In an ideal electric motor, electrical energy transforms seamlessly into a perfectly smooth, continuous rotational force. In reality, the output torque of modern permanent magnet synchronous motors (PMSMs) and switched reluctance motors (SRMs) is marred by microscopic, periodic fluctuations known as torque ripple and cogging torque.
These high-frequency pulsations act as internal hammers, exciting the mechanical resonances of the motor structure and producing the characteristic high-pitched whine or structural vibration observed in electric vehicles and precision robotics, as detailed by Yaojian Li et al. in Experimental study on vibration and noise of the ROEWE electric vehicle and Myung-Gyu Kim et al. in The Development of a Method for NVH Prediction in Electrical Driving Systems Using a Quasi-Transient Excitation Force. To build quiet, precise electric drives, researchers focus on two complementary paradigms: inherent structural cancellation during motor manufacturing, and active disturbance compensation via advanced power electronics control.
The Physics of Torque Pulsation
Why do these ripples occur? The problem stems from both geometric design and electrical imperfections.
Cogging Torque
Even when a motor is completely unpowered, turning the rotor by hand reveals a jerky, resistant sensation, an effect highlighted by Nobuyuki Matsui et al. in A DC Brushless Motor Using a Rare Earth Magnet for a Power Assisted Wheelchair. This is cogging torque, caused by the physical attraction between the rotor's permanent magnets and the steel slots of the stator. As the magnets pass the slots, the magnetic reluctance varies periodically, creating a position-dependent disturbance, as noted by Bao Zhang et al. in Iterative Learning-Based Negative Effect Compensation Control of Disturbance to Improve the Disturbance Isolation of System.
Inverter-Induced Harmonics
When the motor is running, the power inverter introduces non-idealities. For instance, to prevent short circuits, inverters require a brief "dead-time" when switching transistors. As Jie Shi demonstrates in Analysis and Compensation Control of Dead-Time Effect on Space Vector PWM, this dead-time distorts the voltage waveforms, manifesting primarily as 6th-order harmonics in the rotor's reference frame. These electrical harmonics interact with the fundamental magnetic field to produce rapid pulsations in phase current and torque, a phenomenon explored by Yuxi Yang et al. in Interzone duty cycle control of switched reluctance motor with comprehensive current error.
Mitigation via Structural Optimization
One way to eliminate torque ripple is to design it out of the hardware entirely. Since cogging torque is a function of geometry, altering the spatial distribution of the magnetic flux can force individual ripple components to cancel each other out.
Standard Assembly: [Magnet +] [Magnet +] [Magnet +] --> Cumulative Imperfections (High Ripple)
Optimized Assembly: [Magnet +] [Magnet -] [Magnet +] --> Destructive Cancellation (Low Ripple)
For example, Haipeng Liu et al. proposed an assembly approach in A Permanent Magnet Assembling Approach to Mitigate the Cogging Torque for Permanent Magnet Machines Considering Manufacturing Uncertainties that accounts for manufacturing uncertainties. By strategically pairing permanent magnets with opposing magnetic flux variations, the individual imperfections of the magnets destructively interfere, smoothing out the aggregate cogging torque.
Similarly, Feng Xing et al. demonstrated in Toward Reducing Undesired Rotation Torque in Maglev Permanent Magnet Synchronous Linear Motor that optimizing the physical topology of Halbach magnet arrays and coil structures inherently minimizes unwanted rotational forces before they can ever propagate into the mechanical system.
Active Control and Destructive Harmonic Injection
When hardware modifications are insufficient or too costly, control algorithms can be used to dynamically fight torque ripple. The core intuition here is simple: if we know a harmonic voltage causes a specific torque ripple, we can deliberately inject a counter-harmonic voltage to neutralize it.
In Influence of harmonic voltage coupling on torque ripple of permanent magnet synchronous motor, Shengtao Geng et al. uncovered a "negative synergistic effect" between specific combinations of voltage harmonics. While individual harmonics linearly worsen torque ripple, certain coupled harmonic pairs generate rotating magnetic fields that actively oppose and cancel each other out. This destructive interference stabilizes the output torque without reducing the motor's average power delivery capacity.
To execute this in real-time, control systems must accurately track and predict these fast-moving disturbances. Shenhui Du et al. utilized a Fading Memory Kalman Filter in Harmonic Injection Control of Permanent Magnet Synchronous Motor Based on Fading Memory Kalman Filtering to dynamically isolate harmonic disturbances and inject precise compensation currents, drastically reducing experimental torque ripple.
Learning and Observing Periodic Disturbances
Because torque ripple and cogging torque are highly deterministic—repeating predictably as a function of time or rotor position—they are ideal candidates for observer-based and learning-based control frameworks.
Disturbance Observers (DOB)
Instead of modeling every complex, non-linear physical interaction inside the motor, control loops can treat the effects of dead-time and slotting as a single, lumped disturbance. Jie Shi used a Linear Extended State Observer (LESO) to estimate and compensate for dead-time effects without requiring finicky current-polarity sensors. Similarly, Hyeong-Joon Ahn showed in Control Boost of an Active Magnetic Bearing System with DOBs that implementing disturbance observers significantly boosts system robustness against synchronous vibrations.
Iterative Learning Control (ILC)
Since the disturbances repeat every rotation, the controller can "learn" from previous cycles. Dongqi Ma developed an accelerated iterative learning control method in Accelerated Iterative Learning Control of Speed Ripple Suppression for a Seeker Servo Motor to suppress low-speed speed ripples, while Bao Zhang et al. combined ILC with negative effect compensation to isolate position-periodic cogging torque in high-precision tracking platforms.
Cycle N: [Measure Error] --> [Store Profile]
|
v
Cycle N+1: [Apply Anti-Control Profile] --> [Reduced Error]
Alternative Mathematical Formulations
For highly non-linear motors like Switched Reluctance Motors (SRMs), traditional PI control fails to manage the severe current pulsations that cause audible whine. To solve this, Nicholas Moehle reframed the entire problem in Optimal Current Waveforms for Switched-Reluctance Motors by mapping the non-linear motor physics into a Mixed-Integer Convex Program (MICP). This mathematical formulation allows the system to globally optimize and discover the mathematically "perfect" current waveforms that maximize efficiency while inherently minimizing torque ripple across all operating speeds.
Key Foundational Work
Analysis of characteristics of electromagnets with shading coil using the time-periodic finite element method. (1985) — contributed 1 concept elements to this direction.
A DC Brushless Motor Using a Rare Earth Magnet for a Power Assisted Wheelchair. (1997) — contributed 1 concept elements to this direction.