Abstract
The recent increase in oil price and environmental concerns has attracted various research efforts on hybrid electric vehicles (HEVs) which provide promising alternatives to conventional engine-powered vehicles with better fuel economy and fewer emissions. To speed up the design and prototyping processes of new HEVs, a method that automatically generates mathematics equations governing the vehicle system response in an optimized symbolic form is desirable. To achieve this goal, we employed MapleSim™ 4.0, a new physical modeling tool developed by Maplesoft Inc., to develop the multi-domain model of a series-HEV, utilizing the symbolic computing algorithms of Maple software package to generate an optimized set of governing equations. The HEV model consists of a mean-value internal combustion engine (ICE), a chemistry-based Ni-MH battery pack, and a multibody vehicle model. Simulations are then used to demonstrate the performance of the developed HEV system. Simulation results show that the model is viable and the number of governing equations is reduced significantly, resulting in a computationally efficient system.
Introduction
Designing a Hybrid Electric Vehicle (HEV) is a complex cross-disciplinary challenge, requiring the integration of mechanical drivetrain systems, electrical motors/generators, and chemical battery storage. Historically, engineers have faced a "fidelity vs. speed" trade-off. To maintain simulation speeds suitable for real-time testing or design optimization, vehicle models are often simplified to a single "particle mass." However, such simplifications ignore significant power losses and handling nuances caused by 3D multibody dynamics, such as tire-road interaction and suspension movement. This research aims to bridge this gap by creating a mathematically rigorous, multi-domain model that remains computationally efficient.
System Architecture and Multi-Domain Modeling
The researchers developed a high-fidelity model of a series-HEV using MapleSim, a physical modeling tool built on the Maple symbolic engine. The model integrates several distinct physical domains into a single cohesive system:
- Mean-Value ICE Model: Captures throttle and intake manifold dynamics efficiently without simulating individual cylinder strokes.
- Chemistry-Based Ni-MH Battery Model: Utilizes Nernst and Butler-Volmer equations to simulate chemical reactions at the electrodes, providing more accurate State of Charge (SOC) tracking than traditional equivalent circuit models.
- 3D Multibody Vehicle Dynamics: A 14-degree-of-freedom (DOF) model including the chassis, unsprung masses, and wheel spin.
High-Fidelity Tire Dynamics
A key differentiator of this model is the inclusion of realistic road interactions. By incorporating the "Magic Formula" tire model, the simulation accounts for the complex forces and moments acting on the vehicle during various maneuvers. This level of detail is crucial for range-extender vehicles, where power losses in tires and suspensions can significantly impact overall efficiency.
Symbolic Optimization and Efficiency
The "secret sauce" of this approach is symbolic computing. The software identifies repeated terms and algebraic constraints to simplify the massive initial set of equations. In this study, the raw model generated 605 equations, which were symbolically reduced to just 41 governing equations without losing the underlying physics. This reduction makes the model viable for real-time "Hardware-in-the-Loop" (HIL) simulations, where the software must keep pace with physical hardware.
Key Results and Validation
Simulation results validated the model's performance across standard and sinusoidal speed profiles. The model successfully demonstrated:
- Regenerative Braking: The system effectively charged the battery during deceleration.
- SOC Management: An "on-off" controller engaged the ICE to recharge the battery when the SOC dropped below 50%, returning it to 80%.
- Fuel Economy: The series-HEV architecture qualitatively confirmed reduced fuel consumption by maintaining the engine within its most efficient operating range.
Conclusion and Future Work
While the current model assumes 100% efficiency for power electronics, it provides a robust framework for HEV design. Future work will focus on incorporating thermal losses in DC-to-DC converters and using the optimized C/C++ code for advanced sensitivity analysis and real-time control strategies. This methodology demonstrates that chemical, thermal, and mechanical domains can be solved simultaneously in a unified mathematical framework, avoiding the pitfalls of traditional co-simulation.
Summary updated