Indian City Electric Bus Powertrain Optimization Using Active Learning Simulation Methodology

Public transport electrification is going to play a massive role in India's COP26 pledge to achieve net zero emissions by 2070. India plans to electrify 800,000 buses in a push towards 30% EV penetration by 2030. Further encouraged by government incentives under National Electric Bus Program (NEBP), eBus market is expected to grow at a CAGR of ~86% annually over the next 5 years. With most OEMs going for fleet electrification for reducing CO2 emissions and to cater to growing demand in Indian cities for cleaner public transport, improving powertrain efficiency and performance of state-of-the-art eBuses is a natural progression of e-mobility sector development in India. The first step in designing powertrain for an electric city bus is to determine the motor(s) size and transmission specifications (number of gears, gear ratios etc.). Complications arise due to a wider and non-linear operation range of eBus. This study focuses on powertrain optimization for a medium duty electric city bus for an Indian city. FEV's in-house Vehicle Powertrain System Design & Controls (VPSC) tool is deployed for a structured approach. It uses backwards simulation approach to estimate energy consumption and performance of vehicle. An initial (reference) electric drive unit (EDU) is selected via local pre-simulation for detailed flux linked loss maps. This is followed by an iterative AI based Global Optimizer which uses Bayesian algorithm to arrive at subsequent designs until the objective is met within the stated constraints. Performance parameters considered include acceleration and gradeability. The simulation is performed on a high traffic Indian city duty cycle as well as a reference urban VECTO cycle. In both cycles, results show a significant improvement in energy consumption while at the same time improving upon performance with respect to initial pre-simulation selected design of motor and transmission.