Recent Developments of Electrical Drives - Part 19 pot

Setting a value of the energy threshold, and applying a load torque that makes the motor consumes a current over the rated one, it can be observed that the proposed protection permits an

Figure 7 Experimental rig.

In order to obtain the real speed, a simple algorithm is design that allows selecting the integration step and an accurate speed measurement is achieved Since the real speed is introduce into the Simulink design, it is immediate to compare the real speed with the reference and so the control can be rigorously tested

Experimental results

In the experimental results, tests are carried out to see the performance in steady state, transient state, and overcurrent states In this way the different aspects of the proposed scheme can be tested

Concerning the proposed protection two load tests were carried out, one limiting the value

of i qto its rated value and the other with the designed protection (Fig 8) Setting a value of the energy threshold, and applying a load torque that makes the motor consumes a current over the rated one, it can be observed that the proposed protection permits an overcurrent for a certain time, allowing the control to achieve the reference speed (1,000 rpm) in the first seconds When the energy counter reaches the energy threshold, the protection acts and the speed falls to a value so that the current is the rated one

The performance is improved in this way since the target values can be followed even when certain transients overcurrents are required

In order to verify the steady-state behavior of the proposed system a constant load test was carried out during time enough to allow the motor heating The test is carried out considering a target speed of 1,000 rpm and a load torque of approximately half the rated value, and it is made with and without the inclusion of the thermal model in the control scheme Both tests are performed in the same conditions without changing the value of the resistance in the resistor bank The results of both tests are shown in Fig 9, where the evolution of the measured motor speed is displayed The vertical lines are due to the reset

of the incremental encoder position counter

If the thermal effect is neglected and constant parameters are considered, there is a deviation due to the effect of the motor heating On the contrary, if the thermal effect is

Figure 8 Protection test.

considered, parameters are updated properly as the motor temperature increases and the target speed is followed without deviation

A requirement of the dynamical behavior is the response when sudden changes in the target speed occur In order to be demanding with the control features, it is consider a test with a target speed going from 0 to 1,000 rpm in 1.3 s Fig 10 shows that even with this acceleration, the motor follows the target speed just with some oscillations in the starting (7.2 s) and braking (8.5 s) Carrying out the same test without considering the deep-bar effect leads to higher oscillations and poorer transient response and not considering the static friction in the mechanical equation also makes the motor oscillate more in the starting Vector control is used because of the good dynamic performance, and a normal test is also

to apply a sharp load torque to verify the system response In this case during a few seconds

a nominal load torque was applied and in 28th second, approximately, the load torque was

Figure 9 Constant load test.

Figure 10 Speed test.

released and so the motor is instantaneously accelerated (Fig 11) The maximum error in this test is 16 rpm and motor speed goes quickly to the target value (1,000 rpm) In the same test without the deep-bar effect model, this error was 20 rpm, what points out the relevance

of including parameter variation due to this effect

In Fig 12 the estimated and measured quadrature current are shown The evolution is similar, so that the correct information about the load torque is being provided to the speed estimator The rapid change in the quadrature current when changes in the load torque occur

is the clue to obtain a quick response of the control

Until the 28th second the torque is gradually being increased and this information is introduced in the controller thank to the measured quadrature current When the torque is released the quadrature component is also suddenly changed by the control following in this way the target speed

Figure 11 Load test.

Figure 12 Evolution quadrature current in load test.

Conclusions

The scheme that has been proposed improves the performance of the drive in several ways

On the one hand the thermal state estimation can correct the steady-state deviation in the motor speed that otherwise is produced when the motor is heated and parameter detuning occur Another improvement of the scheme is the inclusion of the skin effect estimation and the consideration of the static friction that allow to obtain a good performance both against speed reference or load torque changes Since apart from the control characteristics it is necessary to avoid overcurrents, the proposed protection proves to permit transient currents over the rated value improving the drive performance Not saturating directly the current helps the motor to reach the target values even with high transients torque required All the improvements have been tested experimentally and with high accuracy measure-ments, validating the effectiveness of the proposed solution

List of symbols

b s , b r , b sr Stator-environment, rotor-environment, and stator-rotor convection

coefficients

C s , C r Stator and rotor thermal capacitances

G s , G r , G sr Stator-environment, rotor-environment, and stator-rotor thermal

conductances

i d , i q Direct and quadrature components of stator current space vector

i mr,ω mr Modulus and angular speed of rotor magnetizing current space vector

k H , k F Hysteresis and eddy current coefficients

L r , L m Rotor self-inductance and magnetizing inductance

[2] T Naguchi, I Takahashi, A new quick-response and high-efficiency control strategy of an induction motor, IEEE Trans Ind Appl., Vol IA-22, pp 820–827, 1986

[3] E.Y.Y Ho, P.C Sen, Decoupling control of induction motor drives, IEEE Trans Ind Electron., Vol 35, pp 253–262, 1998

[4] J Holtz, J Quan, Sensorless vector control of induction motors at very low speed using a nonlinear inverter model and parameter identification, IEEE Trans Ind Appl., Vol 38, No 4, 2002

[5] M Wang, E Levi, Evaluation of steady-state and transient behaviour of a MRAS based sen-sorless rotor flux oriented induction machine in the presence of parameter detuning, Elect Mach Power Syst., Vol 27, No 11, pp 1171–1190, 1999

[6] M.N Marwali, A Keyhani, “A Comparative Study of Rotor Flux Based MRAS and Back EMF Based MRAS Speed Estimators for Speed Sensorless Vector Control of Induction Machines”, Proc IEEE Ind Appl Soc Annu Meet IAS’97, New Orleans, LA, 1997, pp 160–166 [7] J Fern´andez Moreno, F P´erez Hidalgo, M.J Dur´an Mart´ınez, Realization of tests to determine the parameters of the thermal model of induction machine, IEE Proc Electr Power Appl., Vol 148, pp 392–397, 2001

[8] M.J Dur´an, J.L Dur´an, F P´erez, J Fern´andez, “Improved Sensorless Induction Machine Vector Control with On-line Parameter Estimation Taking into Account Deep-Bar and Thermal Effects”, 28th Annual Conference of the IEEE Ind Electron Soc IECON, Sevilla, 2002 [9] P.L Alger, Induction Machines, Gordon and Breach Science Publishers, New York, 2nd edition, 1970

[10] W Levy, C.F Landy, M.D McCulloch, Improved models for the simulation of deep bar induc-tion motors, IEEE Trans Energy Convers., Vol EC-5, No 2, pp 393–400, 1990

II-4 WIDE-SPEED OPERATION OF DIRECT TORQUE-CONTROLLED

INTERIOR PERMANENT-MAGNET

SYNCHRONOUS MOTORS

Adina Muntean1, M.M Radulescu1and A Miraoui2

1Small Electric Motors and Electric Traction (SEMET) Group, Technical University of

Cluj-Napoca, P.O Box 45, RO-400110 Cluj-Napoca 1, Romania

adina.muntean@mae.utcluj.ro, mircea.radulescu@mae.utcluj.ro

2Laboratory of Electronics, Electrotechnics and Systems (L2ES), University of Technology of

Belfort-Montb´eliard, rue Thierry-Mieg, F-90010 Belfort, France

abdellatif.miraoui@utbm.fr

Abstract In this paper, an integrated design and direct torque control (DTC) of inverter-fed

inte-rior permanent-magnet synchronous motors (IPMSMs) for wide-speed operation with high torque capability is presented The double-layer IPM-rotor design is accounted for IPMSMs requiring a wide torque-speed envelope A novel approach for the generation of the reference stator flux-linkage magnitude is developed in the proposed IPMSM DTC scheme to insure extended torque-speed en-velope with maximum-torque-per-stator-current operation range below the base speed as well as constant-power flux-weakening and maximum-torque-per-stator-flux operation regions above the base speed Simulation results to show the effectiveness of the proposed DTC scheme are provided and discussed

Introduction

Due to their many positive features, including high torque-to-inertia and power-to-weight ra-tios, fast dynamics, compact design, and low maintenance, inverter-fed interior permanent-magnet synchronous motors (IPMSMs) are viable contenders for industrial drives with high torque capability over a wide-speed range Indeed, PMs being completely embedded inside the steel rotor core, a mechanically robust construction of IPMSMs allowing wide speed-torque envelope is primarily obtained Secondly, the rotor-buried PMs, covered by steel pole-pieces, significantly change the magnetic circuit of the motor, since, on the one hand, the PM cavities create flux barriers within the rotor, thus reducing the permeance in a flux direction that crosses these cavities, and, on the other hand, high-permeance paths are created for the flux across the steel rotor-poles and also in space-quadrature to the rotor-PM flux; this establishes the rotor magnetic saliency Hence, it is a hybrid torque production mechanism in IPMSMs, because in addition to the magnet (or field-alignment) torque due

to the interaction of rotor-PM flux and the armature (stator) mmf, there is also a reluctance torque component due to rotor magnetic saliency Thirdly, IPMSM having a small effective

S Wiak, M Dems, K Kom˛eza (eds.), Recent Developments of Electrical Drives, 177–186.

2006 Springer.

dependence on motor parameters as well as the fast torque response in steady-state and transient operating conditions

In this paper, an integrated design and DTC of VSI-fed IPMSMs for wide-speed op-eration with high torque capability is presented Hence, the paper is organized as fol-lows In “IPMSM Design for WiSpeed Operation,” the double-layer IPM-rotor de-sign is adopted for IPMSMs requiring a wide torque-speed envelope In “DTC of VSI-fed IPMSM for Wide-Speed Operation,” an IPMSM DTC scheme incorporating both the optimized constant-torque and flux-weakening controllers for wide-speed range op-eration is developed Simulation results to validate the proposed IPMSM DTC scheme are presented and discussed in “Simulation Results.” Conclusions are drawn in section

“Conclusions.”

IPMSM design for wide-speed operation

The stator of the considered VSI-fed IPMSM is a typical AC design accommodating a three-phase distributed winding in slots to produce the synchronously-rotating, quasi-sinusoidal armature-mmf wave Conversely, the IPMSM rotor can be designed in different configu-rations However, only two of them with radially-magnetized buried-type IPMs have been accounted as being advantageous for wide-speed operation [8–10] The high-energy rotor-PMs usually consist of sintered-NdFeB blocks inserted after magnetization into the rotor cavities

Fig 1 shows the cross-sectional configurations of both IPMSMs in conjunction with their rated-load magnetic flux distribution obtained from finite element analysis The first IPMSM rotor topology has only one (single-layer) PM per rotor-pole, whereas in the second one, each rotor-PM is splitted up in two layers with iron separation in the radial direction

of the rotor core

The well-known coordinate system (d,q) bounded to the rotor (i.e rotating at

syn-chronous speedω r ) is defined hereafter with the d-axis aligned with the stator PM

flux-linkage vectorψ s0 = ψ PM and the orthogonal q-axis aligned with the back-emf vector

ω r ψ PM(Fig 2) By noticing that the (total) stator flux-linkage vector can be splitted into the flux-linkage (with the stator winding) due to the excitation rotor-PMs,ψ PM , and the armature-reaction flux, which entails the self-inductances L and L (L < L ) of the

(b)

Figure 1 IPMSM cross-sectional design and magnetic flux distribution under rated-load condition

for (a) single- and (b) double-layer IPM-rotor topology, respectively

Figure 2 Different coordinate systems for vector representation of IPMSM quantities.

A comparison between the two IPMSM rotor designs of Fig 1, for constant

rotor-PM volume and for identical magnetic properties, rotor outer diameter, airgap, and stator specifications, has been made in order to select the most suitable structure for high-torque wide-speed operation As result of this comparison based on finite-element magnetic field analysis of both IPMSMs, the double-layer IPM-rotor design has been adopted for motor prototype by the following reasons

1 The d-axis stator self-inductance L sdis low and roughly the same for both single- and double-layer PM-rotor configurations

2 The q-axis stator self-inductance L sq and, correspondingly, the inductance difference

L sq − L sdfor the double-layer IPM rotor is up to 20% greater than for the single-layer

IPM rotor, mainly due to the additional q-axis flux path provided between the two

rotor-PM layers

3 The q-axis stator self-inductance L sqfor the rotor topology with only one PM per pole decreases greatly with the stator-current rising, because of the magnetic saturation, whereas for the double-layer PM-rotor topology this effect is less significant

4 The stator flux-linkage due to the double-layer of rotor-PMs is about 10% greater than

in the case of single-layer IPM rotor

5 The electromagnetic torque developed up to the rated rotor speed by the double-layer IPMSM is about 10% increased in comparison with that produced by a single-layer IPMSM, for the same armature mmf However, the torque performances using flux-weakening at high speeds for both IPMSMs are quite similar

DTC of VSI-fed IPMSM for wide-speed operation

In the DTC scheme for VSI-fed IPMSM, the inner torque controller is based on the ex-pression of the electromagnetic torque given by equation (1) Hence, torque is controlled

by regulating (through inverter voltages) the amplitude|ψ s | and the angle δ of the stator

flux-linkage vector

The d- and q-axis stator flux-linkages are

Figure 3 Block diagram of the VSI-fed IPMSM DTC scheme for wide-speed operation.

From equations (2) and (3), the stator flux-linkage vector modulus can be expressed as

|ψ s | = (ψ2

sd + ψ2

sq)1/2 = [(L sd i sd + |ψ PM|)2+ (L sq i sq)2]1/2 (4)

By differentiating equation (1) with respect to time, for constant stator flux-linkage magni-tude, one obtains

dm e /dt = (3p/2)|ψ s |(|ψ PM | cos δ − ξ|ψ s | cos 2δ)(dδ/dt)/L sq(1− ξ) (5) Equation (5) emphasizes that the electromagnetic torque can be dynamically controlled by means of controlling the rate of change of the angleδ.

There are upper limits of variation for both control quantities,|ψ s | and δ, to achieve stable IPMSM DTC Firstly, since according to equation (1), m e = 0 for δ = 0, the condition for positive slope dm e/dδ around δ = 0 leads to

Secondly, by differentiating equation (1) with respect to δ and equating it to zero, the

maximum allowable angleδ limcan be found as

δ lim= cos−1{|ψ PM |/4ξ|ψ s | − [(|ψ PM |/4ξ|ψ s|)2+ 1/2]1/2 (7)

that is

The block diagram of the proposed DTC scheme for wide-speed operation with high torque capability of a VSI-fed IPMSM is shown in Fig 3 The three-phase stator variables are transformed to theα,β-axes variables of the (α,β) stationary coordinate system shown in

Fig 2

Theα, β stator currents, obtained from current sensors, and the stator voltages u s αand

u s β, calculated from the measured DC-link voltage, are then used for stator flux-linkage

vector and electromagnetic torque estimation Some methods of compensation for the effect

of stator-resistance variation and for the DC offset in the measurements, particularly at