A Control Strategy for an Induction Motor Used for Vehicular Traction and/or Positioning Systems with Variable Speeds

Marian-tefan Nicolae* and Iustin-Radu Bojoi

*University of Craiova, Faculty of Electrical Engineering, Craiova, Romania, snicolae@elth.ucv.ro

Politecnico di Torino / Electrical Engineering Department, Torino, Italy, radu.bojoi@polito.it tronic interface (inverter) should convert the supply DC voltage to an AC voltage, in a 3 or multi-phase waveform with a sinusoidal first harmonic of variable frequency. A proper variation of the stator voltage frequency makes possible the obtaining of the desired torque-speed relation, necessary for the vehicle requirements. In particular: - the induction motor speed is adjusted by varying the three-phase voltage given by the inverter; - the torque adjustment is obtained through the supplying voltage phase variation. This modifies the induction motors slip, the result being a variation of the total absorbed current (or just of one of the currents components) in a bi-dimensional reference frame, set or rotating (vector control) [15] , [16]. The studied scheme uses a Direct-Flux Vector Control (DFVC) approach implemented in the stator flux reference frame[17]. The stator flux is directly controlled by the daxis component of the stator voltage vector, while the torque is controlled by regulating the q-axis component of the stator current vector. The direct flux control makes the proposed scheme very effective in terms of control of the motor voltage in the flux-weakening region, particularly exploited in traction, spindle drives and home appliances. The control strategy must fulfill the torque-speed relationship imposed by the vehicle requirements. Maximum torque production must be obtained for the entire speed range under current and voltage constraints imposed by the inverter [18]. II. MACHINE MODELING In the PTC algorithm the outer speed controller is the same as in the FOC, but the inner loops are replaced by a one-step FS-MPC (Finite State Model Predictive Control) of the stator flux and electromagnetic torque [19]. As in any FS-MPC, this algorithm includes a prediction of the outputs and an optimization stage. Additionally, as the stator flux is not directly measurable, it is necessary to make an estimation before the prediction, resulting in a three-stage algorithm: Flux Estimation, Flux and Torque Prediction, and Cost Function Optimization. The PTC scheme is shown in Fig. 1. In the following sections there will be described the more specific type of PTC, based on the DFVC scheme, in stator flux coordinates. A. The adopted reference frame As a general notation, the motor stator vectors (voltage, flux, current) will be called v , and i respectively, and the subscript s will refer to the stator flux reference frame when associated to those variables.

Abstract The paper presents a control strategy employed to control an induction motor (IM) used in vehicular traction applications with variable speed. This strategy is based on the Direct-Flux Vector Control (DFVC) method. In such applications, deep flux weakening is required, and the maximum torque production must be obtained under current and voltage constraints. Information on the mathematical model and control scheme for DFVC is provided. The reference frame is given by the stator flux, all the calculations being made with respect to this frame. The proposed control scheme is also implementing a limitation of the maximum values of current and voltage, according to specific control laws - MTPV (Maximum Torque per Volt) and MTPA (Maximum Torque per Ampere). Experimental results are presented for a small power IM, accelerating in idle running, with an inertial load and for the steady state respectively. They are showing the efficiency of the proposed solution and a robust operation in flux-weakening conditions. The controller is belonging to the dSPACE family , being able to provide very fast floating point results. The proposed strategy can be used for controlling other types of motors, as the Interior Permanent-Magnet (IPM), Synchronous Reluctance (SyR) and wound rotor synchronous motors, considering the modifications imposed by the magnetic model (load angle, flux orientation a.s.o.).

I. INTRODUCTION In recent decades, control of electrical drives has been widely studied [1]. Linear methods like proportional integral (PI) controllers using Pulse Width Modulation (PWM) and non-linear methods such as hysteresis control have been fully documented in the literature and dominate high performance industrial applications [2],[3]. Regarding the possible vector control schemes, the most widely used linear strategy in medium and low power electrical drives is field-oriented control (FOC) using inner current control loops [4]-[9]. Other solutions are based on direct torque control (DTC) approach with direct flux and torque regulation [10]-[12]. A more specific type of DTC is PTC (Predictive Torque Control), used for the cost function optimization in flux and torque prediction [13]. The IM is a type of AC motor very useful in industrial applications, robust, characterized by a high productivity and efficiency, with a simple design and not very expensive [14]. The IM has a power and a specific torque greater than DC motors, due to the absence of the collector and the greater resulting speed it can reach, with the conventional technologies, values of 15-20,000 rpm. This is an advantage, as DC motors can reach 60007000 rpm. For traction applications, the power conditioning elec-

978-1-4673-1810-5/12/$31.00 2012 IEEE

DIRECT-FLUX VECTOR CONTROL IN STATOR FLUX COORDINATES The Direct-Flux Vector Control in Stator Coordinates (DFVC) method is detailed in [7]. The voltage model (6) and torque equation (7) in the stator flux frame are: III.

v dqs = R s i dqs +
Fig. 1. Block diagram of PTC

0 d d + + dt 0 dt
3 p iqs , 2

(6)

T=

(7)

q qs

ds r d , dm m, m

where is the load angle i.e. the phase angle of the stator flux with respect to the rotor. From (6) it results that the stator flux amplitude can be regulated by means of the ds-axis voltage and the load angle (i.e. torque) by means of the qs-axis voltage. However, the torque expression (7) suggests that the control of the qs-axis current instead of the load angle may lead to a simple torque control scheme [10]. Further manipulation of the quadrature state equation in (6) is necessary. A. Maximum Torque per Voltage Operation (MTPV) The MTPV operation occurs in the flux-weakening region at high speed , being called also voltage limited operation [11]. Given the voltage limit, over a certain speed it is no longer convenient to exploit the full inverter current Imax for obtaining the maximum motor torque, because the pull-out torque limit has been reached. The pull-out torque condition coincides with a specific value of the load angle . Such value will be indicated as max from now on and varies with the motor type. Below the pull-out limit the flux phase angle is < max. The expression of max for the considered motors is obtained by expressing the electromagnetic torque (3) as a function of the flux amplitude and phase, and then posing to zero the torque partial derivative with respect to , that is at constant flux amplitude. B. Pull out torque angle of the IM In rotor flux coordinates, the substitution of r = Lm id in (3) leads to

Fig. 2. Definition of common reference frames for the Induction Motor

The adopted vector reference frames are defined in Fig. 2: stationary frame (,), rotor frame (dm,qm), rotor flux frame (d,q) and stator flux frame (ds,qs). The d-q dynamic model of squirrel cage induction motor with a reference frame fixed to the stator [20] is given by:
v s = Rs i s + ds + j s Ms , dt

(1)

dr (2) + ( s r ) Mr dt The electromagnetic torque is found as: 2 pLm (3) T= (iqs idr ids iqr ) , 3Lr where Rs is the stator resistance, Rr is the rotor resistance, Lm is the magnetizing inductance, and is the synchronous speed and p is the number of pole-pairs. In the following subsections, the magnetic model of the considered motor will be briefly summarized, expressed in the rotor mechanical frame (dm, qm), different by the rotor synchronous frame (d, q). vr = Rr ir +

dq = s idq . 0 Ls
From (2) and (8) the torque expression becomes:

(8)

B. Magnetic Model The current to flux relationship for the IM is given by Eq. (4), where the rotor flux can be derived from the rotor equation (5) [17]:

T=

3 2

1 2 sin 2 . Ls 2

(9)

= k r r + Ls i ,
r
dr ,dqm + r ,dqm = Lm idqm ,

(4) (5)

The torque derivative with respect to is

dt where is the leakage factor, Ls is the stator inductance, r = Lr/Rr is the rotor time constant and kr = Lm/Lr is the rotor coupling factor. As already mentioned, the rotor reference frame (dm,qm) has been adopted in (5).

T =

3 2

1 2 cos 2 . Ls

(10)

The MTPV load angle is obtained when (10) is zero, i.e.

max , IM = 45 .

(11)

IV. DFVC IMPLEMENTATION A. About DFVC The proposed DFVC block diagram is shown in Fig. 3. The control scheme (DFVC block from Fig. 3) is presented in Fig. 4 for a speed-controlled AC drive. The flux and quadrature current references (*, iqs*) are computed from the torque reference according to Eq. (6). Three regulators are needed: two for the flux and current regulation, taking into account the inverter voltage and current limits, as well as one for the load angle limitation, as presented in [21]. All of them are included in the DFVC block from Fig. 3. The flux set-point at low speed can be either a constant value or a function of the torque set-point. In Fig. 4 the maximum torque per Ampere (MTPA) law has been chosen, but simpler functions are possible with no significant side-effects. B. Maximum voltage limitation The maximum voltage limit is respected by limiting the flux reference according to the synchronous speed and the actual DC link voltage (1).

C. Maximum current limitation The motor phase current is limited by saturating the quadrature current reference according to Eq. (12), where Imax is the inverter maximum current.
* 2 2 iqs I max ids

(13)

D. Maximum torque per voltage (MTPV) current limitation The maximum voltage limit is obeyed by limiting the flux reference according to the synchronous speed and the actual DC link voltage (11):

* (Vmax Rs i qs ) | | ,

(14)

* (Vmax Rsiqs ) | | ,

(12)

where Vmax is the inverter maximum voltage that is updated according to the measured DC link voltage. The setting of Vmax determines whether the inverter over-modulation region is exploited or not. Typical settings are Vmax = 0.9 -1.1 vdc/3, where vdc is the measured DC link voltage. + vdc * +

T*

* v

PI

DFVC

* vabc

PWM

vdc
, s
i
dt v

Stator Flux Observer Speed computation m

2 2

iabc
3 3
dt vabc

where Vmax is the inverter maximum voltage that is updated according to the measured DC link voltage [4]-[6]. The setting of Vmax determines whether the inverter overmodulation region is exploited or not. Typical settings are Vmax = 0.9 -1.1 vdc/3, where vdc is the measured DC link voltage. The MTPV operation can be obtained by limiting the load angle to be under max, that depends of the motor type. There is a tight relationship between the iqs current component and the load angle . In particular, the limitation of the iqs current produces a limitation of the load angle. For this reason, the MTPV operation is obtained by limiting the iqs current reference, according to the observed load angle. While the just described Imax saturation keeps the current inside the inverter limit, this further iqs limitation is performed by the same saturation block (see Fig. 4) and represented by the current component iMTPV, as in [16]. This, subtracted to the Imax limit (13), leads to a reduction of the current amplitude below the inverter maximum rating (|i| < Imax), as requested in pull-out torque conditions. The max angle limits have been calculated in [10] and in (11). V. EXPERIMENTAL RESULTS

DT

IM

A. Experimental Setup The experimental tests have been performed with an IM whose parameters are reported in Appendix. The inverter is rated 10A (peak) output current, with 220V/50Hz single phase AC input and passive rectifier. The controller is a floating point dSPACE board. The three phase currents, DC-link voltage and rotor position are measured at a sampling frequency of 10 kHz that is also the PWM switching frequency. The motor position is measured by a standard incremental encoder with 512 pulses per revolution. B. Test on acceleration up to 14,000 rpm in idle running Figures 5-7 depict the experimental data gathered during the test on acceleration up to 14,000 rpm in idle running. The following remarks can be done for Fig. 5: - the reference speed has a fast growth slope; the reference torque generated by the speed regulator increases to 0.4 Nm; the motor flux is increased up to the value corresponding to 0.4 Nm, based on the MTPA law; - at a specific moment (0.2 seconds) the operation point reaches the flux weakening region, because the speed has

Fig. 3. DFVC for IM

vdc

Max voltage

iqs

Vs,max
Reference flux generation

*
* iqs

T*
i
s

2 3p

ids
iqs,max

, iqs control

* v

dq

Max torque current computation

Fig. 4. DFVC Control Scheme

Fig. 5. (a) Mechanical speed, reference speed and the reference speed in idle running; (b) The flux waveform given by MTPA, the flux limitation and the flux given by the observer position; c) Torque current given by the q-axis component, its limit and maximum values; (d) Load angle delta; (e) The reference and filtered electromagnetic torque

Fig. 7. (a) Torque current in the q-axis, its limiting and maximum values; (b) The load angle; (c) The stator a-phase current

increased and there is not enough voltage for an MTPA operation. The flux is decreased, so the flux reference follows the calculated value by the de-weakening block (lim); - the torque current (iqs) has a value depending on the reference torque and the motor flux. Considering that the flux is reduced, the current tends to increase, but at a given time the torque current is limited by the value depending on the maximum motor current (iqsCrtLim); - the mechanical speed profile is very similar with the electrical one, because when there is no load the motor does not lose its alignment, and the slip is small; - when the load angle reaches the value of 45 degrees (at 0.65 seconds), to avoid the instability (pull-out), the torque current is reduced in order to preserve the load angle to the prescribed value. The following remarks can be done for Fig. 6: - in the first part of the acceleration stage, both mechanical power and speed increase. At a given time, the mechanical power records a maximal admitted value, as the load angle is limited relatively quickly. When the motor speed reaches the reference speed, the mechanical power has a descending slope; - the DC link voltage is not constant, depending on the power absorbed from the inverter. The average value decreases as the mechanical power generated at the motor output increases; in addition, a voltage ripple of 100 Hz occurs due to the single-phase rectifier supplying the inverter.

Fig. 7 clearly shows the correlation between the torque current, the mechanical power and the phase current of the motor. At an instant approaching 0.2 seconds, the torque current is limited in order to limit the motor current. Indeed, over this period the machine current is limited to 3A peak value. After a limitation of the load angle, the phase current is limited by a maximum threshold. C. Speed revearsal The behavior of the motor in the case of a speed reversal is illustrated in Fig. 8-10. The reference speed has been set at 1000 rpm and a step load (equal to about the motor rated torque value) has been applied to motor, followed by the load removal. The motor flux has been set at a constant value equal to the rated flux of 0.35 Vs. The following remarks can be done for Fig. 8: - the reference speed given by the slip (the electrical speed) has a very fast increasing slope, and the reference torque generated by the speed regulator has an approximately constant value of 0.35 Vs. The machine flux is increased to the value corresponding to 0.35 Vs based on the MTPA law. The reference speed has an almost constant value of 1,000 rpm, while the mechanical speed has a sudden fall after 0.7 seconds, increasing again after 3.8 seconds, as the DC motor representing the inertial load changes its rotating sense. Due to the significant resistive torque (2 Nm), the motor stops and changes, theoretically, its rotation sense; - due to the significant inertial load, the flux weakening is no longer necessary (the flux doesnt reach lim), so the flux current must not be reduced; - when the DC motor rotation direction is reversed, the flux current decreases suddenly to 0 and the IM rotates along with the DC motor; - the torque current (iqs) has a value depending on the reference torque and the machine flux. Considering that the flux, in this case, is constant, the torque current tends to increase due to MTPV law, but at a given time the torque current is limited to 2 Nm by the value depending on the load angle . From Fig. 9 one can see that: - on the first part of the acceleration, the power is increasing, along with the electrical speed. At a given time, the mechanical power reaches a peak, and then it is limited due to an inertial load, so it starts decreasing;

Fig. 6. (a) Mechanical speed , reference speed and electrical speed in the case of idle-running accelerating; (b) Mechanical power; (c) Motor supply voltage

Fig. 8. (a) Mechanical speed, reference speed and the reference speed when the speed is reversed; (b) The flux waveform given by MTPA, the flux limitation and the flux given by the observer position; c) Torque current given by the q-axis component, its limit and maximum values; (d) Load angle delta; (e) The reference and filtered electromagnetic torque

Fig. 11. (a) Mechanical speed, reference speed and the reference speed when the drive is in a steady state; (b) The flux waveform given by MTPA, the flux limitation and the flux given by the observer position; c) Torque current given by the q-axis component, its limit and maximum values; (d) Load angle delta; (e) The reference and filtered electromagnetic torque

Fig. 9. (a) Mechanical speed , reference speed and electrical speed when the speed is reversed; (b) Mechanical power; (c) The motor supply voltage

- the DC link voltage is not constant it depends on the absorbed inverter power. The average value decreases with the increase of the mechanical power generated at the motor output. Fig. 10 shows the correlation between the torque current, the load angle and the machine phase current. The torque current is limited in order to limit the machine current. Indeed, the a-phase current is limited to 5 A as a peak value. When the speed of the DC motor is reversed, the phase current is reduced to a value of approximately 3 A as peak value, because the torque current is no longer needed for driving the existent inertial load. D. Stable state behavior When the load (a DC motor) is giving a resistive load

that has the same characteristics as the induction motor, an unusual phenomena takes place: the drive will stop, and the characteristics will wary in an arbitrary manner. In Fig. 11 the reference speed given by the slip and the mechanical speed are limited to 1,000 rpm and the rated flux is 35 Vs; the current limits are not exceeded, the load angle has small values, as the motor never really gets to accelerate and the motor torque is always following the rated one (having a sinusoidal form). In Fig. 12 one can see that a great amount of mechanical power is wasted, while the DC-link voltage is having a ripple of about 15 V, but the motor is still stable. Figure 13 shows the correlation between the load angle and the "a"-phase current, as the ripple in the load angle is generated by the current overmodulation.

Fig. 12. (a) Mechanical speed , reference speed and electrical speed for a steady state; (b) Mechanical power; (c) The motor supply voltage

Fig. 10. (a) The torque current in the q-axis, his limiting and maximum values; (b) The load angle; (c) The stator a-phase current

Fig. 13. (a) The torque current in the q-axis, his limiting and maximum values; (b) The load angle; (c) The stator a- phase current

TABLE I RELATIONSHIP BETWEEN DC AND MOTOR LINE CURRENTS Motor type IM Rated torque (Nm) 2.6 Rated phase current (A pk) 4.2 Rated inverter current (A pk) 8 Rated line to line voltage (V pk) 270 Base speed (rpm) 3200 Maximum speed (rpm) 16000 Pole pairs 1

[5] [6]

[7]

[8]

APPENDIX The motor under test is a 700W commercial induction motor (IM) described in Table I. VI. CONCLUSION The current and voltage limits are fully exploited by limiting the flux and torque current references with simple control laws that are independent of the motor parameters and valid also in case of a variable DC link. The control algorithm requires three PI regulators: two are for flux and iqs current control, which controls the torque, and one is for the load angle limitation at high speed. The DFVC control scheme for the Induction Motor presented in this paper is suited for applications concerning vehicular traction and positioning systems with variable speeds. Experimental results are presented for the motor accelerating without any loading, and with an inertial load, showing the feasibility of the proposed control, and in particular the robust flux-weakening operation. ACKNOWLEDGEMENT Special thanks to the academic staff from Politecnico di Torino, who made possible all the experiments. REFERENCES
[1] J. Rodriguez, R. M. Kennel, J. R. Espinoza, M. Trincado, C. A. Silva, C. A. Rojas, "High-Performance Control Strategies for Electrical Drives: An Experimental Assessment", IEEE Transactions on Industrial Electronics, Vol. 59, No. 2, February 2012, pp. 812-820. M. Kazmierkowski, R. Krisnan, F. Blaabjerg, Control in Power Electronics, New York: Academic, 2002. N. Mohan, T. Underland, W. Robbins, Power Electronics, 3rd ed. New York: Wiley, 2003. C. Lascu, I. Boldea, F. Blaabjerg. Variablestructure direct torque control A class of fast and robust controllers for induction

[9]

[10] [11]

[12]

[13]

[14]

[15]

[16]

[17]

[18]

[19]

[2] [3] [4]

[20]

[21]

motor drives, IEEE Trans. on Indust. Electron., Vol. 41, No.4,pp. 785-792, Aug 2004. W. Leonhard, Control of Electrical Drives. Berlin: SpringerVerlag, 1996. B. Drury, W. Drury, The control techniques drives and controls handbook, The Institution of Lectrical Enginnering, (Ed.), 2001. ISBN: 0-85296-793-4 & 978-0-85296-793-5, London, UK. M.P. Kazmierkowski and W. Sulkowski, A novel vector control scheme for transistor PWM inverter-fed induction motor drive, IEEE Trans. On Ind. Electron. Vol.38, Issue 1, 1991, pp. 41-47. D.-Ch. Lee, S.-K. Sul and M.-H. Park, High performance current regulator for a field-oriented controlled induction motor drive, IEEE Trans. On Ind. Applicat., Vol.30, Issue 5, 1994, pp.12471257. Y.-T. Kao and C.-H. Liu, Analysis and design of microprocessorbased vector-controlled induction motor drives, IEEE Trans. On Ind. Electron., Vol. 39, Issue 1, 1992, pp-46-54. J. Holtz, Sensorless control of induction motor drives Proceedings of the IEEE, vol.90, no.8, pp. 1359- 1394, Aug 2002. Takahashi and Y. Ohmori, High-performance direct torque control of an induction motor, IEEE Trans. On Ind. Applicat., Vol. 25,1989, pp. 257264. G. S. Buja, M. P. Kazmierkowski, Direct Torque Control of PWM Inverter-Fed AC Motors A Survey, IEEE Trans. Ind. Electron., Vol.51, No. 4, pp. 744-757. Aug. 2004. R. Kennel, J. Rodriguez, J. R. Espinoza, M. Trincado, "Highperformance speed control methods for electrical machines: An Assessment", Proc. IEEE ICIT, Mar. 14-17, 2010, pp. 1793-1799. M. J. Melfi, S. Evon, R. McElveen, Induction versus permanent magnets motors , IEEE Industry Applications Magazine, Nov/Dec 2009. L. Harnefors, M. Jansson, R. Ottersen and K. Pietilinen, Unified Sensorless Vector Control of Synchronous and Induction Motors, IEEE Trans. Ind. Electron.,vol.50, No.1, February 2003, pp.153160. Pellegrino G., Bojoi I.R., Guglielmi P.: Unified Direct-Flux Vector Control for AC motor drives, Energy Conversion Congress and Exposition (ECCE), 2010 IEEE (Atlanta, Georgia, USA) 12-16 Sept. 2010 pp.8 (pp.1150-1157). G. Pellegrino, R. Bojoi and P. Guglielmi, Unified Direct-Flux Vector Control for AC motor drives, IEEE Trans. On Ind. Applicat., Vol.47, Issue 5, 2011, pp.2093-2102. Z. Q. Zhu, D. Howe, Electrical machines and drives for electric, hybrid, and fuel cell vehicles, Proc. IEEE, vol.95, no.4, pp.746765, 2007. R. Kennel, A. Linder, "Predictive control of inverter supplied electrical drives", IEEE PROC. 31st PESC, June 18-23, 2000, vol. 2, pp. 761-766. N. Kalaiarasi and K.S.R. Anjaneyulu , Adaptive Vector Control of Induction Motor Drives . International Journal of Electrical and Power Engineering, 1: 239-245, 2007. Nicolae M.S., Bojoi I.R., Nicolae P.M., Direct-Flux Vector Control of Induction Motor For Light Traction, EPE-PEMC 2012 ECCE Europe, Novi Sad, Serbia, 4-6 September 2012.