Revolutionizing EV Motors: Taguchi Method Optimizes Star-Delta FSPMM Performance
The electric vehicle (EV) industry is undergoing a seismic shift, driven by relentless demands for higher efficiency, greater power density, and quieter, smoother operation. At the heart of this revolution lies the electric motor, the critical component converting electrical energy into mechanical motion. Among the various motor topologies vying for dominance in next-generation EVs, Fractional Slot Permanent Magnet Machines (FSPMMs) have emerged as a frontrunner due to their inherent advantages: compact size, high power-to-weight ratio, and robust performance. However, unlocking their full potential requires sophisticated design optimization that addresses complex electromagnetic interactions. A groundbreaking study published in the Journal of Henan Polytechnic University (Natural Science) has demonstrated a powerful new approach using the Taguchi method to significantly enhance the electromagnetic performance of a specific, highly promising FSPMM variant – the star-delta connected, interior V-shaped rotor machine.
This research, led by Dr. Chen Zhenfei from the School of Electrical and Power Engineering at Hohai University, alongside colleagues Fan Chenyang, Tang Jun, Wang Qingyan from Jinling Institute of Technology, and Li Jiayu, tackles a fundamental challenge in modern motor design: the intricate interplay between numerous geometric parameters and their collective impact on key performance metrics like torque ripple, cogging torque, efficiency, and magnetic field quality. The team’s work offers not just incremental improvements but a systematic, statistically rigorous methodology for achieving substantial gains, potentially setting a new benchmark for FSPMM development.
The journey begins with a recognition of the limitations of conventional FSPMM designs. While offering advantages over integer-slot machines, traditional FSPMMs often suffer from significant harmonic content in their armature magnetomotive force (MMF). This harmonic richness leads directly to undesirable effects: increased iron losses, elevated eddy current losses within the permanent magnets themselves, and pronounced torque fluctuations during operation. These issues manifest as reduced efficiency, audible noise, mechanical vibration, and compromised control precision – all critical drawbacks for the demanding environment of an electric vehicle. To mitigate these problems, researchers have explored various strategies, including altering stator slot geometry or employing different winding configurations. Previous studies often focused on single aspects, such as the effect of slot opening width on cogging torque or eddy current loss, providing valuable but fragmented insights. The novelty of Chen et al.’s work lies in its holistic approach, combining a specific, advanced winding structure with a comprehensive, multi-objective optimization technique.
The researchers chose a 10-pole, 12-slot FSPMM configuration, a common choice balancing pole count and slot number for good performance characteristics. Crucially, they compared two distinct stator winding arrangements: the conventional double-layer star (Y) connection and a more complex four-layer star-delta (Y-Δ) connection. The rationale behind exploring the Y-Δ configuration stems from its potential to inherently suppress harmonic components in the MMF. In a four-layer Y-Δ setup, the windings are divided into star-connected and delta-connected portions. Under ideal conditions, the current in the delta portion lags the star portion by π/6 radians, and its amplitude is approximately 0.5774 times that of the star portion. This phase and amplitude relationship creates a cancellation effect for certain harmonic orders, effectively smoothing out the overall MMF waveform. Furthermore, to shield the permanent magnets from the detrimental effects of these harmonics and to enhance mechanical robustness for potential high-speed applications, the team opted for an interior “V”-shaped rotor structure, where the magnets are embedded deep within the rotor iron core. This contrasts with surface-mounted designs, which are simpler but less resistant to demagnetization and centrifugal forces.
Initial comparative simulations using 2D finite element analysis (FEA) software revealed a clear superiority of the four-layer Y-Δ configuration over the standard double-layer Y connection. Across a range of rotor parameter variations – specifically, changes in the permanent magnet length-to-width ratio (Lw) and the rotor yoke height (ht), while keeping the total magnet volume constant – the Y-Δ machine consistently demonstrated lower torque ripple, higher electromagnetic torque output, and improved overall efficiency. For instance, under identical rotor parameters, the Y-Δ machine exhibited significantly lower torque fluctuation, translating to smoother acceleration and deceleration crucial for passenger comfort. Simultaneously, it delivered higher average torque, meaning more usable power from the same physical size and magnet volume, a direct benefit for extending vehicle range or improving performance. The efficiency gains were also notable, attributed to the combined effect of higher torque output and lower losses, particularly reduced eddy current losses in the magnets due to the harmonic suppression provided by the Y-Δ winding.
However, even with the superior Y-Δ winding, the performance of the interior FSPMM remains highly sensitive to the precise dimensions of its rotor components. The complex magnetic circuit, involving interactions between the magnets, the rotor yoke, the flux barriers separating magnets of the same polarity and adjacent polarities, and the air gap, creates a scenario where changing one parameter can have cascading, often non-linear, effects on multiple performance metrics. Traditional optimization methods, which involve systematically varying each parameter while holding others constant, become computationally prohibitive when dealing with six or more interacting variables. This is where the Taguchi method, developed by Japanese engineer Genichi Taguchi, proves invaluable. It is a statistical experimental design technique that allows engineers to identify the optimal combination of parameters using a minimal number of experiments, focusing on minimizing the sensitivity of the system to uncontrollable factors (noise).
Chen and her team applied the Taguchi method rigorously to optimize the rotor geometry of their four-layer Y-Δ FSPMM. They identified six key geometric parameters as optimization factors: permanent magnet thickness (hm), rotor yoke height (ht), stator slot opening width (wm), air gap length (δ), width of the flux barrier between magnets of the same polarity (Rs), and width of the flux barrier between adjacent polarity magnets (Ri). The target performance metrics, or optimization objectives, were carefully chosen to represent the most critical aspects of motor performance for EV applications: minimizing torque ripple (TR) and cogging torque (TC), maximizing efficiency (η), and minimizing the total harmonic distortion (THD) of the no-load air gap flux density. THD is a crucial indicator of magnetic field quality; a lower THD signifies a more sinusoidal flux distribution, leading to reduced noise, vibration, and core losses.
To implement the Taguchi method, the researchers defined five discrete levels for each of the six optimization factors, based on practical manufacturing constraints and prior knowledge of the machine’s behavior. Using an L25(5^6) orthogonal array – a specific matrix designed to efficiently explore the parameter space – they conducted only 25 distinct FEA simulations. This represents a massive reduction compared to the 15,625 simulations that would be required for a full factorial analysis (5 levels ^ 6 factors). Each simulation provided data for all four optimization objectives. The results were then subjected to detailed statistical analysis. First, the average value of each objective was calculated for each level of every factor. This revealed trends: for example, increasing the air gap length generally reduced THD and TC but could negatively impact torque and efficiency. Similarly, reducing the slot opening width tended to lower TC and TR. The analysis also showed that some factors had a dominant influence on specific objectives. For instance, the magnet thickness (hm) was found to have the highest weighting (83.96%) on efficiency, while the adjacent flux barrier width (Ri) and rotor yoke height (ht) were most influential on cogging torque.
Crucially, the optimal settings for minimizing one objective often conflicted with those for another. Minimizing torque ripple might require a different combination of parameters than maximizing efficiency. To resolve this multi-objective conflict, the team performed variance analysis, calculating the contribution of each factor’s variation to the overall variation in each objective. By synthesizing these findings, they determined a single, optimal combination of parameter levels that offered the best compromise across all four critical performance metrics: hm at level 1 (4 mm), ht at level 1 (3 mm), wm at level 1 (1.5 mm), δ at level 5 (1.5 mm), Rs at level 1 (1.5 mm), and Ri at level 5 (2.5 mm). This optimized set of parameters represented a significant departure from the initial design, notably featuring a much larger air gap and narrower slot opening.
The true test of the Taguchi method’s effectiveness lay in validating the predicted performance improvement through detailed FEA simulations comparing the original design with the optimized version. The results were compelling and unequivocal. The optimized motor achieved a remarkable 20.24% reduction in torque ripple, a critical metric for smooth, quiet operation. Electromagnetic torque saw a 5.85% increase, directly enhancing the motor’s power capability. The total harmonic distortion of the air gap flux density decreased by 10.52%, indicating a cleaner, more sinusoidal magnetic field, which contributes to lower core losses and reduced acoustic noise. Cogging torque, responsible for the characteristic “cogging” sensation at low speeds, was reduced by 15.52%. Perhaps most importantly for EV manufacturers focused on range and energy consumption, the overall efficiency of the motor improved from 95.30% to 95.54%. While the optimization slightly increased eddy current losses in the magnets (from 4.15W to 5.13W), this was more than offset by reductions in iron losses and the significant gain in output power, resulting in the net efficiency improvement.
This study provides several key takeaways for the EV industry and motor designers. Firstly, it validates the four-layer Y-Δ winding configuration as a superior alternative to conventional double-layer Y windings for FSPMMs, offering tangible benefits in torque smoothness, output power, and efficiency. Secondly, it demonstrates the immense practical value of the Taguchi method for optimizing complex electromechanical systems like interior permanent magnet motors. By drastically reducing the computational burden of design exploration, it enables engineers to find globally optimal solutions more quickly and cost-effectively. Thirdly, the specific findings regarding the impact of individual parameters – such as the strong influence of air gap length on harmonic distortion and cogging torque, or the trade-offs involved in adjusting magnet shape and flux barrier widths – provide actionable design guidelines for future motor development.
Looking ahead, the implications of this research extend beyond the specific 10-pole, 12-slot machine studied. The methodology – combining advanced winding structures with robust statistical optimization – is broadly applicable to other FSPMM configurations and potentially other types of electric machines. As EV manufacturers continue to push the boundaries of performance, efficiency, and cost, tools like the Taguchi method will become increasingly essential for navigating the complex design space. The work by Chen Zhenfei, Fan Chenyang, Tang Jun, Wang Qingyan, and Li Jiayu represents a significant step forward, offering a proven pathway to unlock the full potential of FSPMM technology for the next generation of electric vehicles. Their research underscores that innovation in EVs isn’t just about batteries or power electronics; it’s also deeply rooted in the fundamental design and optimization of the electric motor itself, where meticulous engineering can yield substantial real-world benefits.
Chen Zhenfei, Fan Chenyang, Tang Jun, Wang Qingyan, Li Jiayu. Journal of Henan Polytechnic University (Natural Science). doi:10.16186/j.cnki.1673-9787.2022020030