engineering work to put a direct injection line into the boiler from a greater than average distance. In short, it can be observed that high levels of variation in costing exist due to the nature of highly case - specific conditi ons found at each power plant. CREDIT is capable of adjusting to these variations if sufficient background data is present . This level of predictive power is consistent with the intent of the model, which is to support project decision making by iteratively asses sing project outcomes depending upon the degree of information available to the user. 5.2 Scenario Setup In order to utilize the resulting model, four scenarios surrounding the J.H. Campbell power plant were investigated . These scenarios involve the use of un used or marginal farmland derived from a 4 county area north of J.H. Campbell to produce dedicated energy crops for co - firing. In the analysis, the crops are assumed to be aggregated at a processing depot prior to shipment to the power plant . 5.2.1 J.H. Campbel l Power Plant The Campbell power plant complex is located on a 2,000 - acre site along the Lake Michigan shoreline near West Olive, Mich igan ( Figure 1 8 ) . Originally commissioned in 1962 , it houses three pulverized coal tangentially fired boiler units that feed a 906MW steam turbine ( Figure 1 9 ) . The system uses a blend of bituminous and sub - bituminous coal and does not presently co - fire with biomass or supplement combustion reaction s with natural gas or petroleum (EIA, 2012) . 56 Figure 18 : J.H. Campbell Power Plant (Courtes y of Consumers Energy) Figure 19 : J.H. Campbell Power Plant Generalized Process Flow Diagram (courtesy of Consumers Energy) 5.2.2 Feedstock Collection Michigan is home to several types of biomass that can be readily purchased for negotiable prices such as forest slash, pulpwood and C & D debris . These can easily be run solely through the power plant portion of the model by inputting biomass quantities, physical characteristics, and price in order to determine desired blending rati os, energy outputs, CO 2 reductions, and financial outputs . However, in order to display the full capabilities of CREDIT, the proposed scenario will involve the harvest and collection of farmed woody biomass from local sources. 57 In order to conduct this p reliminary analysis, land availability data was gathered from the national agricultural statistics service for 4 viable counties in Michigan that represent a sample collection area for harvested woody biomass. These are: Isabella, Osceola, Mecosta , and C lare counties ( Table 1 1 ) . Table 11 : 2012 NASS Field Data Derived from (NASS, 2012) County Total Landmass (Acres) Total Farmed Land (Acres) Idle Farmland (Acres) Isabella 366,704 135,682 9,576 Osceola 362,256 53,638 6,362 Mecosta 355,090 71,606 4,701 Clare 361,021 25,356 2,597 Idle farmland is defined as land that was once used for crop production that is not presently in a crop rotation, forested, or used for pastureland. Assuming that a viable biomass processing plant would be able to draw from 80% of land that is presently idle, and 5% of currently cultivated land, it is calculated that 32,900 acres would be available for biomass energy crop production. As no value was found in the literature stating how much la nd conversion can be expected for a given scenario, these values were produced as educated guesses of the author. In order to account for the high degree of uncertainty in this value, a large range of ±30% was applied to this value in the Monte Carlo anal ysis. 58 Figure 20 : Study Collection Area 5.2.3 B iomass Production C apability: The aforementioned 4 county region was selected for 2 reasons. First, it has a sufficient amount of idle farmland to produce a relevant quantity of biomass energy, while its centroid is roughly 100 miles from the power plant. This is critical as it will allow for comparison with studies performed in other states that routinely set their observations at round distances such as 100, 200, or 300 miles. More importantly, the MSU research team has relevant field data relating to feedstock production rates in t his area. Specifically, MSU extension has been performing field trials on switchgrass production in Isabella county since 2009. The accumulation of this data will allow for region - specific assumptions about biomass growth and production rates. These rat es follow. 59 Poplar (5 dt/acre) (Srivastava et al., 2011) Switchgrass (7.73 dt/acre, MSU field trials in Isabella county) Given the total area of interest (total area of the 4 aforementioned counties), the total area of tillable land, and production rat es of each crop, it is possible to calculate the distribution density of biomass in the collection area using the generalized formula: Equation 24 : Biomass Distribution Density Equation Using the distribution density of biomass it is possible to estimate the total distance traveled by collection vehicles in order to aggregate biomass at the biomass depot using the following equation from (De & Assadi, 2009) . Equation 25 : Farm - Gate to Biomass Depot Distance Tra veled Equation (De & Assadi, 2009) Where M biomass is the total mass of biomass being transported in tonnes (wet basis), DD biomass is the previously calculated distribution density in tonnes/ km 2 /yr, and VC is the vehicle capacity of transport vehicles in tonnes. Using these values, and average transport distance of 80 km was calculated for switchgrass and an average distance of 81 km was calcula ted for poplar. 5.3 Scenario Analysis Figure 21 displays the relative differences in processes that will be assessed across all scenarios. For the purposes of this study, the in - field operations are estimated based on the works of previous studies and field trials conducted by other researchers at MSU . CREDIT is designed to process information from 60 f arm - gate to energy generation. However, data from other sources regarding planting, cultivation, and harvest were used by this study in order to better calculate the energy return on investment realized b y these 4 scenarios. Figure 21 : Scenario Outline 5.4 Scenario 1 Pelleted Poplar In the first investigated scenario plantation grown and chipped poplar is delivered to a biomass depot located 100 mile from the J.H. Campbell power pl ant. The biomass is assumed to be purchased from the individual farms at a price of $41.60 per dry ton of biomass ($0.056 per kg) (Srivastava et al., 2011) , and 61 transported via a 40 ton (36.3 tonne) capacity chip van to the centralize d processing depot. At the processing depot, the chipped biomass is dried to a moisture content of 10% and stored on site. At need, the biomass is then hammer milled, pelleted, and placed in storage once again before being transported to the power plant. Once at the power plant, the pelleted biomass is unloaded and stored separately from the coal (thus necessitating the addition of offloading equipment and storage silos . T he pelleted biomass is blended with the coal at a rate of 5% by energy value. As suggested in the literature (Nussbaumer, 2003) , pulverized coal boilers receiving >3% biomass by energy content will likely require a separate injection system for the biomass, d ue to its incompatibility with grinding equipment. This necessitates that a hammer mill be procured capable of reducing the pellets to a particle size of <0.25 micro meters . 5.4.1 Energy Assessment Energy return on investment analysis for scenario 1 was perfo rmed in accordance with the process flow outlined in Figure 6 . R esults are shown in Table 1 2 . In - field operations cultivation , including planting and harvest ing, were derived from literature values established (Dillen, Djomo, Al Afas, Vanbeveren, & Ceulemans, 2013) based on a 16 year study of poplar harvest on margina l cropland. The values reported were scaled to the production level of this scenario assuming linear growth. Uncertainties related to this method are captured in the Monte Carlo analysis . Transport from farm to depot and depot to farm are assumed to be p erformed by a 40 ton chip van. Energy values on a MJ/tonne/km were derived from the literature (Anon, 2010) and applied to the calculated transport distances for both aggregation and final shipping of the biomass feedstock. Electricity values for equipment were scaled based on equipment loading of previously derived literature values and natural gas requirements were calculated based on the energy needs of both the drier and torrefaction systems as detailed in E quation 16. 62 T he large and positive EROI predicted by this model suggests that this project is highly viable from an energy accounting standpoint. This is expected due to the fact that the biomass is lightly processed using these techniques when compared to other energ y sources. Table 12 : EROI Accounting for Scenario 1 LHV MMBTUs/yr GJ/yr In - Field Operations 190,341 200,821 Transport (Farm to Depot) 63,885 67,402 Depot Natural Gas for Drying (Depot) 308,119 325,083 Natural Gas for Torrefaction (Depot) - - Electricity for Drying (Depot) 6,806 7,181 Electricity for Torrefaction (Depot) - - Electricity for Grinding (Depot) 9,037 9,535 Electricity for Pelleting (Depot) 25,826 27,248 Transport (Depot to Power Plant) 203,868 215,092 Energy Return on Energy Invested (EROEI) 5 5 63 Figure 22 : Pelleted Poplar Monte Carlo Analysis 5.4.2 Economic Assessment The results of the economic modeling activities for scenario 1 are detailed in Table 1 3 . The primary metric derived from the economic assessment is the levelized cost of electricity or LCOE. This is the price at which the power plant could produce electricity and return a net present value of $0 assuming a required return on capital invest ment. For this scenario a return of 8% was specified. This yielded a LCOE of $0.04 86 . Monte Carlo analysis of this scenario at 1000 iterations, found the LCOE to be fairly normally distributed (Kurtosis = - 0. 3 , Skewness =0 .1) with a standard deviation of $0.005. 0 10 20 30 40 50 60 70 80 Frequency LCOE $/kWh Pelleted Poplar LCOE at 1000 Iterations 64 The LCOE of $0.04 86 is a favorable metric as new installations of other renewable energy sources such as wind and solar typically range from 10 - 20 cents per kWh. New biomass installations for 2015 are projected to be approximately 10 cents per kWh. Table 13 : Scenario 1 Economic Summary Power Plant Metrics Value Unit Retail Price of Electricity $0.120 $/kwh Simplified LCOE $0.0486 $/kWh Return on Capital Investment 8% NPV of Biomass Depot $ - NPV of Power Plant Modifications $168,205,000 Plant - Gate Feedstock Cost $91 $/tonne (ar) $101 $/tonne (dry) Relative Capital Costs $95 $/kW Capacity Additional Ash Disposal $ - In this analysis, one of the single largest contributors to LCOE was the cost of delivered biomass . Power plant upgrades can initially be costly but become trivial once amortized over the 20 year lifespan , in comparison with operating expenses . Interestingly, when the cost of processing biomass is investigated , it is found that the single highest cost at the processing depot is also the purchase of raw feedstock ( Figure 2 3 ) . 65 Figure 23 : Cost distribution of biomass processing for scenario 1 5.4.3 Emissions Analysis Emissions analyses were performed for both coal only and with the co - firing scenario s . Results are presented in terms of emissions reductions in Table 14 . Negative values reflect a net reduction in air emissions. Raw Feedstock Costs 49% Transport Costs (Farm to Depot) 7% Capital Investments 5% Natural Gas - Drying 5% Electricity Drying 1% Electricity Grinding 1% Electricity Pelleting 3% Operation and Maintainance 5% Tranport Coast (Depot to Plant) 24% Processed Biomass Cost Distribution 66 Table 14 : Scenario 1 Emissions Analys is Power Plant Summary Change in Emission Unit Notes Delta CO 2 (Co - Fire - Base Case) (15,211) kg/hr Gross (1,929) tonne/yr (44,961) kg/hr Net (Assuming Biomass is Carbon Neutral) (5,703) tonne/yr Delta SO 2 (216) kg/hr (Co - Fire - Base Case) (27) tonne/yr As expected, blending of biomass with coal resulted in a net reduction in both carbon dioxide and sulfur dioxide emissions. Carbon emissions were reduced both on a gross basis and a net basis. The gross basis values presented is a straight subtraction of CO 2 emissions expectations between the base case of only coal being fired in the boiler systems, and the scenario under investigation. The net basis assumes that any carbon introduced to the system by biomass is a carbon neutral so urce of energy as defined by (Hartmann & Kaltschmitt, 1999) . This value then is the CO 2 avoided by the replacement of coal. 5.5 Scenario 2 TOP Poplar In the second scenario , plantation grown and c hipped poplar is delivered to a biomass depot located 100 mile from the J.H. Campbell power plant. The biomass is assumed to be purchased from the individual farms at a price of $41.60 per dry ton of biomass ($0.056 per kg) (Srivastava et al., 2011) , a nd transported via a 40 ton (36.3 tonne) capacity chip van to the centralized processing depot. At the processing depot, the chipped biomass is dried to a moisture content of 10% and stored on site. At need, the biomass is then torrefied at 25 0 C. The to rrefaction reaction makes use of natural gas and the produced torrefaction gas to heat the reactor as shown in Figure 12 . Torrefied biomass is then hammer milled, pelleted, cooled , and placed in storage once again before being transported to the power plan t. Once at the power plant, the TOP biomass is unloaded and stored separately from the coal (thus necessitating the addition of offloading equipment and storage silos). Sources vary on whether or not 67 torrefied biomass can be stored with coal due to its s assumption is that separate storage will be required. TOP biomass is blended with the coal at a rate of that at this blending ratio, TOP biomass can be directly mixed with the coal prior to pulverization. Unlike scenario 1, this means the system will not need additional handling equipment, a hammer mill, or a separate boiler injection port. This greatly re duces overall capital costs and project footprint at the power plant site. 5.5.1 Energy Assessment Energy return on investment analysis for scenario 1 was performed in accordance with the process flow outlined in Figure 6 . The results of the energy accounting a re found in Table 1 5 . In - field operations cultivation (including planting) and harvest were derived from literature as discussed for scenario 1 . Uncertainties related to this method in addition to other uncertainties are captured in the Monte Carlo analy sis . Transport from farm to depot and depot to power plant are assumed to occur in the same manner as scenario 1 . Electricity values for equipment were scaled based on equipment loading of previously derived literature values, and natural gas requirements were calculated based on the energy needs of both the drier and torrefaction systems as detailed in E quation 16. F or this scenario CREDIT calculates an anticipated EROI of approximately 6. Though not large, the EROI is non - fractional and thus represents a viable project. However, as Figure 2 4 shows, the EROI distribution over the Monte Carlo analysis is scattered an d not normally distributed (Kurtosis >1). As such uncertainty is relatively high over the range of calculated EROI values . Though this is not ideal and it should be noted that all EROI values over the range are non - fractional and thus represent a valid p roject. 68 Figure 24 : Monte Carlo Histogram - Scenario 2 EROI Table 15 : Scenario 2 Energy Assessment LHV MMBTUs/yr GJ/yr In - Field Operations 186,345 196,604 Transport (Farm to Depot) 61,885 65,292 Depot Natural Gas for Drying (Depot) 301,650 318,258 Natural Gas for Torrefaction (Depot) 108,345 114,310 Electricity for Drying (Depot) 6,720 7,090 Electricity for Torrefaction (Depot) 16,246 17,140 Electricity for Grinding (Depot) 8,344 8,804 Electricity for Pelleting (Depot) 23,937 25,255 Transport (Depot to Power Plant) 179,614 189,502 Energy Return on Energy Invested (EROEI) 4 4 0 10 20 30 40 50 60 3.28497802 3.349997369 3.415016718 3.480036067 3.545055415 3.610074764 3.675094113 3.740113462 3.805132811 3.87015216 3.935171508 4.000190857 4.065210206 4.130229555 4.195248904 4.260268253 4.325287601 4.39030695 4.455326299 4.520345648 4.585364997 4.650384345 4.715403694 4.780423043 4.845442392 4.910461741 4.97548109 5.040500438 5.105519787 5.170539136 5.235558485 More Frequency EROI TOP Poplar EROI at 1000 Iterations 69 5.5.2 Economic Assessment The results of the economic modeling activities for scenario 2 are detailed in Table 1 6 . The primary metric derived from the economic assessment is the levelized cost of electricity or LCOE is the same as used in scenario 1: a net present value of $0 an d a return of 8%. This yielded a LCOE of $0.076. Monte Carlo analysis of this scenario at 1000 iterations, found the LCOE to be fairly normally distributed (Kurtosis = - 0.01, Skewness = - 0.3 8) with a standard deviation of $0.007. Figure 25 : Monte Carlo Histogram - Scenario 2 LCOE Though hi gher than its sister scenario (S cenario 1) , the LCOE of $0.0 5 6 is a favorable metric as new installations of other renewable energy sources such as wind and solar typically range from 10 - 20 cents per kWh with new biomass installations for 2015 are projected to be approximat ely 10 cents per kWh. As with S cenario 1, this is largely due to the fact that this scenario takes advantage of in - place infrastructure whereas other projects largely need to build new. In terms of capital cost at the power 0 10 20 30 40 50 60 70 80 90 100 Frequency LCOE $/kWh LCOE for TOP Poplar at 1000 Iterations 70 plant, S cenario 2 shows a clear advantage over S cenario 1. The similarity of TOP biomass to coal greatly reduces the need for capital investment in new handling, processing, and injection equipment (Bergman, 2005; Van Loo & Koppejan, 2008) . This is reflected in the low relative capital costs ($ 32 /kWh capacity vs. $9 7 /kWh for S cenario 1). Table 16 : Scenario 2 Economic Summary Power Plant Metrics Value Unit Retail Price of Electricity $ 0.120 $/kwh Simplified LCOE $ 0.056 $/kWh Return on Capital Investment 8% NPV of Biomass Depot $ - NPV of Power Plant Modifications $ 154,342,000 Plant - Gate Feedstock Cost $ 1 30 $/tonne (ar) $ 134 $/tonne ( dry ) Relative Capital Costs $ 32 $/kW Capacity As with S cenario 1 the single largest contributor to LCOE was the cost of delivered biomass. In S cenario 2 as the cost of delivered biomass reaches $1 30 per tonne (ar) due to the increased processing costs of torrefaction. Torrefaction scenarios may win out over pelleting alone where distances are long or where biomass characteristics are prohibitive to use in coal fired power plants without extreme modifications. However, in this case the benefits do not outweigh the costs according to CREDIT. As with S cenario 1, when the cost of processing biomass wa s investigated . Th e single highest cost at the processing depot is also the purchase of raw feedstock, though the ratio of feedstock costs to total costs is not as high. Operation and maintenance costs as well as capital investments are greatly increased in this scenario, due to the presence of the torrefaction reactor and its supporting systems ( Figure 2 6 ). 71 Figure 26 : Scenario 2 Processes Biomass Cost Distribution 5.5.3 Emissions Analysis Emissions analyses were performed for both the base case (coal only) along with the co - firing scenario. The comparative results are presented in terms of emissions reductions in Table 1 7 . Raw Feedstock Costs 42% Transport Costs (Farm to Depot) 6% Capital Investments 10% Natural Gas - Drying 4% Natural Gas - Torrefaction 1% Electricity Drying 1% Electricity Torrefaction 2% Electricity Grinding 1% Electricity Pelleting 2% Operation and Maintainance 13% Tranport Coast (Depot to Plant) 18% Processed Biomass Cost Distribution 72 Table 17 : Scenario 2 Emissions Analysis Power Plant Summary Change in Emission Unit Notes Delta CO2 (Co - Fire - Base Case) 44,280 kg/hr Gross 12,400 tonne/yr 1,573 kg/hr Net (Assuming Biomass is Carbon Neutral) (45,230) tonne/yr Delta SO2 (5,737) kg/hr (Co - Fire - Base Case) (122) tonne/yr As expected, blending of torrefied biomass with coal resulted in a net reduction in sulfur dioxide emissions. Carbon emissions were increased on a gross basis but decreased on a net basis. This change from scenario 1 is attributed to the change in chemic al composition of the biomass after torrefaction. During the torrefaction process, the ratio of carbon to oxygen is increased, and in this case resulted in a gross increase in carbon dioxide releases. Because this value is directly correlated to coal dis placement, and each scenario displaces the same amount of coal, net CO 2 reductions are similar across all scenarios. 5.6 Scenario 3 Pelleted Switchgrass In this scenario farm grown and baled switchgrass is delivered to a biomass depot located 100 mile s from the J.H. Campbell power plant. The biomass is assumed to be purchased from the individual farms at a price of $82.23 per wet ton of biomass ($0.091 per kg) (Anon, 2008) , and transported via a 20 ton (18.1 tonne) capacity flatbed trailer to the centralized processing depot. At the processing d epot, the switchgrass is de - baled, dried to a moisture content of 10% hammer milled, pelleted, and placed in storage before being transported to the power plant. Once at the power plant, the pelleted biomass is unloaded and stored separately from the coal (thus necessitating the addition of offloading equipment and storage silos). T he pelleted biomass is blended with the coal at a rate of 5% by energy value. As suggested in the literature (Nussbaumer, 2003) , pulverized coal boilers receiving >3% biomass by energy content will likely require a separate injection system for the biomass, due to its incompatibility with 73 grinding equipment. This necessitates that a hammer mill be pr ocured capable of reducing the pellets to a particle size of <0.25 inches (6.3 mm). 5.6.1 Energy Assessment Energy return on investment analysis for scenario 3 was performed in accordance with the process flow outlined in Figure 6 . The results of the energy a ccounting are found in Table 1 8 . In - field operations cultivation , including planting and harvest ing, were derived from literature values established in the literature (Anon, 2008) . The values reported ISU were scaled to the production level of this scenario assuming linear growth. Unce rtainties rel ated to this and other data collection methods are captured in the Monte Carlo analysis presented in Figure 2 7 . Figure 27 : Monte Carlo Histogram for Scenario 3 EROI Transport from the depot to the power plant is assumed to be performed by a 20 ton chip van. Energy values on a MJ/tonne/km were derived from the literature (Anon, 2010) and applied to the calculated transport distances for both aggregation and final shipping of the bi omass feedstock. 0 5 10 15 20 25 30 35 40 45 50 3.584410244 3.660455941 3.736501637 3.812547334 3.888593031 3.964638727 4.040684424 4.11673012 4.192775817 4.268821514 4.34486721 4.420912907 4.496958603 4.5730043 4.649049997 4.725095693 4.80114139 4.877187087 4.953232783 5.02927848 5.105324176 5.181369873 5.25741557 5.333461266 5.409506963 5.48555266 5.561598356 5.637644053 5.713689749 5.789735446 5.865781143 More Frequency EROI Pelleted Poplar EROI at 1000 Iterations 74 Electricity values for equipment were scaled based on equipment loading of previously derived literature values and natural gas requirements were calculated based on the energy needs of both the drier and torrefaction systems as detailed i n E quation 16. The positive large and positive EROI predicted by this model suggests that this project is highly viable from an energy accounting standpoint ( Table 1 8 ) . This is expected due to the fact that the biomass is lightly processed using these tec hniques when compared to other energy sources. Table 18 : Scenario 3 Energy Accounting LHV MMBTUs/yr GJ/yr In - Field Operations 193,642 204,303 Transport (Farm to Depot) 65,557 69,166 Depot Natural Gas for Drying (Depot) 313,462 330,720 Natural Gas for Torrefaction (Depot) - - Electricity for Drying (Depot) 6,877 7,255 Electricity for Torrefaction (Depot) - - Electricity for Grinding (Depot) 9,135 9,638 Electricity for Pelleting (Depot) 26,094 27,531 Transport (Depot to Power Plant) 207,384 218,802 Energy Return on Energy Invested (EROEI) 5 5 5.6.2 Economic Assessment The results of the economic modeling activities for scenario 3 are detailed in Table 1 9 . The primary metric derived from the economic assessment is the levelized cost of electricity or LCOE. For this scenario a return of 8% was specified. This yielded a LCOE of $0.0 65 . Monte Carlo analysis of this scenario at 1000 iterations, found the LCOE to be fairly normally distributed ( Kurtosis = - 0.4, Skewness = 0. 2 ) with a standard deviation of $0.00 8 ( Figure 2 8 ) . 75 Figure 28 : Scenario 3 Monte Carlo Histogram for LCOE The LCOE of $0.0 65 is a favorable metric for power providers searching for renewable energy capacity as new installations of other renewable energy sources such as wind and solar typically range from 10 - 20 cents per kWh ( Table 1 9 ) . This is somewhat higher than the cost of poplar due primarily to the relatively high price of raw feedstock. Table 19 : Scenario 3 Economic Assessment Power Plant Metrics Value Unit Retail Price of Electricity $ 0.120 $/kwh Simplified LCOE $ 0.0651 $/kWh Return on Capital Investment 8% NPV of Biomass Depot $ - NPV of Power Plant Modifications $ $129,774,000 Plant - Gate Feedstock Cost $ 1 15 $/tonne (ar) $ 128 $/tonne ( dry ) Relative Capital Costs $ 95 $/kW Capacity 0 10 20 30 40 50 60 70 80 90 100 Frequency LCOE $/kWh LCOE For Pelleted Switchgrass at 1000 Iterations 76 As with the previous scenarios the single largest contributor to LCOE was the cost of delivered biomass . Power plant upgrades can initially be costly in this scenario ($ 95 /kW of installed capacity), but once amortized over the 20 year lifespan of the project, they become trivial in comparison with operating expenses. The cost of delivered feedstock is also the single highest contributor of cost to biomass processors as well, as seen in Figure 28 . Figure 28 : Scenario 3 Biomass Depot Processing Cost Distribution Raw Feedstock Costs 59% Transport Costs (Farm to Depot) 6% Capital Investments 4% Natural Gas - Drying 4% Electricity Drying 1% Electricity Grinding 1% Electricity Pelleting 2% Operation and Maintainance 4% Tranport Coast (Depot to Plant) 19% Processed Biomass Cost Distribution 77 5.6.3 Emissions Analysis Emissions analyses were performed for both the base case (coal only) along with the co - firing scenario. The comparative results are presented in terms o f emissions reductions in Table 20 . Table 20 : Scenario 3 Emissions Report Power Plant Summary Change in Emission Unit Notes Delta CO2 (Co - Fire - Base Case) (20,700) kg/hr Gross (2,626) tonne/yr (45,230) kg/hr Net (Assuming Biomass is Carbon Neutral) (5,737) tonne/yr Delta SO2 (209) kg/hr (Co - Fire - Base Case) (27) tonne/yr As expected, blending of biomass with coal resulted in a net reduction in both carbon dioxide and sulfur dioxide emissions. Carbon emissions were reduced both on a gross basis and a net basis. Because this value is directly correlated to coal displacement, and each scenario displaces the same amount of coal, net CO 2 reductions are similar across all scenarios. 5.7 Scenario 4 TOP Switchgrass In t his scenario , farm grown and baled switchgrass is delivered to a biomass depot located 100 mile from the J.H. Campbell power plant. The biomass is assumed to be purchased from the individual farms at a price of $82.23 per wet ton of biomass ($0.091 per kg ) (Anon, 2008) and transported via a 20 ton (1 8.1 tonne) capacity flatbed trailer to the centralized processing depot. At the processing depot, the switchgrass is de - baled, dried to a moisture content of 10% , hammer milled, pelleted, and placed in storage before being transported to the power plant. Once at the power plant, the pelleted biomass is 78 unloaded and stored separately from the coal (thus necessitating the addition of offloading equipment and storage silos). Once at the power plant, the TOP biomass is unloaded and stored separately from th e coal (thus necessitating the addition of offloading equipment and storage silos). Sources vary on whether or not assumption is that separate storage wil l be required. The TOP biomass is blended with the coal at a rate that at this blending ratio, TOP biomass can be directly mixed with the coal prior to pulverization. Unlike scenario 3, this means the system will not need additional handling equipment, a hammer mill, or a separate boiler injection port . This greatly reduces overall capital costs and project footprint at the power plant site. 5.7.1 Energy Assessment Energy return on investment analysis for scenario 4 was performed in accordance with the process flow outlined in Figure 6 . The results of the energy accounting are found in Table 21 . In - field operations cultivation , including planting and har vest ing, were derived from literature values established in the literature (Anon, 2008) . The values reported were scaled to the production level of this scenario assuming linear growth. 79 Table 21 : Scenario 4 Energy Accounting LHV MMBTUs/yr GJ/yr In - Field Operations 214,938 226,771 Transport (Farm to Depot) 76,660 80,881 Depot Natural Gas for Drying (Depot) 347,935 367,091 Natural Gas for Torrefaction (Depot) - - Electricity for Drying (Depot) 7,321 7,724 Electricity for Torrefaction (Depot) 17,699 18,673 Electricity for Grinding (Depot) 9,130 9,632 Electricity for Pelleting (Depot) 26,078 27,514 Transport (Depot to Power Plant) 207,189 218,596 Energy Return on Energy Invested (EROEI) 4 4 5.7.2 Economic Assessment The results of the economic modeling activities for scenario 4 are detailed in Table 21 . The primary metric derived from the economic assessment is the levelized cost of electricity or LCOE where a return of 8% was specified. This yielded a LCOE of $0.08 8. Monte Carlo analysis of this scenario at 1000 iterations found the LCOE to be fairly normally distributed (Kurtosis = - 0.01, Skewness = 0.2) with a standard deviation of $0.009 ( Figure 2 9 ) . 80 Figure 29 : LCOE of TOP Switchgrass Monte Carlo Histogram Though higher than scenario 3 that did not use torrefaction the LCOE of $0.08 2 is still cost competitive with other renewable energy. In terms of capital cost at the power plant, scenario 4 shows a clear advantage over scenario 3. T he similarity of TOP biomass to coal greatly reduces the need for capital investment in new handling, processing, and injection equipment. (Bergman, 2005; Van Loo & Koppejan, 2008) . This is reflected in the low relative capital costs ($ 32 /k Wh capacity vs. $9 7 /kWh for scenario 1). 0 10 20 30 40 50 60 70 80 Frequency LCOE $/kWh LCOE for TOP Switchgrass at 1000 Iterations 81 Table 22 : Scenario 4 Financial Summary Power Plant Metrics Value Unit Retail Price of Electricity $ 0.120 $/kwh Simplified LCOE $ 0 .0 8 17 $/kWh Return on Capital Investment 8% NPV of Biomass Depot $ - NPV of Power Plant Modifications $ 91,331,000 Plant - Gate Feedstock Cost $ 1 53 $/tonne (ar) $ 1 58 $/tonne ( dry ) Relative Capital Costs $ 32 $/kW Capacity As with scenario 2, raw biomass feedstock is still the largest cost at the depot . However, capital costs and O &M associated with torrefaction begin to play a much stronger roll , cumulatively accounting for 44% of the cost of processed biomass ( Figure 30 ) . 82 Figure 30 : Scenario 4 Processed Biomass Cost Distribution 5.7.3 Emissions Analysis Emissions analyses were performed for both the base case (coal only) along with the co - firing scenario. The comparative results are presented in terms of emissions reductions in Table 2 3 . Table 23 : Scenario 4 Emissions Report Power Plant Summary Change in Emission Unit Notes Delta CO2 (Co - Fire - Base Case) 12,847 kg/hr Gross 1,629 tonne/yr (45,230) kg/hr Net (Assuming Biomass is Carbon Neutral) (5,737) tonne/yr Delta SO2 (79) kg/hr (Co - Fire - Base Case) (10) tonne/yr Raw Feedstock Costs 53% Transport Costs (Farm to Depot) 6% Capital Investments 8% Natural Gas - Drying 3% Electricity Drying 1% Electricity Torrefaction 1% Electricity Grinding 1% Electricity Pelleting 2% Operation and Maintainance 10% Tranport Coast (Depot to Plant) 15% Processed Biomass Cost Distribution 83 As expected, blending of torrefied biomass with coal resulted in a net reduction in sulfur dioxide emissions. Carbon emissions were increased on a gross basis but decreased on a net basis. This change from scenario 3 is attributed to the change in chemical composition of the biomass after torrefaction. During the torrefaction process, the ratio of carbon to oxygen is increased, and in this case resulted in a gross increase in carbon dioxide releases. The gross basis values presented is a straight subtraction of CO 2 emissions expectations between the base case of only coal being fired in the boiler systems, and the scenario under investigation. The net basis assumes that any carbon introduced to the system by biomass is a carbon neutral s ource of energy as defined by (Hartmann & Kaltschmitt, 1999) . Because this value is directly correlated to coal displacement, and each scenario displaces the same amount of coal, net CO 2 reducti ons are similar across all scenarios. 5.8 Comparison of Results At a distance of 100 miles, it is observed that pelleting is the superior biomass pre - treatment in terms of cost efficiency. This is largely due to the lower costs of biomass processing. In shor t, it has a higher plant - gate cost but improved shipping costs and a reduced capital requirement as compared to other alternatives. At a distance of 100 miles , however, this difference is not enough for it to economically outperform less intensive process ing techniques. Nor were the additional capital upgrades required for the pelleted biomass scenarios sufficiently large to outweigh the higher plant - gate price of torrefied biomass over the course of a 20 year lifespan. However, a ll scenarios are found to be cost competitive with projected energy production costs. Table 2 3 contains the 2020. All are in the upper half of available option and are more cost effective th an solar PV and s olar t hermal technologies, and under the right conditions, i s competitive with wind. This is largely due to - alone biomass is 84 projected to be greater than $100/MWh. Important to note are a few differences in assumptions made by CREDIT and the EIA . C ritically , CREDIT assumes a more conservative required return on capital % for all listed technologies. This further emphasizes the competitiveness of the studied scenarios. Table 24 : EIA LCOE Projections for 2020 (EIA, 2015) In combination with study ouputs. 48.6 55.6 65.1 - - - 85 6 C ONCLUSIONS CREDIT has been shown to produce approximations of energy production and economic costs that are consistent with those of a decision support tool based on comparisons to exiting case studies . When used to analyze a scenario surrounding the J.H. Campbell p ower plant in western Michigan, wherein biomass from a four county region 100 miles from the power plant was trucked to the facility, CREDIT lower than many other renewable energy options. At a blending rate of 5% biomass by energy value and a biomass transport distance of 100 miles, it was determined that pelletized poplar was the most economical of all studied options at $ 48.6 /MW, followed by pelletized sw itchgrass, TOP poplar, and TOP switchgrass with LCOE values of 55.6 , 65.1, and 81.7 dollars per MWh, respectively. Given the assumptions and parameters of this study, it appears that at a distance of 100 miles, poplar is preferable to switchgrass as a fee dstock and pelleting is preferable to torrefaction as a pre - treatment from an economic perspective. These pricing results indicate that biomass co - firing is a highly competitive option for renewable energy generation. Among options investigated by the US EIA only geothermal is competitive with pelleted poplar in terms of LCOE, while pelleted switchgrass and TOP poplar appear to be on - par with the cost of wind energy . This is logical, as co - firing utilizes existing infrastructure to reduce overall proje ct costs in the near - term and energy companies should investigate to meet renewable energy portfolios. Project financing is heavily influenced by the cost of raw biomass feedstock. Across all scenario investigations, the expense to purchase feedstock wa s the largest cost factor over a 20 year pro - forma analysis, accounting for between 40 and 80% of lifetime project costs at the energy depot and between 65 - 95% of costs at the power plant. As switchgrass is assumed to be nearly twice as expensive as popla r 86 on a per/energy unit basis, the end price of switchgrass derived bioenergy is understandably high in comparison to poplar. Similarly, under these conditions, it was observed that TOP biomass was less cost efficient than dried and pelleted biomass. This is due to its increased plant - to plant - gate costs. There are some conditions made by the scenario analysis that may shift this value as well as the sensitivity to plant - gate feedstock costs. First, a blendin g rate of 5% biomass by energy value was selected to correspond with the availability of biomass in the study region. Biomass blending at higher rates will require increased capital costs for the pelleted biomass scenarios, but little to no increase for t he torrefied scenarios due to the relative change in infrastructure needed. Further, a 20 year project life was selected for this project, thus allowing capital costs to amortize and account for less of the total project cost. If the project were to decr sensitivity to capital costs. This would in - turn begin to favor a feedstock like TOP biomass that has a higher feedstock cost but a lower capital cost. This fact will be particularly important when investig ating scenarios for coal power plants with a limited life expectancy. Finally, the dependence of project success on feedstock cost denotes the need to locate the lowest cost (Tillman, 2000) . Such feedstocks might include construction and demolition debris, mill wastes, forest residues and industrial paper wastes. Such feedstocks may provide a valuable source of energy while being provided at low or no cost to the energy producer. Due to t he sensitivity of the results with respect to feedstock, future work should expand upon the of biomass energy availability in Michigan would be great ly enhanced through integration with GIS - based biomass availability assessments. Efforts such as the Michigan Forest Biomass Inventory 87 (Michigan Tech), the Michigan Waste Biomass Inventory to Support Renewable Energy (Michigan State University) and Natio nal Ag Statistics Services field statistics, would provide data for understanding how much biomass energy is available on a state or national level for biomass co - firing. For the scenario analyses in this study, the NASS survey of unused farmland was util ized to estimate the potential availability of cropland for purpose grown energy crops, but given the ability to integrate existing databases, one might just as easily run several simulations based on the availability of forest residues, construction debri s, and waste biomass. Further, this could be automated in order to produce mapped results of biomass co - firing energy potential in Michigan and beyond. Upon integration with existing GIS databases, it may be useful to develop and implement biomass aggrega tion and pre - treatment optimization sub - routines. These calculations would theoretically allow users to identify the optimum number of biomass collection points as well as the optimum structure of biomass pre - treatment activities based on feedstock type, power plant configuration, distance to the power plant, and biomass blending rates for a given GIS defined collection area. Optimization calculations would be performed with relation to economic performance and environmental considerations. Optimization calculations would provide a useful baseline for the development of biomass supply chain activities, in keeping with the goal of delivering feedstock in the most efficient manner possible. Lessons learned from the development of this tool could also be use d to generate a companion tool for the integration of biomass co - firing at natural gas plants. Increasingly, new energy needs are being met by the installation of natural gas plants in Michigan. Through the use of gasification or through parallel combust ion (depending on the configuration of the natural gas plant), it is technically feasible to 88 portfolio may warrant such a study and to - date no known decision s upport tools have been created for this task. The work completed in this study may also have relevance for public policy. Reduction of baseline greenhouse gas emissions through the production of renewable energy remains a topic of discussion in State and American politics. Integration of this tool with existing databases may allow future works to investigate cost - benefit analyses associated with the utilization of co - firing and the creation of biomass based infrastructure, the implementation of which coul d have implications for energy portfolio diversification and energy security. 89 APPENDICES 90 Appendix A Supplemental Results Charts Figure 31 : Delivered Feedstock Cost Distribution Pelleted Poplar TOP Poplar Pelleted Switchgrass TOP Switchgrass Transport Costs (Depot to Plant) $4,534,001 $3,994,584 $4,534,001 $4,607,854 Operation and Maintainance $944,853 $2,803,461 $1,032,618 $2,468,975 Electricity Pelleting $567,670 $526,153 $567,670 $573,204 Electricity Grinding $198,636 $183,411 $198,636 $200,670 Electricity Torrefaction $- $357,088 $- $389,020 Electricity Drying $149,604 $147,711 $149,604 $160,920 Natural Gas - Torrefaction $- $307,701 $- $- Natural Gas - Drying $875,059 $856,687 $875,059 $988,137 Capital Investments $927,424 $2,267,676 $1,045,250 $2,112,703 Transport Costs (Farm to Depot) $1,420,790 $1,376,309 $1,420,790 $1,704,913 Raw Feedstock Costs $9,403,986 $9,206,552 $14,456,636 $16,324,770 $- $5,000,000 $10,000,000 $15,000,000 $20,000,000 $25,000,000 $30,000,000 $35,000,000 MIllions of Dollars Per Year Delivered Feedstock Cost Distribution 91 Figure 32 : Poplar LCOE Sensitivity to Depot Distance from Power Plant y = 0.0129x + 3.5653 y = 0.0112x + 4.4415 2 4 6 8 10 12 14 16 18 0 200 400 600 800 1000 1200 LCOE ($/MWh) Depot to Powerplant Distance Pelleted TOP 92 Table 25 : Scenario 1 Monte Carlo Statistics EROEI LCOE NPV $/tonne ar Cost of Processed Biomass as Received Cost of Processed Biomass Per Dry Matter Effective Price of Elect. ($/kWh installed) Mean 4.789 0.049 168191324.000 71.032 90.971 101.079 95.815 Standard Error 0.021 0.000 408183.933 0.181 0.204 0.227 0.242 Median 4.785 0.049 168907000.000 71.221 90.754 100.838 95.953 Mode #N/A #N/A 169058000.000 #N/A #N/A #N/A #N/A Standard Deviation 0.680 0.005 12907909.314 5.715 6.452 7.169 7.664 Sample Variance 0.462 0.000 16661412285587 8.000 32.658 41.626 51.391 58.744 Kurtosis - 1.195 - 0.234 - 0.264 - 0.620 - 0.453 - 0.453 - 0.709 Skewness 0.052 0.190 - 0.173 0.033 0.081 0.081 - 0.020 Range 2.387 0.029 69819000.000 28.687 35.026 38.918 36.451 Minimum 3.639 0.036 131249000.000 57.092 73.266 81.406 76.622 Maximum 6.026 0.065 201068000.000 85.779 108.292 120.324 113.073 Sum 4789.3 55 49.045 168191324000.0 00 71032.3 19 90970.806 101078.674 95815.259 Count 1000.0 00 1000.0 00 1000.000 1000.00 0 1000.000 1000.000 1000.000 Confidence Level(95.0%) 0.042 0.000 800996.255 0.355 0.400 0.445 0.476 93 Table 26 : Scenario 2 Monte Carlo Statistics EROEI LCOE NPV $/tonne ar Cost of Processed Biomass as Received Cost of Processed Biomass Per Dry Matter Effective Price of Elect. ($/kWh installed) Mean 4.251 0.056 153659001.000 110.252 130.058 134.081 31.666 Standard Error 0.016 0.000 423299.036 0.210 0.230 0.237 0.115 Median 4.246 0.055 155103000.000 110.379 130.162 134.188 31.508 Mode #N/A #N/A 171462000.000 #N/A #N/A #N/A #N/A Standard Deviation 0.518 0.005 13385890.855 6.649 7.268 7.493 3.623 Sample Variance 0.269 0.000 17918207397497 5.000 44.210 52.825 56.143 13.127 Kurtosis - 1.100 - 0.040 - 0.032 - 0.563 - 0.391 - 0.391 - 1.154 Skewness 0.030 0.382 - 0.367 0.061 0.035 0.035 0.066 Range 2.016 0.035 85114000.000 32.010 38.057 39.234 12.911 Minimum 3.285 0.042 102309000.000 94.066 111.262 114.703 25.399 Maximum 5.301 0.077 187423000.000 126.076 149.319 153.937 38.310 Sum 4251.2 19 55.860 153659001000.0 00 110252. 113 130058.242 134080.662 31666.133 Count 1000.0 00 1000.0 00 1000.000 1000.00 0 1000.000 1000.000 1000.000 Confidence Level(95.0%) 0.032 0.000 830657.249 0.413 0.451 0.465 0.225 94 Table 27 : Scenario 3 Monte Carlo Statistics EROEI LCOE NPV $/tonne ar Cost of Processed Biomass as Received Cost of Processed Biomass Per Dry Matter Effective Price of Elect. ($/kWh installed) Mean 4.681 0.065 128,732,781.000 95.460 115.282 128.091 96.109 Standard Error 0.021 0.000 556,463.377 0.262 0.279 0.310 0.241 Median 4.633 0.065 129,379,500.000 95.262 114.913 127.681 96.587 Mode #N/A #N/A 110,014,000.000 #N/A #N/A #N/A #N/A Standard Deviation 0.665 0.007 17,596,917.055 8.273 8.814 9.794 7.633 Sample Variance 0.442 0.000 309,651,489,834,867. 000 68.437 77.690 95.914 58.260 Kurtosis (1.146) (0.321) (0.334) (0.857) (0.737) (0.737) (0.785) Skewness 0.156 0.207 (0.194) 0.056 0.109 0.109 (0.133) Range 2.357 0.042 102,008,000.000 38.522 43.727 48.585 34.746 Minimum 3.584 0.047 71,105,000.000 77.172 95.445 106.050 77.954 Maximum 5.942 0.089 173,113,000.000 115.694 139.172 154.635 112.700 Sum 4,680.579 65.487 128,732,781,000.000 95,459.724 115,282.228 128,091.364 96,109.230 Count 1,000.000 1,000.000 1,000.000 1,000.000 1,000.000 1,000.000 1,000.000 Confidence Level(95.0%) 0.041 0.000 1,091,971.156 0.513 0.547 0.608 0.474 95 Table 28 : Scenario 4 Monte Carlo Statistics EROEI LCOE NPV $/tonne ar Cost of Processed Biomass as Received Cost of Processed Biomass Per Dry Matter Effective Price of Elect. ($/kWh installed) Mean 4.256 0.082 91127563.000 133.056 153.010 157.742 31.979 Standard Error 0.019 0.000 629170.501 0.307 0.324 0.334 0.115 Median 4.187 0.081 92369500.000 132.868 152.870 157.598 32.075 Mode #N/A #N/A 96289000.000 #N/A #N/A #N/A #N/A Standard Deviation 0.615 0.008 19896118.184 9.714 10.250 10.568 3.650 Sample Variance 0.378 0.000 39585551879882 2.000 94.370 105.072 111.672 13.320 Kurtosis - 1.184 - 0.294 - 0.294 - 0.650 - 0.491 - 0.491 - 1.195 Skewness 0.189 0.247 - 0.240 0.076 0.055 0.055 - 0.042 Range 2.153 0.045 108886000.000 48.633 54.637 56.327 12.921 Minimum 3.234 0.062 31166000.000 110.167 127.547 131.491 25.460 Maximum 5.387 0.107 140052000.000 158.800 182.183 187.818 38.381 Sum 4256.0 59 81.746 91127563000.00 0 133055. 565 153009.626 157741.882 31978.823 Count 1000.0 00 1000.0 00 1000.000 1000.00 0 1000.000 1000.000 1000.000 Confidence Level(95.0%) 0.038 0.001 1234647.359 0.603 0.636 0.656 0.226 96 Appendix B Economic Assumptions Table 29 : Variable Cost Assumptions Default Min Max Source Market Electricity Price ($/kWh) 0.12 0.08 0.14 EIA 2014 Switchgrass Farmgate Feedstock Price ($/dt) 0.05 0.04 0.05 ISU 2014 Poplar Farmgate Feestock Price ($/dt) 0.05 0.04 0.05 Saffron and Chai 2011 Natural Gas Price ($/MJ) 0.003 0.002 0.003 EIA 2014 Specific Transportation Cost ($/km) 2.25 1.80 2.70 Svanberg 2013 Plant O&M Cost ($/kW installed capacity) 47.60 38.08 57.12 Caputo 2009 Depot O&M Costs (% of capital costs) 8% 6% 10% Batizdiari 2013 Excavation 4% 4% 4% Caputo 2009 Engineering 12% 12% 12% Caputo 2009 Contingency 5% 5% 5% Binkley 2015 Financing APR 7% 7% 7% Binkley 2015 Loan (% of total capital costs) 80% 80% 80% Binkley 2015 Grant Funding (% of total capital costs) 10% 10% 10% Binkley 2015 Cash on Hand (% of total capital costs) 10% 10% 10% Binkley 2015 97 Table 30 : Capital Cost Assumptions Base Size Scaling Factor Base Cost Min Cost Max Cost Source(s) Dryer Cost 4,535 (kg/hr) 0.65 $440,000 $352,000 $528,000 (Batidzirai et al., 2013) Torrefier Cost 1,000 (kg/hr) 0.60 $ 2,027,391 $5,148,800 $7,723,200 (Batidzirai et al., 2013) Grinding Cost 4,898 (kg/hr) 0.65 $193,622 $154,898 $232,346 (Srivastava et al., 2011) Cooling Cost 4,535 (kg/hr) 0.60 $29,970 $23,976 $35,964 (Srivastava et al., 2 011) Peripheral Equipment Cost 4,898 (kg/hr) 0.60 $1,152,000 $921,600 $1,382,400 (Srivastava et al., 2011) Pelleting Equipment 2,721 (kg/hr) 0.61 $438,000 $350,400 $525,600 (Srivastava et al., 2011) Building Cost Factor 1 (sq meter) 1 $900.00 $700.00 $800.00 (Srivastava et al., 2011) Storage Cost 1 (cu meter) 1 $13.50 $13.00 $14.00 (Srivastava et al., 2011) Boiler Modifications 1000 (kg/hr) 0.6 $61 $49 $73 (Caputo et al., 2005; De & Assadi, 2009) Plant Storage 1000 (kg/hr) 0.6 $166,625 $133,300 $199,950 (Caputo et al., 2005; De & Assadi, 2009) Plant Handling 1000 (kg/hr) 0.6 $68,052 $54,441 $81,662 (Caputo et al., 2005; De & Assadi, 2009) Plant Conditioning 1000 (kg/hr) 0.6 $16,648 $13,318 $19,978 (Caputo et al., 2005; De & Assadi, 2009) 98 Table 31 : Depreciation Schedule for Equipment and Buildings Depreciation Rate Farm Machinery and Equipment 7 year MACRS Machinery 10 Year MACRS Building 20 Year MACRS Year 1 10.71% 7.50% 3.750% Year 2 19.13% 13.88% 7.219% Year 3 15.03% 11.79% 6.677% Year 4 12.25% 10.02% 6.177% Year 5 12.25% 8.74% 5.713% Year 6 12.25% 8.74% 5.285% Year 7 12.25% 8.74% 4.888% Year 8 6.13% 8.74% 4.522% Year 9 8.74% 4.462% Year 10 8.74% 4.461% Year 11 4.37% 4.462% Year 12 4.461% Year 13 4.462% Year 14 4.461% Year 15 4.462% Year 16 4.461% Year 17 4.462% Year 18 4.461% Year 19 4.462% Year 20 4.461% Year 21 2.231% 99 Appendix C Parameter Look - Up Table Values Table 32 : Miscanthus Fuel Properties (PHYLLIS2, ECN) Unit Minimum Maximum Median Mean Std dev Samples Miscanthus Fuel Properties Proximate Analysis Moisture content wt% (ar) 7.3 49 42 36.47 11.6 32% 37 Ash content wt% (dry) 1.5 7.46 3.2 3.74 1.41 38% 39 Volatile matter wt% (daf) 73.87 94.27 89.55 85.9 10.68 12% 3 Fixed carbon wt% (daf) 5.73 26.13 10.45 14.1 10.68 76% 3 Ultimate Analysis Carbon wt% (daf) 46.73 51.97 49.8 49.63 1.1 2% 47 Hydrogen wt% (daf) 5 6.48 5.67 5.63 0.33 6% 47 Nitrogen wt% (daf) 0.1 1.83 0.49 0.54 0.29 54% 47 Sulphur wt% (daf) 0.02 0.21 0.06 0.06 0.04 63% 45 Oxygen wt% (daf) 40.06 46.78 43.64 43.81 1.48 3% 47 Total (with halides) wt% (daf) 92.86 101.66 100 99.88 1.07 1% 47 Calorific Values Net calorific value (LHV) MJ/kg (daf) 15.59 20.97 18.53 18.55 0.64 3% 46 Gross calorific value (HHV) MJ/kg (daf) 17 22.2 19.77 19.77 0.63 3% 46 HHVMilne MJ/kg (daf) 18.47 20.16 19.28 19.32 0.35 2% 39 Chemical Analyses Halides Chlorine (Cl) mg/kg (daf) 200 3 955.9 2 000.0 2 149.1 1 070.5 50% 45 Fluorine (F) mg/kg (daf) 27.8 29 28.4 28.4 0.9 3% 2 100 Table 33 : Wheat Straw Fuel Properties (PHYLLIS2, ECN) Unit Minimum Maximum Median Mean Std dev Samples Wheat Straw - Fuel Properties Proximate Analysis Moisture content wt% (ar) 0 17.41 9.74 10.24 4.09 40% 23 Ash content wt% (dry) 1.3 13.5 6.45 6.44 2.72 42% 48 Volatile matter wt% (daf) 78.04 84.53 81.58 81.5 1.84 2% 19 Ash content at 550°C wt% (dry) 4.71 10.32 8.02 7.77 2.31 30% 4 Ash content at 815°C wt% (dry) 7.75 9.81 7.9 8.49 1.15 14% 3 Fixed carbon wt% (daf) 15.47 21.96 18.42 18.5 1.84 10% 19 Ultimate Analysis Carbon wt% (daf) 46.35 52.6 49.04 48.86 1.37 3% 38 Hydrogen wt% (daf) 3.2 6.39 5.96 5.87 0.52 9% 38 Nitrogen wt% (daf) 0.29 2.08 0.61 0.72 0.38 53% 40 Sulphur wt% (daf) 0.03 0.46 0.12 0.15 0.09 62% 36 Oxygen wt% (daf) 39.42 47.92 43.73 44.08 1.64 4% 38 Total (with halides) wt% (daf) 0 101.6 100 73.13 44.75 61% 52 Calorific Values Net calorific value (LHV) MJ/kg (daf) 15.2 20.49 18.21 18.11 1.07 6% 36 Gross calorific value (HHV) MJ/kg (daf) 16.63 21.74 19.42 19.35 1.03 5% 34 HHVMilne MJ/kg (daf) 15.18 20.54 19.01 18.96 0.91 5% 38 Chemical Analyses Halides Chlorine (Cl) mg/kg (daf) 207.1 22 775.0 2 793.7 4 335.7 5 221.3 120% 32 Bromine (Br) mg/kg (daf) 10.8 32.3 15.5 19.6 11.3 58% 3 Fluorine (F) mg/kg (daf) 7.2 7.7 7.4 7.4 0.4 5% 2 101 Table 34 : Willow Fuel Properties (PHYLLIS2, ECN) Unit Minimum Maximum Median Mean Std dev Samples Willow Fuel Properties Proximate Analysis Moisture content wt% (ar) 10.23 50.1 11.3 25.25 19.82 78% 5 Ash content wt% (dry) 0.45 4.59 1.6 2.18 1.55 71% 7 Volatile matter wt% (daf) 80.29 86.05 83.2 83.19 2.67 3% 4 Ash content at 550°C wt% (dry) 1.3 1.8 1.55 1.55 0.35 23% 2 Fixed carbon wt% (daf) 13.95 19.71 16.8 16.81 2.67 16% 4 Ultimate Analysis Carbon wt% (daf) 45.29 51 50.54 49.62 2.18 4% 6 Hydrogen wt% (daf) 5.78 6.74 6 6.11 0.34 6% 6 Nitrogen wt% (daf) 0.1 1.12 0.54 0.54 0.39 72% 6 Sulphur wt% (daf) 0.03 0.1 0.05 0.05 0.03 54% 5 Oxygen wt% (daf) 41.64 46.76 42.9 43.57 1.99 5% 5 Total (with halides) wt% (daf) 0 100 98.07 61.75 48.25 78% 9 Calorific Values Net calorific value (LHV) MJ/kg (daf) 17.53 19.12 18.18 18.27 0.58 3% 6 Gross calorific value (HHV) MJ/kg (daf) 18.86 20.59 19.46 19.61 0.64 3% 6 HHVMilne MJ/kg (daf) 17.43 20.93 20.06 19.59 1.32 7% 5 Chemical Analyses Halides Chlorine (Cl) mg/kg (daf) 101.3 337.1 219.2 219.2 166.7 76% 2 Fluorine (F) mg/kg (daf) 10.1 10.1 10.1 10.1 0 0% 1 102 Table 35 : Poplar Fuel Properties (PHYLLIS2, ECN) Unit Minimum Maximum Median Mean Std dev Samples Poplar Fuel Properties Proximate Analysis Moisture content wt% (ar) 4.8 15 9.9 9.9 7.21 73% 2 Ash content wt% (dry) 0.4 2.28 1.1 1.13 0.63 56% 13 Volatile matter wt% (daf) 71.76 86.12 83.43 80.44 7.64 9% 3 Fixed carbon wt% (daf) 13.88 28.24 16.57 19.56 7.64 39% 3 Ultimate Analysis Carbon wt% (daf) 48.3 51.98 49.7 49.91 1.18 2% 10 Hydrogen wt% (daf) 5.8 6.34 6.08 6.09 0.16 3% 10 Nitrogen wt% (daf) 0.1 0.48 0.21 0.26 0.15 58% 8 Sulphur wt% (daf) 0.01 0.05 0.05 0.04 0.02 53% 7 Oxygen wt% (daf) 41.72 45.8 43.96 43.73 1.23 3% 10 Total (with halides) wt% (daf) 0 100 100 76.91 43.85 57% 13 Calorific Values Net calorific value (LHV) MJ/kg (daf) 18.24 19.51 18.78 18.78 0.49 3% 9 Gross calorific value (HHV) MJ/kg (daf) 19.55 20.89 20.11 20.11 0.52 3% 9 HHVMilne MJ/kg (daf) 18.62 21.06 19.68 19.77 0.77 4% 10 Chemical Analyses Halides Chlorine (Cl) mg/kg (daf) 101.2 1 013.5 122.3 413.5 459.6 111% 6 103 Table 36 : Switchgrass Fuel Properties (PHYLLIS2, ECN) Unit Minimum Maximum Median Mean Std dev Samples Switchgrass Fuel Properties Proximate Analysis Moisture content wt% (ar) 8.16 15 11.9 11.72 2.78 0.24 5 Ash content wt% (dry) 1.9 10.11 6.25 6.3 1.38 0.22 34 Volatile matter wt% (daf) 72.91 86.91 84.25 83.23 4.46 0.05 8 Fixed carbon wt% (daf) 13.09 27.09 15.75 16.77 4.46 0.27 8 Ultimate Analysis Carbon wt% (daf) 45.19 53.16 50.63 49.43 2.46 0.05 13 Hydrogen wt% (daf) 5.64 6.53 6.13 6.13 0.35 0.06 13 Nitrogen wt% (daf) 0.4 1.3 0.59 0.64 0.18 0.28 30 Sulphur wt% (daf) 0 0.21 0.13 0.12 0.06 0.45 13 Oxygen wt% (daf) 39.01 48.64 43.72 43.97 2.89 0.07 13 Total (with halides) wt% (daf) 0 101.78 0.61 38.7 49.32 1.27 34 Calorific Values Net calorific value (LHV) MJ/kg (daf) 16.86 18.9 17.66 17.82 0.69 0.04 12 Gross calorific value (HHV) MJ/kg (daf) 18.29 20.22 18.94 19.16 0.7 0.04 12 HHVMilne MJ/kg (daf) 16.91 21.62 19.47 19.51 1.22 0.06 13 Chemical Analyses Halides Chlorine (Cl) mg/kg (daf) 370.3 5 249.9 1 062.7 1 952.3 1 943.6 1 5 Major elements Potassium (K) mg/kg (dry) 3 400.0 3 400.0 3 400.0 3 400.0 0 0 1 Sodium (Na) mg/kg (dry) 33 33 33 33 0 0 1 104 Table 37 : Coal Proximate and Ultimate Analysis (PHYLLIS2, ECN) Proximate Analysis (wt % ar) Ultimate Analysis (wt % moisture & ash free) Fixed Carbon Volatile Matter Moisture Ash C H O N S Net Heating Value (MJ/kg) Select a Value Anthracite 81.8 7.7 4.5 6 91.8 3.6 2.5 1.4 0.7 36.2 Bituminous 54.9 35.6 5.3 4.2 82.8 5.1 10.1 1.4 0.6 36.1 Sub - Bituminous 43.6 34.7 10.5 11.2 76.4 5.6 14.9 1.7 1.4 31.8 Lignite 27.8 24.9 36.9 10.4 71 4.3 23.2 1.1 0.4 26.7 Table 38 : Transportation Conditions Vehichle Capacity - Mass (kg) Volume (m3) Source kg/m3 lb/ft3 condition Hybrid Poplar - Chipped 36,280 139 NREL 275 chipped Willow Wood - Chipped 36,280 139 NREL 275 chipped Hybrid Poplar - Pelleted 36,280 139 NREL 625 pelleted Willow Wood - Pelleted 36,280 139 NREL 625 pelleted Wheat Straw - Baled 15,419 NREL Baled Switchgrass - Baled 15,419 NREL Baled Miscanthus - Baled 15,419 NREL Baled Straw - Pelleted 36,280 139 NREL 625 pelleted Switchgrass - Pelleted 36,280 139 NREL 625 pelleted Miscanthus - Pelleted 36,280 139 NREL 625 pelleted Hybrid Poplar (Torrefied) 36,280 139 NREL 240 Torrefied Willow Wood (Torrefied) 36,280 139 NREL 240 Torrefied Hybrid Poplar (T & P) 36,280 139 NREL 800 TOP Willow Wood (T & P) 36,280 139 NREL 800 TOP 105 Table 39 : Coal Fired Powerplant Data - 1 Utility Name Plant Code Plant Name City Lat Long Feeding Mechanism Max Steam Flow (Thousand Pounds per Hour) Coal Fire Steam Flow (0.1 Tons per Hour) Efficiency 100% Load Efficiency 50% Load Consumers Energy Co 1695 B C Cobb - Boiler 4 Muskegon 43.258768 - 86.242268 Pulverized Fuel 1,050.0 62.6 0.9 0.9 Consumers Energy Co 1695 B C Cobb - Boiler 5 Muskegon 43.258768 - 86.242268 Pulverized Fuel 1,050.0 62.6 0.9 0.9 Consumers Energy Co 1702 Dan E Karn - Boiler 1 Essexville 43.644996 - 83.840074 Pulverized Fuel 1,750.0 105.5 0.9 0.9 Consumers Energy Co 1702 Dan E Karn - Boiler 2 Essexville 43.644996 - 83.840074 Pulverized Fuel 1,750.0 108.4 0.9 0.9 Consumers Energy Co 1710 J H Campbell - Boiler 2 West Olive 42.910296 - 86.20074 Pulverized Fuel 2,550.0 140.5 0.9 0.9 Consumers Energy Co 1720 J C Weadock - Boiler 7 Essexville 43.639927 - 83.844712 Pulverized Fuel 1,050.0 63.0 0.9 0.9 Consumers Energy Co 1720 J C Weadock - Boiler 8 Essexville 43.639927 - 83.844712 Pulverized Fuel 1,050.0 63.0 0.9 0.9 Consumers Energy Co 1723 J R Whiting - Boiler 1 Erie 41.792114 - 83.44948 Pulverized Fuel 690.0 44.3 0.9 0.9 Consumers Energy Co 1723 J R Whiting - Boiler 2 Erie 41.792114 - 83.44948 Pulverized Fuel 690.0 44.3 0.9 0.9 Consumers Energy Co 1723 J R Whiting - Boiler 3 Erie 41.792114 - 83.44948 Pulverized Fuel 850.0 53.9 0.9 0.9 106 Table 40 : Coal Fired Power Plant Data - 2 Utility Name Plant Code Plant /Boiler Name Standard Particulate Rate Max Steam Flow (Thousand Pounds per Hour) Coal Fire Steam Flow (0.1 Tons per Hour) Petroleum Fire Steam Flow (0.1 Barresl per Hour) Gas Fire Steam Flow (0.1 Tons per Hour) Primary Fuel 1 Primary Fuel 2 Air Flow 100% Load (Cubic Feet per Minute) Wet Dry Bottom Fly Ash Reinjection Consumers Energy Co 1695 B C Cobb - Boiler 4 0.180 1,050.0 62.6 0.0 0.0 BIT SUB 370,000 D N Consumers Energy Co 1695 B C Cobb - Boiler 5 0.180 1,050.0 62.6 0.0 0.0 BIT SUB 370,000 D N Consumers Energy Co 1702 Dan E Karn - Boiler 1 0.160 1,750.0 105.5 0.0 0.0 BIT SUB 650,000 D N Consumers Energy Co 1702 Dan E Karn - Boiler 2 0.160 1,750.0 108.4 0.0 0.0 BIT SUB 650,000 D N Consumers Energy Co 1710 J H Campbell - Boiler 2 0.150 2,550.0 140.5 0.0 0.0 BIT SUB 850,000 D N Consumers Energy Co 1720 J C Weadock - Boiler 7 0.180 1,050.0 63.0 0.0 0.0 BIT SUB 340,000 D N Consumers Energy Co 1720 J C Weadock - Boiler 8 0.180 1,050.0 63.0 0.0 0.0 BIT SUB 340,000 D N Consumers Energy Co 1723 J R Whiting - Boiler 1 0.200 690.0 44.3 0.0 0.0 BIT SUB 260,000 D N Consumers Energy Co 1723 J R Whiting - Boiler 2 0.200 690.0 44.3 0.0 0.0 BIT SUB 260,000 D N Consumers Energy Co 1723 J R Whiting - Boiler 3 0.190 850.0 53.9 0.0 0.0 BIT SUB 320,000 D N 107 Appendix D - Screenshots Figure 33 : User Input Screen Capture 108 Figure 34 : Investment Cost Screen Capture 109 Figure 35 : Transportation Mass and Energy Balance Screenshot 110 Figure 36 : Biomass Depot Mass and Energy Balance Screenshot 111 Figure 37 : Co - Firing Mass and Energ y Balance Screenshot 112 Figure 38 : Pro - Forma Screenshot 113 BIBLIOGRAPHY 114 B IBLIOGRAPHY Anon. (2008). Estimated Costs for Production Storage and Transportation of Switchgrass. In I. S. Extension (Ed.), Ag Decision Maker (pp. 8). www.estension.iastate.edu/agdm: Iowa State University Extens ion. Anon. (2010). FreightTransportation Values (Trucks and Trains). In P. C. o. G. C. Change (Ed.), Climate Tech Book (pp. 11). Pew Center on Global Climate Change. Basu, P. (2013). Biomass gasification, pyrolysis, and torrefaction: practical design and t heory. Boston; Amsterdam: Elsevier/AP. Batidzirai, B., Mignot, A., Schakel, W., Junginger, H., & Faaij, A. (2013). Biomass torrefaction technology: Techno - economic status and future prospects. ENERGY, 62 , 196 - 214. Bergman, P. C. A. (2005). Torrefaction fo r biomass co - firing in existing coal - fired power stations. Biocoal . Binkley, D. (2010). The economics of an anaerobic digestion under different electricity purchase agreements: Michigan State University. Binkley, D., Safferman, S., Dong, Y., Smith, J. S., Tekesin, O., & Vincent, A. (2015). Anaerobic Digestion Development Iterative Tool to Evaluate Blended Feedstocks. Water research, Under Review . Bologa, A., Paur, H. - R., Seifert, H., Woletz, K., & Ulbricht, T. (2012). Fine particle generation, evolution a nd control by small biomass combustion. Management of Environmental Quality: An International Journal, 23 (1), 36 - 55. doi: 10.1108/14777831211191584 Caputo, A. C., Palumbo, M., Pelagagge, P. M., & Scacchia, F. (2005). Economics of biomass energy utilization in combustion and gasification plants: effects of logistic variables. Biomass and Bioenergy, 28 (1), 35 - 51. Centre, I. C. C. (2005). Fuels for Biomass Co - Firing. Channiwala, S., & Parikh, P. (2002). A unified correlation for estimating HHV of solid, liqui d and gaseous fuels. Fuel, 81 (8), 1051 - 1063. De, S., & Assadi, M. (2009). Impact of cofiring biomass with coal in power plants A techno - economic assessment. Biomass and Bioenergy, 33 (2), 283 - 293. Demirbas, A. (2005). Potential applications of renewable energy sources, biomass combustion problems in boiler power systems and combustion related environmental issues. Progress in Energy and Combustion Science, 31 (2), 171 - 192. doi: http://dx.doi.org/10.1016/j.pecs.2005.02.002 Dillen, S. Y., Djomo, S. N., Al Afas, N., Vanbeveren, S., & Ceulemans, R. (2013). Biomass yield and energy balance of a short - rotation poplar coppice with multiple clones o n degraded land during 16 115 years. Biomass and Bioenergy, 56 , 157 - 165. doi: http://dx.doi.org/10.1016/j.biombioe.2013.04.019 Dunajski, L., Kruk, W., & Nowak, W. (2013). Preventing circulating f luidised bed agglomeration and deposition during biomass combustion. Archives of Thermodynamics, 34 (1), 3. doi: 10.2478/aoter - 2013 - 0001 ECN. (2015). Phyllis2 - Database for Biomass and Waste . Retrieved from: https:// www.ecn.nl/phyllis2 EIA. (2012). Form EIA - 860 detailed data Available from USA Energy Information Administration Retrieved 7/20/2014, from US Energy Information Administration http://www.eia.g ov/electricity/data/eia860/ EIA. (2015, 4/14/2015). Levelized Cost and Levelized Avoided Cost of New Generation Resources in the Annual Energy Outlook 2015. Annual Energy Outlook 2015. Retrieved 6/20/2015, 2015, from http://www.eia.gov/forecasts/aeo/electricity_generation.cfm EPA. (2007). Biomass Combined Heat and Power Catalog of Technologies. Retrieved 11/25/2014, 2014, from http://www.epa.gov/chp/documents/biomass_chp_catalog.pdf FEMP. (2004). Biomass Cofiring in Coal - Fired Boilers . (DOE/EE - 0288). US Department of Energy Retrieved from http://www.nrel.gov/docs/fy04osti/33811.pdf . Gebreegziabher, T., Oyedun, A. O., & Hui, C. W. (2013). Optimum biomass drying for combustion - A modeling approach. ENERGY, 53 , 67 - 73. doi: 10.1016/j.energy.2013.03.004 Gil, M. V., Oulego, P., Casal, M. D., Pev ida, C., Pis, J. J., & Rubiera, F. (2010). Mechanical durability and combustion characteristics of pellets from biomass blends. Bioresource technology, 101 (22), 8859 - 8867. doi: 10.1016/j.biortech.2010.06.062 Hartmann, D., & Kaltschmitt, M. (1999). Electric ity generation from solid biomass via co - combustion with coal - Energy and emission balances from a German case study. BIOMASS & BIOENERGY, 16 (6), 397 - 406. Jenkins, B. M., Baxter, L. L., Miles Jr, T. R., & Miles, T. R. (1998). Combustion properties of bio mass. Fuel Processing Technology, 54 (1 3), 17 - 46. doi: http://dx.doi.org/10.1016/S0378 - 3820(97)00059 - 3 Joshi, Y., de Vries, H., Woudstra, T., & de Jong, W. (2014). Torrefaction: Unit Operation Modelling and Process Simulation. Applied Thermal Engineering . doi: 10.1016/j.applthermaleng.2013.12.059 Kinney, S. - A. (2012). Wood Bioenergy and Natural Gas Prices. Forest 2 Market. Retrieved 2 - 12 - 15, 2015, from http://www.forest2market.com/blog/Wood - Bioenergy - and - Natural - Gas - Prices Lempp, P. (2013). Biomass Co - Firing: Technology Brief. from http://www.irena.org/DocumentDownloads/Publications/IRENA - ETSAP%20Tech%20Brief%20E21%20Biomass%20Co - firing.pdf Maciejewska, A., Veringa, H., Sanders, J., & Peteves, S. (2006). Co - firing of biomass with coal : constraints and role of biomass pre - treatment. DG JRC Institute for Energy . 116 Murphy, D. J., & Hall, C. A. (2010). Year in review EROI or energy return on (energy) invested. Annals of the New York Academy of Sciences, 1185 (1), 102 - 118. NASS. (2012). Farm s, Land in Farms, Value of Land and Buildings, and Land Use: 2012 and 2017 . Retrieved from: http://www.agcensus.usda.gov /Publications/2012/Full_Report/Volume_1,_Chapter_2_County_ Level/Michigan/st26_2_008_008.pdf Nussbaumer, T. (2003). Combustion and co - combustion of biomass: fundamentals, technologies, and primary measures for emission reduction. Energy & Fuels, 17 (6), 1510 - 1521. O'Donnel, B., Hickner, M., & Barna, B. (2002). Economic Risk Analysis Using Analytical and Monte Carlo Techniques. Retrieved 6/15/2015, 2015, from htt p://ufdcimages.uflib.ufl.edu/AA/00/00/03/83/00154/AA00000383_00154_00094.pdf Park, J., Meng, J. J., Lim, K. H., Rojas, O. J., & Park, S. (2013). Transformation of lignocellulosic biomass during torrefaction. Journal of Analytical and Applied Pyrolysis, 100 , 199 - 206. doi: 10.1016/j.jaap.2012.12.024 Pennington, D. (2015). [MSU Extension Field Plot Summaries]. Sargent, R. G. (2011). Verification and Validation of Simulation Models. Paper presented at the Winter Simulation Conference. Srivastava, A., Abbas, D., Saffron, C., & Pan, F. (2011). Economic Analysis of Woody Biomass Supply Chain Logistics for Biofuel Produciton in Michigan (pp. 60): Department of Biosystems and Agricultural Engineering - Michigan State University. Svanberg, M., Olofsson, I., Flodén, J. , & Nordin, A. (2013). Analysing biomass torrefaction supply chain costs. Bioresource technology, 142 , 287 - 296. doi: 10.1016/j.biortech.2013.05.048 Tillman, D. A. (1997). Biomass Cofiring Guidelines (pp. 61): EPRI. Tillman, D. A. (2000). Biomass cofiring: the technology, the experience, the combustion consequences. Biomass and Bioenergy, 19 (6), 365 - 384. doi: 10.1016/S0961 - 9534(00)00049 - 0 Van Loo, S., & Koppejan, J. (2008). The handbook of biomass combustion and co - firing . Sterling, VA; London: Earthscan. Werther, J., & Ogada, T. (1999). Sewage sludge combustion. Progress in Energy and Combustion Science, 25 (1), 55 - 116. doi: 10.1016/S0360 - 1285(98)00020 - 3 Weststeyn A., E. E. B. (2004). First Torrefied Wood Successfully Co - Fired. Williams, A., Pourkashan ian, M., & Jones, J. M. (2001). Combustion of pulverised coal and biomass. Progress in Energy and Combustion Science, 27 (6), 587 - 610. doi: 10.1016/S0360 - 1285(01)00004 - 1