CN102522948A - Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system - Google Patents

Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system Download PDF

Info

Publication number
CN102522948A
CN102522948A CN2012100042518A CN201210004251A CN102522948A CN 102522948 A CN102522948 A CN 102522948A CN 2012100042518 A CN2012100042518 A CN 2012100042518A CN 201210004251 A CN201210004251 A CN 201210004251A CN 102522948 A CN102522948 A CN 102522948A
Authority
CN
China
Prior art keywords
torque
current
hybrid intelligent
value
stator
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN2012100042518A
Other languages
Chinese (zh)
Other versions
CN102522948B (en
Inventor
徐凯
许强
徐文轩
徐果薇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing Jiaotong University
Original Assignee
Chongqing Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Chongqing Jiaotong University filed Critical Chongqing Jiaotong University
Priority to CN2012100042518A priority Critical patent/CN102522948B/en
Publication of CN102522948A publication Critical patent/CN102522948A/en
Application granted granted Critical
Publication of CN102522948B publication Critical patent/CN102522948B/en
Expired - Fee Related legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention discloses a hybrid intelligent adjusting method of torque hysteresis width in a DTC (Direct Torque Control) system. According to the hybrid intelligent adjusting method, with speed change and stator current change as input signals and increment of the torque hysteresis width as an output signal, a dynamic and adjustable torque hysteresis width value is obtained on the basis of determining a torque hysteresis width basic value, and the torque hysteresis width value changes along with the change of the control process; and the hybrid intelligent adjusting method has stronger adaptability and overcomes the defect that the torque hysteresis width is set into a fixed value in the conventional technology. The hybrid intelligent adjusting method disclosed by the invention has the beneficial effects that: the harmonic content in current is weakened, minimum values of flux linkage and torque pulse of an induction motor are obtained, and noise and vibration of the induction motor are reduced; the problems of high torque pulse and existence of speed steady-state error and periodic pulse of the rotating speed, of the induction motor under the low-speed operation condition, are solved; and the induction motor can obtain a smooth and effective driving performance. Therefore, the hybrid intelligent adjusting method is a fuzzy-neural-network hybrid intelligent control mode with high performance.

Description

The hybrid intelligent control method of torque hysteresis band in the DTC system
Technical field
The present invention relates to a kind of control technology of induction motor, relate in particular to the hybrid intelligent control method of torque hysteresis band in a kind of DTC system.
Background technology
The major advantage that the DTC of induction machine control (direct torque control) is had has: the torque dynamic response is fast, the variation of rotor parameter is had certain robustness; Its weak point mainly shows as: have torque pulsation, performance is obvious especially when low-speed running.
In fact, under the effect of switching voltage vector, when the stator magnetic linkage deviation reaches the upper limit of hysteresis comparator or down in limited time, though there is the new switch vector of impact, the stator magnetic linkage vector can not change direction and amplitude at once yet, and to continue to increase or reduce; In this case, stator magnetic linkage and electromagnetic torque may be gone out deviation range, " overshoot " phenomenon occurs, cause bigger pulsation.
The pulsation of torque can directly have influence on the speed characteristics of drive system: the flip-flop of torque pulsation will have influence on the steady-state error of system, also can cause arriving the prolongation of stable state time, all can be influential in high speed and low speed; The alternating component of torque pulsation can cause the pulsation of speed, and particularly influence is bigger when low speed, during low cruise; The stator magnetic flux rotation is slower, and the alternating component frequency of torque pulsation is lower, and this will cause the periodically pulsing of rotating speed under the low speed situation; Therefore, cause the velocity error periodic velocity fluctuation that on the basis of steady-state error, superposeed again, especially serious is; When low cruise; Little torque pulsation meeting causes the speed relative error very big, even up to 100%, such speed effect is to cause DTC system low-speed performance main reasons for decrease.
Summary of the invention
To the problem in the background technology, the present invention proposes the hybrid intelligent control method of torque hysteresis band in a kind of DTC system, the steps include:
1) with
Figure 2012100042518100002DEST_PATH_IMAGE002
and
Figure 2012100042518100002DEST_PATH_IMAGE004
as two input variables;
Figure 2012100042518100002DEST_PATH_IMAGE006
as output variable, sets up two-dimentional fuzzy reasoning table; Wherein,
Figure 716305DEST_PATH_IMAGE002
is the spinner velocity variable quantity;
Figure 633445DEST_PATH_IMAGE004
is the stator current variable quantity,
Figure 350866DEST_PATH_IMAGE006
be torque hysteresis band variable quantity;
2) confirm the base value
Figure 2012100042518100002DEST_PATH_IMAGE008
of torque hysteresis band according to following formula:
Figure 2012100042518100002DEST_PATH_IMAGE010
In the formula; is the motor number of pole-pairs; is the stator magnetic linkage vector;
Figure 2012100042518100002DEST_PATH_IMAGE016
is the stator voltage space vector; is the minimum switch periods of inverter;
Figure 2012100042518100002DEST_PATH_IMAGE020
is the sampling time;
Figure 2012100042518100002DEST_PATH_IMAGE022
is the motor leakage inductance:
Figure 2012100042518100002DEST_PATH_IMAGE024
; Wherein,
Figure 2012100042518100002DEST_PATH_IMAGE026
is stator inductance;
Figure 2012100042518100002DEST_PATH_IMAGE028
is inductor rotor, and
Figure 2012100042518100002DEST_PATH_IMAGE030
is coefficient of mutual inductance;
3) regularly spinner velocity
Figure 2012100042518100002DEST_PATH_IMAGE032
and stator current
Figure 2012100042518100002DEST_PATH_IMAGE034
are sampled; Calculate the difference of
Figure 799603DEST_PATH_IMAGE032
in current sampling period and last sampling period; Obtain
Figure 551658DEST_PATH_IMAGE002
numerical value in the current period; Calculate the difference of
Figure 311804DEST_PATH_IMAGE034
in current sampling period and last sampling period, obtain
Figure 516520DEST_PATH_IMAGE004
numerical value in the current period;
4), obtain corresponding under
Figure 653103DEST_PATH_IMAGE002
and value conditions in the current period according to two-dimentional fuzzy reasoning table;
5) according to the following formula to calculate the current
Figure 879795DEST_PATH_IMAGE002
and value conditions hysteresis torque width
Figure 2012100042518100002DEST_PATH_IMAGE036
:
Figure 2012100042518100002DEST_PATH_IMAGE038
The output variable of adjustment hysteresis comparator according to
Figure 953241DEST_PATH_IMAGE036
;
6) repeating step 3), 4), 5), recomputate
Figure 320769DEST_PATH_IMAGE036
.
On the basis of aforementioned schemes, the inventor has also proposed a kind of preferred version of two-dimentional fuzzy reasoning table, described two-dimentional fuzzy reasoning table such as following table:
Figure 2012100042518100002DEST_PATH_IMAGE039
Where,
Figure 2012100042518100002DEST_PATH_IMAGE041
is
Figure 378373DEST_PATH_IMAGE004
corresponds to the amount of blur,
Figure 2012100042518100002DEST_PATH_IMAGE043
is
Figure 2012100042518100002DEST_PATH_IMAGE045
corresponds to the amount of blur; NB1, NM1, NS1, ZO1, PS1, PM1, PB1 is
Figure 2012100042518100002DEST_PATH_IMAGE047
subset; NB2, NM2, NS2, ZO2, PS2, PM2, PB2 is
Figure 404229DEST_PATH_IMAGE043
subset; NB3, NM3, NS3, ZO3, PS3, PM3, PB3 is
Figure 2012100042518100002DEST_PATH_IMAGE049
subset;
Figure 923066DEST_PATH_IMAGE049
is
Figure 195916DEST_PATH_IMAGE006
corresponds to the amount of blur.
The obfuscation domain of input variable and output variable chooses [6; 6]; The degree of membership curve of input variable can adopt like Figure 15, the Gaussian curve shown in 16 and express; The degree of membership curve of output variable adopts equally distributed triangular function shown in figure 17 to express, and the subclass value of resulting thus
Figure 859591DEST_PATH_IMAGE043
and
Figure 876088DEST_PATH_IMAGE041
sees the following form:
Figure 2012100042518100002DEST_PATH_IMAGE050
The subclass value of
Figure 2012100042518100002DEST_PATH_IMAGE052
sees the following form:
Figure 2012100042518100002DEST_PATH_IMAGE053
Figure 2012100042518100002DEST_PATH_IMAGE055
;
Figure 2012100042518100002DEST_PATH_IMAGE057
;
Figure 2012100042518100002DEST_PATH_IMAGE059
; Wherein,
Figure 2012100042518100002DEST_PATH_IMAGE061
,
Figure 2012100042518100002DEST_PATH_IMAGE063
are the quantizing factor in the Fuzzy Processing, and
Figure DEST_PATH_IMAGE065
is the scale factor in the Fuzzy Processing.
On the basis of aforementioned schemes; In order further to improve systematic function; The present invention has also done following improvement: the I/O sample that every fuzzy rule in the two-dimentional fuzzy reasoning table all is converted into neural net; Utilize the I/O sample again, adopt the BP learning algorithm that neural net is trained; After neural metwork training is accomplished; Neural net is placed system, be used for the calculating of step 4)
Figure 515728DEST_PATH_IMAGE006
.
Useful technique effect of the present invention is: in the DTC system, adopt the variable hybrid intelligent control method of torque hysteresis band; Can weaken current harmonics; Obtain the minimum value of motor magnetic linkage, torque pulsation, motor can obtain low switching loss, noise and vibration like this.Simultaneously, reducing speed steady-state error and periodically pulsing, under the service conditions of low speed, make motor obtain level and smooth, an efficient driveability, is a kind of high performance hybrid intelligent control mode.
Description of drawings
Fig. 1, existing a kind of Direct Torque Control System of Induction Machine sketch map;
Fig. 2, Direct Torque Control System of Induction Machine sketch map of the present invention;
The principle schematic of Fig. 3, two-dimentional fuzzy reasoning table of the present invention;
Fig. 4, be used to realize the neural network model structural representation of fuzzy control rule part;
Fig. 5, can realize the electrical structure sketch map of a kind of hardware embodiment of function of the present invention;
Fig. 6, DSP main program flow chart of the present invention;
Fig. 7, DSP interrupt service routine flow chart of the present invention;
Fig. 8, employing PC are realized the present invention program's flow chart;
When Fig. 9, employing prior art, the electromagnetic torque pulsation oscillogram under the limit;
When Figure 10, employing the present invention program, the electromagnetic torque pulsation oscillogram under the limit;
When Figure 11, employing prior art, stator current waveforms figure under the limit;
When Figure 12, employing the present invention program, stator current waveforms figure under the limit;
Figure 13, the stator magnetic flux trajectory diagram when adopting prior art;
Figure 14, the stator magnetic flux trajectory diagram when adopting the present invention program;
Figure 15, the corresponding degree of membership curve of input variable
Figure 225058DEST_PATH_IMAGE043
;
Figure 16, the corresponding degree of membership curve of input variable ;
Figure 17, the corresponding degree of membership curve of output variable
Figure 995885DEST_PATH_IMAGE049
.
Embodiment
To the problem in the background technology; The inventor is through concentrating on studies and analyzing discovery: the magnetic linkage in the Direct Torque Control of Induction Machines and the hysteresis band of torque affect magnetic linkage and torque pulsation deeply; Pulsating quantity is to be limited in the bandwidth of hysteresis comparator: the ring value that stagnates is big more, and pulsation is just big more.From this thinking,, just must reduce the bandwidth of hysteresis comparator, but this will increase the switching frequency and the switching loss of inverter, reduce operational efficiency, also improve the requirement to electronic switching technology if pulsation is reduced; In addition; Less fixedly hysteresis band can not be eliminated the pulsation (effect that especially reduces to pulse in low rotation speed area is very little) of magnetic linkage and torque fully; And less fixedly hysteresis band can make systematic function in certain opereating specification, degenerate; Therefore, the bandwidth that reduces hysteresis comparator can not play good effect.
Can know from the analysis of front,, just should reasonably confirm the bandwidth value of hysteresis comparator for reaching not only the limit switch frequency but also reducing magnetic linkage and the purpose of torque pulsation.Common method of the prior art is: the torque hysteresis band is set to a fixing value, and this just exists the bigger problem of magnetic linkage and torque pulsation.In recent years, many new control thoughts, particularly Based Intelligent Control thought began to be applied in the direct Torque Control of induction machine like fuzzy control, ANN Control etc.Correlation based on the magnetic linkage in the aforementioned analysis and torque pulsation and hysteresis band; It is directly relevant with motor torque ripple with the stator current variable quantity to add the spinner velocity variable quantity; The inventor considers the input as Fuzzy Controller of spinner velocity variable quantity and stator current variable quantity; Obtain the increment of torque hysteresis band through fuzzy rule, thus fast, accurately, dynamic adjustments torque hysteresis band value, make induction machine when low cruise; The pulsation of torque be can reduce effectively, speed steady-state error and periodically pulsing reduced; Weaken current harmonics simultaneously, the noise and vibration when reducing the motor operation reaches high performance direct torque control purpose, and concrete scheme is:
1) with
Figure 286052DEST_PATH_IMAGE002
and as two input variables;
Figure 538972DEST_PATH_IMAGE006
is as output variable; Set up two-dimentional fuzzy reasoning table (fuzzy rule is a kind of conventional control method of the prior art, and its ins and outs repeat no more at this); Wherein,
Figure 897272DEST_PATH_IMAGE002
is the spinner velocity variable quantity;
Figure 307525DEST_PATH_IMAGE004
is the stator current variable quantity,
Figure 358657DEST_PATH_IMAGE006
be torque hysteresis band variable quantity;
2) confirm the base value
Figure 221571DEST_PATH_IMAGE008
of torque hysteresis band according to following formula:
Figure 383562DEST_PATH_IMAGE010
In the formula;
Figure 382742DEST_PATH_IMAGE012
is the motor number of pole-pairs;
Figure 607706DEST_PATH_IMAGE014
is the stator magnetic linkage vector;
Figure 957916DEST_PATH_IMAGE016
is the stator voltage space vector;
Figure 985914DEST_PATH_IMAGE018
is the minimum switch periods of inverter;
Figure 105180DEST_PATH_IMAGE020
is the sampling time;
Figure 232536DEST_PATH_IMAGE022
is the motor leakage inductance: ; Wherein,
Figure 573836DEST_PATH_IMAGE026
is stator inductance;
Figure 813187DEST_PATH_IMAGE028
is inductor rotor, and
Figure 111445DEST_PATH_IMAGE030
is coefficient of mutual inductance;
3) regularly spinner velocity
Figure 433317DEST_PATH_IMAGE032
and stator current
Figure 740802DEST_PATH_IMAGE034
are sampled; Calculate the difference of
Figure 569080DEST_PATH_IMAGE032
in current sampling period and last sampling period; Obtain
Figure 303818DEST_PATH_IMAGE002
numerical value in the current period; Calculate the difference of
Figure DEST_PATH_IMAGE067
in current sampling period and last sampling period, obtain numerical value in the current period;
4), obtain
Figure 815516DEST_PATH_IMAGE006
corresponding under
Figure 961513DEST_PATH_IMAGE002
and
Figure 909877DEST_PATH_IMAGE004
value conditions in the current period according to two-dimentional fuzzy reasoning table;
5) according to the following formula to calculate the current
Figure 840542DEST_PATH_IMAGE002
and
Figure 489829DEST_PATH_IMAGE004
value conditions hysteresis torque width
Figure 292700DEST_PATH_IMAGE036
:
The output variable of adjustment hysteresis comparator according to
Figure DEST_PATH_IMAGE069
.
6) repeating step 3), 4), 5), recomputate
Figure 828035DEST_PATH_IMAGE036
.
Accompanying drawing 1 is existing a kind of Direct Torque Control System of Induction Machine, and this control system is mainly by forming with the lower part: stator flux observer, magnetic linkage amplitude are calculated and sector judgement, electromagnetic torque observation, speed control, magnetic flux and torque hysteresis comparator, the selection of switching voltage vector, rectification and inverter, induction motor etc.; Its operation principle is following: the current value that at first arrives according to the space vector of voltage and the sensor of inverter output; After passing through coordinate transform respectively, through stator magnetic linkage amplitude
Figure DEST_PATH_IMAGE071
and the electromagnetic torque value
Figure DEST_PATH_IMAGE073
that stator magnetic linkage calculates, torque calculation obtains reality.The error of given speed and actual rotor speed as input, is obtained given electromagnetic torque
Figure DEST_PATH_IMAGE075
through behind this speed regulator.Given electromagnetic torque
Figure 825553DEST_PATH_IMAGE075
, magnetic linkage value
Figure DEST_PATH_IMAGE077
are compared with Practical Calculation value
Figure 686193DEST_PATH_IMAGE073
,
Figure 933635DEST_PATH_IMAGE071
; Obtain torque error
Figure DEST_PATH_IMAGE079
, magnetic linkage error
Figure DEST_PATH_IMAGE081
; Stagnating through corresponding torque, magnetic linkage, (this step those skilled in the art generally are defined as the chain rate that stagnates than link with it to chain rate again; After the chain rate that stagnates is handled than link; System is the subsequent module output switching signal; Its concrete processing procedure is a prior art, repeats no more at this; The chain rate that stagnates than relevant with the present invention in the link be for the stagnant chain rate of torque; Hysteresis band with a torque error
Figure 879725DEST_PATH_IMAGE079
and a fixed numbers in the prior art stagnates chain rate than handling; The present invention then stagnates chain rate than handling with the currency of torque error
Figure 870815DEST_PATH_IMAGE079
and dynamic change
Figure 651208DEST_PATH_IMAGE036
; This also is the concrete applied links of the present invention in system; Promptly stagnant chain rate than link in, hysteresis band is replaced with dynamic value by fixed value) back produces torque, magnetic linkage switching signal
Figure DEST_PATH_IMAGE083
and
Figure DEST_PATH_IMAGE085
.Meanwhile, can obtain the sector position angle at stator magnetic linkage place by component
Figure DEST_PATH_IMAGE089
,
Figure DEST_PATH_IMAGE091
and the synthetic quantity thereof of stator magnetic linkage in
Figure DEST_PATH_IMAGE087
coordinate.Composite stator magnetic linkage switching signal
Figure 630403DEST_PATH_IMAGE085
, torque switch signal
Figure 860527DEST_PATH_IMAGE083
and sector position angle
Figure 655308DEST_PATH_IMAGE095
; Through looking into switching voltage vector option table, just can obtain inverter control signal
Figure DEST_PATH_IMAGE097
,
Figure DEST_PATH_IMAGE099
and
Figure DEST_PATH_IMAGE101
.Through the three-phase voltage and the electric current of inverter control induction machine, make motor finally reach the control purpose by control requirement output torque.
Under the stator coordinate system, the computing formula of torque is following:
Figure DEST_PATH_IMAGE103
（1）
In the formula;
Figure 38010DEST_PATH_IMAGE089
,
Figure 641903DEST_PATH_IMAGE091
are the component of stator magnetic linkage in
Figure DEST_PATH_IMAGE105
coordinate;
Figure DEST_PATH_IMAGE107
,
Figure DEST_PATH_IMAGE109
are the component of stator current in coordinate, and
Figure 770844DEST_PATH_IMAGE012
is the number of pole-pairs of motor.Stator magnetic linkage can be obtained by computes:
Figure DEST_PATH_IMAGE111
（2）
（3）
In the formula; ,
Figure DEST_PATH_IMAGE117
are the component of stator voltage in
Figure 5123DEST_PATH_IMAGE105
coordinate, and
Figure DEST_PATH_IMAGE119
is stator resistance.
With reference to the PWM voltage inverter, the instantaneous space voltage vector of output can be obtained by computes:
Figure DEST_PATH_IMAGE121
（4）
In the formula,
Figure 906214DEST_PATH_IMAGE097
,
Figure 110930DEST_PATH_IMAGE099
With
Figure 247513DEST_PATH_IMAGE101
Be switching states,
Figure DEST_PATH_IMAGE123
Be direct voltage,
Figure DEST_PATH_IMAGE125
( k=1,2 ..., 6)
It is the switching voltage vector of 6 different directions.
Accompanying drawing 2 is the Direct Torque Control System of Induction Machine sketch mapes that adopt the present invention program.What the present invention was different with prior art is, has increased the torque hysteresis band adjustable part in the frame of broken lines in the system of the present invention, and with the output variable of torque hysteresis band adjustable part as the torque input variable of chain rate that stagnate than link.This part is definite by the hysteresis band base value, hysteresis band incremental adjustments two large divisions forms.Which, by the hysteresis value determines the width of the base module to calculate the torque hysteresis width of the base value
Figure 466792DEST_PATH_IMAGE008
, the torque hysteresis width of the base value
Figure 663418DEST_PATH_IMAGE008
it is a fixed value, the follow-up process is always the same; torque hysteresis width
Figure 355430DEST_PATH_IMAGE036
by the hysteresis Width incremental adjustment module according to the formula
Figure 30125DEST_PATH_IMAGE038
calculated.Wherein, is a dynamic value variable in control procedure; Therefore;
Figure 428877DEST_PATH_IMAGE036
also is a dynamic value; This just feasible torque hysteresis band that finally acts on motor can dynamically be adjusted according to the numerical value of spinner velocity variable quantity and stator current variable quantity.Obviously, this is the not available ability of aforesaid control system of the prior art.
Fuzzy control is the control method (its basic principle is as shown in Figure 3) of existing maturation, but its concrete rule then will specifically be provided with according to different systems with parameter value.The inventor through a large amount of experiments and analysis, has summed up the following fuzzy control rule that is applicable to induction machine DTC control in the long term studies process:
Relation between spinner velocity, stator current and the torque hysteresis band three can be summarized as follows: under the situation that spinner velocity and stator current all increase rapidly, torque this moment will depart from command torque rapidly, and overshoot can be touched ring border, upper strata.At this moment, through inquiry switching voltage vector option table, the reverse voltage vector is with selected, and this voltage will force torque to reduce rapidly in torque response, cause less stress and be lower than the ring that stagnates, and like this, torque pulsation will can be in any more.Therefore; The torque hysteresis band of this moment should be not too small; If with hysteresis band increment
Figure 707204DEST_PATH_IMAGE006
get on the occasion of, just can avoid the generation of this situation.
If spinner velocity and stator current all reduce rapidly, corresponding torque error has bigger overshoot, also can have the excessive phenomenon of torque pulsation this moment.For avoiding the generation of this situation, should consider from the angle that reduces torque error.Therefore; Can torque hysteresis band
Figure 185590DEST_PATH_IMAGE036
be reduced;
Figure 766744DEST_PATH_IMAGE006
gets negative value with the hysteresis band increment, just can reach the purpose that torque hysteresis band
Figure 39593DEST_PATH_IMAGE036
reduces.
According to the above-mentioned analysis that torque pulsation is produced reason, can make corresponding fuzzy control rule by following several kinds of situation:
The situation I: when spinner velocity increases comparatively fast or moderate, and stator current is when also increasing, and should increase ring width this moment;
The situation II: when spinner velocity increases comparatively fast or moderate, and stator current is when reducing, and it is constant that should keep ring width this moment;
The situation III: when spinner velocity reduces comparatively fast or moderate, and stator current is when also reducing, and should reduce ring width this moment;
The situation IV: when spinner velocity reduces comparatively fast or moderate, and stator current is when increasing, and it is constant that should keep ring width this moment;
The situation V: when spinner velocity, stator current did not all change, it is constant that should keep ring width this moment;
The situation VI: when spinner velocity and one of them amount of stator current remain unchanged, and another amount should suitably increase ring width when increasing this moment;
The situation VII: when spinner velocity and one of them amount of stator current remain unchanged, and another amount should suitably reduce ring width when reducing this moment;
Fuzzy reasoning table adopts dual input, single output mode, and input variable (spinner velocity and stator current change) and output variable (torque hysteresis band increment) are carried out Fuzzy processing.And fuzzy quantity is divided into 7 sub-set, the language value of each sub-set is got and is respectively NB, NM, NS, ZO, PS, PM, PB, the implication of its representative is respectively: negative big, negative in, negative little, zero, just little, the center, honest.In order to distinguish each self-corresponding subclass of input variable and output variable; Spinner velocity variation, stator current variation and the pairing subclass of torque hysteresis band increment are used NB2 respectively ... PB2, NB1 ... PB1 and NB3 ... PB3 comes mark, can draw following two-dimentional fuzzy reasoning table thus:
Figure 2012100042518100002DEST_PATH_IMAGE128
Wherein,
Figure 643881DEST_PATH_IMAGE041
is
Figure 660379DEST_PATH_IMAGE004
corresponding fuzzy amount,
Figure 87250DEST_PATH_IMAGE043
be
Figure 531001DEST_PATH_IMAGE002
corresponding fuzzy amount.NB1, NM1, NS1, ZO1, PS1, PM1, PB1 is subset; NB2, NM2, NS2, ZO2, PS2, PM2, PB2 is
Figure 505090DEST_PATH_IMAGE043
subset; NB3, NM3, NS3, ZO3, PS3, PM3, PB3 as
Figure 795257DEST_PATH_IMAGE049
subset;
Figure 675489DEST_PATH_IMAGE049
is corresponds to the amount of blur;
The obfuscation domain of input variable and output variable chooses [6; 6]; The degree of membership curve of input variable can adopt like Figure 15, the Gaussian curve shown in 16 and express; The degree of membership curve of output variable adopts equally distributed triangular function shown in figure 17 to express, and the subclass value of resulting thus
Figure 471723DEST_PATH_IMAGE043
and
Figure 881976DEST_PATH_IMAGE041
sees the following form:
Figure 2012100042518100002DEST_PATH_IMAGE130
The subclass value of
Figure 477649DEST_PATH_IMAGE049
sees the following form:
Figure DEST_PATH_IMAGE131
Figure 278246DEST_PATH_IMAGE055
;
Figure 440237DEST_PATH_IMAGE057
;
Figure 704996DEST_PATH_IMAGE059
; Wherein, ,
Figure 280170DEST_PATH_IMAGE063
are the quantizing factor in the Fuzzy Processing, and
Figure 42589DEST_PATH_IMAGE065
is the scale factor in the Fuzzy Processing.
On the basis of aforementioned schemes, also can every fuzzy rule in the two-dimentional fuzzy reasoning table all be converted into the I/O sample of neural net, utilize the I/O sample again, adopt the BP learning algorithm that neural net is trained; After neural metwork training is accomplished; Neural net is placed system, be used for the calculating of step 4)
Figure 161855DEST_PATH_IMAGE006
.The neural network model structure is shown in accompanying drawing 4, and the contents such as adjustment of the training process of neural net and himself weights are routine techniques of the prior art, repeat no more at this.Adopt the benefit that neural net is carried out to be calculated to be:
(a) can make full use of the robustness of fuzzy control itself: the neural net that obtains has not only realized knowledge and the information translation between general fuzzy control rule and the neural net; Make the inside weights and the threshold value of neural net remember the fuzzy control DECISION KNOWLEDGE; Function with fuzzy reasoning has kept the robustness of fuzzy control itself well;
(b) can regulate torque hysteresis band value apace: the fuzzy control rule of realizing through neural net is not the rule rule through showing to the expression of knowledge, but impliedly is distributed in rule in the whole network.When using, needn't carry out complex search and reasoning, have the ability of the various control laws of parallel processing, as long as just can access hysteresis band increment output valve, thereby fast, dynamically regulate the torque hysteresis band through Distribution calculation at a high speed.
(c) can accurately reflect the dynamic changing process of hysteresis band: the nonlinear fitting ability of neural net self and generalization ability are followed the trail of the internal reasoning process of neural net, make neural net can obtain the nonlinear characteristic of hysteresis band increment.When input was the sample data of training, it can activate respective rule exactly and make reflection.And when the input of the fuzzy rule of input signal and training had difference, neural net can activate a series of dependency rules and carry out non-linear extensive processing, reflects the dynamic changing process of hysteresis band more accurately, makes controller have better control effect.
A kind of hardware configuration embodiment of the present invention adopts inverter bridge to supply power to induction motor shown in accompanying drawing 5, and DC generator adopts the s operation control scheme of DSP+PC as the load usefulness of induction motor.DSP mainboard by TMSLF2812 constitutes is accomplished velocity setting, stator current and DC bus-bar voltage sampling, voltage vector is synthetic and export etc.Rate signal detects through photoelectric encoder, and paired pulses is counted, thereby obtains the value of feedback of speed.Voltage then obtains through on off state and intermediate voltage reconstruct, needs the variable of observation, is transformed into analog signal through oscilloscope observation like stator magnetic linkage, electric current equivalent through the D/A of DSP master control borad.Isolated drive circuit is sent in the vector pulse that calculates, the switching tube of inverter is controlled.Variable torque hysteresis band calculates and in PC, realizes, DSP and PC are through the dual port RAM swap data.
Software design partly comprises contents such as main program and interrupt service routine.Main program accomplish each module of system initialization, interrupt being provided with etc.Accompanying drawing 6 is DSP main program flow charts of a kind of specific embodiment of the present invention.The control algolithm of whole DTC system is accomplished by interrupt service routine, and accompanying drawing 7 is DSP interrupt service routine flow charts of a kind of specific embodiment of the present invention.
Accompanying drawing 8 be a kind of specific embodiment of the present invention realize variable torque hysteresis band calculation flow chart with PC.PC utilizes dual port RAM to read rotating speed and stator current value that master board is sent in real time; Calculate variable torque hysteresis band
Figure DEST_PATH_IMAGE133
through fuzzy-neural hybrid intelligent in the PC; And it is sent into dual port RAM, for the usefulness of DSP master control borad.
Be effect more of the present invention, the DTC system of (the present invention program) comparison that experimentizes when (prior art) and the variable situation of torque hysteresis band during with torque hysteresis band fixing situation.Experiment is at nominal load, carries out under the 5Hz slow-speed of revolution condition.The parameter of testing used induction motor is following:
Rated power
Figure DEST_PATH_IMAGE135
; Rated voltage
Figure DEST_PATH_IMAGE137
; Rated current
Figure DEST_PATH_IMAGE139
; Rated frequency
Figure DEST_PATH_IMAGE141
; Number of pole-pairs
Figure DEST_PATH_IMAGE143
; Rated speed ; Nominal torque
Figure DEST_PATH_IMAGE147
; Stator resistance
Figure DEST_PATH_IMAGE149
; Rotor resistance
Figure DEST_PATH_IMAGE151
; Main inductance
Figure DEST_PATH_IMAGE153
, leakage inductance
Figure DEST_PATH_IMAGE155
.
Accompanying drawing 9 is the fixing stable state electromagnetic torque pulsation oscillograms of torque hysteresis band, electromagnetic torque pulsation oscillogram when accompanying drawing 10 is the present invention program's stable state.Can find out that from comparison owing to adopted the inventive method, the torque pulsation of oscillogram shown in Figure 10 significantly reduces.
Accompanying drawing 11 is fixedly stator current waveforms figure during stable state of torque hysteresis band, stator current waveforms figure when accompanying drawing 12 is the present invention program's stable state.Can be found out that by accompanying drawing 11 when the 5Hz low velocity is moved, adopt the fixing stator current waveforms of torque hysteresis band to distort relatively more severely, harmonic content is bigger, this is the main cause that causes torque pulsation big; Can be found out that by accompanying drawing 12 behind employing the present invention program, stator current waveforms has basically no distortion, harmonic content is less, and the harmonic wave that influences torque pulsation is very little, so torque pulsation is little.
Accompanying drawing 13 is the fixing stator magnetic flux trajectory diagrams of torque hysteresis band, and accompanying drawing 14 is the present invention program's stator magnetic flux trajectory diagrams.From comparison, can find out, adopted the present invention program after, flux path becomes smoothly by coarse, pulsation significantly reduces.
In sum, adopt the steady-state process of the present invention program's DTC system at low speed, its electromagnetic torque pulsation, stator current harmonic wave and flux path pulsation are all less.This method can guarantee that the DTC system has good runnability when low velocity, thereby realizes high performance DTC control.

Claims (3)

1. the hybrid intelligent control method of torque hysteresis band in the DTC system; It is characterized in that: the steps include: 1) with and
Figure 2012100042518100001DEST_PATH_IMAGE004
as two input variables;
Figure 2012100042518100001DEST_PATH_IMAGE006
as output variable, sets up two-dimentional fuzzy reasoning table; Wherein,
Figure 687676DEST_PATH_IMAGE002
is the spinner velocity variable quantity;
Figure 253787DEST_PATH_IMAGE004
is the stator current variable quantity,
Figure 224629DEST_PATH_IMAGE006
be torque hysteresis band variable quantity;
2) confirm the base value
Figure 2012100042518100001DEST_PATH_IMAGE008
of torque hysteresis band according to following formula:
Figure 2012100042518100001DEST_PATH_IMAGE010
In the formula;
Figure 2012100042518100001DEST_PATH_IMAGE012
is the motor number of pole-pairs;
Figure 2012100042518100001DEST_PATH_IMAGE014
is the stator magnetic linkage vector;
Figure 2012100042518100001DEST_PATH_IMAGE016
is the stator voltage space vector;
Figure 2012100042518100001DEST_PATH_IMAGE018
is the minimum switch periods of inverter;
Figure 2012100042518100001DEST_PATH_IMAGE020
is the sampling time;
Figure 2012100042518100001DEST_PATH_IMAGE022
is the motor leakage inductance: ; Wherein,
Figure 2012100042518100001DEST_PATH_IMAGE026
is stator inductance; is inductor rotor, and
Figure 2012100042518100001DEST_PATH_IMAGE030
is coefficient of mutual inductance;
3) regularly spinner velocity and stator current
Figure 2012100042518100001DEST_PATH_IMAGE034
are sampled; Calculate the difference of
Figure 717360DEST_PATH_IMAGE032
in current sampling period and last sampling period; Obtain
Figure 42162DEST_PATH_IMAGE002
numerical value in the current period; Calculate the difference of
Figure 84067DEST_PATH_IMAGE034
in current sampling period and last sampling period, obtain
Figure 177925DEST_PATH_IMAGE004
numerical value in the current period;
4), obtain
Figure 573287DEST_PATH_IMAGE006
corresponding under
Figure 912663DEST_PATH_IMAGE002
and
Figure 459182DEST_PATH_IMAGE004
value conditions in the current period according to two-dimentional fuzzy reasoning table;
5) according to the following formula to calculate the current
Figure 521651DEST_PATH_IMAGE002
and
Figure 427291DEST_PATH_IMAGE004
value conditions hysteresis torque width
Figure 2012100042518100001DEST_PATH_IMAGE036
:
Figure 2012100042518100001DEST_PATH_IMAGE038
The output variable of adjustment hysteresis comparator according to
Figure 336472DEST_PATH_IMAGE036
;
6) repeating step 3), 4), 5), recomputate
Figure 720180DEST_PATH_IMAGE036
.
2. the hybrid intelligent control method of torque hysteresis band in the DTC according to claim 1 system is characterized in that: two-dimentional fuzzy reasoning table such as the following table described in the step 1):
Where,
Figure DEST_PATH_IMAGE041
is
Figure 598750DEST_PATH_IMAGE004
corresponds to the amount of blur,
Figure DEST_PATH_IMAGE043
is
Figure DEST_PATH_IMAGE045
corresponds to the amount of blur; NB1, NM1, NS1, ZO1, PS1, PM1, PB1 as
Figure DEST_PATH_IMAGE047
subset; NB2, NM2, NS2, ZO2, PS2, PM2, PB2 is
Figure 426022DEST_PATH_IMAGE043
subset; NB3, NM3, NS3, ZO3, PS3, PM3, PB3 is
Figure DEST_PATH_IMAGE049
The subset;
Figure 711965DEST_PATH_IMAGE049
is corresponds to the amount of blur;
and
Figure 69762DEST_PATH_IMAGE041
a subset of the values below:
Figure 2012100042518100001DEST_PATH_IMAGE050
The subclass value of
Figure 953536DEST_PATH_IMAGE049
sees the following form:
Figure DEST_PATH_IMAGE053
;
Figure DEST_PATH_IMAGE055
;
Figure DEST_PATH_IMAGE057
; Wherein,
Figure DEST_PATH_IMAGE059
, are the quantizing factor in the Fuzzy Processing, and
Figure DEST_PATH_IMAGE063
is the scale factor in the Fuzzy Processing.
3. the hybrid intelligent control method of torque hysteresis band in the DTC according to claim 1 and 2 system; It is characterized in that: the I/O sample that every fuzzy rule in the two-dimentional fuzzy reasoning table all is converted into neural net; Utilize the I/O sample again, adopt the BP learning algorithm that neural net is trained; After neural metwork training is accomplished; Neural net is placed system, be used for the calculating of step 4)
Figure 446090DEST_PATH_IMAGE006
.
CN2012100042518A 2012-01-09 2012-01-09 Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system Expired - Fee Related CN102522948B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2012100042518A CN102522948B (en) 2012-01-09 2012-01-09 Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2012100042518A CN102522948B (en) 2012-01-09 2012-01-09 Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system

Publications (2)

Publication Number Publication Date
CN102522948A true CN102522948A (en) 2012-06-27
CN102522948B CN102522948B (en) 2013-09-11

Family

ID=46293766

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2012100042518A Expired - Fee Related CN102522948B (en) 2012-01-09 2012-01-09 Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system

Country Status (1)

Country Link
CN (1) CN102522948B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104158455A (en) * 2014-08-25 2014-11-19 东南大学 Driving control system of power robot
CN104242743A (en) * 2013-06-21 2014-12-24 福特全球技术公司 Determination of Hysteresis Controller Band for IPMSM-Based Drive System
CN105048896A (en) * 2015-07-08 2015-11-11 河南科技大学 Brushless DC motor direct torque adaptive fuzzy control method
CN111835252A (en) * 2019-04-17 2020-10-27 华北电力大学（保定） Flexible load vibration and PMSM torque ripple comprehensive suppression method under stator current vector orientation considering electrical loss
CN114465549A (en) * 2021-07-08 2022-05-10 湖南科技大学 Switched reluctance motor direct instantaneous torque control method based on variable hysteresis PWM

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101645687A (en) * 2009-09-11 2010-02-10 山西合创电力科技有限公司 Integrated method of vector and direct torque control of induction wind generator
EP2190113A1 (en) * 2008-11-19 2010-05-26 Abb Research Ltd. Method and apparatus for controlling an electrical Machine with direct torque control

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP2190113A1 (en) * 2008-11-19 2010-05-26 Abb Research Ltd. Method and apparatus for controlling an electrical Machine with direct torque control
CN101645687A (en) * 2009-09-11 2010-02-10 山西合创电力科技有限公司 Integrated method of vector and direct torque control of induction wind generator

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
沈红平: "基于MRAS的异步电机无速度传感器直接转矩控制", 《中国优秀硕士学位论文全文数据库》 *
涂承宇等: "《模糊控制理论与实践》", 31 August 1998, 地震出版社 *
马德中: "采用滞环宽度优化控制的感应电机直接转矩控制", 《微特电机》 *
高圣伟等: "基于模糊逻辑的DTC减小转矩脉动方法研究", 《河北工业大学学报》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104242743A (en) * 2013-06-21 2014-12-24 福特全球技术公司 Determination of Hysteresis Controller Band for IPMSM-Based Drive System
CN104242743B (en) * 2013-06-21 2018-11-13 福特全球技术公司 Based on the determination of the hystersis controller frequency band of the drive system of internal permanent magnet synchronous motor
CN104158455A (en) * 2014-08-25 2014-11-19 东南大学 Driving control system of power robot
CN104158455B (en) * 2014-08-25 2016-08-24 东南大学 A kind of driving control system of Power Robot
CN105048896A (en) * 2015-07-08 2015-11-11 河南科技大学 Brushless DC motor direct torque adaptive fuzzy control method
CN105048896B (en) * 2015-07-08 2018-03-23 河南科技大学 A kind of brshless DC motor Direct Torque adaptive fuzzy control method
CN111835252A (en) * 2019-04-17 2020-10-27 华北电力大学（保定） Flexible load vibration and PMSM torque ripple comprehensive suppression method under stator current vector orientation considering electrical loss
CN111835252B (en) * 2019-04-17 2023-08-11 华北电力大学（保定） Flexible load vibration and PMSM torque pulsation comprehensive suppression method considering electrical loss
CN114465549A (en) * 2021-07-08 2022-05-10 湖南科技大学 Switched reluctance motor direct instantaneous torque control method based on variable hysteresis PWM
CN114465549B (en) * 2021-07-08 2023-05-23 湖南科技大学 Direct instantaneous torque control method of switched reluctance motor based on hysteresis PWM

Also Published As

Publication number Publication date
CN102522948B (en) 2013-09-11

Similar Documents

Publication Publication Date Title
CN107276465B (en) A kind of torque-current neural network switch reluctance motor control method and system
CN102522948B (en) Hybrid intelligent adjusting method of torque hysteresis width in DTC (Direct Torque Control) system
CN109327178A (en) A kind of switched reluctance machines Multi-step predictive control device building method
Lin et al. Adaptive fuzzy-neural-network control for a DSP-based permanent magnet linear synchronous motor servo drive
CN106849793B (en) A kind of Over Electric Motor with PMSM fuzzy Neural Network Control System
Sun et al. Optimal design of terminal sliding mode controller for direct torque control of SRMs
CN110022109A (en) Torque-current neural network model SRM torque pulsation control method and system
CN101571705B (en) Position servo system and method
Meshcheryakov et al. Modeling and analysis of vector control systems for asynchronous motor
CN113285481B (en) Grid-connected converter inductance parameter online estimation method, prediction control method and system
CN109391202A (en) Permanent magnet synchronous motor model prediction-Direct Torque Control
CN109302110A (en) A kind of asynchronous machine broad sense cascade model prediction flux linkage control method and device
CN109742999A (en) A kind of the SRM method for controlling torque and system of dynamic neural network adaptive inversion
Zhang et al. The Brushless DC motor control system Based on neural network fuzzy PID control of power electronics technology
CN109787495A (en) A kind of model-free predictive-current control method of two-level inverter
Wang et al. A novel PID controller for BLDCM speed control using dual fuzzy logic systems with HSA optimization
Krishnan et al. A smart FPGA based design and implementation of grid connected direct matrix converter with IoT communication
CN105871278A (en) Direct feature control-based novel salient-pole permanent magnet synchronous motor control method
CN113517832A (en) Low-voltage servo discrete linear active disturbance rejection control method
Dasari et al. Modified Luo converter based FOPID controller for torque ripple minimization in BLDC drive system
Kalaivani et al. Intelligent control for torque-ripple minimization in switched reluctance motor
CN110518625B (en) Grid-connected inverter direct-current component suppression method with variable learning rate BP-PID control
CN107093971B (en) A kind of moment controlling system and control method of permanent magnet synchronous motor
Sun et al. A single neuron PID controller based PMSM DTC drive system fed by fault tolerant 4-switch 3-phase inverter
Jiao et al. Research on PMSM sensorless control based on improved RBF neural network algorithm

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant
C17 Cessation of patent right
CF01 Termination of patent right due to non-payment of annual fee

Granted publication date: 20130911

Termination date: 20140109