1. Background and Problem Definition.
Background. Coal provides 30.4% of electricity in the United States [1]. Coal power plants are confronted with the challenge of reducing greenhouse gas (GHG) emissions. How can power plants continue to produce a third of the nation’s electricity while reducing emissions? By cofiring coal with sustainable harvested biomass.
The goal of the Clean Power Plan [2] is to make steps towards reducing the carbon emissions from power plants and take action on climate change. This plan regulates greenhouse gas emissions on existing fossil fuel-fired electric generating power plants. The EPA has defined “building blocks” for reducing emissions into a Best System of Emissions Reduction (BSER). The building blocks are as follows:
1. Improving heat rate at affected coal-fired steam electric generating units (EGU).
2. Substituting increased generation from lower-emitting existing natural gas combined cycle units for generation from higher emitting affected steam generating units.
3. Substituting increased generation from new zero-emitting renewable energy generating capacity for generation from affected fossil fuel-fired generating units.
This project relates to pollution prevention control since biomass cofiring reduces the amount of coal the power plant uses and replaces it with a sustainable renewable energy source. Cofiring biomass has the potential to significantly reduce air pollution from coal-fired power plants. Thus, this project lies under the Clean Air Act – Section 103 and it is aligned to the goals of the People, Planet and Prosperity program.
Problem Definition. A recent report from U.S. Department of Energy (DoE) [3] identified that a key barrier to increasing biomass use for electricity generation is the high cost of biomass based on both supply and logistical challenges. Previous cofiring efforts lead by DoE and individual utilities found that at least 10% biomass can be mixed with ground coal to be pulverized in the current coal milling and particle entrainment feed line. Replacing 10% of the coal used by an average coal power plant with an equivalent amount of switchgrass can reduce yearly CO2 emissions by nearly three million tons. This study [3] highlighted as a gap the need for advanced logistical systems that include depots where biomass is processed to achieve a cofiring rate of 20%. To date, there is a lack of integrated sophisticated optimization models that can assist decision makers with the process/logistics design that minimizes the overall cost while quantifies the greenhouse gas reduction benefits of substituting high volumes of biomass (i.e., 20%) in coal power plants. Our proposed project aims to overcome this challenge. The developing of a sophisticated expert system to cofire biomass in power plants will aid not only developed countries but also the developing world to reduce their greenhouse gas emissions and take a proactive role towards climate change.
The long-term goal of this project is to develop a robust response to the following P3 challenge: the nexus of greenhouse gas emissions and electricity generation by (1) developing an expert system that assists coal power plants decision makers and stakeholders to design an optimal production and logistics network to cofire coal with switchgrass.
2. Purpose, Objectives and Scope.
This project focuses on creating novel mathematical models and optimization algorithms to optimize the logistics of growing, harvesting, transporting, and processing enough switchgrass using agricultural lands near coal power plants to supply at least 10% co-firing rate or greater. A mixed-integer optimization model that minimizes biomass production, harvesting and transportation costs was proposed. This project offers an excellent introduction to useful analytical skills students could use in their future professional endeavors.
Biomass supply system logistics is a key element in the assessment of economic and environmental impact of cofiring biomass in power plants and it is particularly relevant for the scaling of the biomass use. The biomass supply systems can be broadly broken into the harvesting and transportation operations with their associated costs. The modeling and optimization of biomass supply systems is key in assessing the feasibility of biomass cofiring in a cost-effective and sustainable manner.
The main goal of this project is to increase public awareness of greenhouse gas emissions for energy production and assess the cost and greenhouse gas reduction benefits of cofiring large amounts of biomass in existing coal power plants. This project is not intended to be an analysis of policies regarding renewable portfolio standards or the optimal use of land/biomass for energy production.
The objectives of this research follow:
1. Modeling and Optimizing Logistics. An optimization model to design a biomass supply system able to meet the cofiring demand at minimum harvesting, storage and logistics costs.
2. Testing the Models in a Real-life Case Study. A case study based on a power plant in Bexar County, which currently purchases coal from Wyoming. The proposed biomass is switchgrass (i.e., an herbaceous biomass) to be grown and harvested in two counties nearby the power plant.
3. Assessing Sustainability considering Climate Change. Four climate change scenarios used to evaluate the impact of climate change in switchgrass yields and logistics.
4. Creating a Preliminary Web-based Tool. This objective materializes the optimization model, data obtained from simulations and collected cost data into a DSS, which will expand the use and applicability of the models and data to a broader community (i.e., students, academicians, industry government and government policy makers).
3. Data, Findings, Outputs/Outcomes
Objective 1. Modeling and Optimizing Logistics. We proposed a Hub-and-Spoke model, which consists of three sets of nodes. The first set of nodes represents the parcels, the second set corresponds to the depots (i.e., a facility to wrap, storage and consolidate biomass bales), and the third one is related to the power plant(s). Figure 1 shows a visual representation of a simplified Hub-and-Spoke model with 4 parcels, 3 depots facilities and 1 power plant. This figure exemplifies the possible connections between parcels and depots (arcs T1), between depots and the power plant (arcs T2), and directly from parcels and the power plant (arcs T3). This optimization model aims to find the parcels to be utilized for the production of switchgrass, the depots that are needed, and the distribution network that minimizes the overall supply chain cost. The proposed mathematical programming formulation can be used to solve instances of various sizes (i.e., local, regional or national case studies). The decision variables correspond to flows from parcels to depots, depots to power plants and from parcels to power plants (in metric tons).
Figure 1. Schematic Diagram of a Hub-and-Spoke Model.
Moreover, binary variables are used to determine how many depots are needed and where to
allocate them. Thus, the problem is modeled as a Mixed-Integer Programming model (MIP).
The objective function is to minimize the expected cost for harvesting, processing and
transporting switchgrass, as well as the cost for opening depots. We model constrains that (1)
restrict the supply of biomass for every parcel to the maximum yield, (2) assure a mass balance in
every depot, (3) assure the biomass demand satisfaction with the inclusion of a slack variable, and
(4) set up a limit on the storage capacity at depots, among others.
Objective 2. Testing the Models in a Real-life Case Study. The case study corresponds to a local power company located in San Antonio, Texas, which operates several power plants including a 1350MW coal power plant [4]. This power plant consumed approximately 6.4 million tons of coal in 2014. The utility company relies on coal to generate about 38% of the electricity. With electricity demand in the U.S. expected to grow by more than 25% through 2030, reducing GHG emissions
is critical.
The first step in the data collection process was to estimate the yield of switchgrass in the two counties nearby the power plant. Next, the costs estimation for producing one metric ton of switchgrass was obtained. Lastly, we determined the distances from the centroid of each parcel to each designated depot, from parcels directly to the power plant and from depots to the power plans and used these distances to estimate the transportation costs using two types of trucks.
Switchgrass Yield. The switchgrass yield was obtained for the counties of Wilson and Atascosa. These counties were divided into 1940 and 2940 parcels, respectively. Each parcel size is 100 Hectares. The United States Department of Agriculture – Agricultural Research Service (USDA-ARS) maintains a software package called ALMANAC, which is capable of simulating crop growth among other agricultural performance metrics [5]. With this software, the growth of switchgrass was simulated on the agricultural lands surrounding the power plant. The coordinates determined from the parcel units were used to calculate the actual yield. The crop used in this ALMANAC analysis was Alamo switchgrass and its corresponding management details. The management for Alamo switchgrass entails initial planting and annual collection of the biomass without hurting the root to allow the plant to grow the following year without the need to re-plant.
Each county was analyzed in five scenarios: one baseline scenario from 2015 to the end of 2024, and two climate change scenarios (A2 & B2) for years 2050 and 2080. The climate change scenarios are the same ones used in the analysis of Behrman et al. [6]. The A2 scenario represents an extremely pessimistic future, which assumes large pollution increase, halted economic growth, and not much progress in the change of technology. A2 also predicts a substantial increase in carbon dioxide levels in the atmosphere as well as in temperature. The B2 scenario corresponds to a moderate increase in CO2 levels in the atmosphere and a smaller increase in temperature.
The yield of the baseline scenarios as well as the climate change scenarios for 2050 and 2080 are shown in Figure 2. As expected, the climate change scenarios show a lower yield compared to the original 2015-2024 scenario.
Water resources. The Soil and Water Assessment Tool (SWAT) hydrologic model (Arnold et al. 1998) was used in the present study to answer the following question: what are the main impacts of climate change on the hydrology cycle in the study area? SWAT is a continuous, semi-distributed river basin model which requires input information regarding climate, topography, soil type, and land use. Two climate scenarios, Historic and Climate Change were generated. The data shows that the changing climate will consistently increase temperatures and reduce rainfall averages for all months of the year, with exception of August. In average, the aggregate results of the projected climates indicate that annual precipitation will be reduced from 29.3 to 26.3 inches and average temperature will increase in approximately in 5º F in all months of the year.
Figure 2. Biomass Yield [tons/hectare] in Four Scenarios.
Harvesting Costs. Harvesting costs in the model include rent, baling, fertilizing, and swathing. These values are based on the cost of producing switchgrass for biomass feedstock [7]. The cost of rent per hectare varies by county. It is assumed that the land will not be irrigated when selecting rent prices [8]. Cost estimations may be seen in Table 1. The harvesting cost is computed according to the following equation, which factors in land size and the yield for the area.
Logistics Costs. Transportation costs are calculated using the distances between the key entities. To estimate such distances, it was assumed that transportation is performed along public roads, which are stored in North America street maps and can be processed with the aid of The Open Source Routing Machine software [9]. The distance matrices obtained are: from parcels to potential depots, from potential depots to the power plant, and directly from parcels to the power plant. There are two separate transportation models: one considers a small truck to transport biomass from the parcel to the depot and the other a larger truck to transport the biomass from the depot to the power plant [10]. The following table is primarily based on the development and implementation of integrated biomass supply analysis and logistics model [11]. The cost figures in Table 2 are then applied into the following equation.
Equations 1 and 2 are summed up to produce the total cost of purchasing one ton of biomass from a particular parcel and having it delivered to the plant.
Additional Restrictions. The annual investment cost is obtained by adding the annual storage cost to the cost of wrapping the biomass (to prevent deterioration) multiplied by the total capacity in storage, this totals to $333,000 per depot location.
The demand of switchgrass needed by the power plant depends on the co-firing rate. The coal power plant in 2015 was produced 97,441,853 MMBtus of energy using 5,019,754 metric tons of coal [12]. Each short ton of coal was releasing 17.61 MMBtus of energy in 2015 [13]. It was noticed that efficiency was improved each year. Switchgrass does not produce as much energy per ton as coal, only releasing 15 MMBtus per short ton [14]. A ratio was computed and applied to calculate the total amount of switchgrass needed to power the plant as 5,893,191 metric tons. Three cofiring rates are studied. The respective amount of metric tons required to produce the coal plant’s energy is shown in Table 3.
In the event there is a shortage in the number of tons of switchgrass available in these two counties, a penalty cost was factored into the model. If there is a switchgrass shortage, additional biomass switchgrass will need to be purchased from Tennessee. Tennessee was selected because of the biomass market already available. Market price for switchgrass in Tennessee is $66 per metric ton [15]. Switchgrass is cheaper per ton in our model because the yield per hectare is greater than in other regions.
Depot Selection. The potential depot locations were selected by evenly spacing approximately 30 potential depot locations for each county. A total of 59 depot locations are used in the model.
Objective 3. Assessing Sustainability considering Climate Change. The numerical experimentation consisted in evaluating 15 scenarios, that is, the baseline case, the 2050 A2 and B2 climate change cases, and the 2080 A2 and B2 climate change with cofiring rates of 10, 15 and 20 percent. Out of all the scenarios tested, the results of the four more insightful scenarios are compared and discussed in this section (see Figures 4-6 to visualize the solutions). Table 4 shows the optimal design for the baseline scenario with 10 and 20%. The last row in Table 4 shows the cost for purchasing the amount of coal necessary to produce the same energy generated with a 10% or 20% cofire rates [13]. Replacing 10% of coal with switchgrass increases the cost by about 37%; however, this cost comparison does not include benefits from reduction of CO2 and health issues discussed in section 4. Including the social cost of carbon in the coal price, provides a positive return on investment, and hence, our logistics design is viable.
Table 4 shows that baseline scenario has the smallest total cost and covers all the demand utilizing the biomass transported within the network (i.e., locally grown and harvested switchgrass and logistics infrastructure), whereas climate change scenario 2050 A2 (not favorable) and B2 (favorable) at 20% cofiring rate require biomass supplied by a third-party vendor (i.e., switchgrass supplier outside the state of Texas) at a cost of $70M and $39M, respectively. Since the yield in climate change scenarios is not sufficient to offer a competitive harvesting cost, some biomass needs to be obtained from a third-party supplier (e.g., in this case, the state of Tennessee). In the baseline scenario, the logistics infrastructure is economical feasible since it consolidates biomass in three depots. The total cost for harvesting switchgrass and investing in a biomass logistics network is $52.7M. Based on the results, scaling-up the cofiring rate is viable since the approximately one fourth of the available parcels are utilized at 20% cofiring rate. Increasing the switchgrass cofiring, and consequently the demand, will allow economies of scale as the number/capacity of the depots is increased. Decreasing the harvesting and logistics costs will potentially lead to a cost-competitive production of clean electricity. To make cofiring a sustainable option, the future climate change scenarios should be more favorable than the current yield estimations so that enough biomass is available.
Objective 4. Creating a Preliminary Web-based Tool. During phase I, significant progress was achieved toward developing the foundation of a web-based tool that materializes the data generated through simulations, optimization models and advanced visualization tools. The alpha version can be found in our project’s website (http://utsaengineer.wpengine.com/kcastillo/p3home/). The tool is divided into three modules. Data Depot module is a repository of the simulation runs of the models ALMANAC, SSURGO and SWAT. These sanitized data sets enable reproducible research. The Visualization module displays the solutions from the optimization model (i.e., optimal supply and logistics network) as shown in Figures 4 and 5. The Discovery Room module aims to serve as an outreach portal were presentation, videos from workshops, papers, and reports generated from this project are posted. The DSS is publicly available and will be an inexpensive mean for coal power plants to gain understanding of the biomass cofiring potential to reduce the GHG emissions.
4. Discussion, Conclusions, and Recommendations. In this section, we discuss the cost analysis of biomass cofiring including not only operational costs (as discussed in section 3) but social and healthcare benefits from reduction of emissions.
Quantitative benefits to people, prosperity and the planet. Carbon dioxide released during loading, unloading, transportation, stacking, and baling of biomass was calculated using estimates from the integrated biomass supply analysis and logistics model [16]. For every metric ton of coal burned, 2.01 metric tons of CO2 are released into the atmosphere. This number was calculated based on an equation from the Energy Information Administration [17]. The value is based on the average carbon content of subbituminous coal at 40 percent [18], and the heating value per pound [12]. Table 5 below summarizes the emissions. Using switchgrass instead of coal reduces the amount of CO2 emissions by about 90 percent.
To quantify the benefits from reductions in emissions. The Social Cost of Carbon (SCC) is considered, which is an estimate of economic damages associated with CO2 emissions, accounting for climate change damages, changes in net agricultural productivity, human health, property damage from increased flood risk, and changes in energy system costs. SCC is estimated using three different discount rates that calculate damages from the time of CO2 release until 2050. In this calculation, three percent was used [19]. In Table 5, net coal emissions were calculated by subtracting the estimated CO2 output of transporting biomass from the amount of CO2 released by burning coal. The total SCC was calculated after converting the 2007 dollars provided in the report for inflation to March 2017 dollars, $43.36 per metric ton of CO2 [20]. Therefore, the social costs of the net emissions ($79.3M) using 20% coal are greater than the cost of cofiring 20% of switchgrass ($52.7M).
The Clean Air Taskforce assessed the negative health impacts and calculates the likelihood of health risks from fine particle pollution from 500 power plants across the United States [21]. This report provided the 2012 estimates of death, heart attacks, asthma attacks, hospital admissions, chronic bronchitis and asthma ER visits at a rate per 100,000 people in Bexar County. When multiplying this rate by the 2012 population of Bexar county: 1,789,834 residents [22], the total number of associated health risks is estimated at $267.4 M. At 20 percent biomass cofiring rate, about six lives would be saved by the reduction of emissions alone.
References
1. EIA Short-term energy outlook http://www.eia.gov/forecasts/steo/report/prices.cfm.
2. EPA Fact Sheet Social Cost of Carbon; 2015.
3. Boardman, R. D.; Cafferty, K. G.; Nichol, C.; Searcy, E. M.; Westover, T.; Wood, R.;
Bearden, M. D.; Cabe, J. E.; Drennan, C.; Jones, S. B. Logistics, Costs, and GHG Impacts of
Utility Scale Cofiring with 20% Biomass; Pacific Northwest National Laboratory (PNNL),
Richland, WA (US), 2014.
4. EPA Request for Action Plan regarding CPS Energy – J.T. Deely Power Plant. 2014.
5. Kiniry, J. R.; Arnold, J.; Xie, Y. Applications of Models with Different Spatial Scales. Agric.
Syst. Model. F. Res. Technol. Transf. 2002, 207–227.
6. Behrman, K. D.; Kiniry, J. R.; Winchell, M.; Juenger, T. E.; Keitt, T. H. Spatial forecasting of
swtchgrass productivity under current and future climate change scenarios. 2013, 23, 73–85.
7. Griffith, A. P.; Epplin, F. M.; Redfearn, D. D. Cost of Producing Switchgrass for Biomass
Feedstock. 2010, 1–7.
8. National Agriculture Statistics Service Quick Stats. Retrieved April 7, 2017; 2016.
9. Open Source Routing Machine http://map.project-osrm.org/ (accessed Jan 1, 2017).
10. Hart, J. Biomass Supply Chain Logistics for Co-Firing Coal Power Plants., 2015.
11. Sokhansanj, S.; Mani, S.; Turhollow, A.; Kumar, A.; Bransby, D.; Lynd, L.; Laser, M. Largescale
production, harvest and logistics of switchgrass (Panicum virgatum L.) – current technology
and envisioning a mature technology. Biofuels, Bioprod. Biorefining 2009, 3, 124–141.
12. U.S. Energy Information Administration. Shipments to J T Deely.
13. U.S. Energy Information Administration. Electricity Data Browser; 2015.
14. Perlack, R. D.; Turhollow, A. F.; Eaton Switchgrass Production in the USA; 2011.
15. Chugh, S.; Yu, T. E.; Jackson, S. W.; Larson, J. A.; English, B. C.; Cho, S.-H. Economic
analysis of alternative logistics systems for Tennessee-produced switchgrass to penetrate energy
markets. Biomass and Bioenergy 2016, 85, 25–34.
16. Sokhansanj, S.; Kumar, A.; Turhollow, A. F. Development and implementation of integrated
biomass supply analysis and logistics model (IBSAL). Biomass and Bioenergy 2006, 30, 838–
847.
17. Hong, B. D.; Slatick, E. R. Carbon Dioxide Emission Factors for Coal; 1994.
18. U.S. Energy Information Administration. Subbituminous and Bituminous Coal Dominate
U.S. Coal Production; 2011.
19. EPA Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility
Generating Units. 2015, 80.
20. Bureau of Labor Statistics CPI Inflation Calculator 2017.
21. Clean Air Taskforce The Toll From Coal: An Updated Assessment of Death and Disease
from America’s Dirtiest Energy Source; 2010.
22. Texas Association of Counties Bexar County Profile; 2015.