1. Trang chủ
  2. » Thể loại khác

Strategy for the instantaneous torque control of permanent magnet brushless DC drives2

9 352 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Tiêu đề Strategy for the instantaneous torque control of permanent magnet brushless DC drives
Tác giả T. S. Low, K. J. Tseng, T. H. Lee, K. W. Lim, K. S. Lock
Trường học National University of Singapore
Chuyên ngành Electrical Engineering
Thể loại Journal article
Năm xuất bản 1990
Định dạng
Số trang 9
Dung lượng 613,85 KB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

Experimental results for drive applications using instantaneous torque control and conventional sinusoidal current control are evaluated and compared.. Such control schemes produce torqu

Trang 1

Strategy for the instantaneous torque control of

K.J Tseng, BEng

T.H Lee, MSc, PhD, BA

K.W Lim, BEng, DPhil

K.S Lock, BSc, PhD, CEng, MlEE

Indexing terms: Torque, Control systems, Brushless DC drives

Abstract: The paper describes an instantaneous

torque control for a permanent magnet motor

driven in a brushless direct current drive configu-

ration The motor has high torque capabilities,

which makes it suitable for direct drive applica-

tions, but unfortunately it suffers from torque pul-

sations when controlled by a conventional

sinusoidal current controller Instantaneous

torque control is aimed at eliminating such torque

pulsation, thereby improving speed and position

control The approach adopted is to design an

instantaneous torque control algorithm based on

a variable structure strategy in the dq rotating ref-

erence frame The switching commands for the

inverter are generated by the digital controller

directly The required instantaneous torque feed-

back information is estimated from knowledge of

the motor parameters and measurements of

instantaneous currents and rotor position The

performance of the proposed controller design and

torque feedback technique are investigated in

computer simulation studies and experimental

implementations Experimental results for drive

applications using instantaneous torque control

and conventional sinusoidal current control are

evaluated and compared

List of symbols

a, b, c

U = control variable

8 = electrical angle

0, = mechanical angle

0 = angular velocity

a = angular acceleration

T, = developed electromagnetic torque

7; = load torque

Ter = torque reference

= inverter vector elements

S = switching function of variable structure system

Paper 7493B ( P l ) , first received 27th November 1989 and in revised

form 1 lth June 1989

Dr Low, Mr Tseng, Dr Lee and Dr Lock are with the Department of

Electrical Engineering, National University of Singapore, 10 Kent

Ridge Crescent, Singapore 051 1, Republic of Singapore

Dr Lim is with the Department of Electncal Power Engineering,

University of New South Wales, Kensington, New South Wales 2033,

Australia

i, , i, , i, = phase currents

J = moment of inertia

D = viscosity coefficient

A- = flux linkage

R = winding resistance

1 = inductance

L ,

M,

E ,

Lij,

= self inductance xth harmonic coefficient

= mutual inductance xth harmonic coefficient

= rotor to stator inductance xth harmonic coeffi-

= dq-transformed inductance xth harmonic coef- cient

ficient

1 I n t r o d u c t i o n

High torque PM motors operated in brushless DC (BLDC) drive configuration are suitable for the direct

drive of industrial servo systems and robots [ l 21 A

cascade control structure is usually adopted for BLDC

drive applications (Fig l), consisting of an outer loop

inverter

controller controller

posltion feedback

Fig 1 Cascade BLDC drive rontrol sfrurfure

that generates the reference torque command to achieve the desired speed or position and an inner loop that con- trols the switching of the inverter so as to cause the

motor to produce the desired torque [3, 41 The inner loop, commonly known as the torque controller or

current controller, also ensures synchronism between the stator feed currents and the rotor flux This function is known as electronic commutation and has been com- pared to that of the commutator and brushes of conven- tional brush-type DC motor

Several inner loop configurations have been in use, or

have been proposed, in an effort to achieve the twin objectives of good torque control and the implementa- tion of electronic commutation Most of these methods rely on analogue current controllers to produce sinus- oidal phase currents, as shown in Fig 2 The controller consists of three independent current control loops, which may be based on the current hysteresis controlled

PWM method, or the subharmonic PWM method [ 5 ]

Trang 2

Such control schemes produce torque pulsation when the

motor flux distribution differs significantly from the ideal

sinusoid and it is well-known that nonsinusoidal flux dis-

processing of signals such as position, velocity and current, which contain information on disturbances and parameter variations, the switching patterns can be opti-

rotor

_ _ _ _ _

C

e

feedback

Fig 2 Conventional current controller

torque estimator

torque producer

and load

torque

referencc

I

1

I

I

L -_

currents position

I information information

r - - - 7

Fig 3

control

Block diagram illustrating the concept of instantaneous torque

tribution is a characteristic of high torque PM motors

16, 71

In a direct drive system, the load is directly coupled to

the shaft of the brushless motor This eliminates the

backlash and high friction inherent in a conventional

geared drive [8, 91 However, the absence of a high gear

ratio requires the motor to operate at low speeds, where

torque pulsation becomes highly undesirable At high

speeds, these ripples are usually filtered out by the system

inertia For low speed operations, there is obviously a

need for instantaneous torque control to minimise the

torque pulsation so as to achieve excellent transient and

steady-state responses Instantaneous torque control

would also permit the fastest possible response for speed

or position control [lo] It is to be noted here that the

desired torque trajectory must be determined by the

outer loop controller

In recent years, research efforts have been directed at

the direct digital control of inverters, in which the switch-

ing signals for the inverter are generated directly by

microprocessors This type of control has been made

feasible only in recent years owing to the advent of high

performance digital signal processors (DSP) A great

advantage of direct digital control is that it has made

possible new control algorithms that are not easy to

implement using conventional analogue current control-

lers It is also easy to modify or upgrade the control

program Another important advantage is that since the

on-off patterns for switching devices are generated by the

mised On the other hand, such optimising algorithms cannot be easily implemented on analogue current con- trollers [ll-151

The direct digital control of an inverter provides the possibility of designing an instantaneous torque control-

ler The concept is illustrated in Fig 3 The combination

of the inverter and PM motor is regarded as a torque- producing unit controlled by a three-bit digital command from the torque controller The instantaneous torque feedback signal can be estimated from knowledge of the instantaneous currents and the rotor position Thus, by designating an appropriate control algorithm, it is pos- sible to eliminate torque pulsation by direct digital control of the inverters

This paper describes the instantaneous torque control

of a BLDC drive, with the objectives of minimisation of

torque ripples and improvement of the transient and steady-state responses of speed and position control The focus is on the digital approach to torque control, whereby the switching signals for the inverter are gener- ated directly by a microprocessor The problem of obtaining a reliable and accurate instantaneous torque feedback is examined The performances of the proposed controller design and torque feedback technique are investigated in computer simulation studies and experi- mental implementations

2 P M motor modelling

The motor is modelled as two interconnected subsystems One is the electromagnetic subsystem, which converts the effects of electrical currents into developed torque (airgap torque), and the other is the mechanical subsystem, which governs mechanical responses to the torque

The mechanical dynamics can be summarised in a motor dynamic equation:

The electromagnetic behaviour of a three-phase synchro- nous machine is modelled as consisting of a field winding

f and three stator windings a, b and c This model is the

abc-model For a PM motor, the rotor may be modelled

as a fictitious winding with a constant current source i, The voltage equation for each of the stator windings can

be written as

d l

v = R i + -

dt

Trang 3

The flux-current relationships can be given as

where I,, is the self inductance of coil x and I,, is the

mutual inductance between coils x and y

The effects of saliency are included by expressing the

stator self and mutual inductances as Fourier functions of

the electrical rotor position 0, which is defined as the

angle between the rotor direct axis and stator phase a

axis [7, 161

The torque developed by the motor can be derived by

using the energy method, which is based on the principle

of the conservation of energy [17] It can be shown that

the developed torque is therefore given by

where P is the number of poles

This abc-model has been studied using the SIMNON

simulation package [18] The parameters for the motor

model are determined experimentally The motor used

has the back EMF waveform shown in Fig 4 The wave-

r

Fig 4 Back EMF waveform

form differs significantly from the ideal sinusoid The

variations in the self and mutual inductances of the

motor phase windings with respect to rotor position are

shown in Figs 5 and 6 , respectively They too differ from

the perfect sinusoid

Fig 5 Variation ofwinding selfinductance with position

While the abc-model is useful for simulation studies, it

does not lend itself to easy manipulation when designing

a controller for the inner loop The dq-axis model has

proved useful in this respect [l5] The model is obtained

by using the dq-transformation, which can be represented mathematically in terms of the electrical angle 0 [17]

Fig 6 Variation of winding mutual inductance with p o s i t m

By applying the transformation rule to the flux- current relationships of eqn 3, the following expressions for the direct and quadrature axis fluxes are obtained

I d = Id, id + Id, i, + ld, i, ( 5 )

I , = I,, i, + id + I,, i, ( 6 )

where the transformed inductances can be written as

id, = L d d o + L d d 6 COS 60 + L d d l 2 cos 120 f ' '

'dq =

Id, = L d f o + L d f 6 COS 60 + L d / l z cos 120 + ' ' '

I,, = L,,o + Lqq6 COS 60 4- L,,,, COS 128 +

(7)

Ldq6 sin 60 + LdqlZ sin 120 + (8)

(9) (10)

L,,, sin 60 + L,,,, sin 120 +

L,,, sin 60 + Lqf12 sin 120 +

(11) (12) The transformed inductances contain harmonics that are multiples of six By transforming the voltage equations from the abc-domain, the direct and quadrature axis volt- ages are obtained:

lqd =

d l d

dt

ud = Rid + - - wi,

d l

dt

where w = dQ/dt is the rotor electrical angular velocity The torque equation is obtained as

3 Torque control

The torque control design is based on the variable struc- ture strategy (VSS), in which the switching signals for the inverter are determined directly by the digital controller There is no need for the analogue PWM hardware associated with conventional current controllers In vari- able structure strategy systems, the control is allowed to change its structure [19]

A hierarchical method is adopted for the design of the VSS systems The design sequence for the ith hyper-

surface begins with the selection of a switching surface

s(x, t ) = 0 This will determine the desired characteristic

of the system in the sliding mode The parameters of the discontinuous control are then determined from the con- ditions that cause sliding mode to exist This ensures that

Trang 4

the control action will steer the state to the sliding

surface After reaching the sliding surface, the state will

be forced to remain (slide) on this plane Then the system

is said to be in the sliding mode, which is a special phe-

nomenon of VSS

The sufficient condition for the sliding mode to exist

on the ith hypersurface is given by the following inequal-

ity:

S,-O

This is known as the existence condition for a sliding

mode This sliding mode can also be shown to be stable

The developed torque of a PM motor modelled using

the dq transformation is seen to be a function of the

values of direct axis current id and quadrature axis

current i, Therefore, the states are selected as id and i,

To select the switching surfaces, the following objec-

tives are considered:

(i) The instantaneous torque response sould he as fast

as possible for optimum transient response of the mecha-

nical outputs

(ii) The desired torque is to be achieved with the

minimum input current This is to minimise ohmic losses,

thereby ensuring maximum efficiency

c41

(iii) Electronic commutation must be maintained

(iv) There should be no steady-state error

(v) The torque loop must be robust and stable

Since the direct axis is always aligned with the rotor

flux axis, the contribution to torque from id (reluctance

effect) is much smaller than that from i, Therefore, by

keeping id close to zero at all times, the total torque is

where K is the torque parameter, which can be assumed

to be constant during the small switching interval at

Objective (ii) is thus achieved This is equivalent to

keeping the torque angle at 90" By regulating the level of

i, to track the torque demand, objective (i) is also

achieved As id and i, are imaginary quantities that rotate

with the rotor, electronic commutation is maintained,

thereby satisfying objective (iii) Objectives (iv) and (v) are

guaranteed if the control inputs satisfy the existence con-

dition of the sliding mode at all times

The switching surfaces are therefore selected as

where id,,, = 0 and T,,, is the torque reference, as deter-

mined by the outer loop controller

The VSS control functions are selected as

1 s ; > o

- 1 s ; < o

sgn (si) =

These selections are based on the concept of vector

control The controller determines the control inputs U,

and U,, which are in the rotating dq-domain, by consider-

ing the positions of the state trajectories in the switching

plane These control inputs are then mapped into the abc

voltage vectors of the inverter

It is necessary to determine V, and V, so that the exis- tence condition of sliding mode (eqn 16) is satisfied Determination of V,:

From eqn 13,

d

U, = Rid + - ( I , ) - 01,

dt

= Rid + - d (I,, i, + I,, i, + I d , if)

dt

- a i, + I,, id + I,, i,)

S, = id

If we multiply eqn 24 by s, and substitute in eqn 20, the

following equation is obtained:

d

' d d ' [ dt

s, S, = A Rid + - (I,, i, + I,, i,)

ld + 1,, i, + I,, id + I,, i,

A sufficient condition for V, to satisfy eqn 16 is therefore

d

By repeating the above exercise for V,, we find that a sufficient condition for V, to satisfy the existence condi- tion for sliding mode is

There are four possible combinations of U, and U, Their

effects can be classified as in Table 1 The forms of the

control functions have been selected so that the mapping

to the actual inverter vectors is simple yet effective The PM motor is driven by a three-phase voltage-fed inverter bridge The inverter consists of three pairs of power MOSFETS Each pair operates in a toggle

Table 1 : Control inputs and actions

Control inputs Control actions

vd = + V d , vq = + V , Increase i d , increase i o

vd = +V, vq = - V , Increase i d , decrease i,

v d = - V d , v q = + V , Decrease i d , increase i q

vd = - V d , vq = - V , Decrease i d , decrease i ,

Trang 5

manner and is controlled by a single bit digital signal

The three-bit control signal is known as a vector There

are eight different vectors possible Table 2 lists the

vectors and line voltages in terms of the DC link voltage

Table 2 : Vectors and line voltages

Vector Line voltages

' a b ' b c

0 0 1 0 -vdc +v,,

0 1 0 -v,= +v,, 0

0 1 1 -v,, 0 +v,,

1 0 0 +v,, 0 -vdc

1 0 1 +v,, -vdc 0

1 1 0 0 +VdC -vdc

K c Vectors 0 and 7 are null vectors as they caused a

short circuit to either the positive or the negative termin-

als of the DC link The remaining six combinations are

active vectors These vectors can be illustrated in a vector

diagram, as in Fig 7

Pictorial representation of inverfer networks

Fig 7

To map the control in the dq-plane to the actual

inverter vector, the electrical rotor cycle is divided into 12

regions Depending on which of the 12 regions the rotor

angle is in, the closest inverter vector to each of the four

control inputs is selected It is shown that as long as the

DC link voltage is large enough that eqns 26 and 27 are

satisfied, the existence condition for sliding mode is not

degraded by this simple mapping The results of the

mapping are tabulated in Table 3

The sampling frequency should be high, at least

10 kHz, so that the chattering of the currents is small

Also, the computation delay should be kept to a

minimum Therefore, the control algorithm requires a

Table 3: Mapping t o inverter vectors

Rotor angle Inverter vectors

345"-15"

1 5 " 4 5 "

450-75"

75"-105"

105"-135"

135"-165

165"-195"

195"-225"

225"-255"

255"-285"

285"-315"

31 5"-345"

V d = +v,

v u = +vu

1 1 0

1 1 0

0 1 0

0 1 0

0 1 1

0 1 1

0 0 1

0 0 1

1 0 1

1 0 1

1 0 0

1 0 0

v,= -v,

vo = +v,

0 1 0

0 1 1

0 1 1

0 0 1

0 0 1

1 0 1

1 0 1

1 0 0

1 0 0

1 1 0

1 1 0

0 1 0

Vd = -v, Vd = +v,

vo = -v, vq = -vo

0 0 1 1 0 1

0 0 1 1 0 0

1 0 1 1 0 0

1 0 1 1 1 0

1 0 0 1 1 0

1 0 0 0 1 0

1 1 0 0 1 0

1 1 0 0 1 1

0 1 0 0 1 1

0 1 0 0 0 1

0 1 1 0 0 1

0 1 1 1 0 1

digital signal processor for its realisation Fig 8 shows the block diagram of the torque controller

m

I

Id I I I U U l e

loop controller

tronsformat ion

from outer controller loop

tronsformat ion

torque est i motor

to outer loop controller

Fig 8 Block diagram of torque controller

3.2 Simulation results

The direct digital torque controller, the inverter and the

PM motor have been simulated using SIMNON The controller sampling period is 100 p s The simulated torque from the motor model is used as the feedback torque signal, i.e the feedback is ideal Therefore, the simulated performance of the controller depends solely

on the control algorithm and is not degraded by any nonideal feedback

Simulation results are shown in Figs 9a-9 The steady state instantaneous torque is constant and contains no torque fluctuation This shows that the control algorithm

is effective as an instantaneous torque control scheme as long as an accurate torque feedback signal can be obtained The transient response is almost immediate As the developed torque is constant, the shaft velocity is also constant

The phase current is not sinusoidal Harmonics are present in the waveform to neutralise the harmonics in the flux distribution Thus, the controller is able to elimi- nate the torque ripples The direct axis current id has been kept close to zero, which is one of the objectives of the controller The entire torque is thus contributed by the quadrature axis current i, The quadrature axis current waveform shows ripples caused by the control action in keeping the instantaneous torque constant Figs 9f and g show the nature of the switching control inputs ud and U,,:

The results indicate that the digital torque control algorithm can have a very good performance if an accu- rate instantaneous torque signal can be obtained to provide the feedback

4 Torque feedback

A prerequisite for the proper functioning of the proposed torque control is the torque feedback information If instantaneous torque control is desired, then the torque feedback must accurately reflect the developed instanta- neous torque It can be directly measured by torque transducers, such as strain gauges [20], however these

Trang 6

are impractical for servo control applications as slip rings are required to transmit the measured signal They are also susceptible to noise Thus, alternative means of obtaining the torque information must be sought The method proposed here is to estimate the torque from knowledge of the motor parameters and measurements of instantaneous currents and rotor position

4.1 Torque estimation

In the torque control scheme, the direct axis current i, is controlled at zero at all times Therefore, by assuming

i, = 0, the torque equation, eqn 15, becomes

dl

01

b

C

0

-1

I :

'

;

,

- 1 0

e

10r

10

-10

-101

9

Fig 9

(I Torque.Nrn

h Speed radls

< Phase current, A

d Direcl-axis current A

Quadrature-are current A

f Direct-axis voltage, V

g Quadralure-axis vollage, V

Simulation results oJrorque control with ideal torque feedback

The estimated torque can be either the average or instan- staneous torque, depending on the degree of simplifica- tion of eqn 28

To estimate the average torque, only the fundamental components of the variation in machine inductances with rotor position are included The transformed inductances become

4'7 = L,,o

r,, = o

I,, = 0

I,, = L * f O

The torque equation simplifies to the following familiar form :

which is termed the estimated average torque

For a motor with sinusoidal rotor flux distribution, this estimated torque is also the instantaneous torque However, if the machine has a nonsinusoidal flux dis- tribution, the equation only gives the average torque For

a constant torque command, the control action would be

to maintain a constant i, This is equivalent to injecting purely sinusoidal phase currents Hence, using the esti- mated average torque for torque control is, in principle, equivalent to the conventional sinusoidal current control

To estimate the instantaneous torque, it is necessary

to take into account all the significant harmonic com- ponents of the variations in inductances with rotor angle Generally, for a high-torque PM motor, the contribu- tions to torque by the stator winding self and mutual inductances are much smaller than that of the rotor PM

flux [7] If this is coupled with the fact that the harmonic components decrease progressively with frequency, it is sufficient to consider only the fundamental components

of the stator winding self and mutual inductances There- fore,

I 44 = L W O

I,, = 0

For the rotor flux, all the significant higher harmonic terms must be considered:

I,, = L,,, sin 68 + L,,l sin 128 +

Therefore, the torque equation can be approximated by

3 P

T = - - [L,,, i, + (6L,,, + Ldf6)i, cos 68

+ (12LqfI2 + L,,,,)i, cos 128 +

3 P

2 2

i,

- - -

- [ K O + K , cos 68 + K , , cos 120 + ' .]i, ( 3 0 )

which i is a multiple of six The estimated instantaneous

torque is found from eqn 30 To take into account any saturation effect, some form of torque harmonic param- eter estimation scheme must be devised, which is beyond the scope of this paper

Trang 7

4.2 Simulation results

The torque controller has been simulated with estimated

average torque feedback to produce sinusoidal currents

The results are given in Figs loa-c The phase current is

microcomputer

- ' b L 0 2 O L 0 0 6 0 8 IO

DC input

0 5 1

01

b

1

C

Fig 10

torque feedback

a Torque, Nrn

h Speed, rad/s

sinusoidal The developed torque contains torque ripples

because of the interaction of the sinusoidal phase current

with the nonsinusoidal rotor flux Fig 10b shows that the

torque pulsation causes speed ripples in the shaft veloc-

ity The torque waveform also shows that the dominant

torque ripple harmonic is six times that of the phase

current frequency

Figs 1 la-c show the results obtained using estimated

instantaneous torque feedback The torque pulsation has

been eliminated Therefore, the shaft velocity shows little

fluctuation The phase current is no longer sinusoidal

The simulation results show that using torque control

with estimated average torque feedback is equivalent to

sinusoidal current control It causes torque pulsation if

the rotor flux distribution is nonsinusoidal The results

Simulation results of torque control with average estlmated

c Phase currenl, A

0

0 2

C

Fig 11

feedback

o Torque Nm c Phase currenl, A

h Speed rad/s

Simulation results of torque control with instantaneous torque

also show that the combination of VSS torque control and estimated instantaneous torque feedback form a viable instantaneous torque controller

5 Implementation of controllers

5.1 DSP-based experimental setup

A DSP-based experimental control system setup has been

constructed to investigate the performance of the pro- posed instantaneous torque controller The block diagram of the experimental setup is shown in Fig 12

( t o osci I loscope)

II 1 /-

digital signal processor

I ;h2 ( t o oscilloscope)

ch 1

Fig 12 DSP-based experimental setup

CT = Hall-etTecl current transducer

The mechanical arrangement consists of an experimental

PM motor coupled to an absolute rotary optical encoder

at one end of the rotor shaft with the other end coupled

to a load via a shaft-torque detector

A voltage-fed three-phase inverter bridge is used to

drive the motor It consists of three pairs of power MOSFETs and their gate-drive circuitry These MOSFETs are capable of switching speeds of above

100 kHz The DC link voltage is adjustable The switch- ing signals to the bridge are sent directly by the digital controller Phase current feedback signals are produced

by Hall-effect current transducers As the stator windings

are star-connected, only two of the currents have to be measured

The heart of the digital control system is a DSP32 digital signal processor [21] This is supported by a microcomputer that serves as a host development system during development of the control application programs and provides input/output (I/O) facilities during real time testing of the programs The DSP32 controller has been physically constructed on a prototyping card mounted in the microcomputer

5.2 Torque control

The flowchart of the torque control program is given in

Fig 13 Both instantaneous torque control and the con-

ventional sinsuoidal current control methods have been

Trang 8

implemented Instantaneous torque control is obtained

by coupling the controller with estimated instantaneous

torque feedback By using the estimated average torque

feedback, sinusoidal current is obtained A torque refer-

ence of 1 Nm is set and the load was adjusted so that the

motor speed is about 10 rev/min

I mapping look up table I

output vector

ri

I wait for next sample - 1

Fig 13 Flowchart of torque control program

Fig 15 shows the torque and phase current produced

by the sinusoidal current control There is torque pulsa-

tion with a frequency six times that of the current fre-

quency Fig 14 shows the torque and phase current

obtained using the instantaneous torque control The

torque ripples have been greatly reduced The phase

current waveform is no longer sinusoidal The experiment

Fig 14

control (T,,, = 1 N m )

Torque: 0.5 Nm/div

Phase current: 1.0 A/div

Torque and phase current produced with instantaneous torque

Time: I 0 0 ms/div

are shown in Figs 16 and 17 The results again show a

great reduction in torque ripples Thus, the instantaneous torque control implemented is effective in eliminating torque pulsations

Fig 15

control (T,,, = 1 Nm) Torque: 0.5 Nm/div Phase current: 1.0 A/div

Torque and phase current produced with sinusoidal current

Time: 100 msldiv

Fig 16

control IT,,, = 2 N m )

Torque: 0.5 Nm/div Phase current: 1.0 A/div

Torque and phase current produced with instantaneous torque

Time, 100 ms/div

Fig 17

control IT,,, = 2 N m )

Torque: 0.5 Nm/div Phase current' 1.0 A/drv

Torque and phase current produced with sinusoidal current

Time: 100 ms/dw

Trang 9

5.3 Speed control

A PI speed control algorithm is implemented on the DSP

controller together with the inner loop controller Both

the sinusoidal current control and the instantaneous

torque control are tested with the speed controller

The time response of the system for a step speed

command for the two different combinations of speed

and inner loop control schemes are shown in Figs 18a

and b The results show that speed ripples are negligible

for speed control with instantaneous torque control

Large variations in speed occur for speed control using

sinusoidal current control at very low speed From the

oscillograms, it is clear that the instantaneous torque

control has improved the dynamic response to speed

control

b Fig 18

U With mslanianeous torque control

h Wilh sinusoidal current conlrol

I rev/min/div, 5Ml ms/div

Experimental results for speed control

6 Conclusions

To overcome the problem of torque pulsation in the

operation of brushless DC drives, an instantaneous

torque control scheme has been designed and simulated

The torque control algorithm is based on a variable

structure strategy and the switching patterns for the

inverter are generated directly by the digital controller

The torque feedback information is derived from know-

ledge of machine parameters, instantaneous currents and

the rotor angle Simulation results have shown the

scheme to be viable

An experimental controller centred on a high per-

formance digital signal processor has been constructed to

implement the control scheme Experimental results show that the instantaneous torque controller has successfully minimised torque pulsation and significantly improved the performance of the speed controller when the latter is coupled with the instantaneous torque controller

7 References

I LOW, T.S., BINNS, KJ., RAHMAN, M.F., and WEE L.B.: ‘A Nd-Fe-B excited permanent-magnet motor - design and per- formance Third International Conference on Electrical Machines and Drives, IEE, London, UK, Nov 16th-18th 1987, pp 246249

2 LOW, T.S ‘Permanent-magnet motors for direct drive applica-

tions’, Autom News, Nov 1987, pp 1 4 1 7

3 RAHMAN, M.F LOW, T.S., and WEE, L.B.: ‘Development of a digitally controlled brushless dc drive system’ Proceedings of the

1986 Conference on Applied Motor Control (CAMC ’86) Minnea-

polis, Minnesota, USA, June 10th-12th 1986, pp 283-288

4 WEE, L.B., LIM, K.W., LOW, T.S., and RAHMAN, M.F : ‘A vari- able structure strategy for motion control’ Proceedings of the 1987 Conference on Industrial Electronics (IECON ’87), Cambridge Massachusetts, USA, November 3rd-6th 1987, pp 167-174

5 HOSHINO, A., KUROMARU, H., and KOBAYASHI, S.: ’AC

ssynchronous servo based on the armature voltage prediction model’ Proceedings of the 1987 Conference on Industrial Elec- tronics (IECON ’87), Cambridge, Massachusetts, USA, Nov 3rd-6th 1987, pp 205-21 1

6 HOANG, L., PERRET, R., and FEUILLET, R ‘Minimization of

torque ripples in brushless dc motor drives’, IEEE Trans., 1986,

IA-22, (4) pp 748-755

7 NG, B.H., LOW,T.S., LIM, K.W., and RAHMAN, M.F:’An mves- tigation into the effects of machine parameters on the torque pulsa- lions in a brushless dc drive’ Proceedings of the 1988 Conference on Industrial Electronics (IECON ’88), Singapore Oct 24th-28th 1988,

8 ASADA, H., KANADE, T., and TAKAYAMA, 1.: ‘Control of a

direct drive arm’, Trans A S M E , 1983,105, pp 136142

9 DAVIES, S., and CHEN, D.: ‘High performance brushless dc

motors for direct drive robot arm’, P C I M , Aug 1985, pp 35-38

IO ILK-SPONG, M., MILLER, T.J.E., MACMIN, S.R., and THORP,

J.S.: ‘Instantaneous toraue control of electric motor drives’ IEEE,

pp 749-754

1987, PE-2, (l), pp 55-6‘1

11 CHAMPENOIS, G., MOLLARD, PH., and ROGNON, J.P ’ ‘ T w o

digital torque control structures for permanent magnets converter fed sinusoidal synchronous machines’ Proceedings of the 1988 Con-

ference on Industrial Electronics (IECON ’W), Singapore, Oct

24th-28th 1988, pp 725-730

12 MATSUI, N., and OHASHI, H ‘DSP-based adaptive control of a brushless motor’ IAS 88, Pittsburgh, Pennsylvania, USA, Oct 2nd-6th 1988, pp 375-380

13 HASHIMOTO, H., KATO, Y., and MIYATA, M: ‘On-off pattern generation based on VSS for dc servo motor’ Proceedings of the

1987 Conference on Industrial Electronics (IECON 27) Cambridge

Massachusetts USA, Nov 3rd-6th 1987, pp 118&11R6

14 HASHIMOTO, H.: ‘Variable structure strategy for motion control systems’ Proceedings of the 1987 Conference on Industrial Elec- tronics (IECON ’87) Cambridge, Massachusetts, USA, Nov 3rd-6th 1987, pp 159-165

15 HASHIMOTO, H., YAMAMOTO, H., YANAGISAWA S., and

HARASHIMA, F.: ‘Brushless servo motor control using variable

structure approach’, IEEE Trans., 1988.24, (I), pp 16&170

16 NG, B.H.: ‘An investigation into the effects of machine parameters

on the torque pulsations in a brushless dc drive’ MEng Thesis

1988, National University of Singapore

17 FITZGERALD, A.E., KINGSLEY, C., and UMANS, S.D.: ‘Electric machinery’ (McGraw-Hill, 1983)

18 SIMNON, ‘User’s guide for MS-DOS computers’ version 1.0,

Department of Automatic Control, Lund Institute of Technology, Aug 1986

19 DeCARLO, R.A., ZAK, S.H., and MATTHEWS, G.P.: ‘Variable structure control of nonlinear multivariable systems: a tutorial’, Proc IEEE, 1988.76, (3) pp 212-232

20 JONES, B.E.: ’Instrumentation, measurement and feedback’ (McGraw-Hill, 1977) pp 55-56

21 ‘WE DSP32 digital signal processor information manual’ (The

AT&T Documentation Management Organization Sept 1986)

22 ‘WE DSP32-SL support software library user manual’ (The AT&T Documentation Management Organization, March 1987)

Ngày đăng: 03/01/2014, 19:52

TÀI LIỆU CÙNG NGƯỜI DÙNG

TÀI LIỆU LIÊN QUAN

🧩 Sản phẩm bạn có thể quan tâm