Optimization-based trip chain emulation for electrified ride-sourcing charging demand analyses

ABSTRACT

Range anxiety remains one of the key concerns for ride-sourcing drivers to adopt battery electric vehicles (BEVs). To investigate the feasibility of using BEVs for ride-sourcing services, we propose an optimization-based methodology to estimate the daily driving trip patterns of ride-sourcing vehicles based on widely available non-identifiable trip data. Furthermore, we investigate the charging needs of electrified ride-sourcing vehicles using agent-based simulation. The methodologies are illustrated through a case study in the city of Chicago. Through sensitivity analysis on driver working hours and initial charging status, we quantify the range of daily average vehicle miles traveled (VMT) per car and identify the hot spots of current public charging demand and potential unsatisfied charging demand. This study can be used to determine the priorities of future charging infrastructure investment to further mitigate range anxiety and promote adoption of electrified ride-sourcing services. 

KEYWORDS:

• Electrified ride-sourcing
• batch matching
• agent-based modeling
• charging demand

Acknowledgments

The efforts of Spencer Aeschliman and Yan Zhou at Argonne National Laboratory (under Contract DE-AC02-06CH11357) are supported by Exelon Corporation, as well as the Vehicle Technology Office, Energy Efficiency and Renewable Energy Office, U.S. Department of Energy. The efforts of Zhaomiao Guo and Chuang Hou are supported by Argonne through contract 8F-30218. Md Rakibul Alam is partially supported by the UCF ORC fellowship. ATEAM model was initially developed under an Argonne-Exelon cooperative research and development agreement (CRADA).

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. Trip chain is all consecutive trips a vehicle takes during a period of time (e.g., one day).

2. Notice that trip-level data can be either real TNC trips data or household travel survey data.

3. Data source: https://data.world/ride-austin.

4. For example, ${\mathbb{T}}_d=[21,37]{\bigcup }_{}[41,57]$ means that a driver works two shifts in a day: from 8 am to 12 pm and from 13 pm to 17 pm if each time step is 15 min and t = 1 corresponds to 3 a.m.

5. Data source: https://data.cityofchicago.org/Transportation/Transportation-Network-Providers-Trips/m6dm-c72p.

6. The reason we assume drivers plan ahead for three trips is because the average TNC trip length in the Chicago area is around 5.0 miles. Three trips are about 15.0 miles, which allows drivers to find a charging station within our study area.

7. The reason we assume 20% SOC as a minimum is because a BEV battery is typically recommended to stay above 20% to maximize battery life, see e.g., https://batteryuniversity.com/learn/article/bu_1003a_battery_aging_in_an_electric_vehicle_ev.

8. Since the total number of trips is constant for different working hour scenarios.