Available online at www.sciencedirect.com

ELSEVIER

Procedía Earth and Planetary Science 1 (2009) 1448-1454

ScienceDirect

Procedia Earth and Planetary Science

www.elsevier.com/locate/procedia

The 6 International Conference on Mining Science & Technology

Back-to-back three-level double-fed induction motor control system

for mine hoist

The stator-flux-oriented vector control strategy for double-fed induction motor was investigated in this paper. Three phase three-level rectifier was used as AC/DC block. This rectifier worked with low harmonic distortion in the grid side. Besides, it maintained a high power factor while the energy flows in bidirectional way. For the DC/AC converter, three-level inverter was adopted. Using stator-flux oriented vector control, the double-fed induction gained a high dynamic performance when the stator side was controlled operating in a unite power factor condition. In this paper a novel simplified space vector pulse width modulation (SVPWM) method was proposed to simplify the calculation. The neutral-point potential balance can be easily realized with this method. The control strategy was verified by experiments.

Keywords: dual three-level; space vector pulse width modulation (SVPWM); stator flux oriented; double-fed

1. Introduction

In recent years, double-fed induction motor has become a new research hotspot because of its unique advantages. With characteristics of both synchronous motor and induction motor, the machine can operate in sub-synchronous, synchronous, super synchronous speed by regulating rotor-side AC exciting at the aim of adjusting reactive power in stator side. As only a part of the overall energy, slip energy is controlled by inverter, so that the inverter whose capacity is smaller than that of the motor could be selected.

With the growing price of non-renewable resources such as oil, every country puts a lot into the research of new energy sources. As a novel green and renew-able energy, the wind energy attracts scholars' great attentions. Double-fed motor has become the first choice of wind turbines owing to many advantages. The industrial and mining enterprises control the winding machine mostly with the method of rotor series connecting resistance. If vector control is used to control the winder by variable frequency speed regulation, energy-saving effect is obvious [1].

Switching AC/AC converter was used as the main circuit by Siemens for bidirectional power flow, but now it gradually withdraws from the market due to its low power factor, and serious pollution to the power grid. Although

* Corresponding author. Tel.:+86-516-83884395; fax: +86-516-83884395.

E-mail address: hanyaofei@126.com.

Han Yao-fei*, Tan Guo-jun, Li Hao, Ye Zong-bin, Wu Xuan-qin

School of Information and Electrical Engineering, China University of Mining and Technology, Xuzhou 221008, China

Abstract

1878-5220/09/$- See front matter © 2009 Published by Elsevier B.V. doi:10.1016/j.proeps.2009.09.223

with good performance, Matrix converter is not suitable in high power application because of higher switching frequency requirement [1].

Using the same switching frequency, three-level inverters can reach less harmonic content of the output voltage than two-level ones can. In practice, considering cost and reliability of device, two-level main circuit structure is mostly adopted when the output voltage is under 690V. At the voltage of above 1000V, the two-level inverter is liable for damage to equipments during a long term operation as a result of the impact on electrical equipment insulation at high dv/dt, as well as a higher harmonic content. Therefore, the three-level or multilevel inverter should be employed. As stated above, back-to-back dual three-level AC/DC/ AC converter with front-end controlled rectification as the main circuit is studied in this paper.

Nomenclature

Vsd, Vsq D-axis and Q-axis flux of the stator

V ^ V rq D-axis and Q-axis flux of the rotor

i rd , i rq D-axis and Q-axis voltage of the stator

ird,irq D-axis and Q-axis current of the rotor

urd,urq D-axis and Q-axis voltage of the rotor

the synchronous angular frequency and slip angular frequency

2. Proposed scheme

2.1. The vector control algorithm of double-fed induction motor

If different orientation method is used to achieve doubly-fed control, control structures and control performances are different. By means of stator voltage vectors control, grid voltage fluctuating easily causes jitter of orientation angle, which influences the control performance. There are two obvious advantages: no cross-coupled current and simple torque equation by adopting stator and rotor current oriented control, whereas expressions for rotor flux are complex. By means of rotor flux oriented control, rotor current itself is controlled variable, which causes observation of rotor flux difficult and effects control performance. Cross-coupled current is less by using stator vectors oriented control; Simple torque equation is the product of two scalars; Certain grid voltage fluctuating is allowed[3]. In consideration of the above, control strategy of rotor flux oriented control was used in this paper. The vector figure for the doubly-fed induction motor is shown in Figure 1.

Where the d-q axes are oriented to the stator flux space vector, the voltage equations of doubly-fed induction motor can be written as[2][3][4][5][6],

Rs LmRs .

usd = p^sd + V sd--L^rd 1

LmRs .

usq = ®1¥ sd--L^rq (2)

urd = Rrird + pv rd- , V rq (3)

u rq = Rrlrq + pV rq + ® , V rd (4)

Electromagnetic torque of motor is given by

T„ =-

3PL„

-V sA,

2 2 Ls 'sd rq The reactive power at stator side can be expressed as

Q » raiVsd(^ -T^d)

The vector control for double-fed induction motor is shown in Fig.2.

Fig. 1. Vector figure for the double-fed induction motor

Fig. 2. Diagram of vector control for double-fed induction motor

Flux observer based on voltage model is widely used because of steady network voltage and low harmonic.

Vsa = |(usa " rsisa )dt

^Vsp = f(usP - VsP )dt

The integral operation introduces the problems of integral drift and the initial value which must be solved in practice. Reference [1] presents the method of obtaining "moving average" to eliminate the influence of initial value, but the accuracy of average value depends on the number of sampling points and coincide degree between operation cycle and practical cycle. In Reference [7], three modified methods are proposed. Removing initial value, the first one brings in phase shift using low-pass filter instead of purely integral part. The latter two require heavier computation, especially in the third method orthogonal compensation is needed. The method based on the new magnetic flux observer in this paper is adopted, which can eliminate initial value and realize integral operation without phase shift. In addition, it is in favor of program realization. Equation (8) is transfer function of the novel stator-flux estimator:

+ 343s w + 9 sœ1 +

Fig. 3. Novel stator-flux estimator and pure integrator's waveforms

Fig. 4. Diagram of controlling three-level PWM rectifier

The diagram of integral effect based on the t novel stator-flux estimator, and pure integral part is shown as Fig.3 under the influence of integral initial value.

2.2. Control arithmetic based on three-level inverters and hardware implementation

Based on the voltage oriented control, feed forward decoupling rectifier and voltage, current dual close-loop feedback system, the control objective of this paper is (1) to maintain a constant DC-bus voltage with a good dynamic response, neutral-point potential balance; (2) to keep the line power factor at unity and make the current sine. The basic principle of rectifier is not given in this paper. Fig.4 shows the control structure of three-level PWM rectifier, waveform of rectifier is presented later.

2.3. Simplified arithmetic based on three-level inverters

In Reference [2], two-level inverter is controlled by current hysteresis control. High switching frequency is required to obtain lower current harmonic, moreover, switching frequency is not fixed. Thus, it is difficult to apply the method to high-power occasions.

The voltage space vector diagram of three-level inverter is more complex than that of two-level inverter. As for the control of voltage space vector, the former methods always divide a sector into four small triangles, and then seek the dwelling time of every effective vector for each small triangle accordingly. However, this calculation method is fussy and can't be easily applied to the three-level or higher-level inverter.

In this paper, a novel space-vector pulse width modulation (SVPWM) algorithm for three-level inverters is adopted [9].

The voltage space vector diagram of the three-level inverter shown in Fig.3 can be regarded as a hexagon composed by six small two-level space vectors, as shown in Fig.5.

Fig. 5. Simplification for three-level space vector diagram

In order to decide the location of the reference voltage space vector, we only need to judge corresponding small hexagon which is expressed as S. After deciding the small hexagon where the reference voltage space vector located, the three-level voltage vector plane can be simplified into two-level plane through coordinate shift. It is necessary to firstly correct the reference voltage space vector before the three-level voltage vector plane is transformed into two-level plane. Modification of reference voltage space vector is shown in Fig.6. Vref stands for the original reference voltage vector, while Vs_ref represents the corrected one.

Fig. 6 Modification of reference voltage space vector

Fig. 7. The alternative index S in overlapped regions

2.4. Control of neutral-point potential balance

When the two capacitors at DC side are different, the neutral-point potential is deviated because of inherent problem in the topology structure of the diode-clamping three-level converter. Even if the two capacitors are exactly the same, the various switch states also have different impacts on the neutral-point potential. The voltage space vectors of different module can be generated by different switching combination. According to the effect on neutralpoint potential by different switching status, these vectors will be divided into three classes: big vector, middle vector and small vector. The middle vectors can balance the neutral-point voltage if capacitance parameter is completely symmetric at the DC side. The middle vectors can also cause neutral-point voltage deviation because of asymmetric parameter in practice. Small vectors will cause fluctuation of neutral-point potential. There is a striking contrast effect between the two different middle vectors corresponding to different switch status vector[5]. For simplification, only the effect of small vectors was considered in this paper. In the following paragraphs, the corresponding short voltage vectors when the switch state causes the capacitor C1 discharged, the capacitor C 2 charged and the rise of neutral-point potential is known as the positive short vectors; Otherwise, it is called the negative short vectors (analyzes by taking the inversion mode of three-level inverter as example).

There exist the regions that are overlapped by adjacent small hexagons as shown in Fig.7. So if the reference voltage vector stays at those regions, S can have any possible values.

At this time the reference voltage space vector can be simplified from three-level plane to either two-level plane of S=1 or S=2. Vs1_ref is the corrected reference voltage vector when the index S has the value of 1, and Vs2_ref is the corrected reference voltage vector when the index S has the value of 2. If the two-level plane of S=1 is selected, the switching sequence decided by reference voltage vector is given as follows: (because the modulation waveform is central symmetric in the whole control cycle, only half of the sequence is listed):

(0-1-1)- (0 0-1)-(1 0-1)-(1 0 0)

V V V V

MN 2 N 8 y1p

T T T T

1N 2N 8 1p

V1N , V2N are respectively voltage space vectors of the corresponding switching sequence (0-1-1) and (0 01), and they are both negative short vectors; Likewise, voltage space vector V8 (1 0-1) is a middle vector, voltage space vector V1P (1 0 0) is a positive short vector voltage space vector. T1N , T2N, T8 , T1p are dwelling times of the corresponding voltage vectors. The voltage vector V1N and V1p have the same output voltage V1, and dwelling times of the voltage vectors V1N, V1P are equal: T1N = T1p . If the above switching

sequence is adopted, the dwelling time of negative short vectors is longer than the positive ones, the current will charge the upper capacitor, while discharging the lower capacitor, and neutral-point potential will decline. If the two-level plane of S=2 is selected, the switching sequence is given as follows:

(0 0-1)-(1 0-1)-(1 0 0)-(1 1 0) T T T T

2N 8 1P 2p

When the dwelling time of negative short vectors is shorter than the positive ones, the upper capacitor will be discharged, while the lower capacitor is charged, and neutral-point potential will rise.

Analyzing the other two-level planes, we can conclude that if the reference voltage vector stays at the overlapped region, the neutral-point potential can be controlled by changing the corresponding value of index S.

It should be noted that the inverter operates as rectifier and inverter, the positive short vectors and the negative ones are completely opposite. While the motor is double fed, the line-side and rotor-side inverters may operate in four quadrants. At the same time, the neutral-point potential should be controlled based on the status of inverters[8].

3. Industrial application

In this paper, dual three-level frequency control system (at 2kHz modulating frequency) is composed of the TMS320F2812 DSP and some relevant peripheral circuit as a controller and the IGBT (5SNA 1200E330100) as

switches. This system has been used in Yunjialing Mine, in Handan, Hebei Province. Fig.8 is the hardware diagram of dual three-level. The parameters of wound rotor induction motor are described below:

Motor type: YR1000-10/1430, rated power: 1000Kw; stator rated voltage: 6000V, stator rated current: 118A, rotor rated voltage: 1166V, rotor rated current: 536A, rated speed: 590r/min.

Fig. 8. Hardware diagram of dual three-level

line voltage of grid side eab(1800V/div )

phase current of grid side Ia(700A/div) line voltage of grid side ebc(1800V/div )

:>OOdQQO0Q0C

phase current of grid side Ib(700A/div)

Fig. 9 (a) Stator line voltage and phase current wave forms; (b) Line voltage and phase current wave forms of grid-side converter

In practical application, the waveforms in the following section are the test results at 1000V line voltage of inverter in line side and 2000V dc-link voltage. In Fig.9(a), the motor operates in sub-synchronous speeds. Usab , I sa, measured by transformer PT, is respectively line voltage and phase current, and the grid voltage is positive-sequence, thus Usab leads Isa by 30° (1.67ms) at unity power factor in stator side. As shown in Fig.9(a), the power factor is controlled to be 0.997 and approximate to 1.

For the voltage and current waveforms in Fig.9(b), current waveforms of D, Q axis in fig.10(a) and (b), using unipolar D/A output for observation, we superimpose a direct offset in order to simultaneously output the plus and minus value of real variables. And the zero point is artificially moved up 2.5 V.

In Fig. 9(b), the input line voltage of the converter in line side is negative sequence, therefore, e

ebc Were

lagging ea , eb by 30° respectively. From the figure, it can be seen that when the doubly-fed motor runs in sub-synchronous speeds, the line-side converter feeds to the grid and still maintains the unity power factor.

Fig. 10(a) shows that in the acceleration stage of motor, the D-axis current of line-side converter rapidly reverses and feeds back energy to the power grid. At this time there are very small fluctuations in the direct voltage and a balance of the neutral-point Potential (cyan for the D-axis rated current of the line-side converter, green for the D-axis feedback current of the line-side converter, blue for upper DC-bus voltage, purple for the lower DC-bus voltage).

dc _ up

DC voltage of loWer capacitor( 400V/div) U

C «UMMKMM M

Reference current of D- axis (214A/div) Id ref

Feedback current of D - axis (214A/div)

Reference current of Q- axis (214A/div) irq n

Feedbackcurrentof Q - axis (214A/div) time (10 s / div)

Fig. 10. (a) The upper side voltage and the lower side voltage of dc link and the D axis' current wave forms of grid-side converter; (b) The D axis' current of the grid-side converter and the Q axis' current of the motor side converter

In Fig.10(b), the motor started to accelerate from stable mode with heavy-load, then ran at constant speed, after its speed been slowed, crawled, and finally the motor stopped(in the figure cyan for the D-axis rated current of the line-side converter, green for the D-axis feedback current of the line-side converter, purple is the Q-axis rated current of the motor-side converter, green is the Q-axis feedback current of the motor-side converter). In the acceleration section, the slip power fed back energy to power grids from the rotor side, and Q-axis rotor current kept constant until its speed achieve the given value, so that the motor can be speeded up at the largest acceleration. The D-axis current of grid-link converter reversed, thus the converter can feed back power to the grid. As the speed rising, slip ratio gradually decreased, and the corresponding feedback slip energy also reduced as well as the D-axis current. When the motor operates in uniform motion, the speed is close to synchronous speed and slip is close to zero, so the slip energy and D-axis current of the line-side converter is also close to zero. In the decelerating section, while the rotor Q-axis current reduced and speed gradually reduced, the line-side converter feeds power to the grid and the D-axis current gradually increased. In the final stage of low-speed crawling, the motor runs at very low speed and the slip power is larger. As the final section of the waveform in the fig. 10 (b) shows, line-side converter feed back power to the power grid.

4. Conclusions

In this paper, a control strategy of vector control for the dual three-level double-fed induction motor is analyzed and studied in details. On the basis of waveforms in practical application, the following conclusions can be drawn:

• There exists fluctuation of the neutral-point potential due to the inherent problem in the topology structure of the three-level converter. As a result, harmonic content of output voltage increases, switches are pressed unbalanced, and lifetime of capacitors is shortened. The above circumstances cause great damage to the equipment. A novel space-vector pulse width modulation (SVPWM) algorithm for three-level inverters is adopted, which is simple to implement. And it is easy to control the neutral-point potential.

• Sinusoidal line-side currents, unity power factor, steady DC-bus voltage can be obtained by using the diode-clamping three-level PWM rectifier.

• With the method of three-level stator flux oriented control, double-fed motor has a rapid torque response and sinusoidal rotor current, and the stator-side power factor is 0.997.

Acknowledgements

Financial support for this work, provided by the Creative Activity Foundation for Postgraduate Students of Jiangsu Province, is gratefully acknowledged. The project number is CX07B_101z.

References

[ 1] S. Wu, Study on control system of double-fed induction motors. Huazhong University of Science and Technology, 2005.

[2] S. Wu, L. Zhou, Study of doubly-fed induction machine with AC/DC/AC converter. Proceedings of the CSEE. 19 (2005) 145-151.

[3] D. Zhou, R. Spee, A.K. Wallace, Model reference adaptive speed control for doubly-fed machines. IEEE. 2 (1993) 1175-1180.

[4] R. Datta,V.T. Ranganathan, Decoupled Control of active and reactive power for a grid-connected doubly-fed wound rotor induction machine without position sensors. IEEE. 4 (1999) 2623-2630.

[5] S. Wu, L.B. Zhou, S.H. Huang, A study on doubly-fed induction machine with bi-directional AC/DC/AC converter. IEEE. 2 (2004) 575581.

[6] B. Hopfensperger, D.J. Atkinson, R.A. Lakin, Stator-flux-oriented control of a doubly-fed induction machine with and without position encoder. IEEE. 147 (2000) 241-250.

[7] J. Hu and B. Wu, New integration algorithms for estimating motor flux over a wide speed range. IEEE Transactions on Power Electronics. 13 (1998) 969-977.

[8] C. Zhan, Study on mathematical model and system simulation of the three-level PWM reversible rectifier. Proceedings of the CSEE. 19 (1999) 45-48.

[9] J.H. Seo, C. Choi, D. Seok, A new simplified space-vector PWM method for three-level inverters. IEEE Trans. Power Electron, 4 (2001) 545-5471.