The transition towards electric mobility has placed the permanent magnet synchronous motor (PMSM) at the heart of modern electric car propulsion systems. Its high power density, efficiency, and excellent torque characteristics make it the preferred choice. However, as consumer expectations for refinement and comfort rise, the issue of vibration and noise generated by the drive system has become a critical performance differentiator and a significant challenge for engineers. In an electric car, the absence of a masking internal combustion engine noise makes the whine, hum, and whirring sounds from the electric motor and gearbox far more perceptible to occupants. These acoustic phenomena are primarily driven by electromagnetic and mechanical excitations, which, if not properly managed, can compromise driving comfort, perceived quality, and long-term system reliability.
This article presents a comprehensive study focused on mitigating vibration and noise in electric car drive systems through targeted electromagnetic and structural optimization. We investigate the root causes of these excitations and propose a dual-strategy approach: optimizing the rotor’s magnetic barrier (often called the flux barrier or spacer bridge) to refine the air-gap magnetic field and, consequently, the radial electromagnetic force waves; and considering the micro-geometry of reduction gear teeth to address mechanical excitation. Our primary focus is on the electromagnetic aspect. We demonstrate that a strategic modification to the spacer bridge geometry can significantly improve the spatial distribution of the air-gap flux density, leading to a reduction in the amplitude of critical force harmonics. This work systematically employs finite element analysis (FEA) for electromagnetic and structural simulation, coupled with a spherical acoustic radiation model, to quantify the impact of our optimizations on electromagnetic performance, structural vibration, and radiated noise. The goal is to provide a clear methodology and tangible results that contribute to the design of quieter, more efficient, and more reliable motors for the next generation of electric cars.
1. Theoretical Foundation of Electromagnetic Vibration and Noise in Electric Car Motors
The dominant source of electromagnetic noise in a PMSM is the radial force acting on the stator teeth. This force causes the stator core and housing to deform periodically, radiating sound into the surrounding medium. The fundamental principle governing these forces is Maxwell’s Stress Tensor. The radial component of the electromagnetic force density (pressure) acting on the stator inner surface can be expressed as:
$$
p_r(\theta, t) = \frac{1}{2\mu_0} \left[ B_r^2(\theta, t) – B_{\theta}^2(\theta, t) \right]
$$
where \( B_r(\theta, t) \) and \( B_{\theta}(\theta, t) \) are the radial and tangential components of the air-gap flux density at position \( \theta \) and time \( t \), and \( \mu_0 \) is the permeability of free space. For most PMSMs, the tangential component \( B_{\theta} \) is significantly smaller than the radial component \( B_r \). Therefore, the equation is often simplified, highlighting the direct relationship between radial force density and the square of the radial flux density:
$$
p_r(\theta, t) \approx \frac{1}{2\mu_0} B_r^2(\theta, t)
$$
This simplification underscores a critical point: any distortion or harmonic content in the radial air-gap flux density is squared, amplifying specific force harmonics. The air-gap flux density \( B_r(\theta, t) \) is the result of the interaction between the magnetomotive force (MMF) produced by the rotor permanent magnets \( F_{pm}(\theta, t) \) and the permeance function of the air-gap \( \Lambda(\theta, t) \), which is modulated by stator slot openings and magnetic saturation:
$$
B_r(\theta, t) = F_{pm}(\theta, t) \cdot \Lambda(\theta, t)
$$
The rotor MMF and the air-gap permeance can be decomposed into their spatial and temporal harmonic series. Their interaction produces a complex spectrum of flux density harmonics. When squared according to the force equation, these flux density harmonics interact to generate radial force waves with specific spatial orders and temporal frequencies. Low-order spatial force waves (e.g., order 0, 2, 4) are particularly problematic as they can more easily excite the global bending and breathing modes of the stator core, leading to higher amplitude vibrations and louder noise. The primary objective in low-noise electric car motor design is to minimize the amplitude of these low-order, high-magnitude force harmonics.
The rotor structure plays a pivotal role in shaping \( F_{pm}(\theta, t) \). In interior permanent magnet (IPM) motors commonly used in electric cars for their high torque and wide speed range, the magnets are embedded within the rotor iron. The bridges of iron that connect the rotor poles and provide mechanical integrity are known as spacer bridges or flux barriers. These bridges are intentionally designed to be magnetically saturated during operation to “isolate” the magnets from each other magnetically. However, their geometry (width, length, shape) critically influences the leakage flux paths and ultimately the spatial distribution of the fundamental and harmonic content of the air-gap field. A suboptimal spacer bridge design can lead to localized flux crowding, increased harmonic distortion in \( B_r \), and consequently, larger amplitudes of problematic radial force harmonics. Therefore, the spacer bridge is a key design parameter for electromagnetic noise control in electric car motors.
2. Methodology: Optimization Strategy and Simulation Framework
2.1 Rotor Spacer Bridge Optimization
The baseline motor model considered is a 48-slot / 8-pole interior PMSM, a configuration popular in electric car traction applications for its balanced performance. The initial spacer bridge design was a conventional straight bridge with a constant width. Analysis of the baseline design revealed a pronounced 4th and 8th spatial order harmonic in the radial electromagnetic force density spectrum, correlating with significant vibration response.
Our optimization strategy focuses on modifying the spacer bridge geometry to smooth the transition of magnetic flux from the magnet cavity into the air-gap, thereby reducing the step-like changes in the permeance function that generate high-order MMF harmonics. The specific intervention is to extend the spacer bridge tangentially outward by 3.5 mm on both sides. This extension effectively increases the area of the saturated iron region guiding the flux, altering the magnetic reluctance path around the bridge. The comparison is conceptualized below:
| Feature | Baseline Design | Optimized Design |
|---|---|---|
| Spacer Bridge Shape | Straight, minimal length | Tangentially extended lobes (+3.5mm each side) |
| Primary Goal | Mechanical integrity, basic flux isolation | Flux guidance & harmonic content management |
| Expected Impact on Field | Sharper flux density transitions | Smoother spatial distribution of B_r |
This modification is parametric and does not fundamentally alter the rotor’s mechanical robustness or the magnet retention capability. Its sole purpose is electromagnetic refinement to benefit the acoustic performance of the electric car drive unit.
2.2 Finite Element Analysis (FEA) Model
A high-fidelity 2D transient electromagnetic FEA model was built for both the baseline and optimized rotor designs. The model solves Maxwell’s equations numerically to compute the time-varying magnetic field distribution. Key simulation parameters and steps include:
- Geometry & Mesh: A cross-section of one pole pair is modeled with anti-periodic boundary conditions. A fine mesh is applied, particularly in the air-gap, spacer bridge regions, and stator teeth, to ensure accurate flux density calculation.
- Material Properties: Nonlinear B-H curves are assigned to the stator and rotor laminations (M470-50A). NdFeB magnets are modeled with a linear demagnetization curve at their operating temperature.
- Operating Conditions: Simulations are run for both no-load (open-circuit) and rated load conditions. For load simulations, three-phase sinusoidal currents are injected corresponding to the motor’s rated torque and speed (a typical operating point for city driving in an electric car).
- Post-Processing: The radial air-gap flux density \( B_r(\theta, t) \) is extracted along a circular path in the middle of the air-gap. The radial electromagnetic force density \( p_r(\theta, t) \) is then calculated using the Maxwell stress method. A 2D Fourier Transform (spatial and temporal) is performed on \( p_r(\theta, t) \) to decompose it into its harmonic spectrum, identifying the amplitude of forces at specific spatial orders \( r \) and temporal frequencies \( f \).
The force harmonic of spatial order \( r \) and frequency \( f \) can be denoted as \( P_{r,f} \). The most critical ones for noise are typically at frequencies \( f = 2mf_1 \), where \( m=1,2,3… \) and \( f_1 \) is the fundamental electrical frequency, and at low spatial orders \( r = 0, 2, 4, … \).
2.3 Acoustic Noise Simulation Model
To translate the electromagnetic forces into radiated noise, a coupled vibro-acoustic simulation is performed. The process involves two main steps:
- Structural Vibration Analysis: The spatial-temporal force distribution \( p_r(\theta, t) \) from the FEA is mapped as a pressure load onto the inner surface of the stator teeth in a structural finite element model of the motor housing assembly. A modal frequency response analysis is conducted to compute the surface vibration velocities \( v(\mathbf{x}, t) \) of the motor housing.
- Acoustic Radiation Analysis: The vibration velocities are used as the boundary condition for an acoustic boundary element model. A sphere with a 1-meter diameter, centered on the motor, is defined as the acoustic evaluation surface. This is a standard practice for approximating the free-field radiation and calculating the sound pressure level (SPL). The acoustic medium is air at standard conditions. The SPL at a point is calculated as:
$$
SPL = 20 \log_{10}\left(\frac{p_{rms}}{p_{ref}}\right) \text{ dB}, \quad \text{where } p_{ref} = 20 \mu Pa.
$$
The overall acoustic power or the average SPL over the sphere provides a single metric to assess the optimization’s effectiveness in reducing noise from the electric car motor.
3. Results, Analysis, and Discussion
3.1 Electromagnetic Force Harmonic Spectrum
The most direct evidence of the optimization’s success is found in the spectral composition of the radial electromagnetic force density. The table below summarizes the amplitudes of key low-order spatial harmonics for both no-load and load conditions at the fundamental forcing frequency (twice the electrical frequency, 2f1).
| Spatial Order (r) | Baseline (No-Load) Amplitude [Pa] | Optimized (No-Load) Amplitude [Pa] | Reduction | Baseline (Load) Amplitude [Pa] | Optimized (Load) Amplitude [Pa] | Reduction |
|---|---|---|---|---|---|---|
| 0 | 1.85e4 | 1.52e4 | 17.8% | 2.21e4 | 1.83e4 | 17.2% |
| 2 | 6.40e4 | 5.05e4 | 21.1% | 8.90e4 | 6.95e4 | 21.9% |
| 4 | 3.15e4 | 2.40e4 | 23.8% | 4.35e4 | 3.25e4 | 25.3% |
The data clearly shows a consistent reduction across all critical low-order spatial harmonics. The 4th order harmonic, often a major excitation source for the stator’s ovalizing mode, shows the greatest percentage reduction. This uniform suppression is attributed to the smoother spatial distribution of \( B_r(\theta) \) achieved by the extended spacer bridge, which mitigates the sharp harmonics in the rotor MMF waveform. The improvement is present under both no-load and load conditions, proving the robustness of the optimization across different operating states of the electric car motor.
3.2 Impact on Torque Performance
A legitimate concern when optimizing for NVH is the potential impact on the primary torque-producing capability. Our analysis shows that the average output torque remains virtually unchanged, which is expected as the fundamental magnetic flux linkage is preserved. However, a key benefit is observed in torque quality: the torque ripple is reduced. Torque ripple \( \Delta T \) is defined as the peak-to-peak variation as a percentage of average torque:
$$
\Delta T = \frac{T_{max} – T_{min}}{T_{avg}} \times 100\%
$$
The simulation results indicate a reduction in torque ripple from approximately 4.2% in the baseline motor to 3.1% in the optimized motor at rated load. This 26% improvement in smoothness is a direct consequence of the reduced interaction between certain stator and rotor magnetic field harmonics. Lower torque ripple not only contributes to finer vehicle drivability (reducing jerks during low-speed maneuvering in an electric car) but also implies lower cyclical stress on the drivetrain components, potentially enhancing durability.
3.3 Vibration and Radiated Noise Results
The reduction in electromagnetic force harmonics directly translates to lower vibration amplitudes on the motor housing. The frequency response function of a point on the housing shows decreased acceleration levels, particularly at the frequencies corresponding to the dominant force harmonics (e.g., 2f1, 4f1, etc.).
The ultimate metric is the radiated sound. The simulated sound pressure level spectrum on the 1-meter sphere is presented below for a key frequency range. The most pronounced noise peaks are associated with the first few multiples of the twice-line-frequency (2f1) components.
| Acoustic Performance Metric | Baseline Motor | Optimized Motor | Improvement |
|---|---|---|---|
| Maximum SPL (at ~200 Hz / 2f1) | 73.02 dB | 65.13 dB | -7.89 dB |
| A-weighted Average SPL (0-2 kHz) | 50.37 dB(A) | 45.27 dB(A) | -5.10 dB(A) |
| Dominant Noise Frequency | 2f1, 4f1 | 2f1 (reduced amplitude) | Peak broadening suppressed |
The results are significant. A reduction of nearly 8 dB in the peak SPL represents a substantial subjective quieting, as a 10 dB reduction is perceived as approximately halving the loudness. The average noise level is also reduced by over 5 dB(A). This demonstrates that the spacer bridge optimization effectively targets the root cause of the tonal electromagnetic noise, leading to a quieter overall acoustic signature for the electric car powertrain. The sound quality improves from a pronounced tonal whine to a much milder broadband hum.
4. Extended Industrial and Economic Implications for Electric Car Manufacturing
The optimization presented has ramifications beyond mere acoustic performance; it touches upon manufacturing efficiency, cost, reliability, and compliance.
Manufacturing Efficiency and Cost: The extended spacer bridge design, while altering the mold for the rotor lamination, does not add complexity to the assembly process. In fact, by reducing the sensitivity of the electromagnetic performance to minor variations in magnet placement or lamination stacking, it can potentially loosen manufacturing tolerances. This can lead to a higher production yield and lower rejection rates. If we consider a simplified cost model where improved yield reduces effective unit cost, the benefit can be quantified. Let \( C_{mat} \) be the material cost, \( C_{lab} \) the labor/assembly cost, and \( Y \) the yield percentage. The effective cost per unit \( C_{eff} \) is:
$$
C_{eff} = \frac{C_{mat} + C_{lab}}{Y}
$$
An increase in yield \( \Delta Y \) from, for example, 97% to 98.5% due to more forgiving magnetic design, directly reduces \( C_{eff} \), improving the cost-competitiveness of the electric car motor.
System Reliability and Warranty: Reduced vibration and torque ripple have a direct positive impact on the mechanical fatigue life of bearings, gears, and housing fasteners. The stress amplitude \( \sigma_a \) on a component is proportional to the vibration acceleration or torque variation. According to the Basquin’s law for high-cycle fatigue, the number of cycles to failure \( N_f \) is related to the stress amplitude:
$$
\sigma_a^b \cdot N_f = C
$$
where \( b \) and \( C \) are material constants. A reduction in \( \sigma_a \) leads to an exponential increase in \( N_f \). This translates to lower failure rates in the field, reduced warranty claims, and enhanced brand reputation for reliability in the electric car market.
Energy Efficiency and Range: While the primary goal was NVH, the reduction in high-frequency electromagnetic force harmonics also implies lower levels of high-frequency magnetic flux variations in the stator iron. These high-frequency variations are a source of additional core losses \( P_{core} \). Although the loss reduction at a single operating point might be small (estimated 0.5-1.5%), over the diverse driving cycles of an electric car, the cumulative effect can contribute to a slight extension of vehicle range or allow for a smaller, lighter battery pack for the same range, creating a virtuous cycle of weight and cost savings.
Regulatory Compliance and Market Appeal: As regulations on vehicle exterior noise (e.g., EU regulation 540/2014) become stricter and now include provisions for quiet vehicles like electric cars to emit warning sounds at low speeds, the management of internal cabin noise becomes a key brand differentiator. A quieter cabin meets higher comfort standards and aligns with the premium, refined image that many electric car manufacturers seek to project.
5. Conclusion and Future Perspectives
This work has demonstrated a focused and effective approach to mitigating electromagnetic vibration and noise in electric car permanent magnet synchronous motors. By strategically optimizing the tangential length of the rotor spacer bridge, we achieved a smoother air-gap magnetic flux density distribution, leading to a significant reduction in the amplitude of low-order radial electromagnetic force harmonics. This electromagnetic improvement was quantitatively proven to reduce structural vibration and radiate acoustic noise, with a peak SPL reduction of nearly 8 dB. Furthermore, beneficial side effects including reduced torque ripple and potential gains in efficiency and reliability were discussed, highlighting the multifaceted value of such an optimization for the overall electric car drive system.
The study reinforces the principle that targeted electromagnetic design, supported by high-fidelity multiphysics simulation, is a powerful tool for NVH engineering in electric vehicles. The spacer bridge serves as a critical tuning parameter in the rotor design landscape.
Looking forward, several avenues for further research and development are apparent. First, the optimization presented here could be integrated into a formal multi-objective optimization (MOO) framework, using algorithms like genetic algorithms or surrogate modeling to simultaneously balance torque density, efficiency, cost, and NVH across the entire speed-torque map relevant to electric car driving cycles. Second, the interaction between the electromagnetic optimization and the mechanical excitation from the gearbox should be investigated in a fully coupled system model to identify potential synergistic or counteracting effects. Third, the application of advanced soft magnetic composite materials or graded permeability alloys in the stator or rotor could provide new degrees of freedom for shaping the magnetic field and controlling force harmonics. Finally, real-time active noise cancellation techniques, using the motor itself as an actuator to generate counter-vibration signals based on current harmonics injection, could be explored as a complementary software-defined solution to the hardware-based optimization described here.
In conclusion, the pursuit of quieter, smoother, and more efficient electric car propulsion systems is an ongoing engineering challenge. Through continued innovation in electromagnetic design, materials science, and control strategies, the goal of making the electric car driving experience not only clean and powerful but also supremely quiet and refined is well within reach.