Event
ME Special Seminar | Probabilistic Graphical Modeling as a Use Stage Inventory Method
Monday, February 10, 2014
10:30 a.m.
DeWALT Seminar Room, 2164 Glenn L. Martin Hall
Kimberly Frye
kfrye@umd.edu
http://www.enme.umd.edu/events/seminars
Mechanical Engineering Special Seminar Series
Probabilistic Graphical Modeling as a Use Stage Inventory Method for Environmentally Conscious Design
Speaker: Cassandra Telenko, Ph.D., Postdoctoral Associate
Singapore University of Technology and Design (SUTD)
Massachusetts Institute of Technology (MIT)
Abstract:
When faced with questions about which product or activity is more environmentally friendly, the correct answer is almost universally, “It depends.” Environmentally conscious design guidelines provide general actions and rules of thumb for reducing environmental impacts, such as energy consumption, of products, but these reductions are conditional and conflicts between guidelines may arise. Such tradeoffs can be evaluated with standard energy life cycle inventories (LCI) that quantify the energy consumption of each stage in a product’s life cycle, such as material extraction and consumer use. LCI studies of different usage contexts are of special interest because a majority of energy consumption can occur during product use. Nevertheless, many energy LCIs assume average scenarios of how a product may be used and do not consider the variable energy consumption of different usage contexts and behaviors. Thus, a central hypothesis of this research is that charactering the usage context of a product as a set of factors can provide a higher fidelity investigation of energy consumption than commonly found in the literature and reveal specific conditions of use that exhibit the highest lifetime energy savings. In this research, a probabilistic graphical model (PGM) was combined with an LCI to represent the usage context as a network of factors characterized by local conditional probability distributions. Using the example of a lightweight vehicle redesign, a PGM was developed using publicly available data sources to describe usage context factors such as driver behavior, alternative driving schedules, and residential density. As evidence for the hypothesis, initial results suggest that targeting lightweight vehicles for usage contexts with high lifetime mileage, such as rural households with more residents, can increase energy savings by 10% more than an average usage scenario.
Bio:
Cassandra Telenko is a postdoctoral associate in the MIT-SUTD International Design Centre with joint appointments at MIT and SUTD. She received a National Science Foundation Graduate Research Fellowship to fund her graduate work at the University of Texas at Austin where she completed her MSE (2009) and PhD (2012) in Mechanical Engineering. She received her B.E. (2007) in Mechanical Engineering from the Cooper Union for the Advancement of Science and Art. Cassandra’s work in environmentally conscious design has resulted in methods for analyzing environmental impacts of design decisions, redesigning products for energy efficiency, modeling usage contexts, and providing actionable guidelines to help designers reduce environmental impacts. These energy and environmental studies include five refereed conference publications and articles in the ASME Journal of Mechanical Design and the Rapid Prototyping Journal. Cassandra’s teaching experience spans engineering design and energy topics at The University of Texas, MIT, and SUTD.
