A Novel Approach to Computing the Permanent Magnet Thickness in PMSMs

Inside the heavy, humming heart of modern industry, from the deep-sea drills of offshore oil rigs to the massive propellers of electric ships, there is a quiet revolution happening in how we build the machines that turn electricity into motion. At the center of this revolution is a type of motor called a permanent magnet synchronous motor, a device that relies on powerful magnets to create force. One of the most critical decisions engineers make when designing these motors is determining exactly how thick those magnets should be. If the magnets are too thin, the motor will be weak and inefficient, failing to do the heavy lifting required. If they are too thick, the motor becomes unnecessarily expensive, wasting rare and costly materials without gaining much extra power. For decades, finding this perfect thickness has been a balancing act between two imperfect methods: one that relies on rough guesses based on old rules of thumb, and another that involves running thousands of computer simulations to test every possible thickness, a process that can take days or even weeks.

A team of researchers from Bohai University in China has now introduced a new way to solve this puzzle, one that combines the speed of a quick guess with the precision of a detailed computer model. They focused their work on a massive motor designed for low-speed, high-torque applications, specifically a 1600-kilowatt machine intended for heavy-duty industrial use. The challenge with this specific motor is that its internal structure is complex, with magnets hidden inside the rotor in a U-shape, surrounded by iron bridges and barriers that cause the magnetic fields to behave in unpredictable ways. In such a complex environment, simple rules often fail, and the slow, brute-force method of testing every option is too time-consuming for modern design schedules. The researchers set out to create a smarter path, a method that could quickly zero in on the correct magnet thickness without wasting time or money.

The new approach works like a conversation between a computer model and a mathematical formula. Instead of guessing the thickness once and hoping for the best, or testing every single possibility from thin to thick, the researchers start with a reasonable estimate. They build a digital twin of the motor in a computer and run a simulation to see how the magnetic fields actually behave with that starting thickness. The computer then measures three specific things that change as the magnets get bigger or smaller: how much magnetic flux leaks away unused, how much resistance the iron core offers to the magnetic field, and the average strength of the magnetic field in the air gap between the rotor and the stator. These are not fixed numbers; they shift depending on the exact shape and size of the motor's internal parts.

Using the data gathered from that first simulation, the researchers feed these real-world values back into their calculation formula to determine a new, more accurate thickness. They then update the computer model with this new thickness and run the simulation again. This cycle repeats, with the computer learning from each step and adjusting the numbers slightly, until the result stops changing significantly. In their study, this iterative process converged on a final magnet thickness of 17.00 millimeters after just a few rounds of calculation. The beauty of this method is that it captures the complex, messy reality of the magnetic fields—like the way the iron core gets saturated or how the magnetic field leaks around the edges—without needing to test hundreds of different designs.

To prove their method worked, the team compared it against the two traditional approaches. They ran the slow, exhaustive computer sweep that tests every possible thickness, which took 1200 seconds and confirmed that 17.00 millimeters was indeed the correct answer. They also tried the old, fast method based on fixed rules, which took only 10 seconds but suggested a thickness of 17.82 millimeters. This older method was wrong because it assumed the magnetic fields behaved in a simple, predictable way that they do not in this complex motor. The new method, however, took 300 seconds to reach the correct 17.00-millimeter answer. This represents a 75 percent reduction in time compared to the exhaustive sweep, while avoiding the error of the quick guess.

The researchers did not stop at the computer screen; they built a physical prototype of the motor using the 17.00-millimeter thickness to see if it performed as predicted. They tested the machine under heavy load, measuring its torque, speed, and efficiency. The results matched the computer simulations almost perfectly, with the measured torque differing by less than one percent from the predicted values. The physical motor delivered the required power efficiently, confirming that the new calculation method had found the sweet spot. Furthermore, the team analyzed the performance of the motor if they had used the thicker, 17.82-millimeter magnets suggested by the old method. They found that the extra material provided almost no benefit in terms of power output but did increase the weight and cost, while also causing slightly more energy loss as heat. The 17.00-millimeter design was not only faster to calculate but also resulted in a more efficient machine that used less rare magnet material.

This work demonstrates that engineers can now design these complex, high-power motors with greater speed and confidence. By letting the computer model teach the calculation formula how the magnetic fields are actually behaving, rather than relying on static assumptions, the design process becomes both faster and more accurate. The result is a motor that is built with the right amount of material, avoiding waste and ensuring peak performance, a small but significant step forward in the engineering of the machines that power our modern world.