Development of an electric driveline model for multiscale road-cargo simulations

Abstract: Currently, the road cargo system with low or zero CO2 emission is under rapid development. Heavy-duty trucks with electrified driveline systems will be the workhorse of future freight. But developing such a brand new and very complex system and adapting it to various application scenarios, such as long-haul freight, city distribution or construction loading, is still a big problem, because there is no previous experience to refer to. There is no standard development procedure or constraint framework for uncertainty either. Simulation on a massive scale with thousands of truck agents will be of great use for developing such a road-cargo system. System engineering will be the guiding methodology for this thesis project about developing a high-performance and multi-adaptive electrified driveline system. Referring to the classical V-shape development methodology, the complex concept will be divided into different levels of subsystems, from the large application scenarios to traffic simulation, driveline system simulation, electric motor and controller blocks development, and the system integration, performance verification and output of the results. The massive scale of traffic simulation will be implemented in AnyLogic, which does not contain any accurate agent model with vehicle dynamic motion during simulation. Thus, a precise vehicle agent model needs to be developed and embedded into AnyLogic’s simulation scenario, so as to make the simulation very close to reality, and to be able to evaluate vehicle concepts as well. The driveline system will be developed in Matlab/Simulink while the information communication between them will be realised in the form of computational calculation functions through the C language program. The development of the driveline model is also progressive. First, an equation-based full glider model was constructed. It simulates the scenario of a heavy-loaded truck driving on a steep slope (30% grade), decelerating from the initial 70 km/h to 0 km/h and then remaining stationary. The second model added the functionality of velocity input and output, enabling information exchange with AnyLogic. It will judge the real-time speed and the desired speed to decide whether to accelerate or decelerate and it uses the “Bang-Bang” control method of the electric motor. But this control mode results in a massive and frequent change in the electric motor output power, leading to extremely high energy consumption and in real life significantly shortened motor lifetime. So a powerful PI controller was introduced to the third Simulink model. The PI controller is embedded in the electric motor and it will replace the “Bang-Bang” control method. The “PID” control method provides a more stable power output so that the truck’s real-time speed can approach the target speed more smoothly. This control system can adapt to a variety of speed inputs and it can decide whether to output full power or partial power, depending on the speed difference. The third version of the Simulink model with PI controller has been verified as an acceptable model through various inputs of different speeds, and it will be converted into a C language program to be embedded in AnyLogic for massive traffic simulation.