As the world grapples with energy and environmental crises, the development of electric vehicles has emerged as a pivotal solution in the transportation sector. Battery electric vehicles (BEVs), in particular, represent a critical pathway toward energy conservation, emission reduction, noise mitigation, and environmental protection. In my research, I focus on leveraging simulation tools to advance the design and optimization of battery EV cars, aiming to reduce development costs, enhance performance reliability, and provide experimental foundations for manufacturing. This article delves into the modeling and simulation of a battery EV car using Simulink, a powerful platform for dynamic system simulation. I will elaborate on the modular modeling approach, integrate various subsystems into a complete vehicle model, and validate the model’s accuracy through simulations under diverse driving conditions, such as the New European Driving Cycle (NEDC), acceleration, and hill-climbing scenarios. The primary goal is to demonstrate the feasibility of the modeling methodology and offer insights for optimizing battery EV car performance, including energy economy and dynamic characteristics. Throughout this work, I emphasize the importance of simulation in refining battery EV car designs, and I will frequently reference key aspects of battery EV car technology to underscore its relevance.
The simulation model of a battery EV car is structured around its core components: the powertrain system, control system, and vehicle dynamics. In Simulink, I adopt a modular approach to build individual subsystems that mirror the real-world architecture of a battery EV car. The overall model structure, as illustrated in Figure 1, comprises several interconnected modules: a driver control model, a vehicle control strategy model, a drive motor model, a reducer model, a wheel model, a vehicle dynamics model, and a battery system model. This hierarchical design allows for systematic development and testing, enabling me to simulate the behavior of a battery EV car under various operational conditions. The driver control model translates driving intentions into throttle and brake signals, which are processed by the control strategy model to generate torque commands for the drive motor. The motor’s output is then transmitted through the reducer to the wheels, propelling the vehicle while the dynamics model calculates performance parameters like speed and distance. Simultaneously, the battery system model manages energy flow, estimating state of charge (SOC) and providing power to the motor. This integrated structure ensures that all aspects of a battery EV car are accurately represented, facilitating comprehensive analysis of its performance.
To begin the modeling process, I first define the key variables and parameters essential for simulating a battery EV car. These parameters are based on a prototype battery EV car, specifically referencing the 2025 BYD Qin L pure electric model, and are stored in a script file for efficient management. The table below summarizes the critical variables used in the model, which include vehicle mass, resistance coefficients, motor characteristics, and battery specifications. This parameterization forms the foundation for all subsequent subsystems, ensuring that the simulation reflects realistic behavior of a battery EV car.
| Variable in Script | Description | Value |
|---|---|---|
| CVW | Curb weight (kg) | 1800 |
| f | Rolling resistance coefficient | 0.015 |
| Cd | Drag coefficient | 0.34 |
| WindAS | Frontal area (m²) | 2.233 |
| Grad | Road gradient (rad) | 0 |
| MTC | Mass conversion factor | 1.2 |
| RollRad | Tire rolling radius (m) | 0.34 |
| GeatR | Reducer gear ratio | 11.886 |
| GeatE | Reducer efficiency | 0.97 |
| MSpdMax | Motor maximum speed (rpm) | 20000 |
| MDrv_1X_rpm | Motor drive speed array (rpm) | Defined based on range |
| MDrv_1Y_Trq | Motor drive torque array (Nm) | Defined based on external characteristics |
| MGen_1X_rpm | Motor generation speed array (rpm) | Same as MDrv_1X_rpm |
| MGen_1Y_Trq | Motor generation torque array (Nm) | -1 * MDrv_1Y_Trq |
| OCV_1X_Pot | Battery OCV curve SOC (%) | [0, 100] |
| OCV_1Y_V | Battery OCV curve voltage (V) | [2.5, 3.2] |
| CellRes | Cell internal resistance (Ω) | 0.025 |
| CellCap | Cell capacity (Ah) | 70.8 |
| BatNum_S | Number of cells in series | 250 |
| BatNum_P | Number of cells in parallel | 1 |
| SOCIni | Initial SOC (%) | 100 |
| SOCMin | Minimum SOC (%) | 15 |
With these parameters established, I proceed to construct the vehicle resistance model, which simulates the forces opposing the motion of a battery EV car. The total resistance $F_r$ consists of rolling resistance $F_f$, air resistance $F_w$, acceleration resistance $F_j$, and gradient resistance $F_i$, as expressed in the following equation:
$$F_r = F_f + F_w + F_j + F_i$$
Expanding this, the resistance can be calculated based on vehicle speed $u_a$, mass $m$, gravitational acceleration $g$, rolling resistance coefficient $f$, gradient angle $\alpha$, drag coefficient $C_D$, frontal area $A$, and mass conversion factor $\delta$:
$$F_r = mgf \cos \alpha + \frac{C_D A u_a^2}{21.15} + mg \sin \alpha + \delta m \frac{du}{dt}$$
In Simulink, I implement this equation by creating separate subsystems for each resistance component. The rolling resistance model computes $F_f = mgf \cos \alpha$, the air resistance model calculates $F_w = \frac{C_D A u_a^2}{21.15}$, the gradient resistance model determines $F_i = mg \sin \alpha$, and the acceleration resistance model derives $F_j = \delta m \frac{du}{dt}$. These subsystems are then integrated using an adder block to output the total resistance force, which serves as a critical input for the vehicle dynamics model. This modular approach allows for easy adjustment of parameters, facilitating sensitivity analysis for a battery EV car under different driving conditions.
Next, I develop the vehicle model, which translates the driving force into motion for the battery EV car. The driving force $F_t$ at the wheels is derived from the motor torque $T_{tq}$, reducer gear ratio $i$, transmission efficiency $\eta_T$, and wheel radius $r$:
$$F_t = \frac{T_{tq} i \eta_T}{r}$$
The vehicle’s acceleration $a$ is then determined by the difference between the driving force and the total resistance, considering the effective mass:
$$a = \frac{F_t – F_r}{m(1 + \delta)}$$
In Simulink, I build a subsystem that takes the wheel driving force as input and outputs vehicle speed and distance through integration. Specifically, I use an integrator block to compute speed from acceleration and another integrator to derive distance from speed. To avoid algebraic loops—a common issue in digital simulations where input depends on output and vice versa—I incorporate a Memory block that delays the signal, ensuring stable computation. This vehicle model subsystem effectively captures the dynamic response of a battery EV car, enabling simulation of its longitudinal motion under various control inputs.
The drive motor model is pivotal for simulating the powertrain of a battery EV car. This module receives torque commands from the control strategy and motor speed signals, then outputs the actual torque and required electrical power. The motor’s external characteristics, including both driving and generating modes, are represented using lookup tables based on the predefined speed and torque arrays. For instance, the driving torque $T_{drv}$ is obtained by comparing the commanded torque with the motor’s driving external characteristic curve, while the generating torque $T_{gen}$ is derived similarly for regenerative braking. The output torque $T_m$ is determined as:
$$T_m = \min(T_{cmd}, T_{drv}) \quad \text{for driving mode}$$
$$T_m = \max(T_{cmd}, T_{gen}) \quad \text{for generating mode}$$
The electrical power $P_e$ required by the motor is calculated using the formula:
$$P_e = \frac{T_m n}{9550}$$
where $n$ is the motor speed in rpm. I encapsulate these calculations into a Simulink subsystem that accurately models the motor behavior, ensuring that the battery EV car’s propulsion system responds realistically to driver inputs and operational conditions.
The reducer model simulates the gearbox that connects the motor to the wheels in a battery EV car. Since many battery EV cars use a fixed-ratio reducer for simplicity and efficiency, this module performs speed reduction and torque multiplication. The wheel torque $T_w$ is derived from the motor torque $T_m$, gear ratio $i$, and efficiency $\eta$:
$$T_w = T_m i \eta$$
Conversely, the motor speed $n_m$ is related to the wheel speed $n_w$ by:
$$n_m = n_w i$$
I implement these equations in a Simulink subsystem that converts signals between the motor and wheel domains, ensuring proper power transmission in the battery EV car model. This reducer model is essential for accurately simulating the torque and speed relationships that influence the vehicle’s acceleration and top speed capabilities.
The wheel model acts as the interface between the drivetrain and the road for the battery EV car. It receives driving torque from the reducer, braking torque from the control system, and vehicle speed, then outputs the wheel driving force and rotational speed. The driving force $F_{wheel}$ is computed as:
$$F_{wheel} = \frac{T_w}{r}$$
where $r$ is the wheel radius. The wheel speed $n_w$ is derived from the vehicle speed $u_a$ using the conversion:
$$n_w = \frac{u_a}{2 \pi r} \times 60$$
This subsystem integrates seamlessly with the vehicle dynamics model, enabling realistic simulation of tire-road interactions in a battery EV car. By accounting for both driving and braking torques, the wheel model ensures that the battery EV car’s motion is accurately represented during acceleration, cruising, and deceleration phases.
The battery system model is crucial for energy management in a battery EV car. It simulates the high-voltage battery pack, including its voltage characteristics, internal resistance, and SOC estimation. The battery pack voltage $V_{bat}$ is determined by the open-circuit voltage (OCV) curve, which depends on SOC, and the internal voltage drop due to current flow. For a battery pack with $N_s$ cells in series and $N_p$ cells in parallel, the voltage is calculated as:
$$V_{bat} = N_s \left( V_{oc}(SOC) – I_{cell} r_0 \right)$$
where $V_{oc}(SOC)$ is the OCV obtained from a lookup table, $I_{cell}$ is the cell current, and $r_0$ is the cell internal resistance. The cell current $I_{cell}$ is derived from the system power $P_{sys}$ and voltage:
$$I_{cell} = \frac{P_{sys}}{V_{bat} N_p}$$
SOC is estimated using the ampere-hour integral method:
$$SOC = SOC_{ini} – \frac{\int I_{cell} dt}{C}$$
where $C$ is the battery capacity. In Simulink, I build a subsystem that incorporates these equations, along with a termination control that halts simulation when SOC falls below a minimum threshold. This model ensures that the energy consumption and range of the battery EV car are accurately predicted during simulations.
The driver control model emulates human driving behavior for the battery EV car. It compares the actual vehicle speed with a target speed profile and generates throttle and brake pedal signals accordingly. I implement a proportional-integral (PI) controller to minimize the speed error $e = u_{target} – u_{actual}$. The control output, which represents pedal positions, is given by:
$$u_{ctrl} = K_p e + K_i \int e dt$$
where $K_p$ and $K_i$ are tuning parameters. This output is then scaled to produce pedal opening signals ranging from 0 to 1. By adjusting $K_p$ and $K_i$, I can simulate different driving styles and ensure that the battery EV car follows the desired speed profile accurately in various scenarios.
The control strategy model serves as the brain of the battery EV car, processing pedal signals to determine torque commands. For driving mode, the demanded torque $T_{dem}$ is computed as the product of throttle opening $\theta_{th}$ and the motor’s driving external torque $T_{drv}(n)$ at the current speed $n$:
$$T_{dem} = \theta_{th} \times T_{drv}(n)$$
For braking, the model distinguishes between regenerative braking and hydraulic braking. The total braking demand $T_{brake}$ is derived from the brake pedal opening $\theta_{br}$ and the maximum braking torque $T_{b,max}$, which is calculated based on vehicle deceleration requirements. The regenerative braking torque $T_{regen}$ is limited by the motor’s generating external characteristic, while the hydraulic braking torque $T_{hyd}$ supplements any shortfall:
$$T_{brake} = \theta_{br} \times T_{b,max}$$
$$T_{regen} = \min(T_{brake}, T_{gen}(n))$$
$$T_{hyd} = T_{brake} – T_{regen}$$
This strategy optimizes energy recovery in the battery EV car, enhancing overall efficiency. I encapsulate these logic in a Simulink subsystem, enabling realistic torque management during simulation.
To simulate real-world driving conditions, I develop a driving cycle model that provides target speed profiles for the battery EV car. This module includes standard cycles such as the New European Driving Cycle (NEDC) and China Light-duty vehicle Test Cycle (CLTC). The speed-time data for these cycles are stored in lookup tables, and a multiplexer switch allows selection between different cycles. For example, the NEDC cycle spans 1180 seconds with varying speed segments, while the CLTC cycle lasts 1800 seconds. By integrating this model, I can assess the performance of the battery EV car under standardized conditions, facilitating comparisons with regulatory benchmarks.
After integrating all subsystems, the complete battery EV car model in Simulink is assembled as shown in Figure 11. This model connects the driver control, control strategy, motor, reducer, wheels, vehicle dynamics, and battery system into a cohesive simulation environment. With this setup, I conduct various simulations to evaluate the performance of the battery EV car. For instance, under the NEDC cycle, I set the simulation time to 1180 seconds and tune the PI controller parameters ($K_p = 5$, $K_i = 0$) to ensure that the actual vehicle speed closely tracks the target profile. The results, plotted as speed versus time, demonstrate good consistency, validating the model’s ability to simulate a battery EV car in urban and extra-urban driving scenarios. Adjusting $K_p$ and $K_i$ allows me to explore different control responses; for example, with $K_p = 0.01$, the speed tracking deteriorates, highlighting the importance of proper controller tuning for a battery EV car.
I also simulate acceleration performance to assess the dynamic capabilities of the battery EV car. By setting a target speed of 50 km/h over 10 seconds, the model outputs an acceleration time of approximately 3.5 seconds, after which the throttle opening drops to around 4%. For a higher target of 250 km/h over 100 seconds, the simulation reveals a 0-100 km/h acceleration time of about 12 seconds and a maximum speed of nearly 190 km/h, limited by motor speed and power constraints. These results align with typical performance metrics for a battery EV car, confirming the model’s accuracy in capturing powertrain limitations.
Hill-climbing ability is another critical aspect evaluated for the battery EV car. I simulate a scenario with a target speed of 100 km/h on a 30% gradient. The model shows that the vehicle accelerates to the target speed in about 45 seconds, with full throttle engagement during acceleration and slight reduction thereafter. When the gradient is increased to 31%, the battery EV car struggles to maintain the target speed, indicating its maximum gradability. These simulations provide valuable insights into the traction and power requirements of a battery EV car in challenging terrains.
To further validate the model, I compare simulation results with real-world data from the prototype battery EV car. The table below summarizes key performance metrics from simulations and actual tests, demonstrating the model’s effectiveness.
| Performance Metric | Simulation Result | Actual Test Result | Deviation |
|---|---|---|---|
| 0-50 km/h acceleration time (s) | 3.5 | 3.1 | +0.4 |
| Maximum speed (km/h) | 190 | 180 | +10 |
| Gradability (30% slope at 100 km/h) | Achievable | Achievable | None |
The minor deviations in acceleration and top speed are attributed to simplifications in the model, such as overestimation of frontal area and neglect of environmental factors like temperature and humidity. Despite these, the overall alignment confirms that the Simulink model accurately represents the behavior of a battery EV car, making it a reliable tool for design optimization.
In conclusion, this research successfully demonstrates the modeling and simulation of a battery EV car using Simulink. By adopting a modular approach, I have constructed detailed subsystems for vehicle dynamics, powertrain, control, and energy management, integrating them into a comprehensive model. Simulations under NEDC, acceleration, and hill-climbing conditions validate the model’s accuracy in predicting the performance and energy economy of a battery EV car. The results show that the model can effectively simulate key characteristics, such as speed tracking, acceleration times, and gradability, with minimal deviations from real-world data. This work underscores the value of simulation in reducing development costs and enhancing the design of battery EV cars. Future efforts could focus on incorporating more environmental variables, refining control strategies for regenerative braking, and extending the model to include thermal effects on battery performance. Overall, this Simulink-based framework provides a robust foundation for optimizing battery EV car technologies, contributing to the advancement of sustainable transportation solutions.