Simulation and Fault Detection in PMSM under Dynamic ... - CiteSeerX

6 downloads 0 Views 414KB Size Report
Abstract: This paper presents a study of PMSM under short circuit in turns at different ... III. Simulation of Short Circuit. The way of Modeling shorted turn faults by ...
Simulation and Fault Detection in PMSM under Dynamic Conditions J. Rosero, J. Romeral, Motion Control and Industrial Applications Group, Technical University of Catalonia C/ Colom. 08222 Terrassa. Catalonia. Spain, e-mail: [email protected] Abstract: This paper presents a study of PMSM under short circuit in turns at different speed and torque values. PMSM in a healthy state and under fault conditions is simulated at different working points using a Flux 2D, twodimensional (2-D) finite element analysis (FEA). Currents and flux density are presented and their harmonic content is obtained. In addition, harmonics content is compared to one obtained from a healthy motor. This paper applies the appropriate tools for determining the fault condition from the current analysis as Empirical mode decomposition (EMD) with Wigner Ville Distribution (WVD). The EMD and WVD are used to fault detection in the operation range of PMSM. Simulated and experimental stator current spectra from healthy and faulty machines are obtained and compared.

I. Introduction The permanent magnet synchronous motor (PMSM) is becoming popular in high performance applications as compared to other types of ac motors. High speed operation and precise torque control are some of advantageous features, which include high torque to current ratio as well as high power to weight ratio, high efficiency, low noise and robustness. In such applications it is advantageous to use a drive capable of continuing operation even with the presence of any single point failure. Such a drive is termed fault tolerant and in order to develop such drive it is necessary to know and understand the faults in motor, which is the aim of the research presented here [1, 2]. A short circuit between turns is the most critical fault in a machine, is quite difficult to detect and almost impossible to remove. Stator winding faults might have a destructive effect on the stator coils. [3]. Stator faults are usually related to insulation failure. There are a number of techniques used to detect turn-to-turn faults, the majority of them based on stator voltages and currents, axial flux and d-q current and voltage component analysis [4, 5]. The best way to analyze the content frequency of nonstationary stator current is with a joint time-frequency analysis. Therefore, a number of time-frequency analysis technologies have been developed for analyzing non-stationary stator current [6, 7] as can be see Fig. 1. Apart from their theoretical advantages, the method proposed for motor fault detection should be really accurate, easy to perform and simple to implement on a Digital Signal Processing (DSP) development tool. The EMD method was motivated by computation of instantaneous frequency defined in terms of Hilbert transform. The local energy and the instantaneous frequency derived from the intrinsic mode function (IMF) [8, 9]. IMF is the decomposition of signal at different frequency ranges. The Wigner-Ville distribution (WVD) is a time-frequency representation [7]. If the analyzed signal contains more than one component, the WVD method suffers from cross term interference as the Pseudo Wigner Ville Distribution (SPWVD) exist for applying the weighting function to the instantaneous correlation. Flux 2D is proposed for simulation of the system PMSM-Control-Drive because involves non-linear

external circuit parts, such as solid state switching devices; also, the electric motor comprises nonlinear magnetic materials. Flux 2D allows the user to consider local phenomena such as saturation, high circulating currents, field current dynamics, Eddy currents and hysteresis losses, non-sinusoidal voltage induced (EMF), etc., whose effects are not important in normal behaviour, but are significative during a fault transient [10, 11].

Fig. 1. Scheme of fault detection techniques analysis in PMSM

In this paper, a PMSM with short circuit is simulated at different speed values using a Flux 2D, twodimensional (2-D) Finite Element Analysis (FEA). Currents and flux density are presented and their harmonics contents are obtained. In addition, the PMSM running at speed variation are analyzed by means of (EMD) with (WVD). Simulations are also compared with experimental results.

IV. Workbench of simulation and experimental The fault analysis for a PMSM of 6000 rpm, 2.3 Nm, 3 pair of poles is based on simulation and experimental tests at different speed values. Simulations and experiments have been carried out for motors with 4, 8 and 12 short turns of stator phase winding. The spectrum has been normalized to a rotor frequency for every case. Numerical simulations were developed with the combination of Flux 2D for the motor model and electric circuit, and Matlab-Simulink for power electronics and control as is shows in Fig. 2.

Fig. 2. Schematic of the finite element analysis (FEA) for a PMSM

The workbench development used to apply and improve motor current signal analysis (MCSA) start with current, a filter that keeps only those components of particular interest; followed by selection of frequency components and finally the software classifies.

III. Simulation of Short Circuit The way of Modeling shorted turn faults by means of Flux 2D can be see in Fig. 3. The fault of short turns of stator phase winding leads to two main effects on the machine flux. The first is that the large current in the shorted turns, leads to an increase in the local leakage flux, particularly slot leakage. This changes the saturation conditions from of the teeth locally. Secondly, the currents induced in the shorted coils oppose the establishment of the main, air-gap flux. They thus reduce that flux and the corresponding main flux path saturation along the winding axis of the shorted coil.

Fig. 3. Circuit scheme for short circuit simulation

Fig. 4. Flux density of a PMSM with short circuit

Therefore, the new frequency component appears in the stator current spectrum as a result of a fault in the stator windings. Only a rise in the rotor slot harmonic frequencies can be expected because under fault conditions a greater number of flux density waves exist in the machine and all of these flux

densities make a contribution at the same frequencies. Thus, there is a greater probability of flux density with the basic number of pole pairs now existing [12] as shown in Fig. 4 and Fig. 5.

Fig. 5. Flux density distribution of a PMSM with short circuit

The simulations of the PMSM in stationary conditions are shown in Fig. 6 through the stator current harmonic. The amplitudes of the current harmonics 1st, 7th, 9th are higher for a faulty motor than for a healthy one. However, at a low speeds the the discrimination of current harmonics between both motors is more difficult to be seen than when motors are running at nominal speed. The experimental results are similar for all speed range. 0 w=1500 rpm

M Healthy M 4 Short turns M 8 Short turns M 12 Short turns

Amplitude ( dB)

-10

-20

-30

-40

-50

-60 0

1

2

3

4

5

6

7

8

9

10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Harmonics

Fig. 6. Stator current harmonics for a PMSM with short circuit, Simulation at 1500 rpm

V. Result The short circuit is analyzed under different speed conditions. Five intrinsic mode functions (IMF) are obtained from the stator current by means of EMD and afterwards WVD is calculated. The results obtained from a machine with short circuit are compared with those obtained from a healthy machine. IMF 3 contains the stator main current harmonic and IMF 1 and 2 together contain the failure harmonic. In this way, the EMD allows you to obtain the IMF that contain the short circuit fault harmonic and thus concentrate our analysis on them, obtaining better resolution and accuracy with fault detection. Now, the biggest harmonic is the 9th fault harmonic (900 Hz in this case); the main harmonic corresponding to power supply, that is always higher to the other ones when it is analyzed with standard MCSA [13], has been eliminated. The simulation with a linear speed change from 1500 rpm to 1000 rpm, starting at 0.2 s and ending at 0.29 s can been seen in Fig. 7 for healthy PMSM. The stator currents decomposition by means of EMD allows to obtain 5 IMF and each one contains its own characteristic frequency range. In this case, the spectrum for healthy machine results clean or empty; on the contrary, for short circuit fault we could locate the fault frequency in a more easy and evident way.

Reassigned smoothed pseudo Wigner-Ville distribution TFRSPWV 600 60

550

Frequency (Hz)

500

50

450 40

400 30

350 300

20

250 10

200 150 0

0.05

0.1

0.15 Time (S)

0.2

0.25

0.3

Fig. 7. SPWVD of IMF 1&2 in healthy PMSM. Speed change from 1500 to 1000 rpm 600 60

550

Frequency (Hz)

500

50

450 40

400 30

350 300

20

Failure Harmonic

250

10

200 150

0.05

0.1

0.15 Time (S)

0.2

0.25

0.3

Fig. 8. SPWVD of IMF 1&2 in PMSM with 12 short circuit turns. Speed change from 1500 to 1000 rpm 600 60

550

Frequency (Hz)

500

50

450 40

400 30

350 300

20

Failure Harmonic

250

10

200 150

0.05

0.1

0.15 Time (S)

0.2

0.25

0.3

Fig. 9. SPWVD of IMF 1&2 in PMSM with 12 short circuit turns. Speed change from 1500 to 1000 rpm, Experimental results

Fig. 8 shows SPWVD of IMF 1 and 2 of the stator currents of a PMSM with 12 short circuit turns in comparison with Fig. 7. Some experimental results are shown in Fig. 9. These figures represent SPWVD of IMF 1 and 2 PMSM stator currents with a 12 turn short circuited. These are the results for low speed where the fault detection is more critical. The results for the whole speed range are also very satisfactory. The method allows remarking the analysis in the characteristic failure and decreasing the computational burden; besides, it maximizes the failure relative value. The quick filter by means of EMD allows eliminating the undesired frequencies for the fault detection and the system can be more

accurate. Now, it presents the short circuit analysis with stator currents of PMSM by means of EMD and IMF. Next, WVD has been applied to IMF 1 and 2.

VI. Conclusion Flux 2D model allows PMSM behaviour to be determined under different fault conditions. Flux 2D is an excellent tool for simulation of the PMSM under fault and for all speed range. This software allows the user to know the behaviour of the motor in different conditions and to analyze what has happened after a failure. Joint EMD and SPWVD show the ability of fault identify within a short time. These transforms do not need many samples, and they can work independently of the speed variations. The featured signature allows differentiating between healthy and faulty conditions, as well as between different degrees of fault. The results can be used for fault detection and diagnosis of such internal faults at low speed and dynamic conditions. This method is suitable to be implemented in embedded on-line applications. On the whole, their precision and reliability is very high. The results show an increase in the spectral resolution and also reliability for the fault diagnosis.

VII. References [1] J. A. Haylock, B. C. Mecrow, A. G. Jack, and D. J. Atkinson, "Operation of a fault tolerant PM drive for an aerospace fuel pump application," in Eighth International Conference on Electrical Machines and Drives (Conf. Publ. No. 444), 1997, pp. 133-137. [2] N. Ertugrul, W. L. Soong, S. Valtenbergs, and H. Chye, "Investigation of a fault tolerant and high performance motor drive for critical applications," in TENCON. Proceedings of IEEE Region 10 International Conference on Electrical and Electronic Technology, 2001, pp. 542-548 vol.2. [3] M. A. Awadallah and M. M. Morcos, "Application of AI tools in fault diagnosis of electrical machines and drivesan overview," IEEE Transactions on Energy Conversion, vol. 18, pp. 245-251, 2003. [4] H. A. Toliyat, S. P. Waikar, and T. A. Lipo, "Analysis and simulation of five-phase synchronous reluctance machines including third harmonic of airgap MMF," IEEE Transactions on Industry Applications, vol. 34, pp. 332-339, 1998. [5] H. A. Toliyat and T. A. Lipo, "Transient analysis of cage induction machines under stator, rotor bar and end ring faults," IEEE Transactions on Energy Conversion, vol. 10, pp. 241-247, 1995. [6] N. Arthur and J. Penman, "Condition monitoring with non-linear signal processing," in IEE Colloquium on NonLinear Signal and Image Processing (Ref. No. 1998/284), , 1998, pp. 4/1-4/5. [7] J. Rosero, J. Cusido, A. Garcia, J. A. Ortega, and L. Romeral, "Broken Bearings Fault Detection for a Permanent Magnet Synchronous Motor under non-constant working conditions by means of a Joint Time Frequency Analysis," in IEEE International Symposium on Industrial Electronics, ISIE 2007, Vigo, Spain, 2007, pp. 3415-3419. [8] B. Liu, S. Riemenschneider, and Y. Xu, "Gearbox fault diagnosis using empirical mode decomposition and Hilbert spectrum," Mechanical Systems and Signal Processing, vol. 20, pp. 718-734, 2006. [9] N. E. Huang, Z. Shen, S. R. Long, M. C. Wu, H. H. Shih, Q. Zheng, N. C. Yen, C. C. Tung, and H. H. Liu, "The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis," Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, vol. 454, pp. 903-995, 1998. [10] A. A. Arkadan and B. W. Kielgas, "The coupled problem in switched reluctance motor drive systems during fault conditions," IEEE Transactions on Magnetics, vol. 30, pp. 3256-3259, 1994. [11] J. Rosero, J. Cusido, A. Garcia, J. A. Ortega, and L. Romeral, "Study on the Permanent Magnet Demagnetization Fault in Permanent Magnet Synchronous Machines," in The 32nd Annual Conference of the IEEE Industrial Electronics Society, IECON06, Paris - FRANCE, 2006, pp. 879-884. [12] G. M. Joksimovic and J. Penman, "The detection of inter-turn short circuits in the stator windings of operating motors," IEEE Transactions on Industrial Electronics, vol. 47, pp. 1078-1084, 2000. [13] J. Rosero, O. Almonacid, M. Amaya, and L. Romeral, "Simulation and Fault Detection of Short Circuit Winding in a Permanent Magnet Synchronous Machines (PMSM)," in ISEF 2007 - XIII International Symposium on Electromagnetic Fields in Mechatronics, Electrical and Electronic Engineering, Prague, Czech Republic, 2007.