Offline insulation testing of power cables is frequently performed using damped AC (DAC)-based partial discharge (PD) testing and cosine rectangular (CR)-based withstand testing. However, existing DAC and CR generators are typically implemented with different circuit topologies, requiring two separate test systems in the field, which results in low test efficiency and high cost. To address this issue, this article proposes a unified DAC–CR generator topology that can generate both DAC and CR voltages in the same test system, thereby satisfying the requirements of cable PD testing and withstand testing. Compared with the conventional DAC topology, the proposed topology only adds one high-voltage switch. Compared with the conventional CR topology, it requires only one unipolar high-voltage DC (HVDC) source to generate a bipolar CR voltage. To implement the proposed topology, two switches with a nominal blocking voltage of 30 kV were designed to include multioutput isolated power supplies, fiber-optic gate driving, insulated gate bipolar transistor (IGBT) voltage equalization, and protection circuits. A DAC–CR prototype is developed based on the proposed topology. Experimental results show that the prototype can generate DAC voltage and excite clear PD signals in a defective cross-linked polyethylene (XLPE) cable joint. Meanwhile, the prototype can generate CR voltages at 0.5 and 1 Hz. These results verify the feasibility of the proposed DAC–CR generator for cable PD testing and withstand testing, providing a novel engineering solution for offline assessment of cable insulation.
Dual three-phase surface permanent magnet synchronous machines (DT-SPMSMs) possess intrinsic fault-tolerance capabilities, making them well-suited for safety-critical applications. Consequently, their performance under faulty operating conditions is of significant interest in both academic and industrial domains. In recent years, numerous studies have examined the single open-phase fault (SOPF) to optimize steady-state performance with respect to torque ripple, peak current, and Joule losses. However, the complexity of the control system increases because active fault-tolerant strategies rely on accurate fault identification and/or software reconfigurations. As a result, a significant number of recent studies have focused on achieving satisfactory post-fault performance without diagnostic procedures or software reconfiguration. These approaches are referred to as passive fault-tolerant strategies. In line with the latest research in this field, this study aims to identify effective control solutions that enable passive fault tolerance in DT-SPMSMs, thereby enhancing drive performance without requiring fault detection or post-fault control (PFC) techniques. Furthermore, this article seeks to unify the recent literature on passive fault-tolerant strategies by proposing a comprehensive approach that considers all previously proposed PFC strategies. This methodology, for the first time, enables the reformulation of PFC strategies in analytical form, thereby proving the equivalence between active and passive strategies under steady-state conditions. The proposed passive fault-tolerant control (PFTC) strategies are validated through experimental testing on a laboratory prototype, and their performance is compared with that of active PFCs.
Control bandwidth is a critical performance metric for permanent magnet synchronous motor (PMSM) servo systems, yet it is fundamentally constrained by the performance trade-offs inherent in linear control strategies. This article proposes a reset-based active disturbance rejection control (RADRC) method to improve the bandwidth and transient performance of a surface-mounted PMSM (SPMSM) speed control system. By introducing a nonlinear reset feedback controller into the observer-based ADRC framework, the proposed RADRC achieves fast transient response with very small overshoot while preserving the disturbance-rejection capability of ADRC. Describing-function-based analysis shows that, compared with its base linear controller (BLC), RADRC increases the gain crossover frequency from 20 to 28 Hz and the phase margin from 30 $^{\boldsymbol{\circ}}$ to 66.6 $^{\boldsymbol{\circ}}$ . A hybrid Lyapunov analysis further establishes global asymptotic stability and excludes nontrivial limit cycles for the nominal continuous-time regulation model. Experiments on a 2-kW SPMSM platform demonstrate low-overshoot tracking, fast disturbance recovery, improved sensitivity characteristics, and modest computational cost. The results also identify increased current harmonics as a practical trade-off of the reset action.
Permanent magnet synchronous machines (PMSMs) generally exhibit nonlinear parameter variations across operating conditions, which are crucial knowledge for high-performance control. Accurate modeling of these nonlinearities needs extensive parameter calibration over wide operating ranges. To reduce this calibration burden, this article proposes an active learning (AL) framework that iteratively selects the informative operating conditions for parameter calibration under limited labeling budgets. First, the electromagnetic and loss equations of PMSMs are used to construct a prediction-residual estimator that quantifies the mismatch between the underlying physical constraints and the model predictions. Building upon this, a hybrid AL framework is developed by considering three complementary criteria: sample diversity, model predictive uncertainty, and physics-informed residual. In this way, the conditions with large residuals but low uncertainty can be identified, which might be overlooked by conventional hybrid AL strategies relying only on diversity and uncertainty. Experimental validation on a PMSM demonstrates that the proposed AL framework achieves the target prediction errors of loss and flux linkages, while reducing the required labeled-sample size by about 20% compared with traditional AL methods. These results indicate the effectiveness of the proposed framework for modeling PMSM nonlinear parameters with reduced calibration costs.
Spacecraft hovering involves strongly coupled nonlinear dynamics, external disturbances, and parameter uncertainties, which pose challenges for control design. To address these issues, this article derives a fully actuated six-degree-of-freedom (6-DoF) spacecraft model using the twistors formulation and designs a fuzzy adaptive control strategy based on the fully actuated system (FAS) approach. Specifically, the known nonlinear coupling terms are compensated directly, and a fuzzy logic system (FLS) is designed to mitigate the adverse effects of external disturbances and parameter uncertainties. The effectiveness of the proposed control strategy is demonstrated through 6-DoF numerical simulations, and its engineering feasibility is further verified on a 3-DoF air-bearing platform. The video of the experiments is available online at https://github.com/paper-video/Fuzzy.
