Design and Control of Drives and Electrical Machines

1. Introduction

The field of design and control of drives and electrical machines has undergone significant transformation in recent years, driven by the accelerating demand for higher energy efficiency, greater power density, enhanced reliability, and seamless integration with renewable energy systems [1,2]. From permanent magnet synchronous machines (PMSMs) to reluctance machines and multiphase systems, electrical machines have become central to modern electrification in transportation, industrial automation, aerospace, and energy infrastructure [3]. In parallel, control systems are being integrated with advanced control strategies capable of handling multivariable interactions, operating constraints, and parametric uncertainties, while sensorless estimation and intelligent condition monitoring are becoming intrinsic functions of modern drives [4,5,6]. Although these developments have significantly expanded system performance boundaries, they have also exposed persistent research gaps: motors, converters, and controllers continue to be optimized independently and sequentially; control performance and computational resource constraints remain difficult to address simultaneously; and observers and diagnostic models may degrade significantly under parameter drift, dynamic loads, noise, and data domain shifts.

This Special Issue, “Design and Control of Drives and Electrical Machines,” brings together nine papers that address these challenges from complementary perspectives. The contributions cover predictive and adaptive control, sensorless estimation, nonlinear motion control, harmonic and torque ripple suppression, data-driven fault diagnosis, and application-oriented control system implementation. The research subjects considered include induction motors, PMSMs, permanent magnet-assisted synchronous reluctance motors (PMaSynRMs), hybrid stepper motors (HSMs), electro-hydrostatic actuators (EHAs), and precision rubidium fountain clock control platform. Taken together, these studies indicate that continued progress in electrical machines and drives depends on coordinated development of hardware implementation, control algorithms, fault diagnosis, and system-level integration.

This perspective is consistent with earlier research emphasizing that the design and control of drives and electrical machines must be understood from the viewpoint of practical implementation constraints, rather than solely as theoretical exercises. It also reflects the transition from isolated motor control toward integrated electromechanical systems capable of operating under uncertain and dynamic conditions. Building on these directions, the present Special Issue illustrates how drive and machine technologies can be developed across both algorithmic and implementation-level contexts, thereby contributing to bridging the gap between theoretical advances and deployable drive technologies.

2. Overview of the Contributions

The first group of contributions addresses predictive control under computational constraints and model uncertainty. Contribution 1 proposes a low switching frequency model predictive control (MPC) method for an induction motor fed by a three-level inverter. A partition optimization procedure reduces the number of candidate voltage vectors to at most three; a boundary-circle limit prevents unnecessary switching actions; and positive and redundant small vectors regulate the neutral point voltage deviation without introducing additional weighting coefficients into the cost function. The results show that the proposed method reduces both computational burden and switching frequency while maintaining satisfactory dynamic and steady-state performance, thereby addressing a longstanding difficulty in multi-objective MPC implementation. Contribution 9 implements velocity-form model predictive control for current regulation in synchronous motors on an STM32 microcontroller. The controller was executed in real time on a low-cost platform using a condensed quadratic programming (QP) formulation and the OSQP solver with a limited number of iterations. Experimental results show that the cost function configuration directly affects the trade-offs among transient response, steady-state error, control increments, and harmonic content. These two studies trace a coherent progression from candidate set reduction and switching frequency mitigation toward embedded optimization and experimental characterization. They also indicate that, following improvements in computational efficiency, systematic investigation is still required regarding recursive feasibility, parameter tuning, and applicability across different motor types and wider operating envelopes.

The second group of contributions focuses on disturbance rejection and robustness enhancement in motor drive systems. Contribution 2 proposes an adaptive model-based active disturbance rejection control (ADRC) with recursive parameter identification based on the recursive least-squares (RLS) method. Through stability analysis and DC motor experiments, it demonstrates that, when model information is reliable, incorporating it into the disturbance compensation framework can improve control performance, while also revealing the applicability boundaries of fixed-gain and purely model-free approaches under parameter uncertainties. Contribution 6 replaces the discontinuous signum function in sliding-mode control (SMC) for PMSM drives with continuous saturation, hyperbolic, and sigmoid switching functions. Simulation and experimental results show that the continuous functions can preserve the fast response of sliding mode control while eliminating noticeable chattering, with the saturation function offering a favorable balance between control performance and computational burden. Contribution 8 addresses torque ripple suppression in an EHA through harmonic injection. After analyzing current harmonics arising from inverter dead-time effects, voltage drops across power semiconductor devices, and machine manufacturing imperfections, the study employs multi-rotational proportional–integral (PI) control to suppress the 5th and 7th current harmonics. Simulation results confirm significant reductions in torque ripple amplitude and total harmonic distortion. Collectively, these three contributions show that high-performance drives should not indiscriminately represent all disturbances as a single lumped disturbance. Instead, broadband uncertainty rejection, nonlinear robustness, and frequency-selective compensation should be coordinated according to the characteristics of the disturbances involved.

The third group of contributions addresses application-specific control systems requiring high precision and reliability, thereby emphasizing complete control system design in which sensing, timing, stability, computing platforms, and application requirements must be coordinated. Contribution 3 develops a robust control platform for a rubidium fountain clock, using PXI hardware and LabWindows/CVI to integrate deterministic timing, arbitrary waveform generation, data acquisition, and servo locking. Although this application is outside the conventional scope of motor drives, its 380 ns signal synchronization, long-term closed-loop operation, and system-level frequency stability evaluation demonstrate that precision control depends not only on feedback algorithms, but also on hardware coordination and reliable data paths. Contribution 5 presents a position sensor-based vector-control scheme for HSMs, which generates the minimum phase current reference based on target torque and stator flux linkage state, employs a Lyapunov-based current controller with only one dominant tunable parameter to track currents directly in the stationary reference frame, and achieves position regulation through nonlinear torque modulation. Stability analysis, simulations, and experiments collectively validate the tracking performance improvements. The common value of these two studies lies in their system-oriented view of control: rigorous feedback design must be supported by deterministic implementation, rational reference generation, and application-oriented performance validation.

The fourth group of contributions focuses on sensorless control strategies for electrical machines. Contribution 4 proposes a hybrid active flux observer for PMaSynRM by combining voltage model and current model. The observer incorporates a PI controller and employs the current model as feedback compensation to correct the stator flux and compensate for back electromotive force errors. Simulation results across low, medium, and high speeds, as well as under loaded conditions, confirm that the proposed method achieves higher estimation accuracy and disturbance rejection capability compared to single-model observers. The key significance of this work lies in treating sensorless control as an information fusion problem, rather than seeking a single observer that dominates across all operating conditions. Its current limitations also reflect common gaps in the field: hybrid observers still require experimental verification at zero and low speeds, during fast torque transients, in the flux-weakening region, under parameter variations, and during sensorless restart. Moreover, online confidence metrics are also needed to determine when the estimated rotor position can be reliably used by the controller.

The fifth group of contributions extends the scope of the Special Issue from dynamic control to motor health management. Contribution 7 proposes a data-driven diagnostic system based on Extreme Gradient Boosting (XGBoost), which constructs an XGBoost model using phase currents, flux, and dq-axis quantities to detect, classify, and assess the severity of inter-turn short-circuit faults, stator open-circuit faults, and permanent magnet demagnetization in PMSMs. The results demonstrate high diagnostic accuracy and computational scalability, with several feature-importance trends consistent with the underlying fault mechanisms. However, the study also reveals a critical next step: the generalization capability needs further validation under dynamic loads, wider operating ranges, laboratory real-fault conditions, and online platforms.

3. Future Research Directions

Based on the contributions to this Special Issue and recent developments in the field, future research may focus on the following directions.

First, control and state estimation should evolve from fixed structures toward physics-informed fusion and operating condition aware adaptation. Fixed models struggle to maintain optimal performance under magnetic saturation, parameter drift, and dynamic load variations. Future approaches should integrate high-frequency signal injection, fundamental frequency models, and data-assisted estimation according to the operating condition and estimation confidence, thereby achieving stable operation over the entire speed range [4,7]. The core objective is not a simple superposition of multiple algorithms, but rather to endow the controller–observer system with the ability to assess information reliability and smoothly switch between model-driven high accuracy and robust estimation modes, while preserving interpretability and stability guarantees.

Second, disturbance rejection and robust control should be further developed toward hierarchical design and uncertainty quantification. Parameter drift, load transients, periodic harmonics, and other disturbances exhibit different time scales and propagation paths, making it difficult for a single controller to simultaneously achieve fast response, noise suppression, and steady-state accuracy. Future research should selectively integrate MPC, ADRC, SMC, and other control approaches, while achieving operating condition adaptivity through online identification and bandwidth scheduling. The key lies in clearly defining the disturbance scope for each control layer and coordinating interlayer stability and real-time implementation under limited computational resources.

Third, greater attention should be given to the interaction between electrical machine design and control. Achievable control performance is ultimately constrained by machine topology, material properties, and power electronic devices. Future studies should therefore adopt a co-design framework in which machine topology, materials, and control algorithms are optimized simultaneously. Additive manufacturing and advanced magnetic materials provide new opportunities for unconventional machine structures, but these emerging topologies will require corresponding innovations in control. This co-design perspective naturally extends to the integration of power electronics, where wide-bandgap devices enable higher switching frequencies and lower losses, while introducing opportunities and challenges for harmonic mitigation and electromagnetic compatibility.

Fourth, condition monitoring should evolve from fault classification toward a closed-loop framework for health aware and fault tolerant operation. Fault classification accuracy alone is insufficient to support industrial applications. Diagnostic models should identify incipient, compound, and cascading faults under varying operating conditions, quantify diagnostic confidence and fault severity, and provide physically interpretable decisions. The ultimate objective is to establish a diagnosis–prognosis–control closed loop in which fault type, remaining useful life, and other health indicators are used to reconfigure current references, thermal limits, modulation strategies, observers, and admissible torque–speed trajectories, thereby preventing secondary damage while maintaining the available system performance [6,8,9].

Finally, future research should place greater emphasis on validation under realistic deployment conditions. Beyond algorithmic performance, unified benchmarks and evaluation protocols should be established to account for temperature variations, component aging, communication delays, and computational constraints. Hardware in the loop, power hardware in the loop, and long duration experimental testing should be employed to verify system reliability and improve the reproducibility and engineering credibility of reported results. Such efforts are essential to ensure that research directions are driven by practical requirements and that promising innovations are systematically validated in relevant application environments.

4. Conclusions

This Special Issue, “Design and Control of Drives and Electrical Machines,” includes nine papers that examine how electrical machines and drive systems can be designed and controlled to meet the requirements of modern applications. The contributions cover low-complexity and embedded MPC, ADRC, sensorless control, chattering and torque ripple mitigation, system-oriented and stability-grounded control design, and data-driven multi-fault diagnosis.

Taken together, the contributions show that future progress in drives and electrical machines should not be judged by algorithmic sophistication alone. Control strategies must be computationally feasible for practical deployment; sensorless and fault-tolerant capabilities must remain robust to uncertainties encountered in real operating environments; machine design and control should be treated as integrated, rather than sequential, activities; and performance should be evaluated under conditions representative of actual deployment. Therefore, the evaluation of controllers should not be limited to rise time and steady-state error, nor should the evaluation of motors be limited to peak efficiency and torque density. The complete drive system must be evaluated in terms of performance, robustness, implementability, and sustainability under realistic operating conditions. This Special Issue presents drive and machine technologies not as isolated theoretical constructs, but as integrated systems that connect advanced control algorithms, reliable hardware implementation, fault diagnostics, and application-aware optimization.

We hope that this collection will serve as a valuable reference for researchers and engineers working on the design and control of drives and electrical machines. We also hope that it will stimulate further research on higher-performance machine designs and on more efficient, reliable, and deployable drive technologies.