Electric Vehicle Creep Booming Noise Solved with Active Damping Optimization
In the rapidly evolving world of electric mobility, where quiet cabins and seamless driving experiences are paramount, a new engineering challenge has emerged: low-frequency booming noise during vehicle creep. Unlike traditional internal combustion engine (ICE) vehicles, which rely on components like torque converters and flexible couplings to absorb drivetrain vibrations, pure electric vehicles (EVs) operate with highly rigid, low-damping drivetrains. This structural characteristic, while efficient, can lead to unexpected NVH (Noise, Vibration, and Harshness) issues—particularly under low-load, low-speed conditions such as creeping.
A recent study published in Noise and Vibration Control has shed light on this phenomenon, offering a practical and effective solution without requiring hardware modifications. The research, conducted by a team from Zhejiang Smart Automobile Intelligence Technology Co., Ltd. and Geely Automotive Research Institute (Ningbo) Co., Ltd., focuses on a real-world case involving a rear-wheel-drive electric SUV experiencing a noticeable 36 Hz booming noise during flat-road creeping at approximately 7 km/h. This issue, though absent on inclines or declines, was persistent enough to degrade perceived vehicle quality and passenger comfort.
The team, led by Shen Long, a senior NVH engineer, identified the root cause not in mechanical defects or poor assembly, but in the instability of the vehicle’s active damping control system—a software-based strategy designed to suppress torque fluctuations. Their findings challenge conventional assumptions and highlight the growing complexity of EV control systems, where software tuning can have profound acoustic consequences.
The Creep Condition Conundrum
Creeping—the slow, idle-like forward motion of a vehicle when the driver’s foot is lifted from the brake—is a common driving scenario, especially in urban environments with frequent stops. In ICE vehicles, engine idle vibration and background noise often mask minor drivetrain oscillations. However, in EVs, the near-silent operation of the electric motor removes this masking effect. As a result, even subtle vibrations become perceptible, transforming what might be an inaudible fluctuation in a gasoline car into a clearly audible low-frequency boom.
The subject vehicle in the study was equipped with an integrated “three-in-one” electric drive system featuring an 8-pole, 48-slot permanent magnet synchronous motor. During flat-surface creeping, occupants reported a distinct low-frequency rumble, subjectively rated as intrusive. Objective measurements confirmed a prominent 36 Hz peak in both cabin noise and drivetrain vibration. The noise level reached 42 dB(A), a level easily detectable in the otherwise quiet EV environment.
What made the issue particularly puzzling was its conditional nature: the booming disappeared when the vehicle crept uphill or downhill. This behavior suggested a strong dependence on load and drivetrain dynamics rather than a fixed mechanical resonance.
Diagnostic Approach and Data-Driven Insights
To pinpoint the source, the engineering team conducted comprehensive on-road testing. They instrumented the electric drive unit with tri-axial accelerometers at key locations: the motor housing, left and right suspension mounts, and front knuckle. A microphone was placed at the driver’s seat, and critical vehicle signals—including motor speed, requested wheel torque, and actual wheel torque—were logged via CAN bus.
Spectral analysis revealed a consistent 36 Hz peak across all measured vibration points, with the highest amplitudes at the motor housing and left mount. This pattern indicated that the vibration originated at the powertrain and was transmitted through the mounting system into the chassis and cabin.
Time-domain analysis provided further clarity. While vehicle speed and requested torque remained stable, the actual wheel torque exhibited significant oscillations—peaking at around 70 N·m. This was a stark contrast to the driver’s input, which called for minimal torque. The periodicity of these torque fluctuations corresponded precisely to a 36 Hz frequency, aligning with the observed noise.
Further investigation revealed a critical insight: the motor, with four pole pairs, naturally produces a 4th-order excitation. At the creeping speed of 550 rpm, the 4th-order frequency calculates to (4 × 550) / 60 = 36.67 Hz—nearly identical to the measured 36 Hz. This confirmed that the booming was linked to electromagnetic torque ripple inherent in the motor’s design, amplified by the drivetrain’s lack of damping.
Active Damping: A Double-Edged Sword
To mitigate such torque fluctuations, many EV manufacturers employ active damping control—a feedback strategy that uses the motor’s fast torque response to counteract unwanted oscillations. The principle is straightforward: detect speed or torque deviations, calculate a corrective torque, and apply it in real time to dampen the oscillation.
However, the Geely team discovered that under certain low-torque, low-speed conditions, this very system could become unstable. Instead of suppressing torque ripple, it was amplifying it.
To test this hypothesis, they disabled the active damping function and repeated the test. The results were striking. With active damping off, the 36 Hz vibration peak at the motor housing dropped from over 1.0 m/s² to just 0.1 m/s². Cabin noise at 36 Hz decreased from 42 dB(A) to 35 dB(A), and actual wheel torque fluctuation fell from 70 N·m to 20 N·m. While some residual noise and vibration remained—indicating the presence of inherent torque ripple—the most significant reduction occurred when the active damping was deactivated.
This counterintuitive outcome led to a critical conclusion: the active damping system, intended as a solution, had become the primary contributor to the booming noise. The control loop was not stabilizing the system but destabilizing it, likely due to phase lag, parameter mismatch, or nonlinearities in the drivetrain (such as backlash or hysteresis) that the control algorithm failed to account for under light-load conditions.
Rethinking Control Logic: A Safer, More Robust Strategy
Armed with this understanding, the team turned to control strategy optimization. Rather than abandoning active damping—a valuable tool in other driving scenarios—they sought to make it more robust and adaptive.
The core of their improvement was a revised control logic that introduced upper limits on the damping torque output, particularly during creep conditions. The rationale was simple: in a system already prone to oscillation, allowing the controller to apply large corrective torques could exacerbate the problem. By capping the maximum damping torque, they prevented the control system from overreacting to small fluctuations.
Two key modifications were implemented:
- A 60% reduction in active damping gain within the motor speed range typical of creeping.
- An 80% reduction in the maximum allowable anti-oscillation torque when creep conditions were detected.
These parameters were not arbitrarily chosen but were derived through iterative real-vehicle calibration, balancing noise suppression against system stability. The goal was not to eliminate all torque ripple—which is physically inherent—but to ensure that any fluctuations decayed quickly rather than growing or sustaining.
Validation and Real-World Impact
After implementing the updated control strategy, the team conducted another round of testing. The results were transformative. The actual wheel torque fluctuation dropped from 70 N·m to just 2 N·m—a 97% reduction. The 36 Hz peak in cabin noise vanished, with the level decreasing by 20 dB(A), effectively eliminating the booming sensation. Vibration measurements at the motor, mounts, and suspension points showed no significant peaks at 36 Hz, confirming that the excitation source had been neutralized.
Subjective evaluations by NVH engineers and test drivers confirmed the objective data: the vehicle now crept smoothly and silently, indistinguishable from higher-end EVs in terms of refinement. The fix was achieved entirely through software—no changes to motor design, gear alignment, or mount stiffness were required.
This outcome underscores a broader trend in automotive engineering: as vehicles become more electrified and software-defined, the line between mechanical performance and control system behavior blurs. What appears to be a hardware defect may, in fact, be a software tuning issue. The ability to diagnose and resolve such problems remotely, through over-the-air updates, represents a significant advantage for manufacturers.
Implications for the EV Industry
The study by Shen Long and colleagues offers more than a solution to a single noise issue—it provides a framework for diagnosing and resolving similar problems across the EV landscape. As automakers race to deliver quieter, smoother electric vehicles, they must contend with the unique dynamics of rigid, low-damping drivetrains. Traditional NVH tools—such as adding mass or stiffening structures—are often insufficient or counterproductive.
Active damping, when properly tuned, remains a powerful tool. But as this research shows, it must be applied with caution. Blindly increasing damping gains can lead to instability, especially in nonlinear or lightly loaded systems. The concept of “limiting the upper bound” of control effort introduces a safety-oriented philosophy to control design, prioritizing stability over maximum performance.
Moreover, the conditional nature of the booming—present on flat roads but absent on slopes—highlights the importance of testing under diverse real-world conditions. Laboratory simulations or standard driving cycles may not capture such edge cases, emphasizing the need for comprehensive on-road validation.
The findings also suggest that future EV control systems should be more adaptive. Instead of fixed damping parameters, intelligent algorithms could adjust damping strength based on road gradient, load, temperature, and driving style. Machine learning techniques could even learn from driver behavior and environmental conditions to preemptively optimize NVH performance.
Looking Ahead: From Reactive Fixes to Proactive Design
While the optimized active damping strategy successfully resolved the immediate issue, the authors emphasize the importance of addressing such problems earlier in the development cycle. Relying on post-launch software fixes, while possible, is less ideal than designing robust control systems from the outset.
Future work, they suggest, should focus on improving the theoretical understanding of torque ripple generation and suppression, refining low-level control algorithms, and developing better simulation tools that can predict these interactions before physical prototypes are built. Integrating NVH considerations into the initial stages of motor and inverter design could prevent many issues from arising in the first place.
Additionally, as EVs adopt more advanced features—such as torque vectoring, regenerative braking coordination, and predictive energy management—the complexity of control systems will only increase. Ensuring that these systems do not interfere with each other or create unintended NVH side effects will be a major challenge.
The case also illustrates the value of interdisciplinary collaboration. Solving the booming noise required expertise in acoustics, mechanical dynamics, control theory, and embedded software. It was not enough to measure noise; one had to understand the entire chain from electromagnetic excitation to structural transmission to human perception.
In conclusion, the resolution of the creep booming issue in this EV represents a significant achievement in applied automotive engineering. It demonstrates that even in the face of complex, software-driven problems, systematic analysis and intelligent control design can deliver elegant, cost-effective solutions. As the automotive industry continues its transition to electrification, studies like this will serve as essential references for engineers striving to deliver the quiet, refined driving experience that consumers expect.
The research not only enhances the quality of a specific vehicle but also contributes to the broader knowledge base of EV NVH engineering. It reminds us that in the pursuit of innovation, attention to detail—and a deep understanding of system interactions—remains paramount.
Shen Long, Zhang Jun, Zhao Mingbin, Qin Bin, Fang Zhen, Zhejiang Smart Automobile Intelligence Technology Co., Ltd. and Geely Automotive Research Institute (Ningbo) Co., Ltd., Noise and Vibration Control, DOI: 10.3969/j.issn.1006-1355.2024.06.033