4:00pm - 4:20pm
Integrating data science methods and Life Cycle Assessment (LCA): Application to the power sector
1Contractor to U.S. DOE, NETL, United States of America; 2U.S. DOE, NETL, United States of America
Data science is radically transforming the scientific landscape. Data-intensive frameworks such as life cycle assessment (LCA) can uniquely benefit from the establishment of large scale and national datasets, and the use of data-science methods. Additionally, the synthesis and application of these data repositories in LCA can aid in improving emissions inventories, technological characterization of product systems, and uncertainty quantification. In this work, a series of case studies are used to explore the intersection of data science and LCA of the power sector. The objective of this work is to provide a broad cross-section of data sources, methods, and approaches that have been applied to characterize the life cycle of power systems, but are generalizable across multiple LCA research domains. Several case studies are presented to highlight the utility of applying data-science methods to LCA of power systems including: (1) regionalized methane emissions from natural gas liquids unloading, (2) characterizing the environmental profile of U.S. hydropower, and (3) time series analysis of the fossil fleet from 2008 to 2016. A detailed description of each case study is provided below:
(1) A ‘bottom-up’ probabilistic model was developed using engineering first principles to quantify methane emissions from natural gas liquids unloading activities for 18 basins in the United States in 2016. For each basin, six discrete liquids unloading scenarios are considered, consisting of combinations of well types (conventional, unconventional) and liquids unloading systems (non-plunger, manual plunger-lift, and automatic-plunger lift). Data used to populate the model is sourced from Drilling Information, the 2016 Greenhouse Gas Reporting Program (GHGRP), and academic literature. (2) An LCA model was developed to quantify the greenhouse gas emissions and water footprint of 2016 U.S. hydropower. Information for hydroelectric powerplants was obtained via the National Inventory of Dams (NIDS) dataset, U.S. Energy Information Administration (EIA), and National Hydropower Asset Assessment Program (NHAA). A lasso linear model using recursive feature elimination and cross-validation was used to develop statistical regressions for hydroelectric carbon dioxide and methane emissions rates. Inverse Distance Weighting (IDW) interpolation was performed on climatological data obtained from NOAA and used to develop reservoir specific water consumption rates. Additionally, several allocation schemes were considered to apportion the emissions and water consumption across multipurpose hydroelectric reservoirs. (3) A data-driven model was developed to quantify the evolution of hourly time-series emissions from the U.S. fossil (coal, gas) fleet from 2008 to 2016. Additionally, this work evaluates fossil assets that have shifted to or from baseload operations over the 2008 to 2016 era; and studies the effect of this shift on plant-level performance metrics (gross efficiency, net capacity factor) and emissions (CO2, SO2, NOX) profiles. The analysis investigates several possible explanatory mechanisms for the shift in plant emissions including the installation of air pollution control equipment, environmental regulations, power plant technology, and changing plant operations. Data used to parameterize the model was obtained from publicly available sources including the Environmental Protection Agency’s (EPA) Air Markets Program Data (AMPD) and EIA. Several opportunities for integration and application to LCA are identified.
4:20pm - 4:40pm
Evaluating Alternative Strategies to Reduce Greenhouse Gas Emissions
University of California, Davis
California’s landmark 2006 Climate Change Solutions Act (Assembly Bill 32) tasked many government entities, including local governments and government agencies, with reducing greenhouse gas (GHG) emissions to 1990 levels by 2020, and 80% below 1990 levels by 2050. Identifying, quantifying, and then selecting among possible strategies to achieve GHG reductions is difficult, especially without a standardized approach for comparison. This study develops a GHG mitigation “supply curve” to support decision-making by the California Department of Transportation (Caltrans). The supply curve supports selection of the most cost-effective strategies for mitigation by undertaking the following process for each strategy: (1) quantify the net GHG emitted or avoided over a strategy’s lifecycle, (2) consider the timing of changes in GHGs, (3) explore the process and difficulty of implementation, and (4) calculate the initial and lifecycle costs of the strategy. A life cycle perspective is required for GHG accounting because benefits achieved during one stage of a strategy’s lifecycle may be reduced or reversed by carbon-intensive upstream or downstream stages. The timeframe for change is also important, both because of public policy targets and because emissions reductions that occur earlier can avert warming in the near-term, potentially avoiding or delaying climate tipping points. Emissions timing can be addressed by using time-adjusted warming potentials in lieu of standard global warming potentials. Similarly, a life cycle perspective is required for understanding costs because of limited funds that public and private entities can invest in GHG reduction strategies and the responsibility of public agencies to consider future costs. By combining time-adjusted GHG reductions and costs, each strategy can be assigned an “emissions reductions per dollar” value by which the strategies can be ordered to determine which of them achieve the biggest “bang for the buck”.
This study applies the above methodology to six mitigation strategies that could be implemented by Caltrans: energy harvesting through piezoelectric technology, efficient maintenance of pavement roughness, automating bridge tolling systems, increased use of reclaimed asphalt pavement, alternative fuel technology for the Caltrans fleet, and installing solar and wind energy technologies within the state highway network. These diverse strategies were selected to test the effectiveness of the evaluation process. Elaborating on the first strategy, studies have indicated potential to produce energy through piezoelectric sensors installed under highway pavements, which produce a voltage from the compression created by passing vehicles. Preliminary results suggest that 100 lane-miles of piezoelectric technology could reduce emissions by approximately 500,000 tons CO2e over 10 years, with initial costs and lifecycle savings in the millions of dollars. This type of output was generated for all strategies. The strategies are presented as a supply curve by placing them in terms of the magnitude of reduction potential and their cost (both initial and lifecycle); a color-coding scheme is also used to communicate confidence in the calculated values, which are affected by the reliability of cost and emissions change information available, and technology maturity.
4:40pm - 5:00pm
Monetized Impacts of Time Resolved Greenhouse Gas Emissions
Colorado State University, United States of America
Life cycle assessment (LCA) is a tool used to compare the environmental impacts of products. Standard LCA methods do not account for the timing of greenhouse gas emissions. As a result, these assessments do not include dynamic climate factors such as increasing atmospheric greenhouse gas concentrations. Furthermore, these assessments are limited to mid-point metrics that do not capture dynamic end-point impacts such as economic damage to society. These shortcomings can lead to an inaccurate comparison of greenhouse gas impacts from different products. This work presents two new methods to address the dynamic shortcomings of LCA. The first method leverages the social costs of greenhouse gases to determine the time-dependent impacts of greenhouse gas emissions. Using these monetized impacts, time-resolved greenhouse gas emissions are weighted based on when they are emitted. The weighted emissions are then summed to determine a present value of emissions relative to today’s environmental and economic conditions. The second method incorporates the social cost of greenhouse gases and time-resolved greenhouse gas emissions into techno-economic analysis. This method pulls societal costs into a techno-economic frame work and evaluates their impact relative to the other costs within production. These two new methods have been demonstrated on time-resolved LCA of electricity generation systems including coal, natural gas, carbon capture and sequestration, nuclear, solar, and wind. Recognizing the inherent uncertainty of future environmental and economic conditions, a range of future scenarios were evaluated. Results from the first method show that accounting for time-resolved monetized impact increases the present value of emissions across all but one of the systems considered. Technologies with large operational greenhouse gas emissions show the largest increase, with coal rising by 32% in the baseline scenario. Technologies with lower operational emissions see a minimal increase, with solar-photovoltaics rising just 1%. Results from the second method show a large impact on levelized cost of electricity (LCOE) for technologies that have significant operational greenhouse gas emissions. Looking at coal, the LCOE in a baseline scenario increases by 88%. Technologies with lower operational emissions see a smaller impact, with the baseline LCOE of nuclear increasing by just 1%. These results quantify the impact of extending analysis beyond mid-point metrics to account for dynamic economic damage and highlight the dramatic impact of large operational emissions extending over the lifetime of a system.
5:00pm - 5:20pm
Life cycle greenhouse gas emissions of U.S. LNG used for international power generation
1ExxonMobil Research & Engineering, United States of America; 2Massachusetts Institute of Technology; 3SeaRiver Maritime; 4Air Products and Chemicals, Inc.; 5IIT Bombay; 6Council on Energy, Environment and Water
The recent growth in U.S. natural gas reserves has led to interest in exporting liquefied natural gas (LNG) to countries in Asia, Europe and Latin America. Here, we estimate the life cycle greenhouse gas (GHG) emissions and life cycle freshwater consumption associated with exporting Marcellus shale gas as LNG for use in power generation in different import markets. The well-to-wire analysis relies on operations data for gas production, processing, transmission, and regasification, while also accounting for the latest measurements of fugitive CH4 emissions from U.S. natural gas activities. To estimate GHG emissions from a typical U.S. liquefaction facility, we use a bottom-up process model that can evaluate the impact of gas composition, technology choices for gas treatment and on-site power generation on overall facility GHG emissions.
Our results show that when U.S. LNG is delivered to Asian nations to generate power in F-class combined cycle power plants (50% efficiency, HHV basis), the life cycle GHG emissions will be 473 kg CO2eq/MWh (80% confidence interval: 452 - 503 kg CO2eq/MWh); life cycle GHG emissions associated with LNG-based power in Europe and South America may be as low as 459 kg CO2eq/MWh, due to shorter marine transportation distances. In markets with existing power plant fleets, the life cycle GHG emissions associated with U.S. LNG are ~54% lower than those associated with locally-produced coal. We also find that new coal power generation technologies (e.g. ultra-supercritical coal fired power), while emitting less than existing coal technologies, produce about two times the GHG emissions associated with newer gas power generation technologies (e.g. H-class combined cycle power plants) fueled by natural gas transported as LNG.