Article pubs.acs.org/est
Using Infrastructure Optimization to Reduce Greenhouse Gas Emissions from Oil Sands Extraction and Processing Richard S. Middleton†,* and Adam R. Brandt‡ †
Los Alamos National Laboratory, Earth and Environmental Sciences, Los Alamos, New Mexico 87545, United States Department of Energy Resources Engineering, Stanford University, Stanford, California 94305-2220, United States
‡
S Supporting Information *
ABSTRACT: The Alberta oil sands are a significant source of oil production and greenhouse gas emissions, and their importance will grow as the region is poised for decades of growth. We present an integrated framework that simultaneously considers economic and engineering decisions for the capture, transport, and storage of oil sands CO2 emissions. The model optimizes CO2 management infrastructure at a variety of carbon prices for the oil sands industry. Our study reveals several key findings. We find that the oil sands industry lends itself well to development of CO2 trunk lines due to geographic coincidence of sources and sinks. This reduces the relative importance of transport costs compared to nonintegrated transport systems. Also, the amount of managed oil sands CO2 emissions, and therefore the CCS infrastructure, is very sensitive to the carbon price; significant capture and storage occurs only above 110$/tonne CO2 in our simulations. Deployment of infrastructure is also sensitive to CO2 capture decisions and technology, particularly the fraction of capturable CO2 from oil sands upgrading and steam generation facilities. The framework will help stakeholders and policy makers understand how CCS infrastructure, including an extensive pipeline system, can be safely and cost-effectively deployed.
■
INTRODUCTION Unconventional oil production has the potential to enhance U.S. energy security and reduce its dependence on overseas imports. Unconventional oil resources include bituminous sands (colloquially known as oil or tar sands), oil shale, and coal-to-liquids. The Alberta oil sands have resources in place of approximately 1.5 T bbl.1 In 2010, the Alberta oil sands produced around 1.5 million barrels of oil a day (M bbl/d), with projected production expansion as high as 5 M bbl/d by 2030.2 Thus, the Alberta oil sands could provide a significant fraction of U.S. oil consumption in coming decades.3 Concerns have been raised over oil sands environmental impacts, including air pollution (GHGs, VOCs, and other air pollutants),4−6 water pollution (hydrocarbon contamination and tailings disposal concerns),7,8 and land use impacts (land disturbance and ecosystem fragmentation).9,10 Oil sands can be extracted from near-surface deposits (i.e., mining) or from deep deposits (i.e., in situ thermal production). However, extracting and upgrading oil from oil sands is more energy intensive than conventional oil production. Oil sands extraction and processing results in well-to-wheels GHG (CO2 eq.) emissions that are 5−30% higher than those from most conventional petroleum feedstocks, although conventional oil that is highly depleted or produced with high flaring volumes can have higher emissions than oil sands operations.4−6 Policies such as the U.S. Energy Independence And Security Act (EISA)11 incentivize reductions of CO2 emissions from unconventional oil production. Canadian federal and provincial © 2012 American Chemical Society
attention is also focused on this issue: regulatory actions (e.g., Specified Gas Emitters Regulation)12 and research efforts13 are aimed at reducing GHG emissions from the oil sands. Oil sands operations emit CO2 from various sources. Mining results in centralized emissions from processing and separation facilities, tailings handling, and upgrading of bitumen to synthetic crude oil, or SCO (upgrading improves hydrocarbon quality and allows transport via pipeline). Mining operations also include diffuse emissions sources such as trucks. In situ production emissions are largely centralized at steam generation facilities.14,15 CO2 capture and storage (CCS) technology involves capturing and compressing CO2 at large industrial sources (e.g., coal-fired power plants), transporting the CO2 in a dedicated pipeline network, and injecting the CO2 into subsurface geologic reservoirs (e.g., depleted oil and gas fields).16 Existing CCS projects currently manage as much as 1 to 3 MtCO2/yr by connecting a single CO2 source with a single injection/storage site; examples include Weyburn (Canada),17 In Salah (Algeria),18 and Slepiner Vest (Norway).19 The world’s largest CO2 capture operation is Exxon’s LaBarge facility in Wyoming (a gas sweetening plant, not a postcombustion source of CO2), which manages roughly 7 Received: Revised: Accepted: Published: 1735
September 11, 2012 December 30, 2012 December 31, 2012 December 31, 2012 dx.doi.org/10.1021/es3035895 | Environ. Sci. Technol. 2013, 47, 1735−1744
Environmental Science & Technology
Article
Figure 1. Study region. The map shows the three major oil sands regions in Alberta, major CO2 emission sources (≥0.5 MtCO2/yr) from the oil sands industry (red circles), and potential sequestration reservoirs (blue circles). Red circles are proportional to emissions, ranging from 0.5 MtCO2/ yr up to 4.5 MtCO2/yr. CO2 storage capacity ranges between 20 MtCO2 and 492 MtCO2.
■
MtCO2/yr;20 the CO2 is used for enhanced oil recovery (EOR). This work brings together four streams of scientific research in order to better understand how the cost of CO2 emissions mitigation varies with the size and extent of CO2 infrastructure in the oil sands region. Effective CO2 mitigation requires efficient investment to minimize cost of compliance and maximize mitigation potential. We develop a suite of optimal infrastructure investment plans for capturing, transporting, and injecting/storing up to 36 MtCO2/yr from the Alberta oil sands industry. Figure 1 illustrates the 22 CO2 emission sources and 16 geologic reservoirs used in this study. The study is divided into four sequential steps. First, we analyze the engineering and economics of capturing CO2 from the oil sands industry.21 Second, we identify a set of spatially dispersed reservoirs capable of securely storing oil sands CO2 emissions. These sites are characterized using a combination of existing acid gas injection data22−25 and the CO2-PENS risk assessment model.26−28 Third, we identify a candidate network of routes where CO2 pipelines could be constructed.29 And fourth, we use the SimCCS optimization framework to integrate CO2 capture, storage, and transportation decisions in order to understand how the oil sands industry might respond to a price on CO2 emissions (such as the Specified Gas Emitters Regulation).26,29−34
ESTIMATING CO2 SOURCE MAGNITUDES AND CONCENTRATIONS Emissions from oil sands operations are calculated using reported energy consumption in oil sands mining and upgrading operations35 as well as steam injected at in situ projects.36 On-site emissions are estimated for each project (mining, upgrading, and in situ) using reported fuel consumption rates. Because this infrastructure model is not a life cycle accounting model, only on site emissions are included (e.g., no embodied emissions are calculated) and no emissions credits are awarded for displacement of energy offsite. That is, we seek to understand the physical emissions at each site, and do not attempt to model the resulting total impact due to offsite induced emissions. Mining and Upgrading Fuel Consumption Data. CO2 emissions from mining operations are modeled using fuel consumption data for 2010. Data are collected for integrated mining and upgrading projects, standalone mines and upgraders, and integrated in situ and upgrading facilities.37 Data collected include: volumes of bitumen mined and processed; volumes of SCO produced; quantities of process gas, coke, SCO, and natural gas consumed, processed and wasted; and electricity generated, consumed, and exported. Not all fuel consumption is expected to produce capture-ready CO2, due to the small scale or dispersed nature of some combustion 1736
dx.doi.org/10.1021/es3035895 | Environ. Sci. Technol. 2013, 47, 1735−1744
Environmental Science & Technology
Article
Figure 2. Comparison of capture cost (y-axis) and amount of CO2 available (x-axis) for the different capture processes. Numbers within each block represent total CO2 (MtCO2/yr) emitted by each process annually (data from literature sources,41,42 see SI for more information on construction of figure). Costs are in $CAD 2008.
sources. Table S2 of the Supporting Information, SI outlines assumptions made about the ability to capture each stream. Regulatory statistics are used in preference to other sources (e.g., corporate sustainability reports) because these data are detailed and are reported consistently across operations. Fuel characteristics vary between projects due to differing technologies applied. Process gas and coke fuel compositions are collected from the literature by project. Natural gas composition is assumed consistent across all in situ and mining projects. From these compositions, heating value and produced CO2 are computed (see SI for calculation details). In Situ Fuel Consumption Data. In situ processes inject steam into the subsurface to reduce oil viscosity, and include steam-assisted gravity drainage (SAGD) and cyclic steam stimulation (CSS). In situ producers report the amount of natural gas, produced gas, and electricity consumed by project.37 These data include natural gas consumed for on site cogeneration. Monthly fuel consumption data for 2010 (1000 m3 of natural gas and produced gas consumed per month) are collected for all projects with available data. CO2 concentrations are adjusted for cases with and without cogeneration (see Table S3 in SI). For projects that do not report fuel consumption data (a small minority of small projects), in situ fuel consumption is modeled using reported 2010 steam oil ratios (SOR) (m3 cold-water-equivalent steam injected per m3 of bitumen produced).35 Enthalpy of steam at SAGD and CSS conditions is assumed to be 2.8 GJ/m3.38−40 Steam generation is assumed to occur in 80−85% efficient (lower heating value basis) once through steam generators (OTSGs) (high and low emissions cases, respectively). Heat recovery from produced fluids is assumed to provide 10 to 30% of the enthalpy of steam40 in the high and low emissions cases, respectively. Resulting modeled demand for natural gas and produced gas ranges from 2.3 to 3.15 GJ/m3 of steam. CO2 Concentration and Capture Cost Variation with Fuel Type and Combustion Process. The cost of CO2 capture varies strongly with the concentration of CO2 in the separated stream. Therefore, each CO2 stream above is assigned an assumed CO2 concentration:41 Benfield H2 unit emits CO2
at 99 mol %, producing at 1.73 MtCO2/y (Syncrude) and 0.23 MtCO2/y (Suncor). Other H2 units emit 18 mol % CO2. Natural gas and process gas combustion emit 9 mol % CO2. Coke combustion is assumed to emit CO2 at 13 mol %, while gasification of residues at the Long Lake project is assumed to emit 44 mol % CO2. A key requirement for modeling is assessing the likelihood of CO2 capture as a function of the size of the CO2 stream. We only include CO2 sources ≥0.5 MtCO2/yr as possible to be captured with current CCS systems. Because it is not possible to discern the size of individual CO2 streams from reported energy consumption data, there is significant uncertainty in the volumes of feasibly captured CO2. A “low availability” case is also run where only 50% of the total CO2 volumes are assumed to be able to captured, as well as low, medium, and high cost scenarios. A wide range of costs of capturing CO2 is provided in the literature on oil sands CCS. Studies for oil sands facilities suggest costs of $20 to $230/tCO2 captured, depending on CO2 concentrations and analyst assumptions.41,42 As this work focuses on infrastructure-scale modeling rather than cost estimation for CO2 capture technologies, we use these literature values. Table S3 of the SI shows the resulting cost ranges and the values selected for use in the model. Importantly, the amount of CO2 produced by each process is not evenly distributed (Figure 2).
■
CO2 INJECTION AND STORAGE Commercial-scale injection and storage of oil sands CO2 requires identifying multiple suitable sequestration sites capable of handling millions of tonnes of CO2 each year. We have identified 16 potential CO2 reservoirs based on actual injection records for acid gas injection22−25 (Figure 1). Acid gas is principally made up of hydrogen sulfide (H2S) and CO2 stripped from natural gas and crude oil during the sweetening process. Due to its environmental impact if released, acid gas is compressed and stored in deep saline aquifers. As of 2003, the Alberta oils sands industry had stored 2.5 Mt CO2 and 2 Mt H2S.43 Acid gas injection and storage is an excellent analog for 1737
dx.doi.org/10.1021/es3035895 | Environ. Sci. Technol. 2013, 47, 1735−1744
Environmental Science & Technology
Article
Figure 3. Cost surface and candidate pipeline network (white lines). Costs range from low (light brown) to high (dark brown). Black lettered circles identify key characteristics in the cost surface: A = Lake; B = First Nation land; C = river; D = transmission line; E = road; and F = pipeline ROW. Inset box 1 centers on the Fort McMurray area. Inset box 2 highlights the Edmonton area.
Storage capacity and injection cost are the first-order criteria for defining reservoir suitability. We use the CO2-PENS risk assessment model26−28 for calculating CO2 capacity and injection cost for each of the 16 reservoirs. CO2-PENS is a dynamic analytical systems model for understanding reservoir performance including costs, capacities, and CO2 leakage risks (such as well bore failure) using data on formation thickness, depth, permeability, and porosity (see Table S3 of the SI). CO2PENS estimates the total volume of CO2 that can be stored over a given time period, the maximum formation injection rate, and the distance required between injection wells. CO2 injection/storage will likely require water to be extracted from the formationCO2-PENS includes this capability.45 Water extraction creates space for the CO2 to be stored, keeps the formation below its fracturing pressure, and allows the CO2 plume to be managed. The water produced per tonne of CO2 injected depends on the reservoir pressure and temperature. Higher pressures increase CO2 density, therefore less water is displaced, while higher temperatures decrease CO2 density thus increasing water production (Table S4 of the SI). In the 16 modeled reservoirs, one tonne of CO2 displaces 1.26−1.78 m3
CO2 injection and storage. For example, acid gas streams can contain up to 90% or more CO2 by volume, are stored in the subsurface under conditions similar to almost-pure CO2, and the acid gas is injected and handled using similar methods as CO2 sequestration. The 16 storage reservoirs are identified, using a screening methodology, from a set of acid gas sites with complete reservoir characteristics (see Table S4 in the SI).22−25 First, only reservoirs used to store acid gas with at least 50% CO2 content are included. Second, storage sites farther than 500 km from oil sands industrial sources are excludedthis excludes sites in British Columbia and the far north of Alberta. Third, small capacity sites are excluded (
