Sa Zhu

Dr. Sa Zhu

orcid

Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
10 October 2017 in IEEE Transactions on Energy Conversion
Verified
Journal Article
Closed access
SZSa Zhu
Sa Zhu
Ming ChengMing Cheng
Ming Cheng
XCXiuhua Cai
Xiuhua Cai
A direct coupling method for coupled field-circuit thermal model of electrical machines is proposed, by which the temperature distributions in the lumped parameter thermal network and finite element region are solved simultaneously without iterative calculation procedure. Each thermal resistance in the lumped-parameter thermal network is seen as a one-dimensional finite element. The lumped-parameter thermal network and the finite element region are linked with the two kinds of boundaries, namely: equivalent convective and equivalent temperature. The users can take advantage of both methods altogether for more accurate and efficient thermal modeling of the electrical machines using the direct coupling method. An outer-rotor permanent magnet machine with an uneven distribution of AC copper loss in the winding region is used to highlight the benefits of the proposed method. The proposed method is verified by comparing with the indirect method and experiments as well.

25 February 2019 in IEEE Transactions on Energy Conversion
Verified
Journal Article
Closed access
SZSa Zhu
Sa Zhu
HWHao Wang
Hao Wang
JZJingwei Zhang
Jingwei Zhang
ZLZhipeng Lu
Zhipeng Lu
Ming ChengMing Cheng
Ming Cheng
A method for fast calculating the high-frequency harmonic loss (HFHL) in the permanent magnet (PM) of interior PM synchronous machine under the pulse width modulated voltage source inverter supply is proposed. The combination of the frozen differential reluctivity tensor method and the linear frequency finite element analysis (FEA) is proposed to investigate the relationship between the high-frequency harmonic voltages (HFHVs) in the rotor reference frame and the flux density variations in the PM, which is described quantitatively with the proposed analytical model using several dimensionless parameters. Then, the PM HFHLs are calculated analytically with the models using the flux density variations as input. The relative errors between the HFHLs calculated analytically and those obtained from the time-stepping FEA are within 15% at eight arbitrary working conditions. The proposed method can be used to calculate the PM HFHL in both 2D and 3D cases, which is testified with the experiment. Quadratic polynomial is used to fit the relationship between the dimensionless parameters and the fundamental currents. With the fitted polynomial and the analytical spectra of the HFHVs obtained from the double Fourier integration method, the PM HFHL map over entire working range can be calculated in one hour.

17 May 2018 in IEEE Transactions on Energy Conversion
Verified
Journal Article
Closed access
Ming ChengMing Cheng
Ming Cheng
JWJingxia Wang
Jingxia Wang
SZSa Zhu
Sa Zhu
Wei WangWei Wang
Wei Wang
This paper investigates the accurate loss calculation and thermal analysis for a nine-phase flux switching permanent magnet (FSPM) machine. First, the 3D finite element method (FEM) model is developed to calculate the PM eddy current loss, in which magnet axial segmentation is taken into account. In order to calculate the iron loss quickly and precisely, the magnetic flux leakage coefficient is proposed in this paper to modify the 2D calculation model of iron loss. Both the dc-biased magnetic induction and the rotational magnetization are taken into account. Then the key parameters for thermal modeling, such as the convective heat-transfer coefficient between external frame and ambient, and the coefficient of convection heat transfer in the end region, are derived, based on which a dissymmetric 3D FEM are used for thermal calculation of nine-phase FSPM machine. Finally, experiments on a prototype machine at the rated generation operation condition are carried out to verify the effectiveness of the developed model.

17 March 2021 in IEEE Transactions on Industrial Electronics
Verified
Journal Article
Closed access
SZSa Zhu
Sa Zhu
BSBin Shi
Bin Shi
The high-frequency iron losses caused by sinusoidal flux density variations with frequencies ranging from 2.5 to 40 kHz, amplitudes from 1 to 70 mT, and dc-biased magnetic inductions from 0 to 1.62 T are measured with the dual-transformer system. Then, in this article, a simple frequency-domain model is proposed based on the linear fitting and two-dimensional interpolation methods for engineering applications. In addition, another two frequency-domain models, respectively, considering and neglecting the excess losses are investigated and compared for the high-frequency iron-loss modeling, which only need low-frequency parameters and contain the physical interpretation to the variation of high-frequency iron-loss coefficients with the frequency, ac flux density amplitude, and dc-biased magnetic induction. The iron losses generated by the unipolar and bipolar pulsewidth modulation (PWM) voltages with different switching frequencies and modulation factors are measured to testify the three models. It is shown that the three models in this article are more accurate than the previously proposed one for calculating the PWM-induced iron loss by considering the variation of model parameters with the dc-biased magnetic induction and ac flux density amplitude, especially when silicon steel sheets reach the saturation.

12 May 2020 in IET Electric Power Applications
Verified
Journal Article
Open access
PDF file
SZSa Zhu
Sa Zhu
JDJianning Dong
Jianning Dong
YLYanru Li
Yanru Li
WHWei Hua
Wei Hua
Normal calculation of the carrier harmonic iron losses (CHILs) brought by the pulse width modulated (PWM) voltage source inverter in interior permanent magnet synchronous motors by the time‐stepping finite‐element analysis (TSFEA) with short time steps is very time‐consuming. In this study, a novel method for fast calculation is proposed, where a combination of the frozen differential reluctivity tensor method and the time‐harmonic finite‐element analysis is proposed to investigate the relationship between PWM voltage harmonics and the corresponding CHILs, with which the CHILs in stator and rotor are fast‐calculated with the voltage harmonics in stator and rotor reference frames, respectively. Besides, the eddy current reaction effect in electrical steel sheets on the CHILs at high frequency is considered by an analytical method and validated with the TSFEA. The results reveal that eight parameters are recommended to calculate the CHILs at different operating conditions, which vary with the fundamental current. Consequently, the CHIL map over the entire operating conditions can be fast‐calculated using the analytical spectra of the PWM voltages. Experiment on a specimen is conducted to validate the accuracy of the proposed method, which shows it depends on whether the equivalent differential permeability of the silicon steel can be accurately modelled.

25 June 2018 in IEEE Transactions on Magnetics
Verified
Journal Article
Closed access
SZSa Zhu
Sa Zhu
Ming ChengMing Cheng
Ming Cheng
YZYing Zhu
Ying Zhu
A new method for fast calculating the permanent magnet (PM) eddy current losses (ECLs) in interior PM synchronous machine caused by the high-frequency harmonic voltages (HFHVs) under pulsewidth modulation (PWM) voltage source inverter supply is proposed, by which time-consuming transient finite element analysis (FEA) is eliminated and the influence of the axial segmentation on the PM ECLs can be taken into consideration. The combination of the fixed differential permeability method and the linear frequency FEA is proposed to investigate the relationship between arbitrary HFHV and the corresponding PM ECL at different load conditions. The analytical model proposed in the previous published paper is used to describe this relationship quantitatively. Finally, the total PM ECLs at different load conditions can be calculated based on the fast Fourier transform results of the input PWM voltages. The relative errors between the total losses obtained with the proposed method and those with time-stepping FEA are within 21% at four typical load conditions, and the proposed method can reduce the computation time for one load condition to few minutes considering arbitrary PM axial segmentation.