Optimal Common-mode Voltage Suppression Pulse-width Modulation Based on Adjacent-Four-Vector Synthesis Principle

In the world of modern electricity, power converters act as the essential translators between the steady flow of direct current from batteries or solar panels and the rhythmic, back-and-forth surge of alternating current that powers our homes and factories. These devices, known as inverters, work by rapidly switching electronic gates on and off to shape the voltage. However, this high-speed switching creates an invisible side effect: a fluctuating voltage that exists between the electrical system and the ground. Engineers call this common-mode voltage. While it does not power the motor or light the bulb, it is a potent source of trouble. In solar farms, this stray voltage can leak into the ground, distorting the current and creating safety hazards. In electric motors, it generates high-frequency vibrations that travel down the motor shaft, grinding away at the bearings and shortening the machine's life. It also acts as a source of electromagnetic noise that can interfere with nearby sensitive equipment.

For decades, engineers have tried to tame this voltage. Some have added bulky filters or extra hardware to the system, but this increases cost and size. Others have tried to change the software that controls the switches, aiming to cancel out the voltage by carefully timing the on-and-off cycles. The most common approach involves a method called space vector pulse-width modulation, which uses a specific set of switching patterns. A popular variation of this method tries to eliminate the most extreme voltage spikes by avoiding certain "zero" states. While this successfully lowers the peak height of the voltage, it often leaves behind a different problem: a slow, rolling fluctuation in the total voltage exposure over time. This lingering fluctuation is what causes the motor bearings to overheat and the magnetic filters to saturate. The challenge has been finding a way to suppress both the sharp peaks and the slow rolling waves without sacrificing the smoothness of the motor's rotation or the efficiency of the power conversion.

A team of researchers from Tongji University and the China University of Mining and Technology has proposed a new strategy to solve this dual problem. They developed a control method based on a principle they call adjacent-four-vector synthesis. To understand the innovation, one must first look at how the old methods worked. Traditional approaches typically use three voltage vectors to build the desired output, much like a painter mixing three primary colors to create a specific shade. When the researchers tried to suppress the common-mode voltage, they found that using only three vectors left them with very little freedom to adjust the outcome; they could lower the peak, but they could not smooth out the rolling waves, or vice versa. The new method breaks this limitation by using four vectors instead of three. By adding this fourth element to the mix, the system gains extra degrees of freedom, allowing the controller to fine-tune the switching sequence with a level of precision that was previously impossible.

The researchers designed a system that selects four specific switching states that sit next to each other in the electrical space, rather than the three closest ones used in older techniques. They then created a mathematical framework to determine exactly how long each of these four states should remain active. This is where the true power of the new method lies. The team realized they could use the extra vector to balance the voltage exposure over time. They developed a step-by-step optimization process. First, they calculated the ideal timing to minimize the slow, rolling fluctuations of the common-mode voltage. Once that baseline was set, they used the remaining flexibility to adjust the switching order, specifically targeting the reduction of electrical noise and current ripples in the motor. This layer-by-layer approach allowed them to tackle the voltage peaks, the long-term voltage exposure, and the output quality all at once.

To verify their theory, the team built a simulation model and a physical laboratory prototype using a standard two-level voltage source inverter. They tested the system under various conditions, comparing their new method against the traditional standard and other existing suppression techniques. The results were striking. In their simulations, the new method reduced the peak-to-peak fluctuation of the common-mode voltage exposure by more than 90 percent compared to the traditional standard. Even when the system was pushed to higher power levels, the reduction remained substantial, cutting the fluctuation by nearly 60 percent. The physical experiments confirmed these findings. When they measured the voltage between the motor and the ground, the traditional method showed a swing of about 48 volts. The new method, along with other suppression techniques, successfully restricted this swing to about 16 volts, a reduction of two-thirds.

Perhaps more importantly, the new method did not come at the cost of performance. In many previous attempts to suppress voltage, engineers had to accept a noisier, rougher output current as the price for a quieter voltage. The researchers found that by carefully selecting the order in which the four vectors were applied, they could actually improve the quality of the output current. When they analyzed the harmonic distortion—a measure of how clean the electrical wave is—their method produced a smoother current than the other advanced suppression techniques they tested. For instance, at a specific operating point, the weighted total harmonic distortion of their method was 0.535 percent, which was lower than the 0.557 percent seen in the competing advanced method. This means the motor would run more quietly and efficiently, with less wear on its internal components.

The study demonstrates that by expanding the toolkit from three vectors to four, engineers can gain the control needed to solve multiple problems simultaneously. The researchers showed that it is possible to limit the peak voltage, smooth out the long-term voltage exposure, and maintain high-quality output without adding extra hardware. Their work suggests that the key to better electric drives and cleaner power grids may not lie in building bigger filters, but in rethinking the fundamental logic of how we switch the power on and off. The findings, validated through both computer models and real-world testing, offer a clear path forward for reducing the hidden costs of electrical noise in modern power systems.