Mobile website
Issue 9,2026 Table of Contents
2026(9):873-882.
DOI: 10.12177/emca.2026.211
Abstract:
[Objective] Black-box optimization algorithms are widely used in various complex engineering design scenarios. However, their inherent computational complexity and internal operational opacity result in intensive computing consumption. In the optimal design of motors, black-box optimization algorithms require numerous iterative computations and finite element simulations to screen optimal structural parameters, which significantly extends the design cycle and restricts engineering application efficiency. Although traditional surrogate models can reduce computational costs, they exhibit poor adaptability and struggle to construct universal databases for multiple operating conditions. Slight changes in design requirements or application scenarios necessitate data recollection and model retraining, leading to inherent engineering drawbacks including weak generalization capability and low data reusability. [Methods] To solve the above problems, a black-box optimization algorithm based on transfer surrogate model (TSM) was proposed. Firstly, a novel TSM was constructed based on transfer learning theory. By autonomously screening high-quality and effective samples from the source domain, the inherent data features were deeply mined to improve the prediction accuracy and generalization ability of the model under the small-sample conditions of the target domain. Subsequently, the optimized TSM was deeply integrated with the iterative optimization algorithm to eliminate redundant computational steps during the optimization process and further enhance the execution efficiency of the optimization task. Finally, simulation tests were conducted to verify the practical effectiveness of the proposed algorithm. [Results] The simulation results showed that, compared with the traditional optimization algorithm, the proposed optimization algorithm based on TSM effectively reduced computational consumption, significantly improved the overall efficiency of motor parameter optimization, and accelerated the iterative convergence speed of the model. The proposed algorithm exhibited superior stability under variable motor design conditions. [Conclusion] The proposed optimization algorithm based on TSM addresses the inherent shortcomings of traditional black-box optimization and conventional surrogate models from the perspectives of data utilization and algorithm integration. It breaks through the limitations of poor scenario adaptability and high retraining costs existing in traditional models. This scheme can efficiently adapt to the variable design requirements of practical engineering scenarios. It provides reliable and efficient technical support for the intelligent optimal design of motors and other complex engineering black-box optimization problems, and possesses excellent engineering applicability and popularization prospects.
[Objective] Black-box optimization algorithms are widely used in various complex engineering design scenarios. However, their inherent computational complexity and internal operational opacity result in intensive computing consumption. In the optimal design of motors, black-box optimization algorithms require numerous iterative computations and finite element simulations to screen optimal structural parameters, which significantly extends the design cycle and restricts engineering application efficiency. Although traditional surrogate models can reduce computational costs, they exhibit poor adaptability and struggle to construct universal databases for multiple operating conditions. Slight changes in design requirements or application scenarios necessitate data recollection and model retraining, leading to inherent engineering drawbacks including weak generalization capability and low data reusability. [Methods] To solve the above problems, a black-box optimization algorithm based on transfer surrogate model (TSM) was proposed. Firstly, a novel TSM was constructed based on transfer learning theory. By autonomously screening high-quality and effective samples from the source domain, the inherent data features were deeply mined to improve the prediction accuracy and generalization ability of the model under the small-sample conditions of the target domain. Subsequently, the optimized TSM was deeply integrated with the iterative optimization algorithm to eliminate redundant computational steps during the optimization process and further enhance the execution efficiency of the optimization task. Finally, simulation tests were conducted to verify the practical effectiveness of the proposed algorithm. [Results] The simulation results showed that, compared with the traditional optimization algorithm, the proposed optimization algorithm based on TSM effectively reduced computational consumption, significantly improved the overall efficiency of motor parameter optimization, and accelerated the iterative convergence speed of the model. The proposed algorithm exhibited superior stability under variable motor design conditions. [Conclusion] The proposed optimization algorithm based on TSM addresses the inherent shortcomings of traditional black-box optimization and conventional surrogate models from the perspectives of data utilization and algorithm integration. It breaks through the limitations of poor scenario adaptability and high retraining costs existing in traditional models. This scheme can efficiently adapt to the variable design requirements of practical engineering scenarios. It provides reliable and efficient technical support for the intelligent optimal design of motors and other complex engineering black-box optimization problems, and possesses excellent engineering applicability and popularization prospects.
2026(9):883-894.
DOI: 10.12177/emca.2026.215
Abstract:
[Objective] In the space vector modulation direct torque control (SVM-DTC) of brushless doubly-fed machine (BDFM), inner-loop torque and flux proportional-integral (PI) controllers are commonly used. However, fixed PI parameters are difficult to adapt to the system’s nonlinearity and frequently varying operating conditions, making it challenging to maintain optimal performance across all operating points. To address this problem, a deep reinforcement learning (DRL)-based SVM-DTC strategy is proposed to eliminate the dependence on manual parameter tuning and improve the system’s dynamic performance and robustness. [Methods] The twin delayed deep deterministic policy gradient (TD3) algorithm was adopted to construct and train an Actor-Critic neural network, which replaced the inner-loop torque and flux PI controllers. The state space consisted of estimated torque, estimated flux, torque error, flux error, integral of torque error, integral of flux error, speed reference, and speed feedback. The action space was defined as the d-q axis components of the control winding reference voltage vector. Meanwhile, a reward function was designed to integrate torque ripple suppression, and flux deviation penalty, guiding the TD3-agent to learn optimal voltage decisions in a continuous action space. [Results] Simulations were carried out under typical operating conditions such as speed step change and sudden load variation to compare the performance of the proposed DRL-based SVM-DTC strategy and the traditional SVM-DTC strategy. The simulation results showed that the proposed method achieved basically consistent speed control performance with the traditional SVM-DTC strategy. [Conclusion] The well-trained Actor network can directly generate the control winding reference voltages according to the system’s real-time states, obviating the parameter tuning process of the inner-loop torque and flux PI controllers. The TD3-based SVM-DTC strategy can achieve high-quality control of torque and flux for BDFM, exhibiting excellent dynamic response and robustness.
[Objective] In the space vector modulation direct torque control (SVM-DTC) of brushless doubly-fed machine (BDFM), inner-loop torque and flux proportional-integral (PI) controllers are commonly used. However, fixed PI parameters are difficult to adapt to the system’s nonlinearity and frequently varying operating conditions, making it challenging to maintain optimal performance across all operating points. To address this problem, a deep reinforcement learning (DRL)-based SVM-DTC strategy is proposed to eliminate the dependence on manual parameter tuning and improve the system’s dynamic performance and robustness. [Methods] The twin delayed deep deterministic policy gradient (TD3) algorithm was adopted to construct and train an Actor-Critic neural network, which replaced the inner-loop torque and flux PI controllers. The state space consisted of estimated torque, estimated flux, torque error, flux error, integral of torque error, integral of flux error, speed reference, and speed feedback. The action space was defined as the d-q axis components of the control winding reference voltage vector. Meanwhile, a reward function was designed to integrate torque ripple suppression, and flux deviation penalty, guiding the TD3-agent to learn optimal voltage decisions in a continuous action space. [Results] Simulations were carried out under typical operating conditions such as speed step change and sudden load variation to compare the performance of the proposed DRL-based SVM-DTC strategy and the traditional SVM-DTC strategy. The simulation results showed that the proposed method achieved basically consistent speed control performance with the traditional SVM-DTC strategy. [Conclusion] The well-trained Actor network can directly generate the control winding reference voltages according to the system’s real-time states, obviating the parameter tuning process of the inner-loop torque and flux PI controllers. The TD3-based SVM-DTC strategy can achieve high-quality control of torque and flux for BDFM, exhibiting excellent dynamic response and robustness.
2026(9):895-903.
DOI: 10.12177/emca.2026.216
Abstract:
[Objective] Wind resource parameters such as wind speed, ambient temperature and humidity, exhibit prominent time-varying characteristics, and their dynamic fluctuations directly affect the output power and temperature rise of wind generators. Addressing this issue, this paper investigates the operating characteristics of permanent magnet wind generator (PMWG) under varying load conditions, aiming to realize fast and accurate monitoring of generator temperature state. [Methods] A 6 MW high power PMWG was taken as the research object. The finite element method was employed to construct the electromagnetic simulation model, and key loss parameters including winding loss, core iron loss and permanent magnet eddy current loss were obtained. Afterwards, thermal network model for the PMWG was established using the thermal network method. Considering the stochastic variation characteristics of wind load during wind turbines operation, constant load and varying load conditions were designed, and the temperature rise characteristics under corresponding conditions were investigated. Finally, the relevant operating conditions were validated using factory type test data and wind farm field test data of the PMWG. [Results] The comparison between the factory type test data of the PMWG and the simulation results showed that the average relative error of each sampling point in the motor temperature rise process under rated load operating condition was 5%, which verified the reliability and accuracy of the thermal network model for the PMWG. The comparison between the measured wind farm operation data and the simulation results indicated that the percentage of average temperature difference under varying load condition was 4.5%. [Conclusion] This study reveals the temperature rise characteristics of high power PMWG under varying load conditions, and provides a reference for temperature rise prediction, early warning of each generator component as well as the setting of control parameters for thermal management systems of PMWG operating in wind farm scenarios.
[Objective] Wind resource parameters such as wind speed, ambient temperature and humidity, exhibit prominent time-varying characteristics, and their dynamic fluctuations directly affect the output power and temperature rise of wind generators. Addressing this issue, this paper investigates the operating characteristics of permanent magnet wind generator (PMWG) under varying load conditions, aiming to realize fast and accurate monitoring of generator temperature state. [Methods] A 6 MW high power PMWG was taken as the research object. The finite element method was employed to construct the electromagnetic simulation model, and key loss parameters including winding loss, core iron loss and permanent magnet eddy current loss were obtained. Afterwards, thermal network model for the PMWG was established using the thermal network method. Considering the stochastic variation characteristics of wind load during wind turbines operation, constant load and varying load conditions were designed, and the temperature rise characteristics under corresponding conditions were investigated. Finally, the relevant operating conditions were validated using factory type test data and wind farm field test data of the PMWG. [Results] The comparison between the factory type test data of the PMWG and the simulation results showed that the average relative error of each sampling point in the motor temperature rise process under rated load operating condition was 5%, which verified the reliability and accuracy of the thermal network model for the PMWG. The comparison between the measured wind farm operation data and the simulation results indicated that the percentage of average temperature difference under varying load condition was 4.5%. [Conclusion] This study reveals the temperature rise characteristics of high power PMWG under varying load conditions, and provides a reference for temperature rise prediction, early warning of each generator component as well as the setting of control parameters for thermal management systems of PMWG operating in wind farm scenarios.
2026(9):904-915.
DOI: 10.12177/emca.2026.220
Abstract:
[Objective] To address the issue of large cogging torque, low motor efficiency and high torque ripple commonly found in canned permanent magnet synchronous motor (CPMSM), this paper aims to analyze the influence of asymmetric permanent magnet structure on the performance of CPMSM. [Methods] Based on the arc-shaped permanent magnet commonly used in engineering, this paper adopted the geometric centerline of a single permanent magnet as the reference. The unilateral eccentric design of the permanent magnet was realized by offsetting the centers of its inner and outer diameters, and three asymmetric permanent magnet structures were proposed. The key motor performances of CPMSM with different permanent magnet structures, including air-gap flux density, induced electromotive force, unilateral magnetic pull of the can sleeve, demagnetization coefficient of permanent magnets, motor losses and operating efficiency, were systematically compared and analyzed under no-load and load conditions. [Results] The results showed that the three asymmetric permanent magnet structures shared basically the same performance variation trend. As the offset of the permanent magnet center increased, the amplitude of air-gap flux density decreased. The amplitudes of cogging torque and can loss decreased accordingly, and the motor efficiency was improved. Without taking manufacturing difficulty and cost into account, the left-side asymmetric structure was the optimal scheme. The motor with this structure obtained the highest efficiency and power factor, the smallest drop in output torque, and the most prominent suppression effect on torque ripple. In addition, the unbalanced magnetic pull caused by the asymmetric permanent magnet exerted a negligible effect on the overall performance of the motor. Compared with the conventional optimization scheme of stator-tooth slotting, the proposed method suppressed cogging torque while maintaining favorable electromagnetic performance, and exhibited remarkable advantages. [Conclusion] The asymmetric permanent magnet structure proposed in this paper achieve remarkable performance in weakening cogging torque, suppressing torque ripple and improving the comprehensive performance of CPMSM. Although this structure brings a certain degree of magnetic-circuit unbalance, it does not alter the overall magnetic-circuit distribution characteristics of the motor, and the motor still maintain excellent electromagnetic performance. This study provides references for the optimal design of CPMSM.
[Objective] To address the issue of large cogging torque, low motor efficiency and high torque ripple commonly found in canned permanent magnet synchronous motor (CPMSM), this paper aims to analyze the influence of asymmetric permanent magnet structure on the performance of CPMSM. [Methods] Based on the arc-shaped permanent magnet commonly used in engineering, this paper adopted the geometric centerline of a single permanent magnet as the reference. The unilateral eccentric design of the permanent magnet was realized by offsetting the centers of its inner and outer diameters, and three asymmetric permanent magnet structures were proposed. The key motor performances of CPMSM with different permanent magnet structures, including air-gap flux density, induced electromotive force, unilateral magnetic pull of the can sleeve, demagnetization coefficient of permanent magnets, motor losses and operating efficiency, were systematically compared and analyzed under no-load and load conditions. [Results] The results showed that the three asymmetric permanent magnet structures shared basically the same performance variation trend. As the offset of the permanent magnet center increased, the amplitude of air-gap flux density decreased. The amplitudes of cogging torque and can loss decreased accordingly, and the motor efficiency was improved. Without taking manufacturing difficulty and cost into account, the left-side asymmetric structure was the optimal scheme. The motor with this structure obtained the highest efficiency and power factor, the smallest drop in output torque, and the most prominent suppression effect on torque ripple. In addition, the unbalanced magnetic pull caused by the asymmetric permanent magnet exerted a negligible effect on the overall performance of the motor. Compared with the conventional optimization scheme of stator-tooth slotting, the proposed method suppressed cogging torque while maintaining favorable electromagnetic performance, and exhibited remarkable advantages. [Conclusion] The asymmetric permanent magnet structure proposed in this paper achieve remarkable performance in weakening cogging torque, suppressing torque ripple and improving the comprehensive performance of CPMSM. Although this structure brings a certain degree of magnetic-circuit unbalance, it does not alter the overall magnetic-circuit distribution characteristics of the motor, and the motor still maintain excellent electromagnetic performance. This study provides references for the optimal design of CPMSM.
2026(9):916-927.
DOI: 10.12177/emca.2026.209
Abstract:
[Objective] This paper takes a 1.5 kW canned permanent magnet synchronous motor (CPMSM) as the research prototype to reveal the influence laws of can materials on the demagnetization characteristics and electromagnetic performance of CPMSM under three-phase short-circuit fault conditions. [Methods] First, the field-circuit coupling calculation method was adopted to comprehensively compare and analyze the key electromagnetic performance indices of CPMSM with SUS430, SUS316, and SUS304 can materials, including three-phase short-circuit current, permanent magnet demagnetization coefficient, air-gap flux density, induced electromotive force, and output torque. Subsequently, the anti-demagnetization mechanism of the motor was revealed by combining the no-load leakage flux coefficient and the operating point of permanent magnets. Finally, a prototype motor was manufactured, and experimental tests were conducted to verify the accuracy and reliability of the established finite element simulation model. [Results] Under three-phase short-circuit fault conditions, the CPMSM with SUS430 can exhibited significantly lower peak d-axis and q-axis currents and extremely slight demagnetization of permanent magnets. Its air-gap flux density, stator flux density, back electromotive force, and output torque remained basically stable before and after the short-circuit fault, with only a slight increase in torque ripple. In contrast, CPMSMs adopting SUS304 and SUS316 cans suffered severe permanent magnet demagnetization. Specifically, the back electromotive force dropped by more than 64%, the average output torque decreased by 54.6%, and the torque ripple was drastically aggravated after the short-circuit fault. [Conclusion] The magnetic SUS430 can effectively improves the anti-demagnetization capability of the CPMSM during three-phase short-circuit faults. The can permeability is identified as the core parameter dominating the demagnetization characteristics of the CPMSM. The research results of this paper can provide a reliable reference for can material selection and anti-demagnetization design of CPMSMs applied in vacuum pumps and other relevant fields.
[Objective] This paper takes a 1.5 kW canned permanent magnet synchronous motor (CPMSM) as the research prototype to reveal the influence laws of can materials on the demagnetization characteristics and electromagnetic performance of CPMSM under three-phase short-circuit fault conditions. [Methods] First, the field-circuit coupling calculation method was adopted to comprehensively compare and analyze the key electromagnetic performance indices of CPMSM with SUS430, SUS316, and SUS304 can materials, including three-phase short-circuit current, permanent magnet demagnetization coefficient, air-gap flux density, induced electromotive force, and output torque. Subsequently, the anti-demagnetization mechanism of the motor was revealed by combining the no-load leakage flux coefficient and the operating point of permanent magnets. Finally, a prototype motor was manufactured, and experimental tests were conducted to verify the accuracy and reliability of the established finite element simulation model. [Results] Under three-phase short-circuit fault conditions, the CPMSM with SUS430 can exhibited significantly lower peak d-axis and q-axis currents and extremely slight demagnetization of permanent magnets. Its air-gap flux density, stator flux density, back electromotive force, and output torque remained basically stable before and after the short-circuit fault, with only a slight increase in torque ripple. In contrast, CPMSMs adopting SUS304 and SUS316 cans suffered severe permanent magnet demagnetization. Specifically, the back electromotive force dropped by more than 64%, the average output torque decreased by 54.6%, and the torque ripple was drastically aggravated after the short-circuit fault. [Conclusion] The magnetic SUS430 can effectively improves the anti-demagnetization capability of the CPMSM during three-phase short-circuit faults. The can permeability is identified as the core parameter dominating the demagnetization characteristics of the CPMSM. The research results of this paper can provide a reliable reference for can material selection and anti-demagnetization design of CPMSMs applied in vacuum pumps and other relevant fields.
2026(9):928-937.
DOI: 10.12177/emca.2026.219
Abstract:
[Objective] To address issues of experience dependence, low efficiency, and possible omission of feasible solutions in motor winding layout design under complex slot-pole combinations, multiphase configurations, and different winding forms, a search-and-pruning-based automatic winding layout method is proposed. The method provides an automated solution framework for the unified generation and screening of concentrated windings, distributed windings, and multiphase windings. [Methods] The winding layout problem was formulated as a constrained coil-side combination search problem. A backtracking-based depth-first search framework was constructed to generate candidate winding layouts through coil-side pairing. Bit-mask representation was used to record slot occupancy, enabling fast conflict detection and feasibility checking. The synthesized fundamental and harmonic phasors were updated through incremental complex summation, which eliminated redundant full-range summation calculations. To narrow down the actual search space, symmetry reduction, a triangle-inequality reachability upper bound, and a projection upper bound along the current synthesized phasor direction were adopted to prune invalid branches that fail to satisfy the target winding factor. [Results] Representative three-phase and five-phase case studies showed that the proposed method can automatically generate feasible winding layout schemes with high fundamental winding factors under different slot-pole combinations. Compared with combinational enumeration method, the proposed method significantly reduced the number of visited search nodes and computation time, with pruning rates exceeding 95% in typical cases. Finite element simulation of a 40-slot 26-pole five-phase permanent magnet synchronous motor showed that the average torque of the winding layout obtained by the proposed method was 14.570 N·m, slightly higher than the 14.547 N·m obtained by the star diagram method. Its torque ripple of 0.067 4% was lower than 0.075 6% obtained by the star diagram method, and the total harmonic distortion of the air-gap flux density waveform reached 9.95%, also lower than the 11.18% given by the star diagram method. [Conclusion] The proposed method effectively reduces the practical search space while ensuring the completeness of feasible solution exploration, thereby yielding winding configurations with a high fundamental winding factor and low low-order harmonic content. It is applicable to rapid design and scheme comparison for complex slot-pole combinations, multiphase configurations and diverse winding forms, providing a reliable front-end design tool for subsequent electromagnetic optimization, harmonic suppression, and torque ripple mitigation.
[Objective] To address issues of experience dependence, low efficiency, and possible omission of feasible solutions in motor winding layout design under complex slot-pole combinations, multiphase configurations, and different winding forms, a search-and-pruning-based automatic winding layout method is proposed. The method provides an automated solution framework for the unified generation and screening of concentrated windings, distributed windings, and multiphase windings. [Methods] The winding layout problem was formulated as a constrained coil-side combination search problem. A backtracking-based depth-first search framework was constructed to generate candidate winding layouts through coil-side pairing. Bit-mask representation was used to record slot occupancy, enabling fast conflict detection and feasibility checking. The synthesized fundamental and harmonic phasors were updated through incremental complex summation, which eliminated redundant full-range summation calculations. To narrow down the actual search space, symmetry reduction, a triangle-inequality reachability upper bound, and a projection upper bound along the current synthesized phasor direction were adopted to prune invalid branches that fail to satisfy the target winding factor. [Results] Representative three-phase and five-phase case studies showed that the proposed method can automatically generate feasible winding layout schemes with high fundamental winding factors under different slot-pole combinations. Compared with combinational enumeration method, the proposed method significantly reduced the number of visited search nodes and computation time, with pruning rates exceeding 95% in typical cases. Finite element simulation of a 40-slot 26-pole five-phase permanent magnet synchronous motor showed that the average torque of the winding layout obtained by the proposed method was 14.570 N·m, slightly higher than the 14.547 N·m obtained by the star diagram method. Its torque ripple of 0.067 4% was lower than 0.075 6% obtained by the star diagram method, and the total harmonic distortion of the air-gap flux density waveform reached 9.95%, also lower than the 11.18% given by the star diagram method. [Conclusion] The proposed method effectively reduces the practical search space while ensuring the completeness of feasible solution exploration, thereby yielding winding configurations with a high fundamental winding factor and low low-order harmonic content. It is applicable to rapid design and scheme comparison for complex slot-pole combinations, multiphase configurations and diverse winding forms, providing a reliable front-end design tool for subsequent electromagnetic optimization, harmonic suppression, and torque ripple mitigation.
2026(9):938-946.
DOI: 10.12177/emca.2026.213
Abstract:
[Objective] The maglev system of the six-phase hybrid excitation flux switching linear motor (SHEFSLM) is nonlinear, strongly coupled and parameter-varying. Although the modular structure adopted by the SHEFSLM can effectively mitigate the magnetic circuit imbalance, external disturbances together with the inherent end effects of linear motors still bring severe challenges to the control of maglev system. To tackle these control performance problems, a variable exponential sliding mode control (VESMC) strategy is proposed for SHEFSLM maglev system. [Methods] Firstly, the electromagnetic thrust equation, levitation force equation and system state equation were derived based on the flux linkage equation, voltage equation and mechanical motion equation of the SHEFSLM maglev system. Secondly, a sliding mode surface was constructed, and a novel variable exponential reaching law was proposed. This reaching law was capable of adaptively adjusting the convergence rate, which enabled rapid convergence when the operating point was far away from the sliding surface and smooth convergence in the vicinity of the sliding surface. Hence, superior system stability was guaranteed and chattering was effectively suppressed. Subsequently, the stability of the closed-loop system was verified via the Lyapunov function, and the convergence time was determined by the controller parameters. Finally, simulation comparisons among the proposed VESMC, sliding mode control (SMC) and proportional-integral (PI) control were carried out. [Results] Simulation results indicated that compared with SMC and PI control, the VESMC proposed in this paper reduced the settling time of no-load startup by 68% and 73.3%, respectively. Under sudden step disturbances, the recovery time was shortened by 68% and 85.5%, respectively. In suppressing end effects, VESMC exhibited better performance than the other two control strategies. [Conclusion] In the presence of disturbances, VESMC exhibits stronger robustness, effectively attenuates the chattering inherent in sliding mode control, and improves the stability and dynamic performance of the system. It offers the advantages of small steady-state error, short settling time, and short recovery time, thereby effectively enhancing the control performance of the maglev system and meeting the control requirements.
[Objective] The maglev system of the six-phase hybrid excitation flux switching linear motor (SHEFSLM) is nonlinear, strongly coupled and parameter-varying. Although the modular structure adopted by the SHEFSLM can effectively mitigate the magnetic circuit imbalance, external disturbances together with the inherent end effects of linear motors still bring severe challenges to the control of maglev system. To tackle these control performance problems, a variable exponential sliding mode control (VESMC) strategy is proposed for SHEFSLM maglev system. [Methods] Firstly, the electromagnetic thrust equation, levitation force equation and system state equation were derived based on the flux linkage equation, voltage equation and mechanical motion equation of the SHEFSLM maglev system. Secondly, a sliding mode surface was constructed, and a novel variable exponential reaching law was proposed. This reaching law was capable of adaptively adjusting the convergence rate, which enabled rapid convergence when the operating point was far away from the sliding surface and smooth convergence in the vicinity of the sliding surface. Hence, superior system stability was guaranteed and chattering was effectively suppressed. Subsequently, the stability of the closed-loop system was verified via the Lyapunov function, and the convergence time was determined by the controller parameters. Finally, simulation comparisons among the proposed VESMC, sliding mode control (SMC) and proportional-integral (PI) control were carried out. [Results] Simulation results indicated that compared with SMC and PI control, the VESMC proposed in this paper reduced the settling time of no-load startup by 68% and 73.3%, respectively. Under sudden step disturbances, the recovery time was shortened by 68% and 85.5%, respectively. In suppressing end effects, VESMC exhibited better performance than the other two control strategies. [Conclusion] In the presence of disturbances, VESMC exhibits stronger robustness, effectively attenuates the chattering inherent in sliding mode control, and improves the stability and dynamic performance of the system. It offers the advantages of small steady-state error, short settling time, and short recovery time, thereby effectively enhancing the control performance of the maglev system and meeting the control requirements.
2026(9):947-962.
DOI: 10.12177/emca.2026.212
Abstract:
[Objective] Among existing modulation strategies, sinusoidal pulse width modulation (SPWM), third harmonic injection pulse width modulation (THIPWM), and space vector pulse width modulation (SVPWM) can realize basic voltage control, but their common-mode voltage (CMV) suppression capabilities are limited, along with issues such as constrained output voltage peak and high switching losses. Although SVPWM further enhances voltage utilization by optimizing switching sequences, its continuous modulation characteristics still results in increased CMV and voltage stress. While CMV can be suppressed by adjusting the distribution of zero vectors, nevertheless, it remains a pressing challenge to reduce voltage stress while maintaining the voltage-boosting capability. [Methods] To address the aforementioned issues, this paper proposed a discontinuous SVPWM (DSVPWM) strategy based on zero sequence signal injection on the basis of the conventional SVPWM method. The proposed strategy reconstructed the switching sequences by eliminating one zero vector state during the charging mode and featured a variable duty ratio. The DSVPWM strategy was applied to the inverter stage of the split-source matrix converter (SSMC). Meanwhile, the rectifier stage of the SSMC adopted the conventional SVPWM strategy to achieve the maximum voltage utilization. [Results] Simulation analyses were performed on the Matlab/Simulink platform. The simulation results demonstrated that the THIPWM strategy produced the highest CMV with a peak value of 533.8 V. The SPWM and SVPWM strategies delivered similar CMV, with peak values of 532.9 V and 532.4 V, respectively. In contrast, the proposed DSVPWM strategy effectively suppressed CMV, reducing its peak value to 289.7 V. In terms of output current quality, the SPWM strategy had the maximum total harmonic distortion (THD) of output current at 4.38%, while the proposed DSVPWM strategy achieved the lowest THD of only 0.44%. The experimental waveforms obtained from the hardware platform were highly consistent with the simulation results, which further verified the effectiveness of the proposed strategy. [Conclusion] Compared with the SPWM, THIPWM and SVPWM strategies, the SSMC CMV can be better suppressed under the DSVPWM strategy proposed in this paper, with lower THD of output current and smaller voltage stress, yet a slight drop in voltage gain.
[Objective] Among existing modulation strategies, sinusoidal pulse width modulation (SPWM), third harmonic injection pulse width modulation (THIPWM), and space vector pulse width modulation (SVPWM) can realize basic voltage control, but their common-mode voltage (CMV) suppression capabilities are limited, along with issues such as constrained output voltage peak and high switching losses. Although SVPWM further enhances voltage utilization by optimizing switching sequences, its continuous modulation characteristics still results in increased CMV and voltage stress. While CMV can be suppressed by adjusting the distribution of zero vectors, nevertheless, it remains a pressing challenge to reduce voltage stress while maintaining the voltage-boosting capability. [Methods] To address the aforementioned issues, this paper proposed a discontinuous SVPWM (DSVPWM) strategy based on zero sequence signal injection on the basis of the conventional SVPWM method. The proposed strategy reconstructed the switching sequences by eliminating one zero vector state during the charging mode and featured a variable duty ratio. The DSVPWM strategy was applied to the inverter stage of the split-source matrix converter (SSMC). Meanwhile, the rectifier stage of the SSMC adopted the conventional SVPWM strategy to achieve the maximum voltage utilization. [Results] Simulation analyses were performed on the Matlab/Simulink platform. The simulation results demonstrated that the THIPWM strategy produced the highest CMV with a peak value of 533.8 V. The SPWM and SVPWM strategies delivered similar CMV, with peak values of 532.9 V and 532.4 V, respectively. In contrast, the proposed DSVPWM strategy effectively suppressed CMV, reducing its peak value to 289.7 V. In terms of output current quality, the SPWM strategy had the maximum total harmonic distortion (THD) of output current at 4.38%, while the proposed DSVPWM strategy achieved the lowest THD of only 0.44%. The experimental waveforms obtained from the hardware platform were highly consistent with the simulation results, which further verified the effectiveness of the proposed strategy. [Conclusion] Compared with the SPWM, THIPWM and SVPWM strategies, the SSMC CMV can be better suppressed under the DSVPWM strategy proposed in this paper, with lower THD of output current and smaller voltage stress, yet a slight drop in voltage gain.
2026(9):963-981.
DOI: 10.12177/emca.2026.217
Abstract:
[Objective] To address the problems that photovoltaic inverter models provided by manufacturers are mostly delivered as black-box dynamic libraries with inaccessible internal control structures, and that detailed electromagnetic transient (EMT) models for large-scale photovoltaic power stations suffer from large model size and low simulation efficiency, this paper proposes an EMT modeling method for photovoltaic units based on test data of manufacturer-supplied black-box models, as well as an equivalent modeling method for photovoltaic power stations following the principle of consistent collection-line losses. It provides a model foundation for the transient characteristic analysis of grid-connected photovoltaic power stations and large-grid simulation calculations. [Methods] First, the manufacturer black-box model was connected to the simulation test system, and fault ride-through tests were carried out under different active power output levels and voltage disturbance amplitudes. The response data including terminal voltage, current, active power, reactive power, d- and q-axis currents, were collected to analyze the dynamic variation laws during fault occurrence, fault duration and fault recovery. Then, the parameter set to be identified for the PSModel inverter model was constructed, including main circuit parameters, normal-operation control parameters, fault ride-through thresholds, current limiting parameters, reactive current support coefficients and time criteria. A parameter tuning process of “feature extraction—staged identification—simulation iterative correction” was adopted, and the normalized error and comprehensive error objective function were introduced to evaluate the deviation between the PSModel model and the manufacturer black-box model. After the unit-level model verification was completed, a detailed EMT model of a 100 MW photovoltaic power station in Anhui province was established based on the PSModel platform independently developed by China Electric Power Research Institute. Furthermore, according to the actual collector network topology, the collector-line structures, including branch-trunk, trunk and radial structures, were equivalently transformed under the principle of active power loss consistency, and the photovoltaic power station equivalent model was formed. Finally, considering the many combinations of load rates, fault types and voltage disturbance amplitudes, representative operating conditions were selected for comparative verification between the detailed model and the equivalent model. Meanwhile, to avoid the limitation of evaluating the equivalent effect only by active power curves, the voltage, current, active power, reactive power and reactive current at the 35 kV bus were further selected as evaluation variables. The sectional average deviation, maximum deviation and weighted average total deviation were used for quantitative analysis. [Results] The established photovoltaic unit model effectively reproduced the active and reactive power dynamic response characteristics of different manufacturer models during fault ride-through. The detailed model and equivalent model of the photovoltaic power station showed basically consistent active power variation trends at the 35 kV bus during fault occurrence, fault duration and fault recovery. The multi-condition error statistics showed that, under the selected symmetrical and asymmetrical disturbance scenarios, the sectional errors and weighted average total deviations of voltage, current, active power, reactive power and reactive current were all lower than the allowable limits specified in GB/T32892—2026. [Conclusion] The proposed method can construct a photovoltaic unit EMT model suitable for engineering simulation when manufacturer control details cannot be fully disclosed. It can also reduce the scale of the photovoltaic power station model and improve simulation efficiency while maintaining the dynamic port characteristics at the 35 kV bus, providing an effective modeling approach for subsequent analysis of the grid-connection security and stability influence of photovoltaic power stations.
[Objective] To address the problems that photovoltaic inverter models provided by manufacturers are mostly delivered as black-box dynamic libraries with inaccessible internal control structures, and that detailed electromagnetic transient (EMT) models for large-scale photovoltaic power stations suffer from large model size and low simulation efficiency, this paper proposes an EMT modeling method for photovoltaic units based on test data of manufacturer-supplied black-box models, as well as an equivalent modeling method for photovoltaic power stations following the principle of consistent collection-line losses. It provides a model foundation for the transient characteristic analysis of grid-connected photovoltaic power stations and large-grid simulation calculations. [Methods] First, the manufacturer black-box model was connected to the simulation test system, and fault ride-through tests were carried out under different active power output levels and voltage disturbance amplitudes. The response data including terminal voltage, current, active power, reactive power, d- and q-axis currents, were collected to analyze the dynamic variation laws during fault occurrence, fault duration and fault recovery. Then, the parameter set to be identified for the PSModel inverter model was constructed, including main circuit parameters, normal-operation control parameters, fault ride-through thresholds, current limiting parameters, reactive current support coefficients and time criteria. A parameter tuning process of “feature extraction—staged identification—simulation iterative correction” was adopted, and the normalized error and comprehensive error objective function were introduced to evaluate the deviation between the PSModel model and the manufacturer black-box model. After the unit-level model verification was completed, a detailed EMT model of a 100 MW photovoltaic power station in Anhui province was established based on the PSModel platform independently developed by China Electric Power Research Institute. Furthermore, according to the actual collector network topology, the collector-line structures, including branch-trunk, trunk and radial structures, were equivalently transformed under the principle of active power loss consistency, and the photovoltaic power station equivalent model was formed. Finally, considering the many combinations of load rates, fault types and voltage disturbance amplitudes, representative operating conditions were selected for comparative verification between the detailed model and the equivalent model. Meanwhile, to avoid the limitation of evaluating the equivalent effect only by active power curves, the voltage, current, active power, reactive power and reactive current at the 35 kV bus were further selected as evaluation variables. The sectional average deviation, maximum deviation and weighted average total deviation were used for quantitative analysis. [Results] The established photovoltaic unit model effectively reproduced the active and reactive power dynamic response characteristics of different manufacturer models during fault ride-through. The detailed model and equivalent model of the photovoltaic power station showed basically consistent active power variation trends at the 35 kV bus during fault occurrence, fault duration and fault recovery. The multi-condition error statistics showed that, under the selected symmetrical and asymmetrical disturbance scenarios, the sectional errors and weighted average total deviations of voltage, current, active power, reactive power and reactive current were all lower than the allowable limits specified in GB/T32892—2026. [Conclusion] The proposed method can construct a photovoltaic unit EMT model suitable for engineering simulation when manufacturer control details cannot be fully disclosed. It can also reduce the scale of the photovoltaic power station model and improve simulation efficiency while maintaining the dynamic port characteristics at the 35 kV bus, providing an effective modeling approach for subsequent analysis of the grid-connection security and stability influence of photovoltaic power stations.
10 Research on Fault Diagnosis Technology for Four-Quadrant Pulse Rectifiers Based on PSO-Optimized CNN
2026(9):982-992.
DOI: 10.12177/emca.2026.208
Abstract:
[Objective] Open-circuit faults of insulated-gate bipolar transistor (IGBT) in four-quadrant pulse rectifiers exhibit strong concealment characteristics, which can inevitably induce grid-side current distortion, power factor degradation, and DC-side voltage fluctuation. Traditional convolutional neural network (CNN) model applied to fault diagnosis is restricted by low diagnosis accuracy and easy trapping in local optimal solutions. To address these limitations, this paper adopts the particle swarm optimization (PSO) algorithm and proposes an intelligent fault diagnosis method integrating wavelet transform and the PSO-CNN. [Methods] Firstly, a single-phase two-level four-quadrant pulse rectifier model was established based on Simulink. Seven graded voltage fluctuation operating conditions and ten different IGBT open-circuit fault states were set to simulate complex and realistic equipment operation scenarios. Subsequently, grid-side current and DC-side voltage signals were collected. A six-layer wavelet decomposition was performed on the signals using the db3 wavelet basis, and normalized feature vectors were constructed by combining the frequency band energy coefficients. Considering the feature differences between current and voltage signals, a dual-channel CNN was built to extract fault features of grid-side current and DC-side voltage respectively. On this basis, the PSO algorithm was introduced to adaptively optimize the key hyperparameters of the CNN. [Results] Weighted F1-score and confusion matrix were taken as quantitative evaluation metrics for model performance, and comparative tests were conducted between the proposed PSO-CNN model and the traditional CNN model. The experimental results showed that compared with the traditional CNN model, the PSO-CNN model increased the weighted F1-score from 0.45 to 0.92. It effectively reduced the misclassification of faults with overlapping features and maintained stable fault diagnosis performance under voltage fluctuation conditions, with only minor recognition deviations in a small number of highly similar fault categories. [Conclusion] The combination of wavelet transform and PSO algorithm significantly enhance the feature extraction ability and classification reliability of the CNN model. The presented intelligent diagnosis method is suitable for real-time condition monitoring of four-quadrant pulse rectifiers, and it provide a feasible and reliable technical reference for intelligent operation and maintenance as well as performance improvement of various power electronic conversion devices.
[Objective] Open-circuit faults of insulated-gate bipolar transistor (IGBT) in four-quadrant pulse rectifiers exhibit strong concealment characteristics, which can inevitably induce grid-side current distortion, power factor degradation, and DC-side voltage fluctuation. Traditional convolutional neural network (CNN) model applied to fault diagnosis is restricted by low diagnosis accuracy and easy trapping in local optimal solutions. To address these limitations, this paper adopts the particle swarm optimization (PSO) algorithm and proposes an intelligent fault diagnosis method integrating wavelet transform and the PSO-CNN. [Methods] Firstly, a single-phase two-level four-quadrant pulse rectifier model was established based on Simulink. Seven graded voltage fluctuation operating conditions and ten different IGBT open-circuit fault states were set to simulate complex and realistic equipment operation scenarios. Subsequently, grid-side current and DC-side voltage signals were collected. A six-layer wavelet decomposition was performed on the signals using the db3 wavelet basis, and normalized feature vectors were constructed by combining the frequency band energy coefficients. Considering the feature differences between current and voltage signals, a dual-channel CNN was built to extract fault features of grid-side current and DC-side voltage respectively. On this basis, the PSO algorithm was introduced to adaptively optimize the key hyperparameters of the CNN. [Results] Weighted F1-score and confusion matrix were taken as quantitative evaluation metrics for model performance, and comparative tests were conducted between the proposed PSO-CNN model and the traditional CNN model. The experimental results showed that compared with the traditional CNN model, the PSO-CNN model increased the weighted F1-score from 0.45 to 0.92. It effectively reduced the misclassification of faults with overlapping features and maintained stable fault diagnosis performance under voltage fluctuation conditions, with only minor recognition deviations in a small number of highly similar fault categories. [Conclusion] The combination of wavelet transform and PSO algorithm significantly enhance the feature extraction ability and classification reliability of the CNN model. The presented intelligent diagnosis method is suitable for real-time condition monitoring of four-quadrant pulse rectifiers, and it provide a feasible and reliable technical reference for intelligent operation and maintenance as well as performance improvement of various power electronic conversion devices.
You are thevisitor
沪ICP备16038578号-3
Electric Machines & Control Application ® 2026
Supported by:Beijing E-Tiller Technology Development Co., Ltd.
沪ICP备16038578号-3
Electric Machines & Control Application ® 2026
Supported by:Beijing E-Tiller Technology Development Co., Ltd.
