New petroleum refining lifecycle model finds the variability in GHG emissions from refining different crudes as significant as magnitude expected in upstream operations
9 December 2012
|Comparison of GHGenius, JACOBS, TIAX, and the new PRELIM gasoline greenhouse gas (GHG) estimates using base case estimates and variations from the scenario analysis. Credit: ACS, Abella and Bergerson. Click to enlarge.|
Researchers at the University of Calgary (Canada) have developed the Petroleum Refinery Life-cycle Inventory Model (PRELIM). PRELIM uses a more comprehensive range of crude oil quality and refinery configurations than used in earlier models and can quantify energy use and greenhouse gas (GHG) emissions with detail and transparency the better to inform policy analysis, the duo suggests.
Using a scenario analysis to explore the implications of processing crudes of different qualities in different refinery configurations, and with a focus on oil sands products, they found differences of up to 14 g CO2eq/MJ of crude, or up to 11 g CO2eq/MJ of gasoline and 19 g CO2eq/MJ of diesel (the margin of deviation in the emissions estimates is roughly 10%). Put another way, “the variability in GHG emissions in the refining stage that results from processing crudes of different qualities is as significant as the magnitude expected in upstream operations”, they found.
Given the magnitude of upstream emissions from the sector, they said, their findings have implications for both policy and mitigation of GHG emissions. A paper on PRELIM and the scenario study is published in the ACS journal Environmental Science & Technology.
As background, they noted that the petroleum refining industry is the second-largest stationary emitter of GHG in the US, and the third-largest in the world. Annual GHG emissions from a large refinery are comparable to the emissions of a typical 500 MW coal-fired power plant. In the US, GHG emissions from refineries in 2010 represented nearly 12% of US industrial sector emissions or 3% of the total US GHG emissions.
This industry faces difficult investment decisions due to the shift toward “heavier” crude in the market, both domestic and imported. For example, in 1990, the fraction of imported crude into the US classified as heavy (at or below API gravity, a measure of density, of 20) was roughly 4%. By 2010 this fraction had increased to 15%. Between 2008 and 2015, it is estimated that more than $15 billion will be spent to add processing capacity specifically for heavy crude blends in US refineries. Each refinery must decide whether and how much they will process heavy crude while considering that processing such crudes requires more energy and results in higher refinery GHG emissions. These major capital investment decisions will impact the carbon footprint of the refining industry for decades to come.
Current and future environmental regulations will also affect the decisions faced by this industry. Life cycle assessment (LCA) has been expanded as a tool to enforce GHG emissions policies. For example, California’s Low Carbon Fuel Standard (CA-LCFS) embeds life cycle assessment within the policy to measure emissions intensity of various transportation fuel pathways through their full life cycle (including extraction, recovery, and transport). Using LCA in this way requires more accurate assessments of the emissions intensity upstream of the refinery for each crude. However, the varying quality of these crudes will also have significant implications for refinery GHG emissions.
...The implications for refinery GHG emissions of processing oil sands (OS) products provide a good case study due to the link between upstream processing decisions and refinery emissions, as well as the wide variety of OS products.—Abella and Bergerson
The two most prominent prominent North American life cycle (LC) tools are Natural Resource Canada’s GHGenius and Argonne National Laboratory’s Greenhouse Gases, Regulated Emissions, and Energy Use in Transportation (GREET). The GREET model and the CA-GREET version, used as the basis of CA-LCFS, do not account for the effects of crude quality at the refinery stage in their calculations (i.e., all crudes will have the same energy requirements and GHG emissions), the developers of PRELIM noted.
GHGenius accounts for crude quality by modifying a default energy intensity value using the average API gravity and sulfur content of an entire refinery crude slate and a regression model based on historic regional refinery performance data.
The models approaches do not decouple the effects of changes in energy requirements due to changes in crude quality and the changes in each refinery’s performance (e.g., process unit efficiencies), nor do they develop a consensus on the impact of allocation (how environmental impacts are split across products in a multiproduct industry), the authors said.
PRELIM can simulate up to ten specific refinery process configurations, each requiring a different amount of energy to process a crude and producing a different slate of final products including transportation fuels as well as heavy fuel oil; hydrogen from the naphtha catalytic reforming proces;, refinery fuel gas; and the possible production of coke or hydrocracking residue.
All the configurations include crude distillation, hydrotreating, and naphtha catalytic reforming processes. The configurations are differentiated by the presence of gas oil hydrocracking; fluid catalytic cracking (FCC); delayed coking; and residual hydrocracking. Supporting unit processes such as steam methane reforming (SMR) and acid gas treatment are also included.
For the scenario analysis, they established a base case to determine the refinery energy use and GHG emissions of a crude in a default refinery configuration based on a set of three broad refinery categories: hydroskimming refinery, medium conversion refinery, and deep conversion refinery. (All 10 refinery configurations in PRELIM fit into one of these three categories.).
The base case assigns each crude (oil sands and conventional) to the appropriate default refinery category, using API gravity and sulfur content of the whole crude as the criteria. Variations examined include changes in assumptions regarding refinery configuration, process energy requirements, energy use for production of hydrogen via SMR, and refinery fuel gas production.
Among their findings for the base case assumptions were that total refinery energy use ranged from 0.06 to 0.24 MJ/MJ of crude (340−1400 MJ/bbl of crude). The resulting GHG emissions of processing crudes of different qualities varied widely, mainly due to differences in hydrogen requirements. Total refinery GHG emissions range from 4 to 18 g CO2eq/MJ of crude being processed (23−110 kg CO2eq/bbl of crude).
For the 12 crudes considered in the base case:
- supply of hydrogen contributes from 0 to 44% of refinery emissions;
- process heating contributes 26−71%;
- FCC catalyst regeneration contributes 0−17%;
- steam contributes 2−7%; and
- electricity contributes 10−21%.
- Up to 48% of the emissions associated with hydrogen requirements result from the SMR (steam methane reforming) process unit.
When the full range of refinery configurations are run for each crude, the emissions changed as much as 12 g CO2eq/MJ of crude. Lighter and sweeter (lower in sulfur) crudes have increased GHG emissions above the base case since the base case assumes a simple hydroskimming configuration, and for heavier crudes (OS and conventional) there are deep conversion configurations in which the GHG emissions are higher or lower than those estimated in the base case.
A wider range of GHG emissions estimates is seen for oil sands products (2.5−26 g CO2eq/ MJ of crude) compared to conventional crudes (2.4−17 kg CO2eq/MJ of crude). Generally, they found, the highest estimates are for bitumen (9.3−26 kg CO2eq/MJ of crude). This represents potential cases such as dilbit being sent to a refinery and the diluent being separated and returned to the OS operation. The synthetic crude oils (SCOs) represented one of the highest and the lowest GHG emissions of all crudes considered. The heavy SCO crude category can have GHG emissions as high as 20 g CO2eq/MJ of crude. Light sweet SCO can have GHG emissions as low as 2.5 g CO2eq/MJ of crude.
The PRELIM application presented in this paper demonstrates that crude quality and the selected process units employed (i.e., the refinery configuration), as well as the energy efficiency of the process units, all play important roles in determining the energy requirements and emissions of processing a crude. The unique amount of hydrogen required to process each crude is dictated by the quality of the crude entering the refinery. It can be the major contributor to refinery energy use and GHG emissions for every crude. Therefore, this should be a key parameter used in estimating emissions. Emissions associated with providing the hydrogen required should also be the focus of emissions reductions at refineries.
This analysis provides insights that can help to inform emissions reductions decisions at refineries. Based on this analysis, the top three ways to reduce GHG emissions at refineries processing heavier crude will be to (1) reduce the amount of hydrogen consumed, (2) increase hydrogen production efficiency (and/or lower GHG emissions intensity of hydrogen production), and (3) capture CO2 from the most concentrated, highest volume sources (i.e., FCC and SMR).
...This analysis substantiates the claim that more accurate assessments of refinery emissions are required to better inform LC-based policies and avoid potential unintended consequences. Putting the refinery emissions variations into context, the variability in GHG emissions in the refining stage that results from processing crudes of different qualities is as significant as the magnitude expected in upstream operations (e.g., in this paper, the variability is up to 14 g CO2eq/MJ of crude, or up to 11 g CO2eq/MJ of gasoline and 19 CO2eq/MJ of diesel—based on the full range of base case crudes). If crudes are run through the same configuration, refinery performance (defined by efficiency of energy use) introduces important variation. The PRELIM application demonstrated up to 43% deviation in the GHG emissions burden attributed to a crude solely by varying the efficiency of the process units in one configuration.
...climate policies based on LCA should equally engage both parts of the supply chain (i.e., crude production/processing/transport and refining stages) to encourage the most cost-effective GHG emissions mitigation pathways. Directives such as the current High Carbon Intensity Crude Oil (HCICO) provision in the CA-LCFS that do not explicitly include these differences in the definition and principles/goals could lead to unintended consequences.—Abella and Bergerson
Jessica P. Abella and Joule A. Bergerson (2012) Model to Investigate Energy and Greenhouse Gas Emissions Implications of Refining Petroleum: Impacts of Crude Quality and Refinery Configuration. Environmental Science & Technology doi: 10.1021/es3018682
TrackBack URL for this entry:
Listed below are links to weblogs that reference New petroleum refining lifecycle model finds the variability in GHG emissions from refining different crudes as significant as magnitude expected in upstream operations: