Lab Affiliation(s):
Sloan Automotive Lab
Advisor:
Prof. John Heywood
Areas of Expertise:
  • carsharing
  • automotive technology
  • energy
Expected date of graduation:
June 1, 2015

Stephen Zoepf

  • PhD

Department: 

  • Data, Systems, and Society

Lab Affiliation(s): 

Sloan Automotive Lab

Advisor: 

Prof. John Heywood

Top 3 Areas of Expertise: 

carsharing
automotive technology
energy

Personal Statement: 

In humankind’s quest to reduce energy consumption, transportation is a crucial area of focus.  Our transportation choices are an integral part of our daily lives, and the range of reliable forms of transportation available to each of us has a profound determining effect on both the economic opportunities we can access, and on our leisure options. Yet, in the United States, the transportation sector consumes more energy than any other end-use sector. The overarching question I pursue in my research is this: “How do we reduce the energy we consume in transportation while simultaneously improving the quality and availability of transportation?”

More specifically, my objective is to deepen our knowledge in three ways.  First, to better understand the real-world capability of new technologies and dispassionately assess both advantages and disadvantages.  Second, to understand how consumers evaluate new technology and add detail to individual decisions about whether or not to use something new.  Finally, to understand where new technology can be applied to provide the greatest benefit and inform relevant corporate and governmental policies.

Real-World Technology Effectiveness

The phrase “your mileage may vary” has long been a part of the vernacular, as customers are wary of products that may not deliver on promised capabilities.  My work develops more robust estimates of the capabilities of new vehicle propulsion technologies, showing that effectiveness is enormously contingent on usage.

Current measurement methods, based on test-cycle performance, are plainly inadequate and with flex-fuel and hybrid vehicles, the challenge of portraying performance now applies to multiple energy sources. We need to become more sophisticated in presenting energy consumption data to consumers.  In the next few years I will develop analytical tools to improve to help researchers build realistic expectations of real-world performance.  I have begun this work with two separate studies (Rodgers, Zoepf, and Prenninger, 2014; Zoepf et al., 2013) in which I found large heterogeneity in energy consumption of plug-in vehicles, partially driven by accessory usage unrelated to propulsion. I also developed a novel model of charging behavior and estimated the change in energy use that would result from changes to charging and vehicle characteristics.

These studies form the building blocks of simulation tools that are less computationally demanding than full vehicle simulation but more informative than conventional tests. Such metrics will be applicable beyond the field of transportation.  In many industries performance measurements are the subject of endless gamesmanship as new benchmarks are developed, and then products are “tuned” to perform well.  I anticipate that the techniques I develop for vehicle performance will be broadly applicable to benchmarks of other types of energy consumption.

Technology Adoption

Consumer adoption is a critical step in moving technology from the research environment to widespread usage, and is the subject of widespread study.  Researchers have reported a wide range of estimates of consumer willingness to pay for energy efficiency: from payback periods as short as three years (Greene et al., 2005; Santini and Vyas, 2005), to reports that consumer discount rate corresponds to their own cost of capital. (Busse et al, 2013)  Such vastly different estimates suggest that we do not fully understand the process of adoption.

I have used stated-preference experiments to investigate how attitudes to technology change in the carsharing market—where consumers invest only a few dollars—in choosing whether to drive an efficient vehicle.  My initial work has shown that carsharing exposes hundreds of thousands of drivers to new technology, and the vehicle technology is less important to users that other service attributes. (Zoepf & Keith) During the course of this work I have developed discrete-choice models that more fully capture indicators of latent variables with machine learning techniques.

This work will contribute to our understanding of the process by which introduction leads to adoption.  We need to develop more nuanced approaches to understanding how consumer awareness translates into a willingness to use.  The carsharing environment offers the chance to study consumers as they experiment with new products and technology in a more natural environment. I will continue to study how the brief exposure to technology in sharing and rental schemes impacts subsequent decisions to make greater commitments buy purchasing or leasing. 

Deployment as a Simulation and Optimization Problem

In public policy, technology goals are often based on oversimplified ideas of how a system will work.  In the transportation space, this has led to policies such as the California Zero Emission Vehicle mandate, which pushes for sales of PEVs with little regard for their usage.  However, my research has indicated that energy usage and fuel displacement are highly context-dependent, indicating we should pay as attention to where and how PEVs are used, not just how many are sold.  If our ultimate goal is the reduction of emissions and energy usage, then we need to deploy our limited technology and resources in effective ways.

I am currently undertaking a research project that investigates when and where deployment of electric vehicles is sensible in carsharing services. This work investigates the tradeoff between the need to keep vehicles earning revenue and need to recharge, providing guidance to deploy expensive assets where they will maximize energy consumption reduction while minimizing service disruption.  This project makes use of both optimization and simulation methodology to calculate benefits both under ideal circumstances and under a variety of more practical scenarios.  Initial results indicate that while a large number of EVs could be used in carsharing services, universal usage is an ineffective approach: small scale, targeted deployments are far more prudent.

This goal of this research is to develop tools that provide quick but effective guidance to program managers and policy makers.  By simplifying complex, energy-intensive systems using a series of heuristics, we can provide guidance in a digestible format to those who are in a position to make regulatory and logistical decisions but who lack the resources to perform analysis themselves.

Expected date of graduation: 

June 1, 2015

CV: 

Thesis Title: 

Technology and Carsharing: User Preferences, Energy Consumption and Fleet Allocation

Thesis Abstract: 

In the past decade a flexible form of urban automobile transportation, carsharing, has allowed more than one million users to rent cars by the minute or hour in cities around the U.S.  These shared vehicle fleets employ large numbers of alternative powertrain  vehicles which are used by hundreds of different drivers, presenting a unique opportunity to study the adoption and usage of vehicle technology in a low-risk environment.  This research uses the carsharing environment as a lens to examine large populations of carsharing users and their interaction with hybrids, and plug-in vehicles to answer three primary questions: (1) ``How does the driving behavior of carsharing members differ from private owners?'' (2) ``What preferences do carsharing users exhibit regarding vehicle technology and other service attributes?'' and (3) ``Are range-limited Plug-in Electric Vehicles capable of meeting the drive-cycle demands of carsharing service?''  

Top 5 Awards and honors (name of award, date received): 

ENI MIT Energy Initiative Fellow, 2014
Barry McNutt Award, 2014, Energy and Alternative Fuels Committees of the Transportation Research Board
Martin Energy Fellowship, 2013-2014
Singapore Global Challenge, 2013, Global Young Scientists Summit@one-north (1st Place)
Recipient of Infinite Mile Award for "Outstanding Service to the Institute"

5 Recent Papers: 

Zoepf, S. and Keith, D.  (In review)  User Decision-Making and Technology Choices in the US Carsharing Market. (PDF)

Rodgers, L., Zoepf, S., and Prenninger, J.  (2014) Analyzing the Energy Consumption of the BMW ActiveE Field Trial Vehicles with Application to Distance to Empty Algorithms.  mobil.TUM International Scientific Conference on Mobility and Transport - Sustainable Mobility in Metropolitan Regions, June 2014. (PDF)

Zoepf, S., D. MacKenzie, D. Keith, and W. Chernicoff. (2013) Charging Choices and Fuel Displacement in a Large-Scale Plug-in Hybrid Electric Vehicle Demonstration. Transportation Research Record: Journal of the Transportation Research Board, No. 2385, Transportation Research Board of the National Academies, Washington, D.C., 2013, pp. 1–10. (PDF)

MacKenzie, D., Zoepf, S. and Heywood, J. (2014) Determinants of U.S. Passenger Car Weight. International Journal of Vehicle Design 65 (1), 73-93. (PDF)

Zoepf, S. and Heywood, J.  (2012).  Characterizations of Deployment Rates in Automotive Technology.  (Paper 2012-01-1057)  Warrendale, PA: SAE International. (Link)

Contact Information:
77 Massachusetts Ave.
31-141
Cambridge
MA
02138
(201) 315-2889