U.S. patent application number 09/741898 was filed with the patent office on 2001-09-27 for method and arrangement for regulating the current in a switched reluctance machine.
Invention is credited to Greif, Andreas.
Application Number | 20010024099 09/741898 |
Document ID | / |
Family ID | 7933666 |
Filed Date | 2001-09-27 |
United States Patent
Application |
20010024099 |
Kind Code |
A1 |
Greif, Andreas |
September 27, 2001 |
Method and arrangement for regulating the current in a switched
reluctance machine
Abstract
A method and an apparatus for regulating the phase current in
the windings of a reluctance machine. Regulation is carried out
using a digital regulator, which operates using a PI characteristic
and presets pulse-width-modulated pulses for a DC chopper
controller. Set values, which are a function of the phase voltage,
are superimposed on the regulator manipulated variable by means of
a pilot control.
Inventors: |
Greif, Andreas; (Muenchen,
DE) |
Correspondence
Address: |
EVENSON, MCKEOWN, EDWARDS & LENAHAN, P.L.L.C.
Suite 700
1200 G Street, N.W.
Washington
DC
20005
US
|
Family ID: |
7933666 |
Appl. No.: |
09/741898 |
Filed: |
December 22, 2000 |
Current U.S.
Class: |
318/701 |
Current CPC
Class: |
H02P 25/0925 20160201;
Y10S 388/906 20130101 |
Class at
Publication: |
318/701 |
International
Class: |
H02P 001/46; H02P
003/18; H02P 005/28; H02P 007/36 |
Foreign Application Data
Date |
Code |
Application Number |
Dec 22, 1999 |
DE |
DE 199 61 798.8 |
Claims
What is claimed is:
1. A method for regulating the phase current in a switched
reluctance machine, whose stator windings in each phase are each
connected to a DC chopper controller, which is connected to a
regulator which processes control error between the required
current value and a measured actual current value and applies
pulse-width-modulated electrical pulses to the DC chopper
controller, comprising the steps of: determining the control error
from the required current values and from actual current values
obtained by sampling at equidistant intervals: digitally forming a
first manipulated variable from the control error using a
proportional-integral characteristic, by linear superimposition of
an integral element and a proportional element and by multiplying
said proportional-integral characteristic by a respective
electrical angular position of the reluctance machine rotor;
linearly superimposing a second manipulating variable on the first
manipulated variable wherein said second manipulated variable is
formed as a pilot control value of a characteristic value by
multiplication by the rotor rotation speed, which characteristic
value is read, as a function of the phase current and as a function
of the electrical angular position, from a characteristic map,
which includes the derivative of the magnetic flux of the
reluctance machine with regard to the electrical angular position,
as a function of the electrical angular position of the rotor and
as a function of the phase current.
2. The method according to claim 1, including the step of storing
characteristic values in a table as a function of the electrical
rotor angle positions, wherein the characteristic values are
determined from a data set with the magnetic flux values of the
reluctance machine as a function of the electrical rotor angular
position and of the phase currents by deriving the flux values with
respect to the rotor angle, by division by a saturation current
which is typical for the transition to the saturated magnetic
state, and by forming the mean values of the respective rotor
position, and in that the pilot value is formed by multiplication
of the characteristic value, which is read as a function of the
measured electrical rotor angular position, by the rotation speed
and the phase current.
3. The method according to claim 1, including the step of
calculating the control error at the time t.sub.K=k*T.sub.A using
the following equation: e(k)=w(k)-x(k) where e (k) is the control
error, W (k) is the required current value, x (k) is the actual
current value, t.sub.K is the time, k is the number of sampling
intervals and T.sub.A is the sampling time, and in that the
manipulated variable is calculated using the following equation:
y(k)=K.sub.p*e(k)+Y.sub.I(k-1)+K.sub.I*e(k-1) where y (k) is the
manipulated variable, K.sub.p is the proportional gain, Y.sub.I is
the integral element of the manipulated variable, K.sub.I is the
product of the proportional gain and the quotient of the sampling
time and the readjustment time of the regulation, and e is the
control error.
4. The method according to claim 1, including the step of setting a
readjustment time of the PI regulation to the time constant of the
phase winding of the reluctance machine, and setting a factor of
the integral element using the following relationship: 26 K 1 = T A
2 K S * T 1 where K.sub.I is the factor for the integral element,
T.sub.A is the sampling interval, K.sub.s is the path time constant
of the controlled system, and T.sub.t is the dead time of the
regulation, while the gain factor is readjusted as a function of
the rotor position using the following relationship: 27 K p ( ) = T
1 ( ) 2 K S * T t where K.sub.p (.gamma.) is the gain factor,
T.sub.1 (.gamma.) is the current-dependent and
rotor-position-dependent time constant of the phase winding,
K.sub.s is the path time constant and T.sub.t is the dead time of
the regulation.
5. The method according to claim 1, including the step of
determining a time constant of the switched reluctance machine
using the following equation: 28 T 1 ( , i ) = ( , i ) i , * R
where T.sub.1 is the time constant, .PSI. is the magnetic flux,
.gamma. is the rotor position, i the phase current and R the pure
resistance of the phase winding.
6. The method according to claim 1, including the step of
determining a time constant T.sub.1 (.gamma.) of the phase winding
for the q-position and for the d-position of the rotor using the
following relationships: 29 T 1 q = ( q , i max ) i max * R T 1 d =
( d , i max ) i max * R and, for the intermediate positions of the
rotor between the q-position and the d-position, are multiplied by
the product of the electrical angular position and the ratio
T.sub.1d/T.sub.1q, where T.sub.1q is the time constant of the phase
winding in the q-position of the rotor, T.sub.1d is the time
constant of the phase winding in the d-position of the rotor, .PSI.
(.gamma..sub.q, i.sub.max) is the magnetic flux of the reluctance
machine when the rotor is in the q-position and the current is a
maximum during operation of the reluctance machine, R is the pure
resistance of the phase winding and .PSI. (.gamma..sub.d,
i.sub.max) is the magnetic flux of the reluctance machine when the
rotor is in the d-position and the current is at the maximum value
at which the reluctance machine is intended to operate.
7. The method according to claim 1, including the step of producing
a start pulse, whenever a switch-on angle for the phase current is
reached, using the following relationship: 30 PWM start = ( n n max
+ i w I max ) * PWM 100 % where PWM.sub.start is the start pulse, n
is the measured rotation speed, n.sub.max is the maximum rotation
speed, I.sub.max is the maximum current of the drive, i.sub.w is
the required current setting and PWM.sub.100% is the pulse-width
pulse for full control.
8. An apparatus for regulating the phase current in a switched
reluctance machine, whose stator windings in each phase are each
connected to a DC chopper controller, which is connected to a
regulator which processes the control error between the required
current value and the measured actual current value and applies
pulse-width-modulated electrical pulses to the DC chopper
controller, said system comprising: a regulator having a
microcontroller to whose input side required current values (e (k))
and actual current values (x (k)) are supplied via an A/D
converter, and to which rotation position signals are supplied
which are produced by a rotation position sensor in the reluctance
machine: means for calculating the manipulated variable (y (k))
from the control error using a PI characteristic is stored in the
regulator, wherein separately calculated proportional and
I-elements are added, and a stored constant is used to calculate
the I-element as the quotient of a constant sampling interval and
the product of twice a path gain of the controlled system and a
dead time of the regulator: means for storing values of a time
constant of the phase winding, as a function of the rotor position,
in a memory in order to determine a gain factor: means for
determining pilot values by multiplication of the rotor rotation
speed by characteristic values stored in a characteristic map, and
superimposed on the manipulated variable of an output of the
regulator wherein said characteristic map contains the derivative
of the magnetic flux of the reluctance machine with respect to the
electrical angular position as a function of the electrical rotor
angle position and the phase current.
9. The apparatus according to claim 8, wherein the characteristic
map comprises a series of characteristic values which are each
determined from the derivative of the magnetic flux values of the
reluctance machine as a function of the electrical rotor angular
position and of the phase currents with respect to the rotor
angular position and by division of these derivative values by a
saturation current, which is typical for the transition to the
saturated magnetic state, and by forming the mean values for the
respective rotor position.
10. The apparatus according to claim 8, further comprising a
programmable logic module for writing a start value to a
pulse-width-modulation register.
Description
BACKGROUND AND SUMMARY OF THE INVENTION
[0001] This application claims the priority of German Application
19961798.8, filed Dec. 22, 1999, the disclosure of which is
expressly incorporated by reference herein.
[0002] The invention relates to a method and an arrangement for
regulating the phase current in a switched reluctance machine,
whose stator windings in each phase are each connected to a DC
chopper controller which is connected to a regulator which
processes the control error between the required current value and
the measured actual current value and applies pulse-width-modulated
electrical pulses to the DC chopper controller.
[0003] An arrangement of the type described above is known (U.S.
Pat. No. 5,754,024). The DC chopper controller in each phase of the
known arrangement comprises a first series circuit of a switching
transistor with a freewheeling diode, and a second series circuit
of a freewheeling diode with a switching transistor. The switching
transistor in the first series circuit is connected to the positive
pole of a DC voltage source, and the switching transistor in the
second series circuit is connected to the negative pole of the DC
voltage source. The freewheeling diodes are reverse-biassed with
respect to the polarity of the DC voltage source. The control
electrodes of the switching transistors, which are IGBTs, are
connected to a pulse-width modulator which has a first input
connected to a clock generator, a second input connected to a
comparator, and a third input to which an on/off signal is applied.
The phase winding is arranged in series with a current sensor
between the points where the switching transistors are connected to
the freewheeling diodes. A first input of the comparator has a
required current value applied to it, and a second input has the
actual current value from the current sensor applied to it. The
required current value and the on signal together with the off
signal for the pulse-width modulator are determined as a function
of the rotor position, measured by a sensor. The pulse-width
modulator starts when it is intended to apply current to the
respective winding, and stops when it is intended to stop current
flowing in the winding once again.
[0004] German Patent DE 43 10 772 C2 discloses a method for
regulating the phase current in a switched reluctance machine,
whose stator windings in each phase are each connected to a DC
chopper controller, which is connected to a regulator which
processes the control error between the required current value and
the measured actual current value and applies pulse-width-modulated
electrical pulses to the DC chopper controller. In the case of the
control circuit disclosed there, the control error between the
required current value and the actual current value is supplied to
a PI regulator.
[0005] European Patent EP 0 684 693 A2 discloses an arrangement for
regulating the phase current of brushless DC machines and switched
reluctance machines, in which the control error is determined from
the required values and from actual current values obtained by
sampling and equidistant intervals.
[0006] A three-point regulator with hysteresis is suitable for
regulating the phase current in the reluctance machine. The output
of the three-point regulator can assume three states, each of which
can be associated with a switching state of a converter or DC
chopper controller. The association with the "on, short-circuit"
and "off" switching states of the current regulator allows the
phase current to be regulated not only in motor operation but also
in generator operation down to zero speed, without the three-point
regulator needing to be switched. If the three-point regulator has
identical switching thresholds when the reluctance machine is being
operated as a motor and as a generator, this, in fact, results in a
higher mean current value in generator operation than in motor
operation. This effect can be minimized by hysteresis loops which
are shifted one above the other. One advantage of a three-point
regulator with hysteresis is its simple structure.
[0007] A disadvantage of the three-point regulator is that the
converter switching frequency caused by the three-point regulator
depends not only on the switching thresholds but also on the rate
of current change in the machine winding, which in turn depends on
the phase voltage, the winding resistance, the present current
value, the phase inductance (which is dependent on the rotor
position) and the rotation speed. Taking account of these
influencing variables, the switching thresholds of the three-point
regulator must be selected such that the maximum switching
frequency of the power semiconductors in the converter is not
exceeded. During operation of the reluctance machine, this results
in switching frequencies which are well below the maximum switching
frequency and are in the audible range. As a result the reluctance
machine produces irritating noises.
[0008] The invention is based on the problem of specifying a method
which can be matched flexibly to different situations that occur
with reluctance machines, and an arrangement for regulating the
current in phase windings of a switched reluctance machine, in
which irritating noise from the reluctance machine, caused by the
switching frequencies of the converter active devices is largely
avoided and in which the phase currents can be set dynamically and
quickly to the predetermined required values.
[0009] According to the invention, with regard to a method of the
type described initially, the problem is solved by determining the
control error from the required values and from actual current
values obtained by sampling at equidistant intervals. Also a first
manipulated variable is formed from the control error digitally
using a proportional-integral characteristic, by linear
superimposition of an integral element and a proportional element
which is multiplied by the respective electrical angular position
of the reluctance machine. Furthermore the first manipulated
variable has a second manipulated variable superimposed on it
linearly, which is formed as a pilot control value of a
characteristic value by multiplication by the rotation speed, which
characteristic value is read, as a function of the phase current
and as a function of the electrical angular position of the rotor,
from a characteristic map, which includes the derivative of the
magnetic flux of the reluctance machine with regard to the
electrical angular position, as a function of the electrical
angular position of the rotor of the reluctance machine and as a
function of the phase current. The method according to the
invention allows the phase currents to be well regulated even at
high rotation speeds and at high pulse-width-modulation
frequencies, as well as allows for rapid changes in the induced
phase voltage.
[0010] One preferred embodiment provides that characteristic values
are stored in a table as a function of the electrical rotor angle
positions. The characteristic values are determined from a data set
with the magnetic flux values of the reluctance machine as a
function of the electrical rotor angular position and of the phase
currents by deriving the flux values with respect to the rotor
angle, by division by a saturation current which is typical for the
transition to the saturated magnetic state, and by forming the mean
values of the respective rotor position. The pilot value is formed
by multiplication of the characteristic value, which is read as a
function of the measured electrical rotor angular position, by the
rotation speed and the phase current. In this embodiment,
relatively little memory capacity is required for storing the
characteristic values. The approximate determination of the
rotational voltage value for the pilot control is not a
disadvantage, because the regulator can quickly compensate for a
relatively small error between the required value and the actual
value.
[0011] In one expedient embodiment, the control error at the time
t.sub.K=k*T.sub.A is calculated using the following equation e
(k)=w (k)-x (k) where e is the control error, W is the required
current value, x is the actual current value, t.sub.K is the time,
k is the number of sampling intervals and T.sub.A is the sampling
time, and in that the manipulated variable is calculated using the
following equation:
y(k)=K.sub.p*e(k)+Y.sub.I(k-1)+K.sub.I*e(k-1),
[0012] where y (k) is the manipulated variable, K.sub.p is the
proportional gain, Y.sub.I is the integral element of the
manipulated variable, K.sub.I is the product of the proportional
gain and the quotient of the sampling time and the readjustment
time of the regulation, and e (k) is the control error. The method
described above allows the manipulated variable to be determined in
a relatively short time from the control error. The regulator
computation time is thus very short. Computation time in this case
refers to the time which passes from reading the actual value via
an A/D converter to the time at which the control signal is applied
to the converter.
[0013] In particular, the readjustment time of the PI regulation is
on the one hand set to the time constant of the phase winding of
the reluctance machine, and the factor of the integral element is
on the other hand set using the following relationship: 1 K 1 = T A
2 K S * T I
[0014] where K.sub.I is the factor for the integral element,
T.sub.A is the sampling interval, K.sub.s is the path time constant
of the controlled system, and T.sub.t is the dead time of the
regulation, while the gain factor is readjusted as a function of
the rotor position using the following relationship: 2 K p ( ) = T
1 ( ) 2 K S * T t
[0015] where K.sub.p (.gamma.) is the gain factor, T.sub.1
(.gamma.) is the current-dependent and rotor-position-dependent
time constant of the phase winding, K.sub.s is the path time
constant and T.sub.t is the dead time of the regulation. Such a
setting process results in the control loop having a good time
response.
[0016] Switched reluctance machines have a time constant which is
dependent on the current and the rotor position and for which: 3 T
1 ( , i ) = ( , i ) i , * R
[0017] where T.sub.1 is the time constant, .PSI. is the magnetic
flux, .gamma. is the rotor position, i the phase current and R the
pure resistance of the phase winding.
[0018] It is particularly advantageous if the time constant T.sub.1
(.gamma.) of the phase winding for the q-position and for the
d-position of the rotor is determined using the following
relationships: 4 T 1 q = ( q , i max ) i max * R T 1 d = ( d , i
max ) i max * R
[0019] and, for the intermediate positions of the rotor between the
q-position and the d-position, are multiplied by the product of the
electrical angular position of the rotor and the ratio
T.sub.1d/T.sub.1q, where T.sub.1q is the time constant of the phase
winding in the q-position of the rotor, T.sub.1d is the time
constant of the phase winding in the d-position of the rotor, .PSI.
(.gamma..sub.q, i.sub.max) is the magnetic flux of the reluctance
machine when the rotor is in the q-position and the current is a
maximum during operation of the reluctance machine, R is the pure
resistance of the phase winding and .PSI. (.gamma..sub.d,
i.sub.max) is the magnetic flux of the reluctance machine when the
rotor is in the d-position and the current is at the maximum value
at which the reluctance machine is intended to operate. This method
allows the time constant of the phase winding to be determined with
sufficient accuracy with a short computation time.
[0020] In a further preferred embodiment, whenever a switch-on
angle for the phase current is reached, a start pulse is produced
by the converter using the following relationship: 5 PWM start = (
n n max + i w I max ) * PWM 100 %
[0021] where PWM.sub.start is the start signal, n is the measured
rotation speed, n.sub.max is the maximum rotation speed, I.sub.max
is the maximum current of the drive, i.sub.w is the required
current setting and PWM.sub.100% is the pulse-width pulse for full
control. Using this start value, the sampling clock rate of the
regulation and the process of switching the phases on and off are
coordinated such that no angular errors result from the
asynchronous relationship between the sampling clock rate and the
switching of the winding phases, which is dependent on the rotation
speed. Furthermore, the reaction time of the regulation is
minimized. In addition, this avoids any discontinuities in the
transition from pulsed operation of the reluctance machine to block
operation.
[0022] In an arrangement for regulating the phase current in a
switched reluctance machine, whose stator windings in each phase
are each connected to a DC chopper controller, which is connected
to a regulator which processes the control error between the
required current value and the measured actual current value and
applies pulse-width-modulated electrical pulses to the DC chopper
controller, the problem is solved, according to the invention, in
that the regulator has a microcontroller to whose input side
required current values and actual current values can be supplied
via an A/D converter, and to which rotation position signals can be
supplied which are produced by a rotation position sensor in the
reluctance machine. A program calculator the control error and the
manipulated variable, using a PI characteristic, is stored in the
regulator. The program has a part for separately calculating the
proportional and I-elements in accordance with the PI
characteristic. The proportional and I elements are added. A
constant is stored for calculating the I-element as the quotient of
a constant sampling interval and the product of twice the path gain
of the controlled system and the dead time of the regulator. In
order to determine the gain factor, values of the time constant of
the phase winding are stored as a function of the rotor position in
a memory. Furthermore, in order to determine pilot values (which
can be formed by multiplication of the rotor rotation speed by
characteristic values and are superimposed on the manipulated
variable of the output of the regulator), a characteristic map,
which contains the derivative of the magnetic flux of the
reluctance machine with respect to the electrical angular position
as a function of the electrical rotor angle position and the phase
current, is stored as a function of the phase current and of the
rotor angular position.
[0023] A considerable saving in memory space is achieved if the
characteristic map includes a series of characteristic values which
have been determined in the following way: differentiation of the
magnetic flux values of the reluctance machine as a function of the
electrical rotor angular position and of the phase currents with
respect to the rotor angular position; division of the
differentiated values by a saturation current which is typical for
the transition to the saturated magnetic state; and formation of
the mean values for the respective rotor position.
[0024] The invention will be described in more detail in the
following text with reference to an exemplary embodiment which is
illustrated in the drawings and from which further details,
features and advantages are evident.
[0025] Other objects, advantages and novel features of the present
invention will become apparent from the following detailed
description of the invention when considered in conjunction with
the accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
[0026] FIG. 1 shows a block diagram of an arrangement for
regulating the phase current in a switched reluctance machine;
[0027] FIG. 2 shows a control loop for regulating the phase
current;
[0028] FIG. 3 shows a timing diagram of pulse-width-modulated
pulses with different duty ratios;
[0029] FIG. 4 shows further details of the structogramm illustrated
in FIG. 2;
[0030] FIG. 5 shows characteristics, which are typical for a
reluctance machine, for the flux as a function of the phase
current, with the rotor rotation position as a parameter;
[0031] FIG. 6 shows approximated characteristics of the flux as a
function of the phase current for a reluctance machine,
[0032] FIG. 7 shows the approximate profile of the time constant of
the reluctance machine as a function of the rotor rotation
position;
[0033] FIG. 8 shows a characteristic, which is typical for the
reluctance machine, of the magnetic flux as a function of the rotor
rotation position with a constant phase current; and
[0034] FIG. 9 shows the partial derivative of the characteristic
shown in FIG. 8 with respect to the rotor rotation position for a
constant phase current.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
[0035] An arrangement for regulating the phase current in a
switched reluctance machine has a regulation and control
arrangement which is denoted by 1 in FIG. 1 and to which, at an
input, the required current values are supplied which are also
stored in the arrangement 1 or are intended to be used as a
reference variable. The regulation and control arrangement 1
produces actuating signals for a converter or inverter 2 which has
two circuit halves which each comprise a series circuit of a
solid-state switch 3, 4 and a freewheeling diode 5, 6, in which the
switch 3 and the diode 5 are connected to the positive pole 7 of a
DC voltage source, and the switch 3 and the diode 6 are connected
to the negative pole 8 of said DC voltage source. A series circuit
comprising one phase winding 9 of the reluctance machine and a
current sensor is connected to the common junction point between
the switch 3 and the diode 5, and the switch 4 and the diode 6. The
converter or inverter 2 is a DC chopper controller, whose switches
are, in particular, IGBTs.
[0036] The reluctance machine may have a number of phases, only one
of which is illustrated in FIG. 1. The number of DC chopper
controllers corresponds to the number of phase windings.
[0037] The rotor 11 of the reluctance machine has a associated
rotation position sensor 12 which emits appropriate signals to the
regulation and control arrangement 1 as a function of the
respective rotation angular position of the rotor 11. The rotation
angular position may be coded in absolute values, so that no
conversion to a digital variable is required. The rotation position
sensor 12 also allows the angular rotation rate to be measured.
[0038] FIG. 2 shows a structogramm of the control loop. The
converter or inverter 2 is the final control element. The current
sensor 10 is the measurement element. The phase winding 9, with its
inductance and pure resistance, forms the controlled system. The
regulation and control arrangement 1 contains a regulator 13 and a
pilot control arrangement 14.
[0039] The regulator 13 is intended to be used to apply a voltage
via the final control element to the phase winding 9, which voltage
is of such a magnitude that the current in the phase winding 9 of
the reluctance machine corresponds to the required value, which is
denoted by w (k) in FIG. 2. The regulator 13 is in the form of a
digital regulator, that is to say it contains a microcontroller,
which is not shown in any more detail.
[0040] Since the converter or inverter 2 has solid-state switches
which operate in the switch mode in order to avoid any power loss,
a pulsed voltage for the phase winding 9 is produced from the
preset DC voltage U.sub.d. Pulse-width-modulated pulses are
produced from the DC voltage. A pulse-width-modulation unit (also
referred to as PWM) is used for this purpose and, in particular, is
integrated in the microcontroller. The constant period duration is
set by means of a program, and the pulse width is preset by means
of a register.
[0041] FIG. 3 shows the pulses for duty ratios of 10%, 50% and 90%
for the period T.sub.PWM. The mean value of the voltage applied to
the phase winding 9 may be varied between 0 and 100% or between
+100% and -100% of the DC voltage U.sub.d, if the forward voltage
drop across the semiconductor switches is ignored. The variation in
the range +100% and -100% is described in detail in Patent
Application 199 43 542.1 from the Applicant, to which reference is
hereby made. The control range of the regulator is in this case
designed such that the regulation can assume values between +100%
and -100%.
[0042] If the regulator output is negative, as is the situation if
the required value is exceeded, logic causes the PWM to switch
backwards and forwards between the two switching states "off" and
"short-circuit", in order to reduce the phase current (Mode
"0"=decrease current). If the regulator output is positive, if the
required current value is undershot, the control logic results in
the PWM switching backwards and forwards between "on" and
"short-circuit" in order to increase the phase current (Mode
"1"increase current).
[0043] The range -100% to 0 is converted to a PWM value of 0 to
+100%, and the range between 0 and +100% is likewise converted to a
PWM value between 0 and +100%.
[0044] This corresponds to a phase voltage of -U.sub.d to 0 in Mode
"0" (decrease current), and a phase voltage of =to +U.sub.d in Mode
"1" (increase current). In this way, the current regulator controls
the mode automatically and there are no discontinuities between the
two operating modes "increase current" and "reduce current.
[0045] The path gain of the control loop illustrated in FIG. 2 is
K.sub.s=K.sub.control*K.sub.mot*K.sub.meas, where K.sub.s is the
overall path gain, K.sub.mot is the gain of the phase winding, and
K.sub.meas is the gain of the measurement device.
[0046] In order to determine the gain of the control element
K.sub.control, the numerical value which must be entered in the
register for 100% pulse-width-modulation is set as follows with
respect to the DC voltage U.sub.d: 6 K control = U 100 % PWM
[0047] The gain of the phase winding 9 is governed by its pure
resistance R so that: 7 K mot = 1 R
[0048] This value indicates the current which will be produced for
a specific applied (DC) voltage when the reluctance machine
rotation speed is zero.
[0049] The current sensor 9 supplies the actual values to the
digital regulator 13 via an A/D converter (ADC). The gain of the
measurement device is then determined as follows: 8 K meas = ADC
max I max
[0050] that is to say the ratio of the maximum output value of the
ADC to the maximum current in the reluctance machine.
[0051] By virtue of the microcontroller, the digital regulator 13
operates as a sampling regulator, in which equidistant sampling
intervals T.sub.A are preferably used. The sampling interval has a
major influence on the dynamic response of the regulation.
[0052] The digital regulator 13 operates using a PI
(proportional-integral) characteristic and requires the so-called
computation time T.sub.R for the time from reading the actual
current values via the A/C converter to the time at which the
pulse-width-modulation unit is operating with the respectively
newly determined pulse-width value.
[0053] The steady-state mean value of the sampling time is added to
this computation time T.sub.R. This results in the dead time
T.sub.t for the regulator 13:
T.sub.1=T.sub.R+T.sub.A/2
[0054] A disturbance variable may occur in the time period from
immediately after to immediately before the sampling by the A/D
converter. However, the shortest effective computation time is the
period of the pulse-width-modulation, since no new PWM value is
transferred until the PWM unit starts a new period.
[0055] FIG. 4 shows the structure of the PI regulator 13 within the
control and regulation arrangement 1. A P element 15 with a gain of
P, and an I-element 16 with a readjustment time of T.sub.n/K.sub.p
are provided.
[0056] The I-element is determined by the constant 9 K 1 = K p * T
A T n
[0057] where T.sub.A is the sampling time.
[0058] The control error e (k) is calculated by the regulation and
control arrangement 1 at each time t.sub.K=k*T.sub.A, based on the
required value w (k) and the actual value of the current x (k), as
follows,
e(k)=w(k)-x(k)
[0059] The P-element y.sub.p (k) is determined from the control
error e (k) as follows:
y.sub.p(k)=K.sub.p*e(k-1)
[0060] The I-element 10 y 1 ( k ) = K 1 1 = 0 k - 1 e ( i )
[0061] is determined separately from the P-element. The P and I
elements are added, thus giving: y(k)=y.sub.p(k)+y.sub.t(k)
[0062] In order to reduce the computation complexity, only the
changes .DELTA.y (K) from the previous value y (k-1) are calculated
and added to this:
y(k)=y(k-1)+.DELTA.y(k)
[0063] The P and I-elements are then given by:
.DELTA.y.sub.p(k)=K.sup.p[e(k)-e(k-1)]
and
.DELTA.y.sub.1 (k)=K.sub.1e(k-1)
[0064] The combined elements y.sub.p (k) and y.sub.I (k) are:
y.sub.p(k)=K.sub.p*e(k)
and
y.sub.I(k)=y.sub.I(k-1)+K.sub.I*e(k-1)
[0065] Since only the error and the I-element which was determined
in the previous sampling interval need ever be stored when using
the procedure described above, this results in a short computation
time.
[0066] In a switched reluctance machine, the electrical time
constant of the individual phase windings is highly dependent on
the rotor rotation position .gamma. and on the phase current i.
This results in a current-dependent and position-dependent time
constant T.sub.1 (.gamma., i) which is determined from the flux
characteristic as a function of the current, that is to say: 11 T 1
( , i ) = ( , i ) i * R
[0067] The time constant T.sub.1 (.gamma., i) can be stored as a
characteristic map and can read in each regulation cycle.
[0068] FIG. 5 shows a typical characteristic of the flux .PSI. of a
reluctance machine as a function of the phase current i, with the
rotation position .gamma. of the rotor as a parameter. The rate of
change of the current is determined from the gradient of a straight
line through the origin and the instantaneous operating point on
the .PSI.-i characteristic. It has been found that an approximate
determination of the time constant leads to good regulation
characteristics, described below:
[0069] In a switched reluctance machine, the magnetic reluctance
varies as a function of the rotation position of the rotor. The
minimum value is reached when a rotor tooth is opposite a stator
tooth which has been excited by current. This position is referred
to as a direct-axis field position or d-position. When the center
of the rotor slot is opposite a stator tooth which has been excited
by current, the magnetic reluctance is at its maximum. This
position is referred to as the quadrature-axis field position, or
q-position. The phase inductance L varies inversely with the
magnetic reluctance. An idealized profile with respect to the rotor
rotation position .gamma. can be assumed for the phase inductance
and thus for the time constant. This profile is shown for the time
constant T.sub.1 in FIG. 7. The time constant T.sub.1 for the
q-position, and the maximum current at which the switched
reluctance machine should be operated are given by: 12 T 1 q = ( q
, i max ) i max * R
[0070] The time constant in the d-position is given by: 13 T 1 d =
( d , i max ) i max * R
[0071] The intermediate values of the time constants between
T.sub.1q and T.sub.1d are defined by a straight line which runs
through the two points T.sub.1q and T.sub.1d. FIG. 7 shows the
profile of the time constants for motor and generator operation of
the reluctance machine.
[0072] This gives good results since the only occasion on which a
sudden change in the required value is applied to the regulator 13
is when a phase is switched on. This occurs in the vicinity of the
q-position for motor operation, and in the vicinity of the
d-position when in generator operation. For all other rotor
positions, the current in general just has to be regulated at a
constant actual value. However, the current may also be preset as a
reference variable.
[0073] The readjustment time T.sub.n of the regulator 13 is set to
be equal to the time constant of the controlled system. This
results in a readjustment time which is dependent on the rotor
rotation position:
T.sub.n(.gamma.)=T.sub.1(.gamma.)
[0074] The regulation is set on the basis of the optimum magnitude.
On this basis, the gain factor of the regulator 13 is calculated to
be 14 K p ( ) = T 1 ( ) 2 K s * T
[0075] since T.sub.t is the sum of the shortest time constants and
there are no further time constants in the control loop.
[0076] For the I-element of the regulator 13: 15 K 1 = T A 2 K s *
T 1
[0077] K.sub.I is thus independent of the rotor position. The phase
current regulator thus has a constant I-element and a P-element
which is readjusted adaptively (as a function of the rotor
position). FIG. 4 shows this regulation structure.
[0078] The phase voltage induced in the phase winding 9 depends on
the rotor rotation position .gamma., the phase current i and the
angular velocity .omega.. The rotating element of the induce phase
voltage u.sub.rot (.gamma., i) is: 16 u rot ( , i ) = ( , i ) *
[0079] In the vicinity of the q-position, the induced voltage is 0,
then initially increases slowly and, beyond a specific electrical
rotor angle, rises very steeply in order then to remain at one
point. This behaviour can be explained with reference to FIG. 8. In
order to prevent the phase current from assuming undesirable values
during the final sampling time as a result of the induced voltages
having steep profiles, the manipulated variable at the output of
the regulator 13 has a pilot control variable superimposed on it,
which corresponds to the present value of the induced voltage in
the phase 9, and is thus made dependent on .gamma., i and
.omega..
[0080] The pilot control value can be read from the characteristic
17 ( , i )
[0081] which is determined and stored for the respective reluctance
machine, as a function of the present phase current, can be
multiplied by the rotation speed or the angular velocity, and can
be added to the output of the regulator 13.
[0082] However, it has been found that there is no need to
determine the induced voltage exactly, since the regulator 13
compensates for relatively small errors between the required value
and the actual value well. The pilot control values can thus be
determined in a simpler manner.
[0083] It is assumed that, in the unsaturated region of the T-i
characteristics, as shown in FIG. 5, the partial derivative 18 i =
const
[0084] is related linearly to the phase current for any given rotor
angle. This linear profile is shown in FIG. 6. In the saturation
region, the .PSI.-i characteristics are assumed to be straight
lines which run parallel, and are likewise illustrated in FIG. 6.
The characteristic map .differential..PSI. (.gamma., i) can thus be
reduced to the profile 19 [ I * ] mean = f ( )
[0085] In the linear region, the induced voltage is then given by:
20 u rot ( , i ) = [ I * ] mean * i *
[0086] In the saturated region, for currents above I.sub.sat: 21 u
rot ( , i ) = [ I * ] mean * i sat *
[0087] In order to determine the profile of 22 [ I * ] mean
[0088] the .PSI. (.gamma., i) profiles are formed for all the phase
current values contained in the 23 ( )
[0089] data set, are divided by the phase current limited to
I.sub.sat, and are averaged. The resultant profile is used to
calculate the induced voltage.
[0090] FIG. 9 illustrates a profile which is typical for a
reluctance machine 24 [ I * ] mean
[0091] Such a characteristic for the respective reluctance machine
is stored in a one-dimensional table, as is illustrated in FIG. 4
by the block annotated pilot control. An appropriate value is read
from the table as a function of the phase current and of the rotor
rotation position, is multiplied by the rotation speed or the
angular velocity, and is added, as a pilot control value, to the
regulator output. The value for the saturation limit I.sub.sat is
set for the d-position in the region of the sharpest curvature of
the T-i characteristic.
[0092] All the phase-current regulators and pulse-width-modulation
units expediently operate synchronously with a common timebase.
[0093] The individual phases must be switched on and off as a
function of the present rotor position and of the
on-and-off-switching angles. Owing to the variable rotation speed
of the GRM, this switching of the phases is completely asynchronous
with respect to the clock.
[0094] In order to avoid the regulator cycle just having been
processed on reaching the switch-on angle, and thus not being
calculated until after the next sampling time T.sub.A in the new
PWM value, a start value is provided which is deliberately matched
to the present operating point of the input drive, and is entered
in the PWM register as soon as the switch-off angle is reached. The
PWM output of the microcontroller is then switched to be inactive,
so that the pulse-width setting does not yet have any effect. This
is ensured by a programmable logic device (PLD). In practice, this
then links the two asynchronous processes to one another (sampling
clock of the regulator and the on- and off-switches for the
phases).
[0095] A further advantage of this start value for the PWM is that
the reaction time of the regulator in response to a phase being
switched on is minimized.
[0096] A further reason in favour of the use of a PWM start value
is the transition from pulsed operation of the GRM to block
operation. In pulsed operation, the regulator has to act as a
limiting element for the phase current. In block operation,
however, the regulator output must be driven at the 100% level
beyond the switch-on angle in order to make the full
intermediate-circuit voltage available for that phase. This is the
only way in which the GRM can be used optimally. This transition
does not take place suddenly, but is dependent on the rotation
speed and the required current value.
[0097] It therefore makes sense not to use a constant start value
for the PWM, but to produce a relationship between the rotation
speed and the required current value. One possible way of achieving
this is: 25 PWM start = ( n n max + i w I max ) * PWM 100 %
[0098] In this case, n.sub.max is the maximum rotation speed and
I.sub.max is the maximum current of the input drive. i.sub.w is the
required current setting, and PWM.sub.100% is the PWM value for
full drive.
[0099] The DC chopper controller illustrated in FIG. 1 operates in
the two-quadrant mode.
[0100] Three of these DC chopper controllers are required to
operate a three-phase GRM and, in a corresponding manner, three
identical current regulators, which are independent of one
another.
[0101] This topology results in the current flowing in one
direction in that phase, although this does not limit the operating
range of the GRM, since the torque is formed independently of the
current direction. Four switching states can be provided with this
embodiment of the converter.
[0102] If both switches 3 and 4 are switched on (switching state:
on), the positive supply voltage +U.sub.d is applied to that phase,
as a result of which the phase current rises. The power is drawn
from the voltage source. If one switch 3 or 4 is open and the other
respective switch is closed (switching state: short-circuit), then
the phase current flows via the corresponding diode 6 or 5,
respectively. The phase is thus short-circuited, and the phase
voltage is 0V, ignoring the forward voltage dropped across the
switches. No energy is exchanged with the source. If both switches
are open (switching state: off), the phase current flows via both
diodes, which means a phase voltage of -U.sub.d. The power is fed
back into the source.
[0103] The regulation method according to the invention can in
principle also be used with other converter topologies.
[0104] The foregoing disclosure has been set forth merely to
illustrate the invention and is not intended to be limiting. Since
modifications of the disclosed embodiments incorporating the spirit
and substance of the invention may occur to persons skilled in the
art, the invention should be construed to include everything within
the scope of the appended claims and equivalents thereof.
* * * * *