Driving Cycle Uncertainty and Energy Consumption of City Buses: Analysis and Optimization

The research presented here studied the effect of driving cycle variation and passenger load uncertainty on the energy consumption of city buses. Furthermore, different methods for reducing the energy consumption were analyzed and compared. The research was conducted with simulation studies. In order to generate a large quantity of varying realistic cycles for a single bus route, a novel driving cycle synthetization algorithm was developed. The algorithm is capable of synthesizing a large number of cycles based on only a handful of measurements by exploring different combinations of events. Cycles generated with the algorithm were employed to compare the energy consumption of different city bus powertrain topologies under uncertainty in the driving cycle and passenger load. Simulated bus powertrain topologies included: compressed natural gas, diesel, parallel hybrid, series hybrid, hydrogen fuel cell hybrid, and battery electric bus. Synthetic driving cycles generated with the novel cycle synthesis algorithm were shown to maintain the statistical properties of the original measured cycles with good accuracy. The presented algorithm could be utilized to further optimize city buses for the routes they will be operated on. Energy consumption results acquired from the simulation studies indicated that battery electric buses are the most robust option against driving cycle uncertainty. Diesel buses appeared to be the most sensitive to the driving aggressiveness. However, the results displayed a strong correlation between energy consumption and driving aggressiveness with all types of powertrains. This suggests that steps should be taken to limit high-speed accelerations of city buses regardless of powertrain type. Battery electric buses were further studied by comparing component-choice-related methods for reducing the energy consumption. The methods included using an aluminum chassis instead of a steel chassis, employing a low-height body for reduced aerodynamic drag, using low-rolling-resistance class C tires, and utilizing an electric heat pump instead of a more conventional electric heater. A novel problem formulation for driving optimization was devised for a nonlinear model predictive controller. The driving optimization algorithm was used to compare the energy savings achievable with predictive driving to those achieved with the component-choice-related methods. Out of all the considered consumption reduction methods, the heat pump produced the highest energy savings in cold conditions. However, the relative effectiveness of the heat pump was significantly influenced by the ambient temperature and driving cycle. The aluminum chassis provided higher consumption reductions than the low-rolling-resistance tires, but the influence of the lighter chassis was highly dependent on the aggressiveness of the driving. On average, the predictive driving achieved higher energy savings than the aluminum chassis. Applying all of the methods simultaneously resulted in an average consumption reduction of more than 30 %.