In recent years, the global automotive industry has been undergoing a significant transformation, with electric vehicles (EVs) emerging as a pivotal solution to address environmental concerns and energy sustainability. As a researcher focused on advancing EV technologies, I have been particularly interested in optimizing the powertrain systems of pure electric vehicles to enhance their performance and efficiency. The challenge lies in balancing multiple objectives, such as driving range, energy consumption, and dynamic performance, which often conflict with each other. In this study, I explore a multidisciplinary approach to parameter matching and optimization, leveraging simulation tools and intelligent algorithms to achieve a harmonious design. The rapid growth of the electric vehicle market, especially in regions like China, underscores the importance of such research. China EV initiatives have driven innovations in battery technology and powertrain design, making it a key player in the global transition to sustainable transportation.
Electric vehicles, particularly pure electric variants, offer zero emissions, simplified structures, and reduced noise compared to traditional internal combustion engine vehicles. However, limitations in lithium-ion battery energy density and the parasitic energy consumption of thermal management systems often constrain the driving range of these vehicles. This makes it crucial to optimize the powertrain parameters, including the motor characteristics, battery capacity, and transmission ratios, through systematic dynamics and multi-objective optimization frameworks. My work builds on existing research in this area, aiming to provide a robust methodology for engineers and designers in the electric vehicle industry, with a focus on applications in the China EV sector.
The core of this research involves using AVL CRUISE, a comprehensive simulation software, to model and analyze the performance of a pure electric vehicle. By integrating multi-objective genetic algorithms, specifically NSGA-II, I aim to optimize key parameters like motor peak power, battery capacity, and reducer gear ratio. This approach allows for a quantitative analysis of trade-offs between driving range under the New European Driving Cycle (NEDC) and acceleration performance from 0 to 100 km/h. The results demonstrate significant improvements, highlighting the effectiveness of combining simulation with optimization algorithms. As electric vehicles become more prevalent, especially in markets like China EV, such methodologies can accelerate the development of high-performance, energy-efficient models.
To begin, I established the basic parameters of the electric vehicle based on common industry standards and performance targets. The vehicle dimensions, mass, and aerodynamic properties were defined to reflect a typical sedan model, akin to those popular in the China EV market. The key performance indicators included a maximum speed of at least 160 km/h, acceleration from 0 to 100 km/h within 11 seconds, a driving range of over 350 km under NEDC conditions, and a maximum gradability of 20%. These targets are aligned with the expectations for modern electric vehicles, ensuring competitiveness in regions like China, where EV adoption is rapidly increasing.
The powertrain of a pure electric vehicle consists of several critical components: the drive motor, battery pack, transmission system, and control units. In this study, I adopted a rear-wheel drive configuration with a single-speed reducer, which simplifies the design and improves space utilization. This layout is common in many electric vehicles, including those manufactured in China, due to its efficiency and cost-effectiveness. The drive motor is typically a permanent magnet synchronous motor (PMSM), chosen for its high power density and reliability. The battery system uses lithium-ion phosphate (LiFePO4) cells, known for their safety and long cycle life, which are widely used in China EV applications.
Parameter matching for the powertrain is a fundamental step in electric vehicle design. It involves determining the optimal values for motor power, speed, torque, transmission ratio, and battery capacity to meet the performance targets. I started by calculating the required motor power based on three key aspects: maximum speed, gradability, and acceleration. The power needed for achieving the maximum speed is derived from the balance between driving resistance and available power. The formula is given by:
$$P_{max1} \geq \frac{1}{\eta_T} \left( \frac{mgf}{3600} v_{max} + \frac{C_D A}{76140} v_{max}^3 \right)$$
where \( P_{max1} \) is the motor power required for maximum speed, \( \eta_T \) is the transmission efficiency (assumed as 0.96), \( m \) is the vehicle mass, \( g \) is gravitational acceleration, \( f \) is the rolling resistance coefficient (0.013), \( C_D \) is the drag coefficient (0.28), \( A \) is the frontal area (2.45 m²), and \( v_{max} \) is the target maximum speed (160 km/h). Based on this, the minimum power required was approximately 50 kW.
For gradability, the power must suffice to climb a maximum slope at a low speed, typically 20 km/h. The equation is:
$$P_{max2} \geq \frac{1}{\eta_T} \left( \frac{mgf \cos \alpha_{max} v_i}{3600} + \frac{C_D A v_i^3}{76140} + \frac{mg \sin \alpha_{max} v_i}{3600} \right)$$
where \( \alpha_{max} \) is the maximum slope angle (corresponding to 20% grade), and \( v_i \) is 20 km/h. This yielded a power requirement of at least 40 kW.
Acceleration performance dictates the power needed to achieve 0-100 km/h within a specified time. Using the equation:
$$P_{max3} \geq \frac{v_f}{3600 \eta_T} \left( \delta m \frac{du}{dt} + mgf + \frac{C_D A v_f^2}{21.15} \right)$$
where \( v_f \) is the target speed (100 km/h), \( \delta \) is the rotational mass conversion factor (1.04), and \( du/dt \) is the acceleration. This calculation resulted in a power requirement of around 83 kW. Thus, the peak motor power was selected as the maximum of these values, set at 90 kW, with a rated power of 40 kW and an overload factor \( \lambda = 2.25 \).
The motor speed and torque were determined based on the vehicle’s top speed and performance characteristics. The peak speed was chosen as 8500 rpm, falling into the medium-speed category for PMSMs, which is common in electric vehicles for its balance of efficiency and cost. The rated speed \( n_e \) and peak speed \( n_{max} \) are related by the constant power zone coefficient \( \beta \):
$$\beta = \frac{n_{max}}{n_e}$$
For this application, \( \beta \) was set to 2.5, giving a rated speed of 3400 rpm. The torque values were calculated as:
$$T_e = \frac{9550 P_e}{n_e}$$
$$T_{max} = \frac{9550 P_{max}}{n_e}$$
where \( T_e \) is the rated torque (142 N·m) and \( T_{max} \) is the peak torque (320 N·m). These parameters ensure that the motor can deliver sufficient force for acceleration and hill climbing while maintaining efficiency.
The transmission system uses a single-speed reducer, and its ratio was optimized to balance speed and torque requirements. The minimum and maximum transmission ratios were derived from the top speed and gradability constraints. The minimum ratio \( i_{min} \) is limited by the motor speed and vehicle speed:
$$i_{min} \leq \frac{0.377 n_{max} r}{v_{max}}$$
where \( r \) is the tire rolling radius (0.319 m). This gave an upper limit of 7.5. The lower limit considers the torque at maximum speed:
$$i_{min} \geq \frac{(mgf + \frac{C_D A u_{a max}^2}{21.15}) r}{T_{v max} \eta_T}$$
where \( T_{v max} \) is the motor torque at maximum speed (86 N·m). This resulted in a lower limit of 4.2. Similarly, the maximum ratio \( i_{max} \) is constrained by gradability and adhesion limits:
$$i_{max} \geq \frac{(mgf \cos \alpha_{max} + \frac{C_D A u_i^2}{21.15} + mg \sin \alpha_{max}) r}{T_{max} \eta_T}$$
and
$$i_{max} \leq \frac{mg r L_b \varphi}{T_{max} L \eta_T}$$
where \( \varphi \) is the adhesion coefficient (0.7), \( L \) is the wheelbase, and \( L_b \) is the distance from the center of mass to the rear axle. The final range for the transmission ratio was 5.8 to 7.3, and an initial value of 6.0 was selected for simulation.
For the battery pack, LiFePO4 cells were chosen due to their high specific power, energy density, and safety, which are critical for electric vehicles in diverse markets, including China EV. The battery capacity was determined based on the driving range under NEDC conditions. The energy consumed during driving is:
$$W_l = P t = \frac{V_a}{3600 \eta_t} \left( mgf + \frac{C_D A V_a^2}{21.15} \right) \left( \frac{S}{V_a} \right)$$
where \( V_a \) is the constant speed (for range calculation), and \( S \) is the target range (350 km). The required battery energy \( E_B \) is:
$$E_B \geq \frac{W_l}{DOD \cdot \eta_{mc} \cdot \eta_{dis}}$$
with depth of discharge (DOD) as 0.85, motor efficiency \( \eta_{mc} \) as 0.92, and discharge efficiency \( \eta_{dis} \) as 0.9. This yielded a minimum energy of 54 kWh. The number of battery cells was calculated based on voltage and power requirements. The total voltage for the system was set to 400 V, using cells with 3.2 V each, so the number of series cells \( N_2 \) is:
$$N_2 = \frac{U_{els}}{U_0} = 125$$
For peak power, the number of cells \( N_3 \) is:
$$N_3 = \frac{P_{e max}}{P_{b max} \eta_m \eta_{mc}}$$
where \( P_{b max} = \frac{2E^2}{9R_{int}} \), with cell voltage \( E = 3.2 \, \text{V} \) and internal resistance \( R_{int} = 2.5 \, \text{m}\Omega \). This gave \( N_3 = 215 \). The total number of cells was determined as the maximum of the series, parallel, and power requirements, resulting in 875 cells arranged in 7 parallel strings of 125 series cells, providing a total capacity of 142 Ah.
The initial parameter matching results are summarized in the table below:
| Category | Parameter | Value |
|---|---|---|
| Drive Motor | Rated Power (kW) | 40 |
| Peak Power (kW) | 90 | |
| Rated Torque (N·m) | 142 | |
| Peak Torque (N·m) | 320 | |
| Peak Speed (rpm) | 8500 | |
| Transmission | Final Drive Ratio | 6.0 |
| Battery | Cell Capacity (Ah) | 20 |
| Number of Cells | 875 | |
| Cell Voltage (V) | 3.2 | |
| Total Voltage (V) | 400 |
With these parameters, I built a simulation model in AVL CRUISE to evaluate the vehicle’s performance. The model included components for the motor, battery, transmission, tires, and vehicle body, configured with the matched parameters. Simulations were conducted for maximum speed, acceleration, gradability, and driving range under NEDC conditions. The results showed that the initial design met all targets: maximum speed of 173.66 km/h, 0-100 km/h acceleration in 10.93 s, maximum gradability of 30.67%, and an NEDC range of 357.3 km. This confirmed the validity of the parameter matching approach for electric vehicles, particularly in the context of China EV standards.
However, to further optimize the design, I employed a multi-objective genetic algorithm (NSGA-II) to refine the parameters. The optimization variables were the motor peak power \( P_{max} \), transmission ratio \( i \), and battery capacity \( C \). The objectives were to minimize the 0-100 km/h acceleration time and energy consumption per 100 km, while maximizing the NEDC driving range. The multi-objective function was formulated as:
$$f(x) = \omega_1 \frac{f_1(x)}{t_e} + \omega_2 f_2(x) + \omega_3 \frac{f_3(x)}{v_e} + \omega_4 f_4(x)$$
where \( f_1(x) \) is the acceleration time, \( f_2(x) \) is the gradability, \( f_3(x) \) is the maximum speed, and \( f_4(x) \) is the energy consumption improvement. The weights \( \omega_1 = 0.3 \), \( \omega_2 = 0.2 \), \( \omega_3 = 0.2 \), and \( \omega_4 = 0.3 \) were assigned to prioritize acceleration and range, reflecting common priorities in electric vehicle design, such as in China EV models. The constraints included:
$$g_1(x) = 11 – \frac{1}{3.6} \int_{u_1}^{u_2} \frac{\delta m}{F_t – (F_f + F_w)} du \geq 0$$
$$g_2(x) = 0.377 \frac{r_d n_{max}}{i} – 160 \geq 0$$
$$g_3(x) = \tan \left[ \arcsin \left( \frac{F_t – F_w}{mg \sqrt{f^2 + 1}} \right) – \arctan(f) \right] – 0.2 \geq 0$$
$$g_4(x) = \frac{100 P_c}{\eta_{euc}} – 12.51 \leq 0$$
The variable bounds were \( 85.5 \leq P_{max} \leq 94.5 \), \( 5.8 \leq i \leq 7.3 \), and \( 133 \leq C \leq 147 \). Using MATLAB’s genetic algorithm toolbox, I configured NSGA-II with a population size of 50, crossover probability of 0.8, mutation probability of 0.02, and maximum generations of 50. After 34 generations, the algorithm converged to an optimal solution: \( P_{max} = 89.3 \, \text{kW} \), \( i = 6.7 \), and \( C = 142 \, \text{Ah} \).
To validate the optimization, I updated the CRUISE model with these parameters and re-ran the simulations. The results showed notable improvements: the NEDC driving range increased to 375 km (a 4.9% gain), and the 0-100 km/h acceleration time decreased to 10.01 s (an 8.4% improvement). The maximum speed slightly reduced to 160.7 km/h, but it still met the target, while the gradability increased to 35.9%. The comparison before and after optimization is detailed in the table below:
| Parameter | Before Optimization | After Optimization | Change Rate |
|---|---|---|---|
| 0-100 km/h Acceleration Time (s) | 10.93 | 10.01 | -8.4% |
| Maximum Speed (km/h) | 173.66 | 160.7 | -7.4% |
| Maximum Gradability (%) | 30.67 | 35.9 | +17% |
| NEDC Driving Range (km) | 357.3 | 375 | +4.9% |
These results demonstrate the effectiveness of integrating simulation with multi-objective optimization for electric vehicle design. The trade-offs between different performance metrics were effectively managed, leading to a balanced solution that enhances both range and acceleration. This approach is particularly valuable for the electric vehicle industry, including China EV manufacturers, who face intense competition and evolving consumer demands. By leveraging tools like AVL CRUISE and algorithms like NSGA-II, engineers can rapidly prototype and optimize designs, reducing development time and costs.
In conclusion, this study highlights a systematic methodology for parameter matching and optimization in pure electric vehicles. The use of AVL CRUISE for simulation and NSGA-II for multi-objective optimization proved to be a powerful combination, yielding significant improvements in key performance indicators. As the adoption of electric vehicles continues to grow, especially in markets like China EV, such research provides a foundation for developing more efficient and high-performing models. Future work could explore additional factors, such as thermal management, regenerative braking, and real-world driving cycles, to further enhance the robustness of electric vehicle designs. Ultimately, this contributes to the global shift toward sustainable transportation, aligning with environmental goals and technological advancements in the electric vehicle sector.