Electrified powertrain simulation and validation of a fuel cell electric vehicle

Abstract

In this study, electrified powertrain simulation of a fuel cell electric vehicle and the verification of its outputs with experimental data was focused. The validation study was provided with the test data of the Toyota Mirai (2017) vehicle published by Argonne National Laboratory. Rapid population growth brings with it increasing urbanization and industrialization. As a result of this, the harmful effects observed on our ecosystem due to human-induced effects are becoming more and more evident. Rising average global temperatures and air pollution are among its most important direct consequences. Various regulations and strategic targets have been established by the United Nations and its sub-organizations to control the negative results of human-induced activities directly or indirectly and for a sustainable future. As a result, increasingly stringent emissions targets in the field of transportation can be shown. The acceleration of this transformation and emission targets increases as the suitability and applicability of the technology is ensured. Battery electric vehicles, hybrid electric vehicles, and fuel cell electric vehicles are increasingly coming to the fore in meeting these targets. Electric vehicles stand out not only as a solution to environmental problems but also with their high performance and efficiency features. Battery electric vehicles belonging to the electric vehicle ecosystem are the most common and simplest structure. They provide the necessary energy from a high-voltage battery and have an electric motor for traction. Hybrid electric vehicles, on the other hand, have an internal combustion engine besides electric motor, unlike battery electric vehicles. They can work in many different paradigms depending on the intended use. Fuel cell electric vehicles provide their energy from the fuel cell and battery system with various energy management system algorithms. In this study, Toyota Mirai(2017), which is a fuel cell electric vehicle that provides its main power with the fuel cell system, was chosen. The power required by the vehicle is provided by low-voltage or high-voltage buses. In the vehicle, auxiliary devices are supplied with low voltage directly from the battery. On the other hand, while high voltage components are mainly fed by the fuel cell system, the battery can also be fed according to the energy management strategy. The system also includes DC/DC converters to change the voltage level and an inverter to supply this generated DC to the AC electric motor. The forward approach methodology is used to perform the drive cycle simulations. This method is used because it gives more accurate results and offers a more realistic approach than the other method (backward approach). This approach requires a driver model. In order to achieve this, a PI controller is used to keep the speed of the vehicle at the reference speed level in the simulation. This controller gives acceleration or braking commands just like a real driver. It calculates the required torque by taking into account the vehicle loads obtained by using the coast-down coefficients and demands it from the electric motor. Thus, the electric motor operates at a certain operating point. This operating point corresponds to a point in the contoured efficiency map of the electric motor. The calculated total electrical power is requested between the fuel cell and the battery system with a rule-based controller according to the energy management strategy. This controller has been developed based on the requested power amount and SoC level by examining the tests conducted by the Argonne National Laboratory. The Rint battery equivalent circuit model is used to calculate the resistance, current, voltage, and charge state of the high-voltage battery. The Nernst voltage is calculated to obtain the maximum theoretical cell voltage of the fuel cell stack. Then the activation, ohmic, and concentration losses are calculated. By subtracting these losses from the Nernst voltage, the polarization curve is obtained. Using ideal gas equations, the gas flow is modeled depending on the partial pressures in the manifold. With these equations, the partial pressures of nitrogen, oxygen, and water vapor in the cathode volume are calculated. An equation is used between the utilized substance and the current. Thus a relation is established between reactants and current. Similarly, an equation is used for the amount of utilized hydrogen for the anode part. In addition, the liquid water formed in the cathode volume is calculated according to the relative humidity. The required amount of air required by the system is provided with the help of a compressor. The fuel cell system response delay is modeled using a low-pass filter. Finally, the temperature changes of the system components are monitored by establishing the heat balance equation throughout the simulation. The instantaneous acceleration of the vehicle is calculated using Newton's second equation of motion. Net force is obtained by subtracting resistive forces which slows the vehicle from the forces that provides traction. Using vehicle mass and net force instantaneous acceleration is found. The instantaneous velocity is found by integrating this value. Extensive reports and raw test data published by Argonne National Laboratory were reviewed to validate the simulation outputs. First, the characteristic curve of the fuel cell system, the stack and system efficiencies according to the power it provides, and the reactant consumptions are compared. The results were considered to have an appropriate accuracy. Then, in order to examine the characteristic delay response of the fuel cell system, instantaneous power requests and power cuts were examined in the simulation as well as in the test data. Later, the maximum power test conditions were established in the simulation environment to evaluate the response of the vehicle's fuel cell system, battery, and electric motor at high loads. When both simulation and test results were examined, it was observed that it could supply 110kW for 30 seconds and then 75kW continuously. Since the maximum power test was also carried out with full power at 25% grade level, at the same time, gradeability analyses of the vehicle could be made with these data. Accordingly, when the vehicle speed is examined in the test data, it is observed that the highest speed value is 72.5 km/h and it can reach 44 km/h continuously. On the other hand, when the same conditions were repeated in the simulation environment, the highest speed and the speed it could reach continuously were obtained as 79 km/h and 48.4 km/h, respectively. When the vehicle's acceleration of 0-80 mi/h on flat ground was examined, it is seen that it reaches 16.7s both in the test and in the simulation. However, when the acceleration of 0-100 km/h is examined, the target speed was reached in 8.9s in the simulation environment, while this value was measured as 9.6s in the tests. Last but not least, Steady State Speed, WLTP and UDDS driving cycles were compared in detail with many data test outputs related to the electric motor, fuel cell system and battery, and the results were evaluated. The comparisons were generally found to be quite satisfactory in terms of simulation reliability and matched significantly. Observed deviations are explained in detail. One of the unexpected findings was that during every driving cycle the vehicle demanded power from the fuel cell system even though it did not need it for traction. It has been observed that this power charges the battery regardless of the state of charge level. When the hydrogen consumption in the test is examined; while the consumption was measured as 192.6g in the test data in the WLTP driving cycle, it was calculated as 190.3g in the simulation. Similarly, this evaluation was made for the UDDS driving cycle, the measurement result obtained in the test was 75.7g, while the consumption calculated in the simulation was 75.1g. Accordingly, the error was found to be 1.19% and 0.80%, respectively.