CN109802611A - A kind of method for controlling torque of internal permanent magnet synchronous motor - Google Patents
A kind of method for controlling torque of internal permanent magnet synchronous motor Download PDFInfo
- Publication number
- CN109802611A CN109802611A CN201910052702.7A CN201910052702A CN109802611A CN 109802611 A CN109802611 A CN 109802611A CN 201910052702 A CN201910052702 A CN 201910052702A CN 109802611 A CN109802611 A CN 109802611A
- Authority
- CN
- China
- Prior art keywords
- current
- torque
- shaft current
- phase
- neural network
- 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.)
- Pending
Links
Landscapes
- Control Of Ac Motors In General (AREA)
- Control Of Motors That Do Not Use Commutators (AREA)
Abstract
The present invention proposes a kind of method for controlling torque of internal permanent magnet synchronous motor, includes: the three-phase current i measured under electric machine operation stateA,iB,iC；By three-phase current iA,iB,iCTransformation obtains biphase current iD,iQ；By the biphase current iD,iQTransformation obtains d-q shaft current id,iq；By d-q shaft current id,iqIt is input in trained neural network, the torque predicted；The operation of PI torque controller is passed through into the torque of neural network output and torque command, export q shaft current；The q shaft current obtains the d shaft current of corresponding MTPA value after MTPA operation；D-q shaft current is input in PI current controller, the corresponding voltage value of d-q shaft current value is obtained；D-q shaft current is worth corresponding voltage value and exports three-phase voltage by Park inverse transformation；By three-phase voltage by three-phase alternating voltage needed for the output control motor operating of SVPWM inverter.The present invention forms feed-forward compensation system using the torque of LSTM neural network prediction, using LSTM Neural Network Based Nonlinear capability of fitting is strong, precision is high, has the advantages that memory capability to historical data.
Description
Technical field
The present invention relates to motor fields, and in particular to a kind of method for controlling torque of internal permanent magnet synchronous motor.
Background technique
Internal permanent magnet synchronous motor (IPMSM), which has, safeguards that simple, light-weight, small in size, high-efficient, power density is high
The features such as, it is widely applied in fields such as robot, electric vehicle, numerically-controlled machine tool, convertible frequency air-conditioners.Especially in servo drive
Field, internal permanent magnet synchronous motor gradually replace direct current generator, stepper motor and become servo-drive developing direction.
With the rapid development of power electronic technique, microelectric technique, sensor technology and motor control theory, built-in type permanent-magnet is synchronous
The research and extension application of electric machine control system receives the most attention of researcher.
Internal permanent magnet synchronous motor the parameter of electric machine it is non-linear, uncertain etc. due to, accurately control motor
Torque has certain difficulty.Since the direct measurement cost of torque is higher, and influenced by accuracy of instrument and response speed
It is larger, therefore the observation to motor torque is usually realized by algorithm.In general, the observation of torque have direct computing method,
Full rank and depression of order dragon Burger observer, model reference adaptive method and Kalman filter and Neural Network Observer etc., card
Thalmann filter can calculate most automatically according to statistical property of the error of state variable estimation, measurement noise and system noise etc.
Excellent feedback gain, relative to common imperial Burger observer, suppression of the Kalman filter to measurement error and interference etc.
Ability processed is stronger, and some is to the steady of the system for forming feedforward compensation using the load torque observed in current research achievement
It is qualitative to be analyzed；Some points out the feedforward compensation for introducing observation torque in the controller, forms two-degree-freedom controller, can
To improve the response speed and robustness of controller；Also by emulation and experimental verification, it is indicated that rushed in same load torque
It hits down, the fluctuation of revolving speed greatly reduces after introducing feedforward compensation.
Existing torque observation has DSP direct computing method, full rank and depression of order dragon Burger observer, model reference adaptive method
With Kalman filter and BP neural network controller etc., Kalman filter can according to state variable estimate error, survey
Statistical property of noise and system noise etc. is measured to calculate optimal feedback gain automatically, relative to common Long Baigeguan
Device is surveyed, Kalman filter is stronger to the rejection ability of measurement error and interference etc., the disadvantage is that it is larger directly to calculate error
And dynamic property is poor, and Long Baige observer is poor to the rejection ability of measurement error and interference etc., Kalman filter
It is poor for torque ripple minimization ability.Although BP neural network overcomes NONLINEAR CALCULATION complexity, controls slow ask
Topic, but it is weaker for the rejection ability of torque ripple, i.e., when motor torque fluctuates BP neural network controller have compared with
Big torque burr phenomena.
Summary of the invention
In view of the foregoing deficiencies of prior art, the purpose of the present invention is to provide a kind of internal permanent magnet synchronous motors
Method for controlling torque, just solve due to the parameter of electric machine it is non-linear and it is uncertain caused by motor torque be difficult to accurately estimate and
The problems such as control.
In order to achieve the above objects and other related objects, the present invention provides a kind of torque control of internal permanent magnet synchronous motor
Method processed, this method comprises:
Measure the three-phase current i under electric machine operation stateA,iB,iC；
By the three-phase current iA,iB,iCTransformation obtains biphase current iD,iQ；
By the biphase current iD,iQTransformation obtains d-q shaft current id,iq；
By the d-q shaft current id,iqIt is input in trained neural network, the torque predicted；
The torque that the neural network is exported and torque command pass through the operation of PI torque controller, export q shaft current；Institute
Q shaft current is stated after MTPA operation, obtains the d shaft current of corresponding MTPA value；
D-q shaft current is input in PI current controller, the corresponding voltage value of d-q shaft current value is obtained；
D-q shaft current is worth corresponding voltage value and exports three-phase voltage by Park inverse transformation；
By three-phase voltage by three-phase alternating voltage needed for the output control motor operating of SVPWM inverter.
Optionally, described by the three-phase current iA,iB,iCTransformation obtains biphase current iD,iQIn step, the three-phase
Electric current iA,iB,iCIt converts to obtain biphase current i by ClarkD,iQ。
Optionally, described by the biphase current iD,iQTransformation obtains d-q shaft current id,iqIn step, by two-phase electricity
Flow iD,iQIt converts to obtain d-q shaft current i by Parkd,iq。
Optionally, the neural network is LSTM neural network.
As described above, a kind of method for controlling torque of internal permanent magnet synchronous motor of the invention, has below beneficial to effect
Fruit:
A kind of method for controlling torque of internal permanent magnet synchronous motor proposed by the present invention, uses neural network prediction
Torque forms feed-forward compensation system, using LSTM Neural Network Based Nonlinear capability of fitting is strong, precision is high, has memory to historical data
The advantages that ability, introduces the feedforward compensation of prediction torque in the controller, forms two degrees of freedom closed-loop control, control can be improved
The response speed of device, and can be good at solving the problems, such as that existing torque observer is poor for torque ripple minimization ability.
Detailed description of the invention
In order to which the present invention is further explained, described content, with reference to the accompanying drawing makees a specific embodiment of the invention
Further details of explanation.It should be appreciated that these attached drawings are only used as typical case, and it is not to be taken as to the scope of the present invention
It limits.
Fig. 1 is a kind of flow chart of the method for controlling torque of internal permanent magnet synchronous motor of the present invention；
Fig. 2 is three phase static shafting and two-phase static axial system；
Fig. 3 is two-phase static axial system and two-phase rotary axis；
Fig. 4 is LSTM neural network structure figure；
Fig. 5 is controller principle block diagram；
Fig. 6 is the three-phase current of PI model；
Fig. 7 is the three-phase current of LSTM model；
Fig. 8 is the motor torque of PI model；
Fig. 9 is the motor torque of LSTM model.
Specific embodiment
Illustrate embodiments of the present invention below by way of specific specific example, those skilled in the art can be by this specification
Other advantages and efficacy of the present invention can be easily understood for disclosed content.The present invention can also pass through in addition different specific realities
The mode of applying is embodied or practiced, the various details in this specification can also based on different viewpoints and application, without departing from
Various modifications or alterations are carried out under spirit of the invention.It should be noted that in the absence of conflict, following embodiment and implementation
Feature in example can be combined with each other.
It should be noted that illustrating the basic structure that only the invention is illustrated in a schematic way provided in following embodiment
Think, only shown in schema then with related component in the present invention rather than component count, shape and size when according to actual implementation
Draw, when actual implementation kenel, quantity and the ratio of each component can arbitrarily change for one kind, and its assembly layout kenel
It is likely more complexity.
As shown in Figure 1, the present invention proposes a kind of method for controlling torque of internal permanent magnet synchronous motor, this method include with
Lower step:
S1 measures the three-phase current i under electric machine operation stateA,iB,iC；
S2 converts three-phase current to obtain biphase current iD,iQ。
Specifically, three-phase current converts to obtain biphase current i by ClarkD,iQ, wherein Clarke transformation is three phase static
Shafting is to the transformation of two-phase static axial system, also known as 3S/2S transformation.As shown in Fig. 2, ABC winding is symmetrical three-phase windings, the number of turns
It is NA；D-q axis winding is orthogonal two-phase winding, and the number of turns is ND；D axis winding direction is consistent with A phase winding axis.
After each winding is passed through electric current, its fundamental wave magnetomotive force is only counted, the magnetomotive force direction and winding axis that forward current generates
Unanimously；Electric current can be random waveform and any instant value.Magnetomotive force is equivalent be different the physical basis converted between shafting and
Basic principle.Because only that in this way, just will not influence energy converting between mechanical in motor.
If three-phase current iA,iB,iCThe fundamental wave synthesis magnetomotive force and biphase current i of generationD,iQThe fundamental wave synthesis magnetic of generation is dynamic
Gesture is equal, then has:
If NA=kND, formula (1), (2) then can transform to:
The internal permanent magnet synchronous motor three-phase windings that the present invention uses are without the neutral conductor, it may be assumed that
i0=k'(iA+iB+iC)=0 (5)
To make transformation meet power constraint independent of time, thenI.e.From
And two-phase static axial system electric current can be obtained are as follows:
Since three-phase current is symmetrical, it may be assumed that
iA+iB+iC=0 (7)
Formula (6) are brought into obtain:
S3 is by biphase current iD、iQTransformation obtains d-q shaft current id、iq。
Specifically, by biphase current iD、iQIt converts to obtain d-q shaft current i by Parkd、iq。
Wherein, Park transformation is two-phase static axial system to the transformation of two-phase synchronous rotary shafting, and also known as 2S/2R is converted.
As shown in figure 3, it is identical as phase winding the number of turns of two-phase synchronous rotary shafting to set two-phase static axial system, using D axis as spatial parameter
Axis, d-q shafting is with an angle speed ωrIt rotates counterclockwise, space phase angle is θr(electrical angle), according to the equivalent original of magnetomotive force
Then, then have:
id=iDcosθr+iQsinθr (9)
iq=-iDsinθr+iQcosθr (10)
It can obtain:
Formula (11) is Park transformation expression formula, while can satisfy power constraint independent of time.
The d-q shaft current i that S4 will be obtainedd、iqIt is input in trained neural network, the torque predicted.
Specifically, the d-q shaft current i that will be obtainedd、iqIt is input in trained LSTM neural network, is predicted
Torque.Wherein, LSTM (Long Short-TermMemory) is shot and long term memory network, is a kind of time recurrent neural network,
It can learn the information relied on for a long time, structure chart is as shown in Figure 4:
The basic component units of LSTM neural network be the middle section Fig. 4 shown in, be made of 3 parts: input gate, forget door and
Out gate.Wherein input gate is determined by currently inputting, and is forgotten door and is codetermined by currently inputting to export with last moment, out gate
Then codetermined by the output of input gate and forgetting door.
The operation of PI torque controller is passed through in the torque and torque command that S5 exports the LSTM neural network, exports q axis
Electric current；The q shaft current obtains the d shaft current of corresponding MTPA value after MTPA operation；
Obtained d-q shaft current is input in PI current controller obtains the corresponding voltage value of d-q shaft current value again by S6；
S7 is then by the obtained corresponding voltage value of d-q shaft current value by Park inverse transformation (inverse process of Park transformation)
Export three-phase voltage；
S8 is finally by gained three-phase voltage by three-phase alternating current needed for the output control motor operating of SVPWM inverter
Pressure is finally reached the effect of control motor operation.
A kind of arrow of internal permanent magnet synchronous motor based on shot and long term Memory Neural Networks (LSTM) provided by the invention
Amount control torque observer, functional block diagram are as shown in Figure 5:
The vector controlled torque observer mainly includes PI torque observer, MTPA module, PI current controller, Park change
/ inverse transform block, SVPWM inverter, decoupling device, IPMSM and LSTM neural network module are changed, wherein TrefRefer to for torque
It enables.
The principle specific workflow is as follows:
System torque order T is given firstref, while LSTM neural network module receives the d-q axis electricity that current system transmits
Stream, and export prediction torque Te；
Then by the T of the LSTM neural network module fitting in current operating systemeWith torque command TrefIt inputs together
To PI torque controller；
PI torque controller is according to the T of inputrefThe output T of instruction and LSTM neural network moduleeAdjust iqref(q axis electricity
Stream order), it is tabled look-up and is found corresponding to i by MTPA moduleqrefIdref(order of d shaft current) output；
PI current controller receives iqrefAnd idref, and v is exported under the action of decoupling devicedref(order of d shaft voltage) and
vqref(order of q shaft voltage)；
Later by vdrefAnd vqrefBy Park inverse transformation (inverse process of Park transformation) output three-phase voltage；
Three-phase alternating current needed for obtained three-phase voltage is input to SVPWM inverter again and exports control motor operating
Voltage；
Then IPMSM operating status is regulated and controled by gained three-phase alternating voltage, the built-in type permanent-magnet used due to the present invention
Synchronous motor three-phase windings are zero without the neutral conductor, i.e. three-phase current vector sum, therefore need to only take wherein biphase current ia、ibWith angle speed
Spend ωr?；
Then by biphase current ia、ibAnd angular velocity omegarInput Park conversion module obtains id、iq, then Park converted
The d-q shaft current i arrivedd、iqIt is input to LSTM neural network module and obtains prediction torque Te；
Above step is repeated, i is adjusted by PI torque controllerqrefUntil the actual prediction of motor exports TeEqual to current
Torque command Tref(Te=Tref), meanwhile, PI current controller adjusts vdrefAnd vqref, until the actual current in motor is equal to
D-q shaft current order (idref=id,iqref=iq), the output torque of motor is equal to torque command T at this timeref, final control electricity
Machine operation.
Experimental verification is done in present invention emulation, and simulation parameter is as follows:
PI torque controller Proportional coefficient K p is 0.01, and integral coefficient Ki is 300；
PI current controller Proportional coefficient K p is 50, and integral coefficient Ki is 40；
Motor number of pole-pairs is 3；
Unit resistance is 51.2m Ω；
Maximum current 120A, DC voltage 120V；
Maximum (top) speed 1500rpm；
Torque capacity 100Nm.
In emulation, use conventional PI control device as the reference experiment of LSTM torque observer.In this moment torque of t=2s
Order has intercepted the part 1.95s-2.05s from obtained emulation data to illustrate effect for 10Nm from 0Nm jump
Data, and mapped with MATLAB, result figure is as shown in Fig. 6-Fig. 9:
The above-described embodiments merely illustrate the principles and effects of the present invention, and is not intended to limit the present invention.It is any ripe
The personage for knowing this technology all without departing from the spirit and scope of the present invention, carries out modifications and changes to above-described embodiment.Cause
This, institute is complete without departing from the spirit and technical ideas disclosed in the present invention by those of ordinary skill in the art such as
At all equivalent modifications or change, should be covered by the claims of the present invention.
Claims (4)
1. a kind of method for controlling torque of internal permanent magnet synchronous motor, which is characterized in that this method comprises:
Measure the three-phase current i under electric machine operation stateA,iB,iC；
By the three-phase current iA,iB,iCTransformation obtains biphase current iD,iQ；
By the biphase current iD,iQTransformation obtains d-q shaft current id,iq；
By the d-q shaft current id,iqIt is input in trained neural network, the torque predicted；
The torque that the neural network is exported and torque command pass through the operation of PI torque controller, export q shaft current；The q
Shaft current obtains the d shaft current of corresponding MTPA value after MTPA table lookup operations；
D-q shaft current is input in PI current controller, the corresponding voltage value of d-q shaft current value is obtained；
D-q shaft current is worth corresponding voltage value and exports three-phase voltage by Park inverse transformation；
By three-phase voltage by three-phase alternating voltage needed for the output control motor operating of SVPWM inverter.
2. a kind of method for controlling torque of internal permanent magnet synchronous motor according to claim 1, which is characterized in that in institute
It states the three-phase current iA,iB,iCTransformation obtains biphase current iD,iQIn step, the three-phase current iA,iB,iCBy
Clark converts to obtain biphase current iD,iQ。
3. a kind of method for controlling torque of internal permanent magnet synchronous motor according to claim 1, which is characterized in that in institute
It states the biphase current iD,iQTransformation obtains d-q shaft current id,iqIn step, by biphase current iD,iQIt is converted by Park
To d-q shaft current id,iq。
4. a kind of method for controlling torque of internal permanent magnet synchronous motor according to claim 1, which is characterized in that described
Neural network is LSTM neural network.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910052702.7A CN109802611A (en) | 2019-01-21 | 2019-01-21 | A kind of method for controlling torque of internal permanent magnet synchronous motor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910052702.7A CN109802611A (en) | 2019-01-21 | 2019-01-21 | A kind of method for controlling torque of internal permanent magnet synchronous motor |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109802611A true CN109802611A (en) | 2019-05-24 |
Family
ID=66559879
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910052702.7A Pending CN109802611A (en) | 2019-01-21 | 2019-01-21 | A kind of method for controlling torque of internal permanent magnet synchronous motor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109802611A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112217428A (en) * | 2020-09-18 | 2021-01-12 | 江苏科技大学 | Underwater robot propulsion system position-sensor-free control method |
| WO2023124921A1 (en) * | 2021-12-31 | 2023-07-06 | 深圳先进技术研究院 | Torque estimation method and apparatus for permanent magnet synchronous motor, and device and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104300866A (en) * | 2014-10-10 | 2015-01-21 | 四川长虹电器股份有限公司 | Motor control method based on SVPWM |
| CN107147342A (en) * | 2017-05-31 | 2017-09-08 | 合肥申芯电子技术有限责任公司 | A kind of induction motor parameter identification system and method |
| JP2018078762A (en) * | 2016-11-11 | 2018-05-17 | 国立大学法人 名古屋工業大学 | Vibration control system using embedded magnet synchronous motor |
-
2019
- 2019-01-21 CN CN201910052702.7A patent/CN109802611A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104300866A (en) * | 2014-10-10 | 2015-01-21 | 四川长虹电器股份有限公司 | Motor control method based on SVPWM |
| JP2018078762A (en) * | 2016-11-11 | 2018-05-17 | 国立大学法人 名古屋工業大学 | Vibration control system using embedded magnet synchronous motor |
| CN107147342A (en) * | 2017-05-31 | 2017-09-08 | 合肥申芯电子技术有限责任公司 | A kind of induction motor parameter identification system and method |
Non-Patent Citations (1)
| Title |
|---|
| 李涉川等: "基于卷积神经网络的永磁同步电机转矩观测器", 《集成技术》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112217428A (en) * | 2020-09-18 | 2021-01-12 | 江苏科技大学 | Underwater robot propulsion system position-sensor-free control method |
| WO2023124921A1 (en) * | 2021-12-31 | 2023-07-06 | 深圳先进技术研究院 | Torque estimation method and apparatus for permanent magnet synchronous motor, and device and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Antidisturbance speed control for induction machine drives using high-order fast terminal sliding-mode load torque observer | |
| Jiang et al. | Improved deadbeat predictive current control combined sliding mode strategy for PMSM drive system | |
| Kuang et al. | Speed control of load torque feedforward compensation based on linear active disturbance rejection for five-phase PMSM | |
| Sami et al. | Integral super twisting sliding mode based sensorless predictive torque control of induction motor | |
| Guo et al. | Model reference adaptive control of five-phase IPM motors based on neural network | |
| Rehman et al. | A new current model flux observer for wide speed range sensorless control of an induction machine | |
| Azza et al. | Development and experimental evaluation of a sensorless speed control of SPIM using adaptive sliding mode-MRAS strategy | |
| Yang et al. | Fast integral terminal sliding mode control with a novel disturbance observer based on iterative learning for speed control of PMSM | |
| Nie et al. | Deadbeat-direct torque and flux control for wound field synchronous machines | |
| Zhang et al. | Force ripple compensation in a PMLSM position servo system using periodic adaptive learning control | |
| Zhang et al. | Minimising torque ripple of SRM by applying DB‐DTFC | |
| Kemmetmüller et al. | Optimal torque control of permanent magnet synchronous machines using magnetic equivalent circuits | |
| CN109802611A (en) | A kind of method for controlling torque of internal permanent magnet synchronous motor | |
| Özdemir | A new stator voltage error-based MRAS model for field-oriented controlled induction motor speed estimation without using voltage transducers | |
| Accetta et al. | Space‐vector state dynamic model of SynRM considering self‐and cross‐saturation and related parameter identification | |
| Kandoussi et al. | Real time implementation of a new fuzzy-sliding-mode-observer for sensorless IM drive | |
| Wang et al. | An improved deadbeat predictive current control based on parameter identification for PMSM | |
| Vijayapriya et al. | Enhanced method of rotor speed and position estimation of permanent magnet synchronous Machine based on stator SRF-PLL | |
| Lazarini et al. | Sensorless three-phase induction motor direct torque control using sliding mode control strategy laboratory set-up for motor speed control teaching | |
| Wang et al. | Improving torque control accuracy and dynamics for high power or high speed induction machine drives that inherently operate at low switching-to-fundamental frequency ratios | |
| Virani et al. | Modelling and Control of PMSM Drive by Field Oriented Control For HEV | |
| Sayouti et al. | Real-time DSP implementation of DTC neural network-based for induction motor drive | |
| Carpiuc | Model–Based Control and Real–Time Simulation of a Four–Phase PMSM Traction Drive | |
| Fan et al. | The improved direct torque control of a new self-decelerating permanent-magnet in-wheel motor for electric vehicles | |
| Du et al. | A High-Fidelity and Computationally Efficient Model for an Electrically Excited Synchronous Generator Based on Current–Flux Linkage Neural Networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20190524 |