Date of Award

5-2018

Degree Type

Thesis

Degree Name

Master of Science in Industrial Engineering (MSIE)

Department

Industrial Engineering

Committee Chair

Hua Cai

Committee Member 1

Seokcheon Lee

Committee Member 2

Andrew Liu

Abstract

The choice of battery range (all-electric driving range) for the battery electric vehicles (BEVs) is an important issue for both the BEV adopters and the BEV makers. This study proposes a Mix-integer Nonlinear Programming (MINLP) model to identify the minimum BEV battery range that can satisfy given travel demands for a driver or vehicle owner, considering the opportunities to charge at existing public charging stations and the uncertainties in charging decision making. I conducted a Stated Preference survey to study the charging decision making under the key factors of dwell time, remaining battery energy, distance of the upcoming trip, charging cost, and the distance from the nearest charging stations. The collected survey data is analyzed using the Logistic Regression and the Latent Class (LC) model to generate the coefficients of the key factors in determining whether to charge at a stop. These two methods provide better suggestion on the battery range than the often-used simplified charging rules. The simplified charging rules are likely to underestimate the minimum required battery range. The survey result shows significant heterogeneity in charging behaviors between different clusters of drivers. Using trip data of taxis and private vehicles in Beijing as a case study, I found that the battery range estimated using the charging preferences from the LC model is larger using the Logistic Regression model. For the taxis in Beijing, BEVs with a battery range of 200 miles are able to satisfy the travel demand for about 90% of the drivers. For private vehicle, a range of 300 miles is needed to cover the travel demands of 90% of the drivers, while a 100-mile range battery is able to satisfy the need for 80% of the private drivers.

Download

Share

COinS