Optimizing the Logistics of a Mobile Fast Pyrolysis System for Sustainable Bio-crude Oil Production

An ASABE Conference Presentation Paper Number: 1009174 Optimizing the Logistics of a Mobile Fast Pyrolysis System for Sustainable Bio-crude Oil Production Miae Ha, Doctorate student, Water Management and Hydrological Sciences Program Marisa L. Bumguardner, Masters student, Water Management and Hydrological Sciences Program Clyde L. Munster, ASABE Member, Professor, Department of Biological and Agricultural Engineering Donald M. Vietor, Professor, Department of Soil and Crop Sciences Sergio Capareda, ASABE Member, Assistant Professor, Department of Biological and Agricultural Engineering Marco A. Palma, Assistant Professor, Department of Agricultural Economics Tony Provin, Associate Professor, Department of Soil and Crop Sciences Texas A&M University, College Station, TX, USA Annual International Meeting Pittsburgh, Pennsylvania, USA June 20-23, 2010 1

Abstract The GIS methods used to identify optimum locations for mobile pyrolysis units in the North Central (NC) region of the U.S. are presented in this paper. Optimum locations were based on feedstock availability. The feedstocks used in the study were corn stover and bioenergy sorghum. For corn stover, 10 year (1999-2008) average corn grain production values were determined for each county in the NC region. Feedstock harvest rates were limited to 25% of the available corn stover leaving 75% for erosion control and soil improvement. LaSalle County, Illinois, was used as a pilot study for the corn stover analysis. For bioenergy sorghum, it was assumed that the production rate was 15 Mg/ha and that bioenergy sorghum would be planted in place of grain sorghum. It was assumed that 100% of the bioenergy sorghum would be harvested for pyrolysis feedstock. ArcGIS and ModelBuilder were used for this feedstock assessment study. A square grid was placed over a map of fields planted with corn and sorghum in 2008. The size of the harvest grid was 6,200 x 6,200 m for the corn stover and 20,000 x 20,000 m for the bioenergy sorghum. The sizes of the harvest grids were based on, 1) the mobile pyrolysis unit requires feedstock at a rate of 80,000 lbs/day, and 2) the mobile pyrolysis unit would remain in place for 6 months. It was also assumed that the mobile pyrolysis unit would be located in the center of the grid and that the average feedstock hauling distance from the field to the mobile unit would be one half the grid size. Therefore hauling distances were 3,100 m for corn stover and 10,000 m for bioenergy sorghum. These short feedstock hauling distances demonstrate a primary advantage of mobile pyrolysis units over a central bioenergy plant. The top 100 locations for corn stover feedstock availability were determined with 57 sites in Illinois, 29 in Nebraska, and 14 in Iowa. The top 50 locations for bioenergy sorghum feedstock availability were also determined with 35 sites in Nebraska and 15 in South Dakota. Keywords. Mobile pyrolysis units, feedstock logistics, GIS, corn stover, energy sorghum, bio-oil, biochar. Introduction The overall goal of this project is to develop a comprehensive decision making tool to optimize the use of a fleet of mobile pyrolysis units in the North Central U.S. to produce bio-oil from agricultural feedstocks. The concept is to use mobile pyrolysis units to convert low density biomass to high density bio-oil to minimize the cost of feedstock logistics. The mobile pyrolysis units would be placed as close as possible to the feedstock sources thereby reducing feedstock transportation, handling, and storage costs when compared to a centralized bioenergy plant. The GIS based decision making tool will be interfaced with an economic model and weather forecasting programs to identify optimum locations for the mobile pyrolysis units as well as the schedule to move the units. The economic model will assess all of the costs required to purchase, setup, operate, and move the mobile pyrolysis units as well as compensation from bioenergy production. The weather forecasting programs will help to identify optimum locations for the mobile units based on projected meteorological conditions in specific regions. Mobile units will be placed in areas expecting the highest crop production and to avoid areas experiencing wet weather that prevents the transport of feedstocks from the fields to the mobile units and increases feedstock drying. The Geographic Information System (GIS) methods used to determine 2

optimum feedstock locations based on feedstock production in the North Central U.S. is presented in this paper and is the first component of this decision making tool. Pyrolysis The process of fast pyrolysis is the thermal degradation of biomass in the absence of oxygen. The products of this process are bio-oil, and gas (also referred to as syngas), and biochar. The pyrolysis process typically involves drying and grinding the feedstock to increase heat transfer rates. After the feedstock is prepared it is rapidly heated to high temperatures and vapors, aerosols, and biochar are produced. The vapors and aerosols are then cooled and condensed forming bio-oil (Bridgwater et al., 1999). Many agricultural crops can be used in the pyrolysis process including corn stover, sorghum, switchgrass, rice, wheat, sugar cane, straw and miscanthus (Mohan et al. 2006). Wood wastes from forestry, leather wastes and sewage sludge have also been utilized for pyrolysis. The bio-oil product is usually a dark brown, organic liquid with high oxygen content (Czernik and Bridgwater, 2004). A charcoal substance referred to as biochar or simply char is a by-product from pyrolysis. This biochar retains some of the nutrients present in the biomass before pyrolysis (Laird, 2008). The gases that cannot be condensed are known as synthesis gas or syngas. Syngas is mainly carbon monoxide and hydrogen and can be used to generate electrical power. The Mobile Pyrolysis Concept This project utilizes a fleet of mobile pyrolysis units instead of a large centralized pyrolysis plant. Using mobile units takes the pyrolysis equipment directly to the feedstock source; minimizing the distance feedstocks have to be transported. Agricultural feedstocks generally have low bulk densities and high water content. Therefore, the further they have to be transported the less cost effective it becomes to convert their energy to usable fuels. The mobile pyrolysis unit converts the energy in the feedstock into a high density bio-oil that then can be transported to the closest oil refinery. Therefore, transporting bio-oil long distances to a refinery is cheaper than transporting feedstock similar distances to centralized pyrolysis plants. Since the syngas can be used to dry the feedstock and power the pyrolysis system this leaves only the biochar byproduct to be utilized. Biochar Utilization One possible use of the biochar is to return it to the feedstock production fields as a soil amendment by land application. Biochar contains some of the plant nutrients which potentially could be utilized for feedstock production. According to Lehmann et al., (2006) crop growth is reduced only when there are very high applications of biochar. The bulk density of biochar is low and can lower the bulk density of clay soils and increase the water holding capacity of sandy soils (Laird, 2008). Biochar also retains a relatively large amount of carbon and has the potential to sequester carbon in the soil for long time periods. Carbon emissions are avoided by this sequestration and carbon trading could be introduced through these avoided emissions (Lehmann et al, 2006). Land application will require the biochar to be transported back to the harvested fields. Since biochar has a low bulk density, short transport distances from a mobile pyrolysis unit is cheaper than long transport distances from a centralized pyrolysis plant. 3

Mobile Pyrolysis Units The mobile pyrolysis unit assumed for this project is a fluidized bed system 12 inches in diameter. The unit processes feedstocks at a rate of 40 tons per day when the feedstock is at 10% or less moisture content. This system can produce bio-oil at a rate of 50 gallons per ton of feedstock and a biochar production rate of 10 tons per day (S. Capareda, personal communication, 2009). The pyrolysis units are mounted on trailers for easy transport. The mobile pyrolysis units will require approximately one acre of land including space for temporary storage of feedstock, bio-oil, and biochar along with space for large tractor trailers to maneuver. Study Location The location analyzed in this GIS project was the North Central U.S. and included the states of Montana, Wyoming, North Dakota, South Dakota, Nebraska, Iowa, Wisconsin, Illinois, and Indiana. La Salle County in Illinois was chosen for a pilot study based on the fact that in 2007, La Salle County was number one in corn production in Illinois. Corn areas planted and harvested for grain and grain yield and production from 1999 to 2008 in LaSalle County are tabulated in Table 1. Table 1. Corn for grain production from 1999 to 2008 in La Salle County, IL (USDA, 2010c). Year Planted all purposes (acres) Harvested (acres) Yield (bushel) Production (bushel) 1999 276,000 273,000 142 38,766,000 2000 284,000 281,600 152 42,803,200 2001 278,000 275,500 159 43,804,500 2002 283,000 279,600 136 38,025,600 2003 288,000 285,500 162 46,251,000 2004 305,000 302,400 179 54,129,600 2005 311,000 308,600 139 42,895,400 2006 298,000 293,900 182 53,489,800 2007 348,000 345,000 186 64,170,000 2008 336,000 334,000 192 64,128,000 Average 300,700 297,910 163 48,846,310 Feedstocks The feedstocks analyzed in the project are corn stover and energy sorghum. Corn stover was chosen because of its high production rates throughout the North Central region. Energy sorghum was chosen for analysis because it is designed specifically for the production of biofuels. It was assumed that energy sorghum would be planted in locations where grain sorghum is now planted. Energy sorghum does not flower in temperate climates and accumulates large amounts of biomass with a yield of approximately 15 Mg/ha (Heggenstaller et. al, 2008). Methodology 4

Overview of GIS Analysis Tool ArcGIS and ModelBuilder were used determine the optimum locations for the mobile pyrolysis units in the North Central region. Optimum locations were based on feedstock availability. A square grid was created to calculate feedstock availability systematically throughout the region. The initial size of the grid was based on the amount of feedstock needed to supply a mobile pyrolysis unit for six months. The amount of feedstock available for pyrolysis in each grid cell was calculated by, 1) summing up the planted feedstock area, and 2) multiplying the planted area by the 10-year average feedstock production rate. Grid cells were then ranked based on feedstock availability. After the grid cells were ranked, the distance required to transport the bio-oil to the nearest oil refinery will be determined. It was assumed that the mobile pyrolysis units were located in the center of the grid cell. It was also assumed that the average distance required to transport feedstock to the mobile pyrolysis unit and biochar back to the feedstock production fields was one half the grid cell size. GIS Methods - Feedstock Availability The locations of actual planted corn and sorghum fields for the North Central (NC) region were obtained from the 2008 cropland data layer (CDL) database as GIS raster files from the spatial analysis research section of National Agricultural Statistics Service (NASS). The spatial analysis research section annually provides the CDL with crop specific digital data layers in GIS raster formats (USDA, 2010a). The CDL program uses imagery from the Resourcesat-1 AWiFS and the Landsat 5 TM satellites, which produces digital categorized geo-referenced images exported to GeoTiff format for use in the ArcGIS interface. The geographic coordinate system of CDL raster data is GCS_WGS_1984, and the projected coordinate systems are WGS_194_UTM_Zone_14N/15M/16N for North Central region. The University of Illinois at Urbana-Champaign Institute of Natural Resources Sustainability performed CDL accuracy assessment of the remote sensing applications. Their analysis showed strong evidence of land cover classification success with 97.6% accuracy and a corresponding omission error of only 2.4% and an overall Kappa coefficient of 0.95 (Luman and Tweddale, 2008). CDL raster data for corn and sorghum fields planted in the NC region were converted to polygon shape files. Figure 1 shows the locations of the grain corn fields planted in 2008 in the North Central region. 5

Figure 1. Location of corn fields in the North Central states in 2008 (USDA, 2010a). Feedstock Grid Cells Square grid cells were developed to determine feedstock availability. The size of the grid cells were calculated using the following assumptions. LaSalle County, Illinois, was used as a pilot study for the grid cell analysis. Corn Stover The 10 year average (1999 to 2008) of corn for grain in LaSalle County was 164 bushel/acre. It was assumed that one bushel of corn produced 56 pounds of corn grain (USDA, 2010b). Therefore, for LaSalle County average grain production was 9,182 lbs/ac. It was also assumed that one pound of corn for grain was equivalent to one pound corn of stover biomass (Pordesimo et. al, 2004). However, only 25% of the corn stover was assumed to be available for pyrolysis thereby leaving 75% of corn stover in the field for erosion prevention and soil improvement. Therefore, for LaSalle County, an average of 2,295 lbs/ac of corn stover was assumed to be available as a feedstock for the mobile pyrolysis units. Since the mobile pyrolysis unit utilizes feedstock at a rate of 80,000 lbs/day (at 10% moisture content) this is equivalent to a utilization rate of 34.85 ac/day (80,000 lbs/day 2,295 lbs/ac). On an annual basis, the area of corn stover required to supply a mobile pyrolysis unit for one year in LaSalle County is 12,721 acres (34.85 ac/day x 365 days). This initial study focused on the movement of the mobile pyrolysis unit twice a year (every 6 months). Therefore, the area 6

of corn stover required for 6 months in LaSalle County was 6,325 acres. The grid size of 6,200 x 6,200 m used in the GIS analysis was chosen based on the area needed to harvest corn stover feedstock from 6,325 acres in LaSalle County. Bioenergy Sorghum It was assumed that a high biomass cultivar of sorghum (bioenergy sorghum) would be grown in fields where grain sorghum is currently grown. Figure 2 shows the locations of grain sorghum fields planted in 2008 in the North Central region. The basic assumption for bioenergy sorghum is that 100% would be harvested each year. The biomass production rate was assumed to be 15 Mg/ha or 13,388 lbs/ac (Heggenstaller et. al, 2008). The mobile pyrolysis unit utilizes 5.98 ac/day of sorghum feedstock (80,000 lbs/day 13,388 lbs/ac). The area of bioenergy sorghum required to supply a mobile pyrolysis unit for one year was 2,182 acres (5.98 ac/day x 365 days). Therefore the area of bioenergy sorghum required for 6 months was 1,091 acres. The grid used in the bioenergy sorghum analysis was 20,000 x 20,000 m. The grid size was based on the total area needed to accumulate bioenergy sorghum feedstock from1,091 acres. Figure 2. Location of sorghum fields in the North Central states in 2008 (USDA, 2010a). GIS Procedures ESRI ArcMap 9.3 version and ArcGIS Desktop 9.3 Service pack 1 were used for the GIS analyses. This software was run on a Dell precision T7500 workstation with Quad Core Intel R 7

Xeon R Processor E5504 2.0GHz, 4M L3, 4.8GT/s. The memory is 4GB and the graphic card is 1.0GB to support high resolution images in ArcGIS. ModelBuilder A square grid based on the area required to supply feedstock to a mobile pyrolysis unit for 6 months in LaSalle County, Illinois, was placed over the theme with the planted fields (corn and sorghum) as shown in Figure 3. The grid size for corn stover was determined to be 6,200 x 6,200 m and the gird size for the bioenergy sorghum was 20,000 x 20,000 m. The feedstock production area was calculated in each grid cell as follows, 1) the grid cell theme was overlain over the theme with the feedstock shape files, 2) feedstock fields that overlapped two or more cells were subdivided into fields that did not overlap cells, 3) the area of the feedstock fields within each cell was determined. This procedure was automated using ModelBuilder in ArcGIS as shown in Figure 4. (A) Grid (B) Corn field locations (C) Merged grid and corn field locations Figure 3. GIS methods were used to calculate feedstock availability for mobile pyrolysis units using the following methods, (A) a square grid was developed (the size of the grid cells was based on 10 year feedstock production rates), (B) planted feedstock fields as polygon shape files from NASS were used, and (C) the grid and feedstock shape files were merged and shape files that overlapped multiple grid cells were subdivided into shape files that do not overlap multiple cells. Model Builder was used to automate these procedures. Figure 4. A flowchart of the Model Builder procedure used to determine feedstock availability for mobile pyrolysis units. 8

Pyrolysis Logistics Hauling Distances A basic assumption was that the mobile pyrolysis units were located in the center of each grid cell. The average distance required to transport feedstock to a mobile pyrolysis unit was assumed to be half the grid cell dimension. Similarly, the average distance required to transport biochar back to the feedstock production fields was assumed to be half the grid cell size. Results Corn Stover Corn stover feedstock production rates for mobile pyrolysis varied from 2,184 to 2,464 lbs/ac in the North Central states assuming a 25% harvest rate. As shown in Figure 5 and Table 2, the top 100 locations for corn stover feedstock production in the North Central states were located in Illinois (57 sites), Nebraska (29 sites), and Iowa (14 sites). The highest corn stover feedstock location based on the grid size of 6,200 x 6,200 m was in Hall County, NE. The average corn production rates range from 156 bu/ac to 176 bu/ac on a county basis for biomass production top 100 cells in the North Central region. Since the grid size was 6,200 x 6,200 m, the average hauling distance for feedstock from the field to the mobile pyrolysis unit was 3,100 m. Likewise, the average hauling distance of biochar from the mobile pyrolysis unit to the feedstock production fields was 3,100 m. Figure 5. Results of the GIS analysis of corn stover availability for mobile pyrolysis in the North Central states for 2008. The locations of the top 100 corn stover feedstock cells based on a grid size of 6,200 x 6,200 m are shown. The grid size is based on a six month move time for the mobile pyrolysis unit. 9

Table 2. The location of the top 100 corn stover feedstock cells for mobile pyrolysis based on a grid size of 6,200 x 6,200 m in the North Central states. No. of Top 100 Feedstock Cells Illinois 57 Nebraska 29 Iowa 14 Bioenergy Sorghum The feedstock production rate for bioenergy sorghum was 13,388 lbs/ac with the basic assumption of 100% harvest. As shown in Figure 6 and Table 3, the top 50 locations for sorghum feedstock production in the North Central states were located in Nebraska (35 sites) and South Dakota (15 sites). The highest sorghum feedstock location based on the grid size of 20,000 x 20,000 m was in Thayer County, NE. Since the grid size was 20,000 x 20,000 m, the average hauling distance for feedstock from the field to the mobile pyrolysis unit was 10,000 m. Likewise, the average hauling distance of biochar from the mobile pyrolysis unit to the feedstock production fields was 10,000 m. Figure 6. Results of the GIS analysis of bioenergy sorghum availability for mobile pyrolysis in the North Central states for 2008. The locations of the top 50 bioenergy sorghum feedstock cells based on a grid size of 20,000 x 20,000 m are shown. The grid size is based on a six month move time for the mobile pyrolysis unit. 10

Table 3. The location of the top 50 bioenergy sorghum feedstock cells for mobile pyrolysis based on a grid size of 20,000 x 20,000 m in the North Central states. No. of Top 50 Feedstock Cells Nebraska 35 South Dakota 15 Conclusions GIS procedures were developed to determine the optimum locations (based on feedstock availability) for mobile pyrolysis in the North Central states for corn stover and bioenergy sorghum feedstocks. The GIS methods are semi-automated allowing the feedstock analysis to be performed quickly over large regions. The average transport distance of feedstock from the field to a mobile pyrolysis unit (based on a six month move time) was 3,100 m for corn stover and 10,000 m for bioenergy sorghum. These very short transport distances greatly reducing feedstock hauling costs and demonstrate the utility of using mobile pyrolysis units to densify bioenergy feedstocks. In addition, these short hauling distances will also facilitate the movement of biochar back to the feedstock production fields. In one year a mobile pyrolysis unit can densify approximately14,000 tons of biomass to 657,000 gal of bio-oil and 4,000 tons of biochar. The mobile units significantly reduce feedstock and biochar hauling logistics. In addition, the conversion of biomass energy from low density feedstocks to high density bio-oil greatly enhances the economic potential of bioenergy production from corn stover and energy sorghum in the North Central states. This is the first step needed to establish a network of pyrolysis stations in the NC region to produce bio-oil in a sustainable system using mobile pyrolysis units. The procedure to schedule the location and move times for the mobile pyrolysis units includes, 1) assessment of optimum locations based on feedstock availability, 2) determination of feedstock and biochar transport distances based on harvest grid size, 3) distance required to move the pyrolysis unit to a new location, 4) distance from mobile pyrolysis setup location to the nearest oil refinery (Figure 7), 5) economic assessment of costs and benefits for each location, and 6) assessment of future weather conditions at each location. 11

Figure 7. The location of oil refineries that produce more than 2,000 barrels per day in the North Central region in January 2008 (USDOE, 2009). Future work The economic model will be integrated with the GIS analysis to locate optimum pyrolysis locations based on probabilities of potential returns on investment. The movement of mobile pyrolysis units will also be integrated with weather prediction models. This will allow weather information and crop production forecasts to be factored into the decision making process to identify optimum locations for the mobile pyrolysis units. Acknowledgements The authors are thankful for grant support from the Department of Energy, Sun Grant Initiative, North Central region. References Bridgwater, A.V., D. Meier, D. Radlein. 1999. An overview of fast pyrolysis of biomass. Organic Geochemistry. 30(12) 1479-1493. Czernik, S., and A. V. Bridgwater. 2004. Overview of applications of biomass fast pyrolysis oil. Energy Fuels. 18(2) 590-598. 12

Heggenstaller, A.H., R.P. Anex, M. Liebman, D.N. Sundberg, and L.R. Gibson. 2008. Productivity and nutrient dynamics in bioenergy double-cropping systems. Agronomy Journal. 100(6) 1740-1748. Laird, D. A. 2008. The charcoal vision: a win-win-win scenario for simultaneously producing bioenergy, permanently sequestering carbon, while improving soil and water quality. Agronomy Journal. 100(1) 178-181. Lehmann, J., J. Gaunt, and M. Rondon. 2006. Bio-char sequestration in terrestrial ecosystems a review. Mitigation and Adaptation Strategies for Global Change. 11(2) 403-427. Luman, D. and T. Tweddale. 2008. Assessment and potential of the 2007 USDA_NASS cropland data layer for statewide annual land cover applications. Illinois Natural History Survey Technical Report. (49) Mohan, D., C.U. Pittman, and P.H. Steele. 2006. Pyrolysis of wood/biomass for bio-oil: a critical review. Energy and Fuels. 20(3) 848-889. Pordesimo, L.O., W.C. Edens, and S. Sokhansanj. 2004. Distribution of aboveground biomass in corn stover. Biomass and Bioenergy. 26(4) 337-343. USDA. 2010a. Cropland Data Layer. National Agricultural Statistics Service. Research and Development Division. Available at: http://www.nass.usda.gov/research/cropland/sars1a.htm. Accessed 20 October 2009. USDA. 2010b. Plant Nutrient Content Database. USDA Natural Resources Conservation Service. Available at: http://www.nrcs.usda.gov/technical/ecs/nutrient/tbb1.html. Accessed 21 October 2009. USDA. 2010c. Quick STATS. USDA National Agricultural Statistics Service. Available at: http://www.nass.usda.gov/quickstats/pulldata_us_cnty.jsp. Accessed 20 October 2009. USDOE. 2009. Ranking of U.S. Refineries. Energy Information Administration. Independent Statistics and Analysis. Available at: http://www.eia.doe.gov/neic/rankings/refineries.htm. Accessed 12 November 2009. 13