New ADRC-Based Control Strategy Enhances Performance of Electric Vehicle Motors
A groundbreaking advancement in electric vehicle (EV) motor control has emerged from research conducted at Shenyang University of Chemical Technology, offering a promising solution to long-standing challenges in the performance and efficiency of interior permanent magnet synchronous motors (IPMSMs). The study, led by Associate Professor Kong Xiaoguang and co-authored by Luan Zhaoyu, introduces a novel control strategy that integrates Active Disturbance Rejection Control (ADRC) with Maximum Torque Per Ampere (MTPA) and leading-angle flux-weakening techniques. Published in the Journal of Dalian Polytechnic University, this research presents a comprehensive approach to improving the dynamic response, stability, and speed range of IPMSMs—key components in modern EV powertrains.
As the global automotive industry accelerates its shift toward electrification, the demand for high-performance, energy-efficient, and reliable electric motors has intensified. Among various motor types, IPMSMs have gained widespread adoption due to their high power density, excellent efficiency, and robust torque characteristics. However, these advantages come with inherent complexities. IPMSMs are inherently nonlinear, multivariable systems with strong coupling between control variables. Their performance is highly sensitive to internal parameter variations—such as changes in stator resistance or inductance due to temperature fluctuations—and external disturbances like load torque variations. Traditional Proportional-Integral (PI) controllers, while widely used for their simplicity, often struggle to maintain optimal performance under such dynamic conditions. They typically exhibit poor speed response, significant overshoot, and limited disturbance rejection capability, especially during rapid acceleration or sudden load changes—common scenarios in real-world driving.
The limitations of PI control become particularly evident when an EV motor operates near or beyond its rated speed. At high speeds, the back electromotive force (EMF) generated by the permanent magnets increases proportionally with rotor speed, requiring higher terminal voltage to maintain current flow. However, the inverter supplying the motor has a finite DC-link voltage, imposing a hard limit on the maximum achievable stator voltage. Once this voltage ceiling is reached, further speed increases are impossible without reducing the magnetic flux—a process known as flux weakening. Conventional control strategies, such as the simple Id = 0 control, fail to exploit the reluctance torque inherent in IPMSMs, leading to suboptimal efficiency and a limited high-speed operating range. Moreover, implementing effective flux-weakening control often requires precise knowledge of motor parameters and complex calculations, making it vulnerable to parameter mismatches and model inaccuracies.
To address these challenges, Kong and Luan propose a control architecture that fundamentally rethinks how disturbances are handled within the motor control loop. At the heart of their strategy is the ADRC controller, which replaces the conventional PI controller in the speed regulation loop. Unlike model-based control methods, ADRC does not rely on an exact mathematical model of the motor. Instead, it treats all uncertainties—both internal (such as parameter drift and unmodeled dynamics) and external (such as load torque disturbances)—as a single “total disturbance.” This total disturbance is estimated in real time using an Extended State Observer (ESO), a core component of the ADRC framework. Once estimated, this disturbance is actively compensated for in the control signal, effectively canceling out its impact on the system. This allows the closed-loop system to behave like a simple, predictable integrator, significantly enhancing its robustness and dynamic performance.
The elegance of ADRC lies in its ability to decouple the control problem from the intricacies of the motor’s physics. By continuously observing and counteracting disturbances, the controller ensures that the motor’s actual speed closely tracks the desired reference, even in the face of abrupt changes. This is achieved through a three-part structure: a Tracking Differentiator (TD) that generates a smooth reference trajectory for the speed command, the ESO that estimates both the system state and the total disturbance, and a Nonlinear State Error Feedback (NLSEF) unit that generates the final control output based on the error between the desired and estimated states. This structure enables faster response times, minimal overshoot, and superior disturbance rejection compared to traditional PI control, which often requires careful tuning and can become unstable under varying operating conditions.
Complementing the advanced speed control is an optimized torque production strategy that maximizes efficiency across the entire speed range. Below the motor’s rated speed, the researchers implement MTPA control. This technique calculates the optimal combination of direct-axis (Id) and quadrature-axis (Iq) currents to produce the maximum possible torque for a given stator current magnitude. By minimizing the current required for a specific torque output, MTPA significantly reduces copper losses (which are proportional to the square of the current), thereby improving the motor’s efficiency during typical urban driving conditions where high torque at low to medium speeds is frequently required. This is particularly crucial for extending the driving range of EVs, a primary concern for consumers and manufacturers alike.
Above the rated speed, where the inverter voltage reaches its limit, the control strategy seamlessly transitions to a flux-weakening mode to enable operation at higher speeds. The researchers adopt a “leading-angle” flux-weakening method, which is both simple and effective. Instead of relying on complex calculations based on precise motor parameters, this method adjusts the angle of the stator current vector relative to the quadrature axis. By increasing this angle (introducing a “leading” component), the direct-axis current (Id) is made more negative (demagnetizing), which counteracts the permanent magnet flux. This reduction in effective flux lowers the back EMF, allowing the inverter to push more current into the motor and thus sustain higher speeds without exceeding the voltage limit. The leading-angle method is inherently robust because it uses feedback from the actual stator voltage to adjust the current angle, making it less sensitive to parameter variations and more reliable in real-world applications.
The integration of ADRC, MTPA, and leading-angle flux weakening creates a synergistic control system. The ADRC ensures precise and stable speed regulation, the MTPA maximizes efficiency at lower speeds, and the flux-weakening technique extends the operational speed range. This holistic approach addresses the full spectrum of performance requirements for an EV motor: high efficiency, wide speed range, rapid dynamic response, and strong robustness against disturbances.
To validate their theoretical framework, Kong and Luan conducted a series of comprehensive simulations using MATLAB/Simulink. The simulation model was built using realistic parameters for a typical EV IPMSM, including stator resistance, d-axis and q-axis inductances, permanent magnet flux linkage, and moment of inertia. The test scenario was designed to mimic challenging real-world driving conditions. The motor was initially commanded to reach its rated speed of 3,500 revolutions per minute (rpm). At 0.15 seconds into the simulation, the speed reference was abruptly increased to 5,500 rpm, pushing the motor into the flux-weakening region. Then, at 0.3 seconds, a significant load torque of 10 N·m was suddenly applied to simulate the vehicle climbing a hill or accelerating hard.
The simulation results were compelling. The motor equipped with the ADRC-based control strategy demonstrated a rapid and smooth response to the initial speed command, quickly reaching 3,500 rpm with minimal overshoot. When the speed reference was increased to 5,500 rpm, the motor transitioned into the flux-weakening region and achieved the new target speed efficiently. Crucially, the speed response was stable, with almost no overshoot, and the system quickly settled to the new operating point. When the 10 N·m load torque was applied, the speed deviation was minimal, and the controller restored the motor to its target speed rapidly, showcasing its excellent disturbance rejection capability.
A direct comparison with a conventional PI-controlled system highlighted the superiority of the ADRC approach. The PI controller exhibited a much larger speed overshoot during the initial acceleration and a significant dip in speed when the load was applied, taking longer to recover. The torque output from the PI-controlled motor also showed larger fluctuations, particularly during the transition into the flux-weakening region and after the load step. In contrast, the ADRC-controlled motor produced a much smoother torque profile, with smaller transient spikes and a quicker return to steady state. The d-axis current waveform, which is critical for flux weakening, was also more stable under ADRC, indicating better control over the demagnetizing process and reduced stress on the motor and inverter.
These results confirm that the proposed control strategy not only meets but exceeds the performance demands of modern electric vehicles. The enhanced stability and robustness mean a smoother, more comfortable driving experience, with less vibration and noise. The improved efficiency translates directly into longer driving ranges, a key selling point for EVs. The extended speed range allows for higher top speeds or more efficient cruising at highway velocities. Furthermore, the reduced sensitivity to parameter variations and external disturbances makes the control system more reliable and easier to implement across different motor units, which may have slight manufacturing tolerances.
The implications of this research extend beyond the laboratory. As automakers strive to differentiate their EV offerings in a competitive market, advancements in motor control represent a critical frontier. While much attention is given to battery technology, improvements in powertrain efficiency and performance through sophisticated control algorithms can yield significant gains. Kong and Luan’s work provides a practical and effective blueprint for next-generation motor controllers. The ADRC framework, with its model-independent nature, is particularly well-suited for mass production, where consistency and reliability are paramount.
Moreover, this research aligns with the broader trend in automotive engineering toward more intelligent and adaptive systems. The ability of ADRC to “learn” and compensate for disturbances in real time is a step toward truly intelligent powertrains that can adapt to varying road conditions, driver behavior, and vehicle loading. Future developments could integrate this control strategy with vehicle dynamics systems, battery management systems, and predictive driving algorithms to create a fully optimized, holistic energy management system.
In conclusion, the study by Kong Xiaoguang and Luan Zhaoyu from Shenyang University of Chemical Technology represents a significant leap forward in electric motor control technology. By combining the disturbance-rejection prowess of ADRC with the efficiency of MTPA and the simplicity of leading-angle flux weakening, they have created a control strategy that is not only theoretically sound but also demonstrably effective in simulation. This work provides a valuable contribution to the field of electric vehicle engineering, offering a clear path to more efficient, powerful, and reliable electric motors. As the world moves toward a sustainable transportation future, innovations like this will be essential in driving the performance and adoption of electric vehicles.
Kong Xiaoguang, Luan Zhaoyu, Journal of Dalian Polytechnic University, DOI:10.19670/j.cnki.dlgydxxb.2024.0313