CN109150054B - LSSVM decision classification-based brushless direct current motor position sensing method - Google Patents
LSSVM decision classification-based brushless direct current motor position sensing method Download PDFInfo
- Publication number
- CN109150054B CN109150054B CN201811180628.9A CN201811180628A CN109150054B CN 109150054 B CN109150054 B CN 109150054B CN 201811180628 A CN201811180628 A CN 201811180628A CN 109150054 B CN109150054 B CN 109150054B
- Authority
- CN
- China
- Prior art keywords
- lssvm
- classification
- classifier
- rotor
- layer
- 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.)
- Active
Links
Images
Classifications
-
- H—ELECTRICITY
- H02—GENERATION; CONVERSION OR DISTRIBUTION OF ELECTRIC POWER
- H02P—CONTROL OR REGULATION OF ELECTRIC MOTORS, ELECTRIC GENERATORS OR DYNAMO-ELECTRIC CONVERTERS; CONTROLLING TRANSFORMERS, REACTORS OR CHOKE COILS
- H02P21/00—Arrangements or methods for the control of electric machines by vector control, e.g. by control of field orientation
- H02P21/14—Estimation or adaptation of machine parameters, e.g. flux, current or voltage
- H02P21/18—Estimation of position or speed
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Control Of Motors That Do Not Use Commutators (AREA)
Abstract
The invention provides a brushless direct current motor position sensing method based on Least Square Support Vector Machine (LSSVM) decision classification, aiming at the problem of brushless direct current motor rotor position detection. The method is characterized in that the voltage and the current of a brushless direct current motor stator are used as the input of a decision LSSVM, rotor position information is used as the output, the position of a direct current motor rotor is divided into 6 areas, the rotor position is represented by the sequence number 1-6 of the area where the rotor position is located, 15 LSSVM two classifiers are arranged, and hierarchical classification decision is carried out; training an SVM network by a grid optimization method to determine LSSVM optimal parameters, applying a trained network model to motor operation, collecting motor stator voltage and current as input of the LSSVM, determining final rotor position information by hierarchical decision, calculating a logical commutation signal by rotor position, and determining on-off of a corresponding switching tube corresponding to each area, namely the commutation logical signal.
Description
Technical Field
The invention relates to a rotor position sensing method in the field of brushless direct current motors, in particular to a brushless direct current motor position sensing method based on Least Square Support Vector Machine (LSSVM) decision classification.
Background
The brushless direct current motor controls the electronic commutation circuit through a rotor position signal to enable each winding of the stator armature to conduct commutation continuously, so that a stator magnetic field and a rotor permanent magnetic field always keep a space angle of about 90 degrees, and torque is generated to push the rotor to operate.
The rotor position information of the traditional brushless direct current motor is measured by a position sensor, and a position detection device needs to be installed, but the brushless direct current motor with the position detection device has the following defects: the motor volume is increased, which is not beneficial to the miniaturization of the motor; the position sensor is arranged in a limited space in the motor, so that the position sensor is difficult to install and maintain; difficult to adapt to harsh environments; the sensor is complicated to wire and is prone to introduce interference. Therefore, brushless dc motors without position sensors have become a focus of research.
The position measurement technology of the existing rotor position sensorless device mainly comprises a back electromotive force method, a current detection method and an intelligent algorithm. The back electromotive force is in direct proportion to the speed, so that a zero-crossing signal cannot be obtained by detecting the back electromotive force when the rotating speed is very low or even zero; the implementation of the current method mainly depends on the accuracy of the current sensor, and generally, the sensor can better obtain a current magnitude signal, and small changes of a current phase and a waveform are difficult to detect. Therefore, the counter-potential method and the current method have limitations when used alone.
Disclosure of Invention
The technical problem is as follows: the rotor signal detection methods with position sensors and various devices without position sensors have limitations, so that the method is difficult to be applied to occasions with higher requirements on the operation of the motor.
The technical scheme is as follows: in order to solve the problems, LSSVM decision classification is used in rotor position detection, and the motor can stably run according to requirements through current speed control. For a brushless direct current motor, a mapping model between a rotor position signal and motor voltage and current is established, and LSSVM decision classification is adopted to realize the mapping. And the motor voltage and current are used as the input of the LSSVM, and the rotor position information is used as the output, so that the judgment of the motor rotor position is realized.
The invention discloses a rotor position detection system based on an LSSVM (least squares support vector machine). according to the working principle of a brushless motor, the induction flux linkage of a A, B, C phase winding is a function of an electrical angle theta, and the induction flux linkage of a A, B, C phase winding has a certain relation with a three-phase end voltage and current, so that the electrical angle theta, namely a rotor position signal, can be predicted according to the three-phase end voltage and current.
The LSSVM is a binary model, and the purpose of the LSSVM is to find a hyperplane to segment samples, and the segmentation principle is interval maximization. The general form of the multivariate nonlinear classification model is described as follows: y is i＝sgn(g(xi) ) itIn (1)
i is 1,2, …, N represents the input quantity of LSSVM classification prediction model, yiRepresenting the model target output quantity. Mapping function
And mapping the samples of the original input space into a high-dimensional feature space omega by using a kernel function, and performing linear classification on the sample data in the feature space by using the mapping function.
The LSSVM classification function may be expressed as: g (x) ═ ω · Φ (x)) + b (1)
Wherein the vector ω ∈ RnThe offset b ∈ R. The SVM minimizes the sample error while using the principle of structure risk minimization, the function fitting problem can be described as an optimization problem:
in the formula
Is a function that maps input data to a high-dimensional feature space; omega epsilon to Rn，eiAs an error, ei，b∈R，C>And 0 is a penalty coefficient used for controlling the smoothness of the solution, the larger the value of the penalty coefficient is, the stronger the penalty degree of the error is, and T is a transposition. Converting the model into dual space according to equation (1) to solve, the following Lagrange function is obtained: l (ω, b, e, α) ═ L
In the formula of alphaiE.g. R is Lagrange multiplier, for omega, e respectivelyi,b,αiDerivation of the deviationAnd let the partial derivative be 0:
elimination of omega, eiAnd (4) finishing to obtain a linear equation system:
wherein y is [ y ]1…yN]，α＝[α1…αN]，ET＝[1…1]G is a kernel function satisfying Mercer's theorem:
solving the formula (4) to obtain α, b, and obtaining the nonlinear regression function as:
Determining a classification result according to the signs of g (x), wherein: optimal lagrange multiplier alphai(ii) a b is a bias term.
The LSSVM is a learning method based on kernels, and kernel function selection has important influence on the LSSVM performance. To this end, the present invention builds classifiers using the 3 kernel functions shown in table one, respectively.
Table one used 3 kinds of kernel functions
| Name of Kernel function | Expression of kernel functions |
| Polynomial kernel function | GP(xi,x＝(xTxi+1)u(u∈N) |
| RBF kernel function | GR(xi,x＝exp(-||x-xi||2σ2(σ≠0∈R) |
| Sigmoid kernel function | GS(xi,x＝tanha(xTxi+c)(a,c∈R) |
In table I, GP，GR，GSRespectively representing a polynomial kernel, an RBF kernel and a Sigmoid kernel. U in the polynomial kernel function is used for setting the highest term degree of the polynomial kernel function; sigma in the RBF kernel function is a width parameter of the function, and the radial action range of the function is controlled; and a and c in the Sigmoid kernel function are used for setting parameters in the kernel function.
Determining a classification result according to the signs of g (x), wherein: optimal lagrange multiplier alphai(ii) a b is a bias term. G (x)iAnd x) is a kernel function, and commonly used local kernel functions comprise a radial basis function, a global kernel function polynomial kernel function, a sig function and the like. The invention provides a brushless direct current motor position sensing method based on LSSVM decision classification, wherein A, B phase voltage u is used as a phase voltagea(k),ub(k) Current ia(k),ib(k),ia(k-1),ib(k-1) as input to LSSVM, S (K) as rotor position signal, using it as output of LSSVM, each SVM output corresponding to a region sample (not), making continuous decision until each category is separated. G j(xiX) is a kernel function, j is 1,2, …, and L is LEach classifier, i, is 1,2, …, m. m is the number of the support vectors and is automatically generated by an LSSVM algorithm, K is a time sequence, and K is the position of the rotor.
The invention provides a position detection algorithm based on LSSVM decision classification, which mainly comprises an LSSVM classification modeling and model operation part.
The LSSVM classification modeling part mainly comprises the following steps:
step 1: inputting and outputting detection signals to the same type brushless direct current motor acquisition system with the position sensor: A. phase B voltage ua(k),ub(k) Current ia(k),ib(k),ia(k-1),ib(k-1) as the input of the LSSVM, and S (K) as the rotor position signal, which is taken as the output of the LSSVM, the rotor of the DC motor rotates 0-360 degrees in electrical angle into 6 regions of 60 degrees, and the rotor position is represented by the sequence number 1-6 of the region.
Step 2: setting 15 LSSVM2 classifiers, wherein the first layer 1 classifier designates the samples with class non-1 (2-6 classes) as positive samples, the samples with class non-6 (1-5 classes) as negative samples, the second layer 2 classifier designates the samples with non-2 classes along the non-1 region (2-6) classes as positive samples, designates the samples with non-6 classes as negative samples, and obtains the second classifier, and in so on, we can obtain 15 such two classes of classifiers as shown in the second table.
Table two-layer decision classifier
The training data and the testing data are measured by a brushless direct current motor with a position sensor, and 5000 groups of measured training data and 2500 groups of measured testing data are subjected to normalization processing;
determining a kernel function GjThe kernel function selects one of a polynomial, radial basis function, and sig function.
Step 3: and respectively selecting training samples of corresponding classes according to the output of the classifier, wherein the positive sample is +1, the negative sample is-1, training the 15 LSSVM by adopting a Vapnik algorithm, and obtaining an optimal model parameter penalty coefficient C, a relaxation variable xi and a kernel function parameter by adopting a grid optimization method cross validation test to obtain 15 optimal LSSVM two classifiers.
The LSSVM decision classification operation part mainly comprises the following steps:
step 1: collecting related voltage and current input signals in real time and normalizing;
step 2: inputting related voltage and current input signals into a well established LSSVM classifier to obtain a classification result of a region Ki where a rotor is located;
when the classification belongs to the classes 1, 2, … and 6, two or more results are obtained, the result of the last classification is kept unchanged; when the classification result belongs to the classes 1, 2, … and 6, the last classification result is kept unchanged.
And Step3, performing real-time control according to the classified rotor position. And determining the on-off of each corresponding switch tube in each area by calculating a logic commutation signal through the position of the rotor.
The conversion relation between the rotor position information and the phase conversion logic signal is shown in table three, wherein 1 represents on, and 0 represents off.
Conversion relation table for three-rotor position information and phase conversion logic signal
Has the beneficial effects that: the position sensing method has the advantages of good dynamic performance, high robustness and the like. The operation speed of the algorithm is high, and the reaction speed of the controller is improved.
Drawings
Fig. 1 is a block diagram of a brushless dc motor position sensing based on LSSVM decision classification.
The specific implementation mode is as follows:
the invention provides a brushless direct current motor position sensing method based on LSSVM one-to-many classification, which combines a system structure diagram and has the following specific implementation scheme:
step 1: inputting and outputting detection signals to the same type brushless direct current motor acquisition system with the position sensor: A. b phase voltageua(k),ub(k) Current ia(k),ib(k),ia(k-1),ib(k-1) as the input of the LSSVM, and S (K) as the rotor position signal, which is taken as the output of the LSSVM, the rotor of the DC motor rotates 0-360 degrees in electrical angle into 6 regions of 60 degrees, and the rotor position is represented by the sequence number 1-6 of the region.
Step 2: setting 15 LSSVM two classifiers, wherein the first layer 1 classifier determines the samples with the category not being 1(2-6 categories) as positive samples, the samples with the category not being 6(1-5 categories) as negative samples, the second layer 2 classifier determines the samples with the category not being 2 categories as positive samples along the region not being 1(2-6 categories), and determines the samples with the category not being 6 as negative samples to obtain a second classifier, and in so on, 15 classifiers can be obtained as shown in the fourth table.
Tabular four-layer decision classifier
The training data and the testing data are measured by a brushless direct current motor with a position sensor, and 5000 groups of measured training data and 2500 groups of measured testing data are subjected to normalization processing; determining the excitation function as G being a radial basis function
Where x is the input data, xiIs the radial base center, σiThe radial base radius is 1,2, … m, where m is the number of central vectors. j is the corresponding classifier rank, j is 1,2, …, 15.
Step 3: respectively selecting training samples of corresponding classes according to the output of the classifier, wherein the positive sample is +1, the negative sample is-1, training 15 LSSVM is carried out by adopting a Vapnik algorithm, and an optimal model parameter penalty coefficient C, a relaxation variable xi and a radial base radius sigma are obtained by adopting a grid optimization method and reserving a cross validation testiAnd obtaining 15 optimal LSSVM two classifiers. Parameter grid by indexThe changes were chosen as follows:
penalty coefficient C: 106-10-1(ii) a Relaxation variable ξ: 10-5-10-1Radial base radius σi：10-5-10-1
A second part: the LSSVM classification model operation part mainly comprises the following steps:
step 1: collecting related voltage and current input signals in real time and normalizing;
step 2: inputting related voltage and current input signals into the established LSSVM classifier to obtain the region K where the rotor is located iClassifying results;
when the classification belongs to the classes 1, 2, … and 6, two or more results are obtained, the last classification result is kept unchanged; when the classification result belongs to the classes 1, 2, … and 6, the last classification result is kept unchanged.
Step3, judging the position S (K) of the rotor through an LSSVM classification network, namely the area K of the rotoriAnd real-time control is carried out according to the position of the rotor. And determining the on-off of the corresponding switch tube in each area by calculating a logic commutation signal through the position of the rotor.
While the foregoing is directed to the preferred embodiment of the present invention, other and further embodiments of the invention may be devised without departing from the basic scope thereof, and the scope thereof is determined by the claims that follow.
Claims (1)
1. A brushless direct current motor position sensing method based on Least Square Support Vector Machine (LSSVM) decision classification is characterized in that a LSSVM decision classifier is used for obtaining a position signal of a rotor of a brushless direct current motor, so that errors of the motor caused by the position sensor are reduced, and the size of the motor is reduced; the method adopts an LSSVM two-classification structure, determines final classification output through hierarchical decision, and mainly comprises an LSSVM classification modeling and model operation part:
1) The LSSVM classification modeling part mainly comprises the following steps:
step 1: inputting and outputting detection signals to a same-model brushless direct current motor acquisition system with a position sensor, and inputting A, B phase voltage ua(k)、ub(k) And current ia(k)、ib(k)、ia(k-1)、ib(K-1) as input to the LSSVM, taking the rotor position signal s (K) as output of the LSSVM, each LSSVM output corresponding to a region sample, wherein K is a time sequence and K is the rotor position, and making a decision continuously until each category is classified; dividing the rotor of the DC motor into 6 areas with an electric angle of 0-360 degrees of rotation, wherein each area is 60 degrees, the area where the rotor is located is represented by a serial number of 1-6, and performing normalization processing on the measured training data and the measured test data;
step 2: 15 LSSVM two classifiers are arranged in total, the first layer is a two-classifier, and the two-classifier comprises six classes from class 1 to class 6, wherein samples in the class other than 1 are determined as positive samples, and samples in the class other than 6 are determined as negative samples; a first second classifier of the second layer determines a sample which is not in a class 2 in the positive samples obtained by the first layer as a positive sample, and determines a sample which is not in a class 6 in the positive samples obtained by the first layer as a negative sample; applying positive sample data obtained by the first second classifier at the second layer to a first second classifier at the third layer, applying negative sample data obtained by the first second classifier at the second layer and non-1-class sample data in the second classifier at the second layer to a second classifier at the third layer, applying non-5-class sample data in the second classifier at the second layer to a third second classifier at the third layer, and so on to obtain 15 classifiers; determining a kernel function G jSelecting one of a polynomial, a radial basis and a sig function by the kernel function;
step 3: respectively selecting training samples of corresponding classes according to the output of the classifiers, wherein the positive sample is +1, the negative sample is-1, training 15 LSSVM two classifiers by adopting a Vapnik algorithm, and obtaining an optimal model parameter punishment coefficient C and a kernel function parameter by adopting a grid optimization method cross validation test to obtain 15 optimal LSSVM two classifiers;
2) the LSSVM decision classification model operation part mainly comprises the following steps:
step 1: collecting related voltage and current input signals in real time and normalizing;
step 2: inputting related voltage and current input signals into the established LSSVM two classifiers to obtain a classification result of the region where the rotor is located;
when the classification belongs to the classes 1, 2, … and 6, two or more results are obtained, the result of the last classification is kept unchanged; when the classification result belongs to the classes 1, 2, … and 6, the classification result of the last time is kept unchanged;
step 3: performing real-time control according to the classified rotor position; and calculating a logic commutation signal through the position of the rotor to determine the on-off of the switching tube corresponding to each area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811180628.9A CN109150054B (en) | 2018-10-09 | 2018-10-09 | LSSVM decision classification-based brushless direct current motor position sensing method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811180628.9A CN109150054B (en) | 2018-10-09 | 2018-10-09 | LSSVM decision classification-based brushless direct current motor position sensing method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109150054A CN109150054A (en) | 2019-01-04 |
| CN109150054B true CN109150054B (en) | 2022-07-15 |
Family
ID=64811507
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811180628.9A Active CN109150054B (en) | 2018-10-09 | 2018-10-09 | LSSVM decision classification-based brushless direct current motor position sensing method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109150054B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102510253A (en) * | 2011-11-07 | 2012-06-20 | 江苏大学 | Sensorless controller of bearingless synchronous reluctance motor and method for controlling sensorless controller |
| CN103501148A (en) * | 2013-09-24 | 2014-01-08 | 江苏大学 | Method for controlling operation of non-radial displacement sensor of bearingless permanent magnetic synchronous motor |
-
2018
- 2018-10-09 CN CN201811180628.9A patent/CN109150054B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102510253A (en) * | 2011-11-07 | 2012-06-20 | 江苏大学 | Sensorless controller of bearingless synchronous reluctance motor and method for controlling sensorless controller |
| CN103501148A (en) * | 2013-09-24 | 2014-01-08 | 江苏大学 | Method for controlling operation of non-radial displacement sensor of bearingless permanent magnetic synchronous motor |
Non-Patent Citations (3)
| Title |
|---|
| Evaluation of SVM Speed and Position Observers for Sensorless PMSM in Start-up Region;Xiaoquan Lu等;《7th IET International Conference on Power Electronics, Machines and Drives (PEMD 2014)》;20140619;第1-6页 * |
| 动车组制动控制系统故障诊断方法研究;丁国君;《中国博士学位论文全文数据库工程科技II辑》;20131015;第35-38页 * |
| 永磁无刷直流电机驱动的电动压缩机控制研究;马跃;《中国优秀硕士学位论文全文数据库工程科技II辑》;20180415;第25-36页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109150054A (en) | 2019-01-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Abid et al. | Robust interpretable deep learning for intelligent fault diagnosis of induction motors | |
| Maraaba et al. | Convolutional neural network-based inter-turn fault diagnosis in LSPMSMs | |
| Soualhi et al. | Detection and diagnosis of faults in induction motor using an improved artificial ant clustering technique | |
| Husari et al. | Incipient interturn fault detection and severity evaluation in electric drive system using hybrid HCNN-SVM based model | |
| Ondel et al. | A method to detect broken bars in induction machine using pattern recognition techniques | |
| CN112036435A (en) | Brushless direct current motor sensor fault detection method based on convolutional neural network | |
| CN111538960B (en) | Alternating current asynchronous motor fault diagnosis method based on improved fuzzy C-means clustering | |
| Peng et al. | Progressively balanced supervised contrastive representation learning for long-tailed fault diagnosis | |
| Jun et al. | Blind source separation method for bearing vibration signals | |
| CN103633903A (en) | Self-detection method for positions of rotors of switched reluctance motors | |
| CN108183639A (en) | A kind of brshless DC motor least squared classified speed regulating method | |
| Bandyopadhyay et al. | A combined image processing and Nearest Neighbor Algorithm tool for classification of incipient faults in induction motor drives | |
| CN109150054B (en) | LSSVM decision classification-based brushless direct current motor position sensing method | |
| CN111769776A (en) | Current reconstruction system and method for hub motor controller | |
| Verma et al. | Neural speed–torque estimator for induction motors in the presence of measurement noise | |
| CN109120191B (en) | Brushless direct current motor position sensing method based on LSSVM hierarchical classification | |
| CN107947650B (en) | Brushless direct current motor position sensorless control method based on extreme learning machine classification | |
| Hemalatha et al. | Sensorless speed and position control of permanent magnet BLDC motor using particle swarm optimization and ANFIS | |
| Rai et al. | A comparative performance analysis for loss minimization of induction motor drive based on soft computing techniques | |
| CN108712116B (en) | Brushless direct current motor position sensorless control method based on extreme learning machine | |
| Skylvik et al. | Data-driven fault diagnosis of induction motors using a stacked autoencoder network | |
| Ali et al. | Experimental investigation of machine learning based fault diagnosis for induction motors | |
| CN109120192A (en) | Based on a pair of polytypic Brushless DC Motor Position method for sensing of LSSVM | |
| CN114474053B (en) | Robot terrain recognition and speed control method and system | |
| Banerjee et al. | Hybrid intelligent predictive control system for high speed BLDC motor in aerospace application |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |