Expressing the energy consumption of electric buses with mesoscopic traffic variables

This paper focuses on energy consumption modeling approaches for traffic and examines how they deviate when applied to evaluate Battery Electric Buses (BEBs), in a try to identify an approach that combines simplicity with accuracy. To do so, the paper exploits three of them: a micro, a meso and a macro one. The microscopic approach relies on a detailed power-based vehicle model that uses second-by-second vehicle speed profiles as traffic activity input, and it serves here as a reference tool. The approach of average speed was employed to represent the macroscopic one that uses a single traffic activity input. For the mesoscopic case, a new function had to be developed that would require traffic inputs on a level-of-detail in between the macroscopic and microscopic scale. A statistical analysis on several standardized driving cycles was conducted to select such inputs, leading to a relationship that associates consumption with two stop-related variables (number and duration). The mesoscopic and macroscopic models could then be evaluated, by comparing their consumption estimations with the detailed microscopic calculations over the same cases (real-world urban traffic of Athens & Hong-Kong, and traffic measures). While the macroscopic results revealed well-known limitations in accuracy of the average speed approach, as it deviated from the microscopic model by 10 % for urban traffic and 20 % for measures, the mesoscopic one closely matched the microscopic model (max 5 % error). Thus, for BEBs, a mesoscopic approach with only two activity inputs (stop-related variables) can satisfy requirements from energy modeling for valid estimations and simplicity in use. With these characteristics, the approach presents exploitation potential in multiple applications of urban transportation systems.