Mobile website
Volume 52,Issue 7,2025 Table of Contents
2025, 52(7):710-720.
DOI: 10.12177/emca.2025.060
Abstract:
[Objective] The traditional methods for extracting fault characteristics of disconnector are not reliable enough, and the fault identification accuracy is not high. To address this issue, a fault diagnosis method for disconnector based on multi-characteristic map and convolutional neural network (CNN)-black-winged kite algorithm (BKA)-least squares support vector machine (LSSVM) is proposed. [Methods] Firstly, three characteristic maps were constructed in the frequency domain, time domain, and time-frequency domain using the Markov transition field (MTF), Gramian angular field (GAF), and short-time Fourier transform (STFT) respectively. Then, three CNN models were established separately, three characteristic maps were input, effective fault characteristics were extracted through convolution, pooling, and other steps. And the t-distributed stochastic neighbor embedding algorithm was used to reduce the dimension of the characteristic data in the fully connected layer of the CNN model. Finally, the extracted characteristic vectors were fused and spliced, the BKA-optimized LSSVM was used instead of the Softmax layer, and the fused characteristic vectors were input into the BKA-LSSVM for fault identification. [Results] Through on-site fault simulation tests, vibration signal data of the disconnector in four different states were collected, and comparative analysis was carried out. The results showed that the model proposed in this paper has higher accuracy, stronger reliability and generalization ability compared with other fault diagnosis models, and the average diagnostic accuracy of the proposed model for 8 runs reached 97.08%. The experimental results verified the feasibility of the proposed model. [Conclusion] The proposed method in this paper extracts the characteristic maps from multiple dimensions, which overcomes the limitations of the single dimension method. By introducing the CNN-BKA-LSSVM model, it can better extract the key characteristics and improve the precision and accuracy of fault identification. The proposed method provides a reliable theoretical basis and technical reference for the fault diagnosis of disconnectors, and also provides new ideas for the maintenance of disconnector equipment, which has important application value.
[Objective] The traditional methods for extracting fault characteristics of disconnector are not reliable enough, and the fault identification accuracy is not high. To address this issue, a fault diagnosis method for disconnector based on multi-characteristic map and convolutional neural network (CNN)-black-winged kite algorithm (BKA)-least squares support vector machine (LSSVM) is proposed. [Methods] Firstly, three characteristic maps were constructed in the frequency domain, time domain, and time-frequency domain using the Markov transition field (MTF), Gramian angular field (GAF), and short-time Fourier transform (STFT) respectively. Then, three CNN models were established separately, three characteristic maps were input, effective fault characteristics were extracted through convolution, pooling, and other steps. And the t-distributed stochastic neighbor embedding algorithm was used to reduce the dimension of the characteristic data in the fully connected layer of the CNN model. Finally, the extracted characteristic vectors were fused and spliced, the BKA-optimized LSSVM was used instead of the Softmax layer, and the fused characteristic vectors were input into the BKA-LSSVM for fault identification. [Results] Through on-site fault simulation tests, vibration signal data of the disconnector in four different states were collected, and comparative analysis was carried out. The results showed that the model proposed in this paper has higher accuracy, stronger reliability and generalization ability compared with other fault diagnosis models, and the average diagnostic accuracy of the proposed model for 8 runs reached 97.08%. The experimental results verified the feasibility of the proposed model. [Conclusion] The proposed method in this paper extracts the characteristic maps from multiple dimensions, which overcomes the limitations of the single dimension method. By introducing the CNN-BKA-LSSVM model, it can better extract the key characteristics and improve the precision and accuracy of fault identification. The proposed method provides a reliable theoretical basis and technical reference for the fault diagnosis of disconnectors, and also provides new ideas for the maintenance of disconnector equipment, which has important application value.
2025, 52(7):721-731.
DOI: 10.12177/emca.2025.064
Abstract:
[Objective] In-wheel motors have compact space and harsh heat dissipation conditions, and the interior permanent magnets are highly susceptible to irreversible demagnetization under high temperature and large armature current surges, which seriously threatens the reliability and performance of the motors. To address this issue, this study designs a new type of FeCo-based combined magnetic pole to improve the anti-demagnetization ability and comprehensive electromagnetic performance of in-wheel motors. [Methods] Firstly, the effects of different temperatures and armature currents on the demagnetization characteristics of motor parmanent magnet were systematically analyzed, and the easily demagnetized area of permanent magnet was determined. Then, a new type of high coercivity FeCo-based material was used to replace the traditional NdFeB in the easily demagnetized area to form a combined magnetic pole. The effects of different combination ratios of FeCo-based and NdFeB in the combined magnetic pole on the electromagnetic performance were investigated through simulation, and the optimal combination ratio of FeCo-based and NdFeB was determined. Finally, the electromagnetic performance of single magnetic pole motor and combined magnetic pole motor was compared and analyzed. [Results] The motor with combined magnetic pole performed better in various aspects such as rated torque, peak torque and efficiency. At an operating temperature of 110 ℃, the rated torque of the combined magnetic pole motor increased by 4 N·m, peak torque increased by 7 N·m, and maximum efficiency increased by 0.4% compared to the single magnetic pole motor. At a high temperature of 160 ℃, the rated torque of the combined magnetic pole motor increased by 8 N·m, torque fluctuation decreased by 0.8%, and peak torque increased by 14 N·m compared to the single magnetic pole motor. At peak conditions, the critical demagnetization temperature of the combined magnetic pole was 15 ℃ higher than that of the single pole motor. [Conclusion] The designed new type of FeCo-based combined magnetic pole effectively enhances the anti-demagnetization ability of the motor under high temperature and high current operating conditions, and significantly improves the comprehensive electromagnetic performance of the motor, providing an effective technical approach to solve the problem of easy demagnetization of in-wheel motors at high temperatures.
[Objective] In-wheel motors have compact space and harsh heat dissipation conditions, and the interior permanent magnets are highly susceptible to irreversible demagnetization under high temperature and large armature current surges, which seriously threatens the reliability and performance of the motors. To address this issue, this study designs a new type of FeCo-based combined magnetic pole to improve the anti-demagnetization ability and comprehensive electromagnetic performance of in-wheel motors. [Methods] Firstly, the effects of different temperatures and armature currents on the demagnetization characteristics of motor parmanent magnet were systematically analyzed, and the easily demagnetized area of permanent magnet was determined. Then, a new type of high coercivity FeCo-based material was used to replace the traditional NdFeB in the easily demagnetized area to form a combined magnetic pole. The effects of different combination ratios of FeCo-based and NdFeB in the combined magnetic pole on the electromagnetic performance were investigated through simulation, and the optimal combination ratio of FeCo-based and NdFeB was determined. Finally, the electromagnetic performance of single magnetic pole motor and combined magnetic pole motor was compared and analyzed. [Results] The motor with combined magnetic pole performed better in various aspects such as rated torque, peak torque and efficiency. At an operating temperature of 110 ℃, the rated torque of the combined magnetic pole motor increased by 4 N·m, peak torque increased by 7 N·m, and maximum efficiency increased by 0.4% compared to the single magnetic pole motor. At a high temperature of 160 ℃, the rated torque of the combined magnetic pole motor increased by 8 N·m, torque fluctuation decreased by 0.8%, and peak torque increased by 14 N·m compared to the single magnetic pole motor. At peak conditions, the critical demagnetization temperature of the combined magnetic pole was 15 ℃ higher than that of the single pole motor. [Conclusion] The designed new type of FeCo-based combined magnetic pole effectively enhances the anti-demagnetization ability of the motor under high temperature and high current operating conditions, and significantly improves the comprehensive electromagnetic performance of the motor, providing an effective technical approach to solve the problem of easy demagnetization of in-wheel motors at high temperatures.
2025, 52(7):732-742.
DOI: 10.12177/emca.2025.052
Abstract:
[Objective] To effectively suppress the electromagnetic vibration and noise generated during the operation of permanent magnet synchronous motors and enhance their silent performance to meet the application requirements of electric vehicles, precision equipment, etc. A fractional-slot interior permanent magnet synchronous motor with air-gap eccentricity with a rated power of 2 kW is taken as the research object in this paper. And its electromagnetic vibration characteristics and optimization methods are systematically analyzed. [Methods] Firstly, through electromagnetic analytical calculations and two-dimensional transient field finite element analysis, the improvement effects of the eccentric rotor on radial electromagnetic force waves at key orders and resulting vibrations were evaluated from the motor design perspective. Secondly, a field-circuit coupled co-simulation model was established to comparatively analyze electromagnetic vibration characteristics under proportional integral control, maximum torque per ampere control and model predictive current control strategies. Finally, a vibration test platform for the prototype was constructed to obtain the spectral characteristics of the vibration acceleration of the prototype under different control strategies and verify the reliability of the simulation model. [Results] The research results showed that with the eccentric rotor structure, the amplitudes of the radial electromagnetic force waves at key orders such as the 2nd and 4th orders of the motor decreased, weakening the vibration response concentrated near the modal frequency of the 2nd order. Among them, the peak value of the vibration acceleration under the rated load decreased by 47.9%. [Conclusion] Rotor eccentricity design can effectively reduce the risk of structural resonance. Model predictive current control can better suppress the additional vibration caused by current harmonics. This study provides an effective reference for the optimization of electromagnetic vibration and noise of motors.
[Objective] To effectively suppress the electromagnetic vibration and noise generated during the operation of permanent magnet synchronous motors and enhance their silent performance to meet the application requirements of electric vehicles, precision equipment, etc. A fractional-slot interior permanent magnet synchronous motor with air-gap eccentricity with a rated power of 2 kW is taken as the research object in this paper. And its electromagnetic vibration characteristics and optimization methods are systematically analyzed. [Methods] Firstly, through electromagnetic analytical calculations and two-dimensional transient field finite element analysis, the improvement effects of the eccentric rotor on radial electromagnetic force waves at key orders and resulting vibrations were evaluated from the motor design perspective. Secondly, a field-circuit coupled co-simulation model was established to comparatively analyze electromagnetic vibration characteristics under proportional integral control, maximum torque per ampere control and model predictive current control strategies. Finally, a vibration test platform for the prototype was constructed to obtain the spectral characteristics of the vibration acceleration of the prototype under different control strategies and verify the reliability of the simulation model. [Results] The research results showed that with the eccentric rotor structure, the amplitudes of the radial electromagnetic force waves at key orders such as the 2nd and 4th orders of the motor decreased, weakening the vibration response concentrated near the modal frequency of the 2nd order. Among them, the peak value of the vibration acceleration under the rated load decreased by 47.9%. [Conclusion] Rotor eccentricity design can effectively reduce the risk of structural resonance. Model predictive current control can better suppress the additional vibration caused by current harmonics. This study provides an effective reference for the optimization of electromagnetic vibration and noise of motors.
2025, 52(7):743-757.
DOI: 10.12177/emca.2025.058
Abstract:
[Objective] Aiming at the problem of unfixed switching frequency when finite control set model predictive control (FCS-MPC) is applied to neutral-point-clamped (NPC) inverters, this paper proposes a fixed switching frequency three-vector optimal sequence model predictive control (FSF-TOS-MPC) strategy. [Methods] Firstly, spatial sector division was used to simplify the optimization algorithm, narrowing the range of basic voltage vector involved in model predictive optimization, thus reducing the amount of computation. Then, through the polarity discrimination of the neutral-point potential of the DC side and the three-phase currents of the load side, the three-vector optimal switching sequence partition pre-selection was carried out. And the Karush-Kuhn-Tucker condition was introduced to solve the duration of each voltage vector in the switching sequence, to ensure that the duration of each vector was constant greater than zero, and the switching frequency was fixed. Finally, multi-objective control was carried out by constructing a cost function to achieve reference current tracking and DC side neutral-point potential balancing. [Results] The conventional FCS-MPC, optimized FCS-MPC and the FSF-TOS-MPC proposed in this paper were compared under different operating conditions by simulation. Simulation results showed that the FSF-TOS-MPC strategy can ensure the stable operation of the inverter with fixed switching frequency, thus effectively suppressing the total harmonic distortion of the output current, improving the dynamic response characteristics of the system, and achieving the neutral-point potential balance. [Conclusion] The FSF-TOS-MPC strategy proposed in this paper has superior control performance, and provides an effective solution to the problem of unfixed switching frequency in the model predictive control of NPC inverters.
[Objective] Aiming at the problem of unfixed switching frequency when finite control set model predictive control (FCS-MPC) is applied to neutral-point-clamped (NPC) inverters, this paper proposes a fixed switching frequency three-vector optimal sequence model predictive control (FSF-TOS-MPC) strategy. [Methods] Firstly, spatial sector division was used to simplify the optimization algorithm, narrowing the range of basic voltage vector involved in model predictive optimization, thus reducing the amount of computation. Then, through the polarity discrimination of the neutral-point potential of the DC side and the three-phase currents of the load side, the three-vector optimal switching sequence partition pre-selection was carried out. And the Karush-Kuhn-Tucker condition was introduced to solve the duration of each voltage vector in the switching sequence, to ensure that the duration of each vector was constant greater than zero, and the switching frequency was fixed. Finally, multi-objective control was carried out by constructing a cost function to achieve reference current tracking and DC side neutral-point potential balancing. [Results] The conventional FCS-MPC, optimized FCS-MPC and the FSF-TOS-MPC proposed in this paper were compared under different operating conditions by simulation. Simulation results showed that the FSF-TOS-MPC strategy can ensure the stable operation of the inverter with fixed switching frequency, thus effectively suppressing the total harmonic distortion of the output current, improving the dynamic response characteristics of the system, and achieving the neutral-point potential balance. [Conclusion] The FSF-TOS-MPC strategy proposed in this paper has superior control performance, and provides an effective solution to the problem of unfixed switching frequency in the model predictive control of NPC inverters.
2025, 52(7):758-768.
DOI: 10.12177/emca.2025.063
Abstract:
[Objective] Current sampling serves as a critical component to achieve efficient and stable operation of the motor, and its accuracy directly determines the motor control performance. In brushless direct current motor (BLDCM) control systems, conventional multi-sensor schemes not only increase the hardware costs, but also degrade the closed-loop control accuracy due to sensor parameter drift or characteristic differences. Aiming at the problem of current reconstruction blind zones of single current sensor in the low index modulation regions and sector boundary areas under space vector pulse width modulation (SVPWM), this paper proposes a midpoint sampling phase-shifting method in the second half cycle of pulse width modulation (PWM). [Methods] By adjusting the phase of the PWM pulse, the phase compensation set two times the minimum sampling time after the falling edge as the actual sampling window and implemented the current sampling operation at the midpoint of the window. This method not only ensured that the minimum sampling time requirement was met, but also reserved a sufficient margin for the signal aberration caused by the switching noise, which effectively reduced the uncertainty of sampling timing and current transient fluctuations. [Results] Simulation and experimental results demonstrated that the proposed method effectively reduced the current reconstruction blind zones in low index modulation regions and sector boundary areas. Compared with the uncompensated phase-shifting method, the accuracy of reconstructed phase current was significantly improved under the proposed compensation method, and the reconstruction error was stabilized at a lower level. The maximum error was reduced from 6.8 A to 1.38 A, which was a reduction of 79.71%, and the total harmonic distortion was reduced from 26.38% to 3.37%, which was a significant effect of harmonic suppression. [Conclusion] The method proposed in this paper effectively addresses the challenge of single-sensor current reconstruction in SVPWM-driven BLDCM systems, while reducing the hardware costs,and provides a feasible reference for the current sampling design of motors.
[Objective] Current sampling serves as a critical component to achieve efficient and stable operation of the motor, and its accuracy directly determines the motor control performance. In brushless direct current motor (BLDCM) control systems, conventional multi-sensor schemes not only increase the hardware costs, but also degrade the closed-loop control accuracy due to sensor parameter drift or characteristic differences. Aiming at the problem of current reconstruction blind zones of single current sensor in the low index modulation regions and sector boundary areas under space vector pulse width modulation (SVPWM), this paper proposes a midpoint sampling phase-shifting method in the second half cycle of pulse width modulation (PWM). [Methods] By adjusting the phase of the PWM pulse, the phase compensation set two times the minimum sampling time after the falling edge as the actual sampling window and implemented the current sampling operation at the midpoint of the window. This method not only ensured that the minimum sampling time requirement was met, but also reserved a sufficient margin for the signal aberration caused by the switching noise, which effectively reduced the uncertainty of sampling timing and current transient fluctuations. [Results] Simulation and experimental results demonstrated that the proposed method effectively reduced the current reconstruction blind zones in low index modulation regions and sector boundary areas. Compared with the uncompensated phase-shifting method, the accuracy of reconstructed phase current was significantly improved under the proposed compensation method, and the reconstruction error was stabilized at a lower level. The maximum error was reduced from 6.8 A to 1.38 A, which was a reduction of 79.71%, and the total harmonic distortion was reduced from 26.38% to 3.37%, which was a significant effect of harmonic suppression. [Conclusion] The method proposed in this paper effectively addresses the challenge of single-sensor current reconstruction in SVPWM-driven BLDCM systems, while reducing the hardware costs,and provides a feasible reference for the current sampling design of motors.
2025, 52(7):769-777.
DOI: 10.12177/emca.2025.062
Abstract:
[Objective] To address the demands of low-altitude heavy-duty unmanned aerial vehicle propulsion systems for high power density and robust anti-saturation capability, this study designs an axial flux permanent magnet machine (AFPMM) with a soft magnetic composite (SMC) stator core. The research aims to reveal the influence of stator material properties on machine inductance characteristics and provide theoretical support for enhancing the electromagnetic performance of AFPMM. [Methods] Firstly, a three-dimensional electromagnetic field model of AFPMM was established based on the finite element method. Then, the magnetic circuit characteristics of stators made of SMC, grain-oriented (GO) silicon steel, and non-oriented (NO) silicon steel were comparatively analyzed. Finally, the nonlinear variation rules of dq-axis inductances with current angle and current amplitude were quantified, and the saturation mechanisms of machine inductance under different stator materials were systematically investigated through magnetic flux density distribution analysis. [Results] The results showed that the machine with SMC stator has isotropic properties and optimal anti-saturation capability. The machine with GO silicon steel stator has anisotropic characteristics with the highest initial inductance as well as output torque, but the inductance drop was significant at high currents. The inductance characteristics of machine with NO silicon steel stator was between SMC stator and GO silicon steel stator. [Conclusion] Although the machine with GO steel stator offers higher torque density, its performance necessitates optimized magnetic circuit design to suppress cross-saturation effects. The machine with SMC stator can effectively balance the dq-axis inductance and alleviate local saturation, which is suitable for high-frequency and complex magnetic circuit working scenarios. This study verifies the advantages of SMC stator in meeting the practical demands of unmanned aerial vehicles and provides theoretical guidance for material selection and electromagnetic design in high-performance motor.
[Objective] To address the demands of low-altitude heavy-duty unmanned aerial vehicle propulsion systems for high power density and robust anti-saturation capability, this study designs an axial flux permanent magnet machine (AFPMM) with a soft magnetic composite (SMC) stator core. The research aims to reveal the influence of stator material properties on machine inductance characteristics and provide theoretical support for enhancing the electromagnetic performance of AFPMM. [Methods] Firstly, a three-dimensional electromagnetic field model of AFPMM was established based on the finite element method. Then, the magnetic circuit characteristics of stators made of SMC, grain-oriented (GO) silicon steel, and non-oriented (NO) silicon steel were comparatively analyzed. Finally, the nonlinear variation rules of dq-axis inductances with current angle and current amplitude were quantified, and the saturation mechanisms of machine inductance under different stator materials were systematically investigated through magnetic flux density distribution analysis. [Results] The results showed that the machine with SMC stator has isotropic properties and optimal anti-saturation capability. The machine with GO silicon steel stator has anisotropic characteristics with the highest initial inductance as well as output torque, but the inductance drop was significant at high currents. The inductance characteristics of machine with NO silicon steel stator was between SMC stator and GO silicon steel stator. [Conclusion] Although the machine with GO steel stator offers higher torque density, its performance necessitates optimized magnetic circuit design to suppress cross-saturation effects. The machine with SMC stator can effectively balance the dq-axis inductance and alleviate local saturation, which is suitable for high-frequency and complex magnetic circuit working scenarios. This study verifies the advantages of SMC stator in meeting the practical demands of unmanned aerial vehicles and provides theoretical guidance for material selection and electromagnetic design in high-performance motor.
2025, 52(7):778-787.
DOI: 10.12177/emca.2025.056
Abstract:
[Objective] In medium- and high-speed domain sensorless control drive systems of permanent magnet synchronous motor (PMSM), the traditional linear extended state observer (LESO) suffers from observation delays in back electromotive force estimation due to bandwidth limitations, which leads to significant errors in rotor position and speed estimation. Aiming at this problem, a PMSM sensorless control strategy based on phase lead correction LESO (PLC-LESO) is proposed. [Methods] Firstly, the design method of the traditional LESO-based sensorless control scheme was introduced, and its inherent observation delay problem in back electromotive force estimation was systematically analyzed using frequency domain analysis. Then, a phase lead correction unit was introduced and the PLC-LESO was designed to observe the back electromotive force. Finally, in order to verify the feasibility and effectiveness of the proposed control strategy, a simulation model was built based on Matlab/Simulink, and the rotor position and speed information was estimated using normalized phase-locked loop to compare and analyze the observation results of the traditional LESO and PLC-LESO. [Results] The simulation results showed that, under identical observer bandwidth setting, the rotor position estimation error of the traditional LESO was 0.6 rad, while the PLC-LESO was able to accurately track the back electromotive force with a rotor position estimation error of only 0.005 rad. When the bandwidth of the traditional LESO was increased to 30 000, its back electromotive force tracking performance was improved, but the fluctuation of the rotational speed estimation was intensified, resulting in resulting in poor system stability. Whereas, the PLC-LESO not only tracked the reverse potential accurately without increasing bandwidth, but also controlled the speed fluctuation within ±10 rpm. [Conclusion] Compared with the traditional LESO, the proposed PLC-LESO exhibits superior control performance in the estimation of the back electromotive force tracking, which improves the accuracy of the system in estimating the rotor position and speed information.
[Objective] In medium- and high-speed domain sensorless control drive systems of permanent magnet synchronous motor (PMSM), the traditional linear extended state observer (LESO) suffers from observation delays in back electromotive force estimation due to bandwidth limitations, which leads to significant errors in rotor position and speed estimation. Aiming at this problem, a PMSM sensorless control strategy based on phase lead correction LESO (PLC-LESO) is proposed. [Methods] Firstly, the design method of the traditional LESO-based sensorless control scheme was introduced, and its inherent observation delay problem in back electromotive force estimation was systematically analyzed using frequency domain analysis. Then, a phase lead correction unit was introduced and the PLC-LESO was designed to observe the back electromotive force. Finally, in order to verify the feasibility and effectiveness of the proposed control strategy, a simulation model was built based on Matlab/Simulink, and the rotor position and speed information was estimated using normalized phase-locked loop to compare and analyze the observation results of the traditional LESO and PLC-LESO. [Results] The simulation results showed that, under identical observer bandwidth setting, the rotor position estimation error of the traditional LESO was 0.6 rad, while the PLC-LESO was able to accurately track the back electromotive force with a rotor position estimation error of only 0.005 rad. When the bandwidth of the traditional LESO was increased to 30 000, its back electromotive force tracking performance was improved, but the fluctuation of the rotational speed estimation was intensified, resulting in resulting in poor system stability. Whereas, the PLC-LESO not only tracked the reverse potential accurately without increasing bandwidth, but also controlled the speed fluctuation within ±10 rpm. [Conclusion] Compared with the traditional LESO, the proposed PLC-LESO exhibits superior control performance in the estimation of the back electromotive force tracking, which improves the accuracy of the system in estimating the rotor position and speed information.
2025, 52(7):788-799.
DOI: 10.12177/emca.2025.059
Abstract:
[Objective] Power electronic components in DC motor systems are prone to soft faults under prolonged high-frequency switching operation. To address the problems in soft fault diagnosis, including insufficient fusion of time-frequency domain features and low recognition accuracy, this paper proposes a fault diagnosis method based on the LSTM-CAM-Transformer model by combining the long short-term memory (LSTM) network and the multi-scale time-frequency domain cross-attention mechanism (CAM). [Methods] Firstly, the collected fault signal was preprocessed, and the parameters of the variational mode decomposition (VMD) algorithm were optimized using the Rime optimization algorithm (RIME), which accurately obtained the optimal combination of the decomposition modals K and the penalty factor α, and effectively removed the noise and interference components in the signal. Then, the 5-dimensional time domain parameters and 5-dimensional frequency domain parameters of each intrinsic modal function were extracted, which were used as the feature vectors for fault diagnosis. Finally, a multi-scale time-frequency domain CAM was utilized to strengthen the information interaction between the time domain and frequency domain of the feature vectors, allowing the model to further mine the time-frequency domain features of the signals. [Results] The model proposed in this paper was compared with the other four models to verify its superiority. The experimental results showed that compared with the other four models, the LSTM-CAM-Transformer model proposed in this paper has the fastest convergence speed, better stability and generalization, and outperforms the other four models in diagnostic accuracy, F1 score, loss value and recall. [Conclusion] The LSTM-CAM-Transformer model proposed in this paper effectively solves the problem of insufficient fusion of time-frequency domain features in the traditional method by integrating the signal preprocessing strategy of RIME-based improved VMD signal preprocessing strategy with the CAM time-frequency feature enhancement mechanism, which provides an efficient and reliable new method for the soft fault diagnosis of the power electronic equipment in the DC motor system.
[Objective] Power electronic components in DC motor systems are prone to soft faults under prolonged high-frequency switching operation. To address the problems in soft fault diagnosis, including insufficient fusion of time-frequency domain features and low recognition accuracy, this paper proposes a fault diagnosis method based on the LSTM-CAM-Transformer model by combining the long short-term memory (LSTM) network and the multi-scale time-frequency domain cross-attention mechanism (CAM). [Methods] Firstly, the collected fault signal was preprocessed, and the parameters of the variational mode decomposition (VMD) algorithm were optimized using the Rime optimization algorithm (RIME), which accurately obtained the optimal combination of the decomposition modals K and the penalty factor α, and effectively removed the noise and interference components in the signal. Then, the 5-dimensional time domain parameters and 5-dimensional frequency domain parameters of each intrinsic modal function were extracted, which were used as the feature vectors for fault diagnosis. Finally, a multi-scale time-frequency domain CAM was utilized to strengthen the information interaction between the time domain and frequency domain of the feature vectors, allowing the model to further mine the time-frequency domain features of the signals. [Results] The model proposed in this paper was compared with the other four models to verify its superiority. The experimental results showed that compared with the other four models, the LSTM-CAM-Transformer model proposed in this paper has the fastest convergence speed, better stability and generalization, and outperforms the other four models in diagnostic accuracy, F1 score, loss value and recall. [Conclusion] The LSTM-CAM-Transformer model proposed in this paper effectively solves the problem of insufficient fusion of time-frequency domain features in the traditional method by integrating the signal preprocessing strategy of RIME-based improved VMD signal preprocessing strategy with the CAM time-frequency feature enhancement mechanism, which provides an efficient and reliable new method for the soft fault diagnosis of the power electronic equipment in the DC motor system.
2025, 52(7):800-811.
DOI: 10.12177/emca.2025.055
Abstract:
[Objective] Accurate multi-regional load forecasting (MRLF) is a critical foundation for ensuring grid stability and optimizing the economic dispatch of modern power systems. Traditional load forecasting methods face challenges in capturing the dynamic spatial dependencies among multi-regional loads. To address the problem that traditional methods are unable to extract dynamic spatial features among loads, a MRLF model based on decoupling and adversarial graph attention network (DAGAT) is proposed.[Methods] The model effectively extracted the dynamic spatial features between multi-regional loads through graph attention networks. Firstly, the regional load series was decomposed into trend and fluctuation components using discrete wavelet transform(DWT). Secondly, a two-channel feature extraction module was designed to extract the spatio-temporal features of the multi-regional load series based on the different characteristics of trend and fluctuation components. In addition, a generative adversarial network architecture based on zero-sum game was introduced for adversarial training of spatio-temporal prediction models, the resulting adversarial loss was weighted and combined with a traditional forecasting loss function as the loss function of the model. Finally, simulation validation was performed based on real multi-regional load data from the New York independent system operator. [Results] Comparison and analysis with mainstream models showed that the DAGAT model proposed in this paper was more superior than traditional machine learning models, prediction models that only capture temporal dependencies, and prediction models that express spatial correlations using static weight matrices. The superiority of the model stemmed from the fact that the graph attention network captured the spatial correlation between multi-regional loads by learning dynamically changing weights, and the DWT clearly represented the characteristics of the multi-regional load series, and the joint loss further optimized the model parameters. [Conclusion] DAGAT effectively improves the accuracy of MRLF and solves the problem that dynamic spatial features between multi- regional loads cannot be captured, which is effective and superior in engineering practice.
[Objective] Accurate multi-regional load forecasting (MRLF) is a critical foundation for ensuring grid stability and optimizing the economic dispatch of modern power systems. Traditional load forecasting methods face challenges in capturing the dynamic spatial dependencies among multi-regional loads. To address the problem that traditional methods are unable to extract dynamic spatial features among loads, a MRLF model based on decoupling and adversarial graph attention network (DAGAT) is proposed.[Methods] The model effectively extracted the dynamic spatial features between multi-regional loads through graph attention networks. Firstly, the regional load series was decomposed into trend and fluctuation components using discrete wavelet transform(DWT). Secondly, a two-channel feature extraction module was designed to extract the spatio-temporal features of the multi-regional load series based on the different characteristics of trend and fluctuation components. In addition, a generative adversarial network architecture based on zero-sum game was introduced for adversarial training of spatio-temporal prediction models, the resulting adversarial loss was weighted and combined with a traditional forecasting loss function as the loss function of the model. Finally, simulation validation was performed based on real multi-regional load data from the New York independent system operator. [Results] Comparison and analysis with mainstream models showed that the DAGAT model proposed in this paper was more superior than traditional machine learning models, prediction models that only capture temporal dependencies, and prediction models that express spatial correlations using static weight matrices. The superiority of the model stemmed from the fact that the graph attention network captured the spatial correlation between multi-regional loads by learning dynamically changing weights, and the DWT clearly represented the characteristics of the multi-regional load series, and the joint loss further optimized the model parameters. [Conclusion] DAGAT effectively improves the accuracy of MRLF and solves the problem that dynamic spatial features between multi- regional loads cannot be captured, which is effective and superior in engineering practice.
2025, 52(7):812-822.
DOI: 10.12177/emca.2025.054
Abstract:
[Objective] In the field of economic scheduling of virtual power plants, centralized algorithms are currently dominant, neglecting their distributed structural forms, leading to an over-reliance on central nodes and inefficient handling of distributed energy resources. At the same time, conventional distributed optimization algorithms are complex and slow in operation, makeing it difficult to respond to the dynamic changes in the power system in a timely manner, which seriously restricts the efficient operation of virtual power plants. And the increase in distributed units leads to the generation of a large number of information exchanges, which is a serious waste of communication resources. Aiming at this problem, this paper proposes an optimal scheduling strategy that conforms to the distributed structure of virtual power plant. [Methods] Firstly, an economic scheduling strategy based on distributed neural dynamics algorithm was proposed, aiming to solve the fast consistency problem in virtual power plants. Secondly, a dynamic distributed event-triggered mechanism was designed to save communication resources. Finally, modelling was performed based on Matlab simulation software. The effectiveness of economic scheduling and the superiority of dynamic distributed event-triggered mechanism were proved by analysing the consistency of fast convergence and comparing the periodic event-triggered mechanism. [Results] Simulation results showed that the optimization algorithm proposed in this paper can effectively and quickly achieve incremental cost consistency, and the triggered mechanism has obvious advantages over the traditional mechanisms, effectively saving resources. [Conclusion] This paper focuses on the economic scheduling of virtual power plants under the dynamic distributed event-triggered mechanism. An economic scheduling optimization algorithm of virtual power plant based on distributed neural dynamics was proposed, which effectively solved the problem of information congestion caused by centralized optimization algorithm. By designing a dynamic distributed event-triggered mechanism that adapts to the distributed structure, the communication resource is saved.
[Objective] In the field of economic scheduling of virtual power plants, centralized algorithms are currently dominant, neglecting their distributed structural forms, leading to an over-reliance on central nodes and inefficient handling of distributed energy resources. At the same time, conventional distributed optimization algorithms are complex and slow in operation, makeing it difficult to respond to the dynamic changes in the power system in a timely manner, which seriously restricts the efficient operation of virtual power plants. And the increase in distributed units leads to the generation of a large number of information exchanges, which is a serious waste of communication resources. Aiming at this problem, this paper proposes an optimal scheduling strategy that conforms to the distributed structure of virtual power plant. [Methods] Firstly, an economic scheduling strategy based on distributed neural dynamics algorithm was proposed, aiming to solve the fast consistency problem in virtual power plants. Secondly, a dynamic distributed event-triggered mechanism was designed to save communication resources. Finally, modelling was performed based on Matlab simulation software. The effectiveness of economic scheduling and the superiority of dynamic distributed event-triggered mechanism were proved by analysing the consistency of fast convergence and comparing the periodic event-triggered mechanism. [Results] Simulation results showed that the optimization algorithm proposed in this paper can effectively and quickly achieve incremental cost consistency, and the triggered mechanism has obvious advantages over the traditional mechanisms, effectively saving resources. [Conclusion] This paper focuses on the economic scheduling of virtual power plants under the dynamic distributed event-triggered mechanism. An economic scheduling optimization algorithm of virtual power plant based on distributed neural dynamics was proposed, which effectively solved the problem of information congestion caused by centralized optimization algorithm. By designing a dynamic distributed event-triggered mechanism that adapts to the distributed structure, the communication resource is saved.
You are thevisitor
沪ICP备16038578号-3
Electric Machines & Control Application ® 2025
Supported by:Beijing E-Tiller Technology Development Co., Ltd.
沪ICP备16038578号-3
Electric Machines & Control Application ® 2025
Supported by:Beijing E-Tiller Technology Development Co., Ltd.
