Engineering Papers⌕ Search

Engineering topics

Yang, Hung-Chia

Publications and source records attributed to Yang, Hung-Chia.

Teaching Freight Mode Choice Models New Tricks Using Interpretable Machine Learning Methods

Understanding and forecasting the intricate freight mode choice behavior under various industry, policy, and technology contexts is essential in freight planning and policymaking. Numerous models have been developed in prior studies to provide insights into freight mode selection, the majority of which use discrete choice models such as multinomial logit (MNL) models. However, logit models often rely on linear specifications of independent variables, despite potential nonlinear relationships in the data. Moreover, there often lacks a heuristic and efficient approach to identify such complex relationships to define the logit model specifications. To fill this gap, we developed an MNL model for freight mode choice using the insights from state-of-the- art machine learning (ML) models. ML models can capture the nonlinear nature of the complex decision-making process, and recent advances in 'explainable AI' have greatly improved their interpretability. The interpretable ML methods help enhance the performance of MNL models and advance knowledge of freight mode choice. Specifically, the influential factors and their relationship with individual modes are identified using SHapley Additive exPlanations (SHAP) to improve the MNL's performance. The workflow is demonstrated in a case study of Austin, Texas, and the SHAP results reveal multiple nonlinear relationships predicted by ML models. Incorporating those relationships into MNL model specifications improves the interpretability and accuracy of the MNL model compared to a conventional MNL model. Findings from this study can be used to guide freight planning and inform policymakers and practitioners on how key factors affect freight decision-making.

ADVANCED PROPULSION SYSTEMS,MATHEMATICS AND COMPUT↗

Decarbonizing the US Energy System

Recent rapid and unexpected cost reductions in decarbonization technologies have accelerated the cost-effective decarbonization of the US economy, with greenhouse gas (GHG) emissions falling by 20% from 2005 to 2020. The literature on US economy-wide decarbonization focuses on maximizing long-term GHG emissions reduction strategies that rely mostly on renewable energy expansion, electrification, and efficiency improvements to achieve net-zero GHG emissions by 2050. While these studies provide a valuable foundation, further research is needed to properly support decarbonization policy development and implementation. In this review, we identify key decarbonization analysis gaps and opportunities, including issues related to cross-sectoral linkages, spatial and temporal granularity, consumer behavior, emerging technologies, equity and environmental justice, and political economy. We conclude by discussing the implications of these analysis gaps for US decarbonization pathways and how they relate to challenges facing major global emitters.

08 HYDROGEN↗

A Strategic Condition Framework for Energy Innovation

Building upon previous work in innovation, technology, and economic fields, we propose a conceptual framework to address the interplay dynamics of public policy and the value chains of clean energy technology innovation systems, by focusing on market structural and strategic conditions in relation to the rate and direction of knowledge diffusion within activity-based value chains. Our framework extends beyond socio-technical transition models by considering such nuances as: 1) the implications of geography, 2) the locus of knowledge types, 3) the dynamics of financial and knowledge flow, and 4) emphasizing the importance of both market structure and institutional conditions. At the firm level, we consider dynamics of interactions between suppliers and customers, innovation activities, the competitive environment, and the firm’s broader strategic relationship to governance activities and structures; we allow for the fact that each of these aspects of a firm’s identity may have a strategically relevant geographic dimension (e.g., service territories, locational concentration of input resources, etc.). We provide a proof-of-concept application of the U.S. utility-scale solar photovoltaic (PV) and onshore wind sectors, including their interactions with the U.S. power market, which we in present in a web-based information platform (Energy I-SPARK, ei-spark.lbl.gov).

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Firm Synthesizer and Supply-chain Simulator (SynthFirm) v1.0

SynthFirm is a large-scale agent-based freight demand model which generates a complete synthetic population of firms in the U.S. and the business-to-business commodity flows between them. Using publicly available data sources as inputs, SynthFirm simulates detailed firm and fleet characteristics, commodity production and consumption, formation of supply chains, and selection of shipping modes, all of which are essential drivers of commodity flow at a disaggregate level.

Xu, Xiaodan↗

Dishwashers in the Residential Sector: A Survey of Product Characteristics, Usage, and Consumer Preferences

Data on consumer purchasing decisions, usage, and behaviors relating to residential appliances help to inform technical and economic analyses related to the energy and water used by those appliances, including dishwashers. Existing publicly available data for dishwashers include two regularly conducted national surveys that describe dishwasher ownership and usage. The United States (US) Department of Energy’s (DOE) Energy Information Administration’s Residential Energy Consumption Survey (RECS) records the presence of a dishwasher in the home, the numbers of times per week the dishwasher is operated, and the dishwasher age, along with household demographic characteristics. The US Census Bureau’s American Housing Survey (AHS) also records the presence of a dishwasher in the home along with household demographic characteristics.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

SMART Mobility. Mobility Decision Science Capstone Report

The U.S. Department of Energy’s Systems and Modeling for Accelerated Research in Transportation (SMART) Mobility Consortium is a multiyear, multi-laboratory collaborative, managed by the Energy Efficient Mobility Systems Program of the Office of Energy Efficiency and Renewable Energy, Vehicle Technologies Office, dedicated to further understanding the energy implications and opportunities of advanced mobility technologies and services. The first three-year research phase of SMART Mobility occurred from 2017 through 2019, and included five research pillars: Connected and Automated Vehicles, Mobility Decision Science, Multi-Modal Freight, Urban Science, and Advanced Fueling Infrastructure. A sixth research thrust integrated aspects of all five pillars to develop a SMART Mobility Modeling Workflow to evaluate new transportation technologies and services at scale. This report summarizes the work of the Mobility Decision Science Pillar. The Mobility Decision Science Pillar sought to fill gaps in existing knowledge about the human role in the mobility system including travel decision-making and technology adoption in the context of future mobility. The objective was to study how underlying preferences, needs, and contextual factors might constrain or hasten future transportation system scenarios. For information about the other Pillars and about the SMART Mobility Modeling Workflow, please refer to the relevant pillar’s Capstone Report.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Survey of Hearth Products in U.S. Homes

There are over 7 million gas-fired fireplaces currently installed in US households. On an annual basis, these use substantial energy. However, the details of the fireplace energy use and the factors that influence it are currently not well understood. Fireplaces are a type of hearth product, which is a product category that primarily consists of fireplaces, stoves, and gas log sets. For the purpose of this study, the fuels used in hearth products are primarily natural gas, propane, or electricity. They may be vented or unvented. This study reports the results of a web survey of 2,100 respondents in the United States performed in February 2016. The responses were cleaned and weighted using the raking method to form a nationally representative population. The reported data include hearth product characteristics, usage information, and repair and maintenance practices. The hearth product characteristics include the hearth product type, fuel type, ignition system type, features, venting, and installation details. The usage information includes seasonal usage of the main burner and standing pilot (if present), daily usage, and the primary utility (whether decorative or for heating). These raw data are further processed and combined with values from the literature to estimate the annual operating hours and energy use and to assess how these are impacted by the hearth product type, features, age, and the main heating appliance, if present. Based on the survey responses, the estimated average annual hours of usage was 234 for the main burner, and 4,593 for the standing pilot. The results presented provide the most comprehensive data regarding hearth products in the United States published to date. These new data allow for an improved understanding of hearth products’ energy use, which in turn may facilitate the development of more informed analyses, and ultimately more efficient hearth products and reduced energy use. These new data also provide insight into topics not previously studied, such as the effect of hearth product features on energy use.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Confronting Regulatory Cost and Quality Expectations. An Exploration of Technical Change in Minimum Efficiency Performance Standards

The dual purpose of this project was to contribute to basic knowledge about the interaction between regulation and innovation and to inform the cost and benefit expectations related to technical change which are embedded in the rulemaking process of an important area of national regulation. The area of regulation focused on here is minimum efficiency performance standards (MEPS) for appliances and other energy-using products. Relevant both to U.S. climate policy and energy policy for buildings, MEPS remove certain product models from the market that do not meet specified efficiency thresholds.

99 GENERAL AND MISCELLANEOUS↗

U.S. Residential Miscellaneous Refrigeration Products: Results from Amazon Mechanical Turk Surveys

Amazon Mechanical Turk was used, for the first time, to collect statistically representative survey data from U.S. households on the presence, number, type and usage of refrigerators, freezers, and various “miscellaneous” refrigeration products (wine/beverage coolers, residential icemakers and non-vapor compression refrigerators and freezers), along with household and demographic information. Such products have been poorly studied to date, with almost no information available about shipments, stocks, capacities, energy use, etc. A total of 9,981 clean survey responses were obtained from five distinct surveys deployed in 2012. General refrigeration product survey responses were weighted to demographics in the U.S. Energy Information Administration’s Residential Energy Consumption Survey 2009 dataset. Miscellaneous refrigeration product survey responses were weighted according to demographics of product ownership found in the general refrigeration product surveys. Model number matching for a portion of miscellaneous refrigeration product responses allowed validation of refrigeration product characteristics, which enabled more accurate estimates of the penetrations of these products in U.S. households. We estimated that there were 12.3±1.0 million wine/beverage coolers, 5.5(–3.5,+3.2) million residential icemakers and 4.4(–2.7,+2.3) million non-vapor compression refrigerators in U.S. households in 2012. (All numerical results are expressed with ranges indicating the 95% confidence interval.) No evidence was found for the existence of non-vapor compression freezers. Moreover, we found that 15% of wine/beverage coolers used vapor compression cooling technology, while 85% used thermoelectric cooling technology, with the vast majority of thermoelectric units having capacities of less than 30 wine bottles (approximately 3.5 cubic feet). No evidence was found for the existence of wine/beverage coolers with absorption cooling technology. Additionally, we estimated that there were 3.6±1.0 million hybrid refrigerator-wine/beverage coolers and 0.9±0.5 million hybrid freezer-wine/beverage coolers in U.S. households. We also obtained estimates of miscellaneous refrigeration product capacities, lifetimes, purchase and installation costs, repair frequencies and costs, and maintenance costs. For wine/beverage coolers, we also obtained information on the penetration of built-in units, AC/DC operating capability, the use of internal lights, and distributions of door opening frequencies. This information is essential to develop detailed estimates of national energy usage and life-cycle costs, and would be helpful in obtaining information on other plug-load appliances. Additional information not highlighted in the main report was presented in Appendices.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Energy Efficiency and Minimum Standards: a Market Analysis of Recent Changes in Appliance Eenergy Efficiency Standards in the United States

The objective of this paper is to examine the effect of energy efficiency standards on prices and market shares of major household appliances in the United States. The household appliance market is analyzed using point-of-sale (POS) retail data of clothes washers and dryers from NPD Group. Energy efficiency information for each clothes washer model was obtained from the Federal Trade Commission (FTC) appliance energy database. Using these data, we conclude that energy efficiency standards had an overall negative impact on clothes washer price, and products with a mid-low efficiency rating experienced the greatest short run price drop following the January 2004 and 2007 new standard effective dates. In neither instance is there any evidence of an increase in prices of affected appliances. In both cases we analyse the price effects of the standard both relative to dryers as a counterfactual group. Additionally, we show that market shares for clothes washers shifted towards greater efficiency over the course of the period between 2002 and 2011.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

U.S. Residential Miscellaneous Refrigeration Products: Results from Amazon Mechanical Turk Surveys

Amazon Mechanical Turk was used, for the first time, to collect statistically representative survey data from U.S. households on the presence, number, type and usage of refrigerators, freezers, and various “miscellaneous” refrigeration products (wine/beverage coolers, residential icemakers and non-vapor compression refrigerators and freezers), along with household and demographic information. Such products have been poorly studied to date, with almost no information available about shipments, stocks, capacities, energy use, etc. A total of 9,820 clean survey responses were obtained from four distinct surveys deployed in 2012. General refrigeration product survey responses were weighted to demographics in the U.S. Energy Information Administration’s Residential Energy Consumption Survey 2009 dataset. Miscellaneous refrigeration product survey responses were weighted according to demographics of product ownership found in the general refrigeration product surveys. Model number matching for a portion of miscellaneous refrigeration product responses allowed validation of refrigeration product characteristics, which enabled more accurate estimates of the penetrations of these products in U.S. households. We estimated that there were 12.3±1.0 million wine/beverage coolers, 5.5(–3.5,+3.2) million residential icemakers and 2.9(–2.5,+4.5) million non-vapor compression refrigerators in U.S. households in 2012. (All numerical results are expressed with ranges indicating the 95% confidence interval.) No evidence was found for the existence of non-vapor compression freezers. Moreover, we found that 15% of wine/beverage coolers used vapor compression cooling technology, while 85% used thermoelectric cooling technology, with the vast majority of thermoelectric units having capacities of less than 30 wine bottles (approximately 3.5 cubic feet). No evidence was found for the existence of wine/beverage coolers with absorption cooling technology. Additionally, we estimated that there were 3.6±1.0 million hybrid refrigerator-wine/beverage coolers and 0.9±0.5 million hybrid freezer-wine/beverage coolers in U.S. households. We also obtained estimates of miscellaneous refrigeration product capacities, lifetimes, purchase and installation costs, repair frequencies and costs, and maintenance costs. For wine/beverage coolers, we also obtained information on the penetration of built-in units, AC/DC operating capability, the use of internal lights, and distributions of door opening frequencies. This information is essential to develop detailed estimates of national energy usage and life-cycle costs, and would be helpful in obtaining information on other plug-load appliances. Additional information not highlighted in the main report was presented in Appendices.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

U.S. residential consumer product information: Validation of methods for post-stratification weighting of Amazon Mechanical Turk surveys

We present two post-stratification weighting methods to validate survey data collected using Amazon Mechanical Turk (AMT). Two surveys focused on appliance and consumer electronics devices were administered in the spring and summer of 2012 to each of approximately 3,000 U.S. households. Specifically, the surveys asked questions about residential refrigeration products, televisions (TVs) and set-top boxes (STBs). Filtered data were assigned weights using each of two weighting methods, termed “sequential” and “simultaneous,” by examining up to eight demographic variables (income, education, gender, race, Hispanic origin, number of occupants, ages of occupants, and geographic region) in comparison to reference U.S. demographic data from the 2009 Residential Energy Consumption Survey (RECS). Five key questions from the surveys (number of refrigerators, number of freezers, number of TVs, number of STBs and primary service provider) were evaluated with a set of statistical tests to determine whether either method improved the agreement of AMT with reference data, and if so, which method was better. The statistical tests used were: differences in proportions, distributions of proportions (using Pearson’s chi-squared test), and differences in average numbers of devices as functions of all demographic variables. The results indicated that both methods generally improved the agreement between AMT and reference data, sometimes greatly, but that the simultaneous method was usually superior to the sequential method. Some differences in sample populations were found between the AMT and reference data. Differences in the proportion of STBs reflected large changes in the STB market since the time our reference data was acquired in 2009. Differences in the proportions of some primary service providers suggested real sample bias, with the possible explanation that AMT user are more likely to subscribe to providers who also provide home internet service. Differences in other variables, while statistically significant in some cases, were nonetheless considered to be minor. Depending on the intended purpose of the data collected using AMT, these biases may or may not be important; to correct them, additional questions and/or further post-survey adjustments could be employed. In general, based on the analysis methods and the sample datasets used in this study, AMT surveys appeared to provide useful data on appliance and consumer electronics devices.

32 ENERGY CONSERVATION, CONSUMPTION, AND UTILIZATI↗

Can Standards Increase Consumer Welfare? Evidence from a Change in Clothes Washer

We study prices and sales of individual clothes washer models before, during and after a 2007 standard that banned manufacture (but not sale) of low-e ciency units and increased the threshold for Energy Star certi cation. While quantities sold of washer models banned from manufacture decreased sharply, prices for banned models increased only modestly. At the same time, sales of higher-e ciency units rose markedly while prices for high-e ciency units declined. On average, washer e ciency increased but prices changed little. A simple welfare analysis indicates that consumer welfare loss from banned washers was far outweighed by gains from lower-priced high-e ciency units. While a full cost-bene t analysis is not feasible with the available data, we estimate a lower-bound gain in consumer surplus equal to 6-16 percent of total sales. This result may accord with earlier theoretical research that shows quality standards can increase welfare in monopolistically competitive industries that possess increasing returns to scale (Ronnen, 1991). Thus, if energy e ciency is a close proxy for quality, energy e ciency standards may increase competition, market e ciency and welfare.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗

Incorporating Experience Curves in Appliance Standards Analysis

The technical analyses in support of U.S. energy conservation standards for residential appliances and commercial equipment have typically assumed that manufacturing costs and retail prices remain constant during the projected 30-year analysis period. There is, however, considerable evidence that this assumption does not reflect real market prices. Costs and prices generally fall in relation to cumulative production, a phenomenon known as experience and modeled by a fairly robust empirical experience curve. Using price data from the Bureau of Labor Statistics, and shipment data obtained as part of the standards analysis process, we present U.S. experience curves for room air conditioners, clothes dryers, central air conditioners, furnaces, and refrigerators and freezers. These allow us to develop more representative appliance price projections than the assumption-based approach of constant prices. These experience curves were incorporated into recent energy conservation standards for these products. The impact on the national modeling can be significant, often increasing the net present value of potential standard levels in the analysis. In some cases a previously cost-negative potential standard level demonstrates a benefit when incorporating experience. These results imply that past energy conservation standards analyses may have undervalued the economic benefits of potential standard levels.

29 ENERGY PLANNING, POLICY, AND ECONOMY↗