Modelling of Hybrid Electric Vehicle Powertrains - Factors That Impact Accuracy of CO_2 Emissions

摘要

Modelling is widely used for the development of hybrid electric vehicle (HEV) powertrain technologies, since it can provide accurate prediction of fuel consumption and CO_2 emissions, for a fraction of the resources required in experiments. For comparison of different technologies or powertrain parameters, the results should be accurate relative to each other, since powertrains are simulated under identical model details and simulation parameters. However, when CO_2 emissions of a vehicle model are simulated under a driving cycle, significant deviances may occur between actual tests and simulation results, compromising the integrity of simulations. Therefore, this paper investigates the effects of certain modelling and simulation parameters on CO_2 emission results, for a parallel HEV under three driving cycles (NEDC, WLTC and RTS95 to simulate real driving emissions (RDE)). A sensitivity analysis on battery state of charge levels (SOC), control systems, component data resolutions, warm-up phase, time-step, driver controller behavior and 0D vs 1D simulation parameters is carried out and their effect on CO_2 emission results are investigated. While any change in one of the parameters may result in either a lower or higher CO_2 value, their cumulative effect on simulation results may result in significant differences of up to +-15%. Unfortunately, it is not hard to overlook the effect of these parameters and conduct powertrain simulations without taking this into account. By identifying key parameters and quantifying their effect on simulation results, this paper aims to improve the accuracy of HEV powertrain simulations to provide more reliable results.