Uncertainty in electric bus driving cycles

Electric vehicles are in the brink of breakthrough in the automotive industry, and batteryelectric city buses are no exception. However, the costs are still high compared to conven-tional buses with internal combustion engines. A thorough electric city bus system plan-ning is an effective way to curb the high costs, where, accurate knowledge about the spe-cific energy consumption of the bus route is required. This study utilized route energy consumption sensitivity analysis to create a reliable en-ergy consumption forecast for a light traffic semi-urban bus route. The analysis was per-formed with the Monte Carlo method where the number of stops was addressed as theuncertainty. The uncertainty was modeled with a normal distributed probability densityfunction. The process of creating and utilizing the energy consumption sensitivity analysis is alsopresented. First, ten base driving cycles were measured in actual buses operating an actualbus line. Artificial driving cycles were synthesized from the measured cycles, based on theMonte Carlo sampling method. The fluctuating accelerations and speeds of different runson the route are addressed by compiling the synthesized cycles from parts of the measuredcycles. The synthesized driving cycles were simulated with a validated energy flow modelof a battery electric bus. As a result, an energy consumption distribution for the route wasacquired. The energy consumption distribution includes all the outcomes obtained with the MonteCarlo method. The distribution also functions as an energy consumption forecast for theroute. It was found out that the distance specific energy consumptions in the route fellbetween 0.483 kWh/km and 0.722 kWh/km. Which means, there is a 49.4 % variationbetween the extremes, when the number of bus stops varies. However, a closer result anal-ysis proved that there is an 80 percent chance that the predicted consumption falls be-tween 0.561 kWh/km and 0.642 kWh/km. Thus, it is denoted only 14.4 % variation ismost likely present between the extremes. Additionally, the study suggested that loweraccelerations result in lower consumptions. Driving cycles with less than average acceler-ations resulted in 5.6 % lower consumptions compared to cycles with higher than averageaccelerations. Thus, it can be concluded that the number of bus stops greatly affected the energy con-sumption. In addition, Monte Carlo method proved to be worthwhile tool to clarify theeffects of uncertainty.