CN116455278A - Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control - Google Patents
Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control Download PDFInfo
- Publication number
- CN116455278A CN116455278A CN202310176216.2A CN202310176216A CN116455278A CN 116455278 A CN116455278 A CN 116455278A CN 202310176216 A CN202310176216 A CN 202310176216A CN 116455278 A CN116455278 A CN 116455278A
- Authority
- CN
- China
- Prior art keywords
- output
- neural network
- layer
- rotating speed
- motor
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 76
- 230000001360 synchronised effect Effects 0.000 title claims abstract description 23
- 238000000034 method Methods 0.000 title claims abstract description 22
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 36
- 230000006870 function Effects 0.000 claims description 33
- 230000004913 activation Effects 0.000 claims description 21
- 210000002569 neuron Anatomy 0.000 claims description 9
- 230000003044 adaptive effect Effects 0.000 claims description 5
- 230000009466 transformation Effects 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 description 4
- 230000001105 regulatory effect Effects 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 238000013461 design Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 238000004088 simulation Methods 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013016 damping Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- H02P21/0003—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control
- H02P21/0014—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control using neural networks
-
- 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
- H02P21/0003—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control
- H02P21/0017—Model reference adaptation, e.g. MRAS or MRAC, useful for control or parameter estimation
-
- 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
- H02P21/22—Current control, e.g. using a current control loop
-
- 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
- H02P25/00—Arrangements or methods for the control of AC motors characterised by the kind of AC motor or by structural details
- H02P25/02—Arrangements or methods for the control of AC motors characterised by the kind of AC motor or by structural details characterised by the kind of motor
- H02P25/022—Synchronous motors
-
- 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
- H02P27/00—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage
- H02P27/04—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage
- H02P27/06—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters
- H02P27/08—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters with pulse width modulation
- H02P27/085—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters with pulse width modulation wherein the PWM mode is adapted on the running conditions of the motor, e.g. the switching frequency
-
- 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
- H02P27/00—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage
- H02P27/04—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage
- H02P27/06—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters
- H02P27/08—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters with pulse width modulation
- H02P27/12—Arrangements or methods for the control of AC motors characterised by the kind of supply voltage using variable-frequency supply voltage, e.g. inverter or converter supply voltage using dc to ac converters or inverters with pulse width modulation pulsing by guiding the flux vector, current vector or voltage vector on a circle or a closed curve, e.g. for direct torque control
-
- 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
- H02P2207/00—Indexing scheme relating to controlling arrangements characterised by the type of motor
- H02P2207/05—Synchronous machines, e.g. with permanent magnets or DC excitation
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Control Of Electric Motors In General (AREA)
Abstract
The invention provides a method for self-setting pid parameters of a permanent magnet synchronous motor control system based on neural network self-adaptive control, which comprises the following steps: step 1: designing a neural network structure of a self-adaptive controller of the rotating speed ring; step 2: utilizing a gradient descent algorithm to adjust the neural network weight of the neural network self-adaptive controller on line; step 3: and (5) feedback control of the motor rotation speed. According to the invention, the self-setting of the pid parameters of the permanent magnet synchronous motor rotating speed ring control system can be realized, and the dynamic performance of motor operation is improved; the self-adaptive capacity of a motor control algorithm can be improved, and dependence on system model parameters is eliminated.
Description
Technical Field
The invention belongs to the field of permanent magnet synchronous motor control, and particularly relates to a method for self-setting of pid parameters of a permanent magnet synchronous motor control system based on neural network self-adaptive control.
Background
Along with the development of power electronics technology, computer technology and automatic control technology, the speed regulation technology of the alternating current motor has also been greatly developed, and the performance of the alternating current motor shows great superiority. In the AC speed regulating system, compared with the asynchronous motor speed regulating, the synchronous motor speed regulating system has the advantages of high power factor, small capacity of the frequency converter and small moment of inertia, and in the high-power AC transmission system, the advantages of the synchronous motor speed regulating system are obvious. With the development and application of permanent magnet materials, the permanent magnet synchronous motor (Permanent Magnet Synchronous Motor, PMSM) is receiving more and more attention due to the characteristics of high power density, large torque inertia ratio and high dynamic response speed.
The controllers of the rotating speed ring and the current ring of the traditional permanent magnet synchronous motor are PI controllers, but the parameter setting of the PI controllers can only be applied to a certain specific working range, when the working state of the motor changes, the control effect of the PI controllers is poor, better performance cannot be provided, and the neural network can be adjusted through learning rules, so that the system characteristics are optimized.
Disclosure of Invention
In order to solve the problem that the PI controller of the traditional permanent magnet synchronous motor can only be applied to a certain specific working range, and the control effect of the PI controller is poor and better performance cannot be provided when the working state of the motor changes, the invention provides a method for self-setting the pid parameters of a permanent magnet synchronous motor control system based on neural network self-adaptive control.
The technical scheme of the invention is as follows:
the invention firstly provides a method for self-setting pid parameters of a permanent magnet synchronous motor control system based on neural network self-adaptive control, which comprises the following steps:
1) Designing a neural network structure of a self-adaptive controller of a rotating speed ring, determining the output of a neural network input layer according to the structure of a motor control system, determining the output of a neural network hidden layer according to the output of the input layer and a hidden layer activation function, and determining the output of a neural network output layer according to the output of the neural network hidden layer and the activation function;
2) Acquiring motor rotating speed and rotor position information from a motor sensor, taking the motor reference rotating speed, the motor actual rotating speed and the rotating speed difference of the motor reference rotating speed and the motor actual rotating speed as the input of a rotating speed loop neural network self-adaptive controller, and utilizing a gradient descent algorithm to adjust the neural network weight of the neural network self-adaptive controller on line; the output value of the neural network output layer is the gain link parameter k of the rotating speed ring pid p Integral link parameter k i Differential link parameter k d And then obtaining the output of the neural network self-adaptive controller by using an incremental pid algorithm:
u(k)＝u(k-1)+k p (e(k)-e(k-1))+k i e(k)+k d (e(k)-2e(k-1)+e(k-2))
where k represents the discrete amount of time, the controller output u (k) is the q-axis reference current i q * ，
3) Inputting the q-axis reference current and the actual q-axis current value into a current loop feedforward pi controller, thereby obtaining q-axis reference voltages and d-axis reference voltages; and according to the q-axis reference voltage, the d-axis reference voltage and the rotor position information, the motor rotating speed is controlled by adjusting the switch of the inverter bridge of the permanent magnet synchronous motor control system in a svpwm modulation mode through coordinate transformation, and the feedback control of the motor rotating speed is realized.
As a preferred solution of the present invention, the neural network structure of the adaptive controller for designing the rotation speed ring in the step 1) specifically includes:
selecting three layers of neural networks, wherein each neural network comprises an input layer, a hidden layer and an output layer, and the input nodes correspond to the operation of a motor control systemThe reference rotation speed of the motor, the actual rotation speed of the motor and the rotation speed difference; the output nodes correspond to three parameters kp, ki, kd of the rotating speed ring pid; the number j of the input layer nodes is 3, the number i of the output layer nodes is 3, and the number of the hidden layer nodes is 3
As a preferred scheme of the present invention, in the step 1), the input layer output is determined according to the structure of the motor control system, specifically:
for a motor control system, the output of the input layer of the selective neural network is:
wherein x (j) represents the input quantity of the neural network, namely the reference rotating speed of the motor, the actual rotating speed of the motor and the rotating speed difference, and j represents a neuron node; the corner mark (1) represents the first layer of the neural network.
As a preferred solution of the present invention, in the step 1), the hidden layer output is determined according to the input layer output and the hidden layer activation function, specifically:
the hidden layer input obtained by the input layer output is:
wherein omega ij Representing connection weights between corresponding neurons; o (O) j Output for the input layer; the corner mark (2) represents a hidden layer;
the output of the hidden layer is:
wherein g (x) represents the activation function of the hidden layer,
as a preferred embodiment of the present invention, the determining the output layer output according to the hidden layer output and the activation function in the step 1) specifically includes:
the input of the output layer is as follows:
the output of the output layer is:
wherein omega li Representing connection weights between corresponding neurons of the hidden layer and the output layer; o (O) i Is the output of the hidden layer; the corner mark (3) represents an output layer; f (x) represents the activation function of the hidden layer, since the output layer output corresponds to the pid parameter k p 、k i 、k d Thus, f (x) is a non-negative activation function, f (x) =relu (x) =max (0, x);
as a preferred scheme of the present invention, the on-line adjustment of the neural network weight of the neural network adaptive controller in the step 2) specifically includes:
the performance index function is taken as follows:
wherein r (k) represents a reference rotational speed, and y (k) represents an actual rotational speed;
correcting the connection weight function of the neural network by using a gradient descent algorithm, and adding an inertia term to enable the search to be quickly converged to the global minimum, so that the method is obtained:
wherein eta is learning rate, alpha is inertia coefficient,solving a first term of an expression for the variable quantity of the connection weight between the hidden layer and the output layer:
the actual motor control pid algorithm employs a discrete incremental algorithm that, for the incremental algorithm,
then, it is possible to obtain:
the input-output functions of the previous layers can then be obtained:
order theThe above can be written as:
similarly, the hidden layer weight calculation formula is as follows:
wherein the method comprises the steps of
The weights may then be updated based on the collected data.
As a preferred embodiment of the present invention, in the step 2), the learning rate η is updated specifically:
learning rate η (k) =σ×η (k-1)
Wherein the change factor σ=2 λ ，λ＝sign(g(k)*g(k-1))；
Wherein λ reflects the gradient direction; therefore, the learning rate gradient directions for the output layer and the hidden layer are respectively:
and updating the learning rate according to the above.
Compared with the prior art, the invention has the advantages that: the invention replaces the traditional pid controller with the neural network self-adaptive controller, reduces the frequency and improves the control precision of the system; meanwhile, the gradient descent algorithm is utilized to adjust the weight of the neural network on line to realize the self-adaption function, and the self-adaption capability of the control system is improved.
Drawings
FIG. 1 is a flow chart of the system of the present invention;
FIG. 2 is a schematic diagram of the system of the present invention;
FIG. 3 is a schematic illustration of a neural network according to the present invention;
FIG. 4 is a main body loop process of the neural network algorithm of the present invention;
FIG. 5 is a graph of the speed of the present invention at idle versus a conventional pid algorithm;
FIG. 6 is a graph of the speed of the present invention versus a conventional pid algorithm for direct loading;
FIG. 7 is a graph of the speed of the present invention versus a conventional pid algorithm at variable load.
Detailed Description
The invention is further illustrated and described below in connection with specific embodiments. The described embodiments are merely exemplary of the present disclosure and do not limit the scope. The technical features of the embodiments of the invention can be combined correspondingly on the premise of no mutual conflict.
Referring to fig. 2, the overall flow of permanent magnet synchronous motor control mainly includes the following:
(1) the operation state of the motor is acquired to include the rotor rotation angular velocity omega r Information such as a position thetar, and the like, the position information is taken as a basis of coordinate transformation, and the rotational speed information is used for determining the angular velocity omega r Converting the rotating speed N per minute and sending the converted rotating speed N per minute into a rotating speed ring;
(2) based on the current rotation speed information N and the expected reference rotation speed N ref The difference is sent to the PI controller to enable the actual rotation speed to track the reference rotation speed, namely a rotation speed ring;
(3) the output quantity of the rotating speed ring PI regulator can be used as the reference value i of q-axis current in the next link qref And then with the actual q-axis current value i q Comparing and sending the voltage to a PI regulator to track, and then combining a d axis to carry out PI regulation so as to obtain q axis and d axis reference voltage v q * And v d * This is the current loop;
(4) from the q-axis and d-axis reference voltages and rotor position, v can be obtained by coordinate transformation α * And v β * Then, the switch of the inverter bridge is adjusted in a svpwm modulation mode so as to control the rotating speed of the motor; thereby realizing feedback control of the motor rotation speed.
The invention mainly aims at the problems that the PI controller of the traditional permanent magnet synchronous motor can only be applied to a certain specific working range and the control effect of the PI controller is poor when the working state of the motor changes. The flow chart of the invention is shown in fig. 1, and comprises the following steps:
1) Designing a neural network structure of a self-adaptive controller of a rotating speed ring, determining the output of a neural network input layer according to the structure of a motor control system, determining the output of a neural network hidden layer according to the output of the input layer and a hidden layer activation function, and determining the output of a neural network output layer according to the output of the neural network hidden layer and the activation function;
2) Acquiring motor rotating speed and rotor position information from a motor sensor, taking the motor reference rotating speed, the motor actual rotating speed and the rotating speed difference of the motor reference rotating speed and the motor actual rotating speed as the input of a rotating speed loop neural network self-adaptive controller, and utilizing a gradient descent algorithm to adjust the neural network weight of the neural network self-adaptive controller on line; the output value of the neural network output layer is the gain link parameter k of the rotating speed ring pid p Integral link parameter k i Differential link parameter k d And then obtaining the output of the neural network self-adaptive controller by using an incremental pid algorithm:
u(k)＝u(k-1)+k p (e(k)-e(k-1))+k i e(k)+k d (e(k)-2e(k-1)+e(k-2))
where k represents the discrete amount of time, the controller output u (k) is the q-axis reference current i qref ，
3) Inputting the q-axis reference current and the actual q-axis current value into a current loop feedforward pi controller, thereby obtaining q-axis reference voltages and d-axis reference voltages; and according to the q-axis reference voltage, the d-axis reference voltage and the rotor position information, the motor rotating speed is controlled by adjusting the switch of the inverter bridge of the permanent magnet synchronous motor control system in a svpwm modulation mode through coordinate transformation, and the feedback control of the motor rotating speed is realized.
The invention establishes a mathematical model of the permanent magnet synchronous motor to determine the input and output physical quantity of the neural network self-adaptive controller. Firstly, writing a motion equation of a three-phase permanent magnet synchronous motor:
wherein J is moment of inertia, B is damping coefficient, T e Is electromagnetic torque, T L For load torque, ω m Is the mechanical angular velocity of the motor; p is p n I is the pole pair number of the motor d 、i q Respectively representd. Current component of q-axis, L d 、L q Inductances denoted d, q axes respectively, ψ f Is a permanent magnet flux linkage;
the current equation is written in the dq coordinate system of the rotor magnetic field orientation synchronous speed rotation as follows:
wherein u is d 、u q Voltage components of d and q axes are respectively represented; omega e Is the electrical angular velocity of the motor;
as can be seen from the above, the stator current i d 、i q Generating cross-coupled electromotive forces in the d-axis and q-axis directions, respectively, if the stator current i d 、i q Completely decoupled, then the above equation becomes:
u d * 、u q * respectively representing voltage components of d and q axes after current decoupling; the voltages of d and q axes can be obtained by adopting a conventional pi regulator:
wherein v is d * And v q * Is the reference voltage after pi regulation. Therefore, for the current loop, the current error can be used as the input of the neural network self-adaptive controller, and the d and q voltages are used as the output of the controller; for the rotating speed loop, the rotating speed error can be used as the input of the neural network self-adaptive controller, and the q-axis current can be used as the output of the controller. The invention provides a neural network self-adaptive controller mainly aiming at a rotating speed ring. For the current loop, the neural network self-adaptive controller can be designed by referring to the method of the invention, and the traditional self-adaptive controller can also be adopted.
As shown in fig. 4, the invention mainly designs a self-adaptive controller of a rotating speed loop neural network, learns the neural network, adjusts network parameters on line, realizes the self-adaptive adjustment of pid control parameters, and specifically comprises the following steps:
step 1: a neural network structure is determined.
As shown in fig. 3, three layers of neural networks are selected, namely an input layer, a hidden layer and an output layer, wherein input nodes correspond to state quantities of operation of a motor control system, namely motor reference rotation speed, motor actual rotation speed and rotation speed difference; the output nodes correspond to the three parameters kp, ki, kd of the pid controller. Therefore, the number of input layer nodes is j=3, the number of output layer nodes is i=3, and the number of hidden layer nodes isSo take l=5.
Step 2: and determining the output of the input layer according to the structure of the motor control system.
For this control system, the input layer outputs of the select neural network are:
wherein x represents the input quantity of the neural network, namely the reference rotating speed of the motor, the actual rotating speed of the motor and the rotating speed difference, and j represents the neuron node; the corner mark (1) represents a first layer of the neural network;
step 3: and determining hidden layer output according to the input layer output and the hidden layer activation function.
The hidden layer input is obtained by the input layer output:
wherein omega ij Representing connection weights between corresponding neurons; o (O) j Output for the input layer; the corner mark (2) represents a hidden layer; the output of the output layer is:
wherein g (x) represents the activation function of the hidden layer,
step 4: and determining output layer output according to the hidden layer output and the activation function.
The output layer function is similar to the equation above:
wherein omega li Representing connection weights between corresponding neurons of the hidden layer and the output layer; o (O) i Is the output of the hidden layer; the corner mark (3) represents an output layer; f (x) represents the activation function of the hidden layer, since the output layer output corresponds to the pid parameter, f (x) is a non-negative activation function, f (x) =relu (x) =max (0, x);
step 5: next, each layer of connection weight is updated.
The performance index function is taken as follows:
the gradient descent algorithm is utilized to correct the connection weight function of the network, and an inertia term is added to enable the search to be quickly converged to the global minimum, so that the method can be obtained:
wherein eta is the learning rate, alpha is the inertia coefficient, and the first term of the expression is solved:
wherein for the followingSince the relative variation of y (k) and u (k), i.e
May also be replaced approximately by a sign function, i.eTherefore, the operation can be simplified, and on the other hand, a tiny amount can be added on the denominator in the actual operation to avoid that the denominator is 0 at the initial moment; the actual motor control pid algorithm may employ a discrete incremental algorithm for which +.>
Then, it is possible to obtain:
the input-output functions of the previous layers can then be obtained:
order theThe above can be written as:
similarly, the hidden layer weight calculation formula is as follows:
wherein the method comprises the steps of
The weights may then be updated based on the collected data.
Step 6: updating the learning rate.
The additional inertia item faces the difficulty in selecting the selection rate, thereby generating contradiction between convergence speed and convergence. Thus, introducing a learning rate adaptive design is also contemplated.
η(k)＝σ*η(k-1)
Wherein σ is a variation factor σ=2 λ ，λ＝sign(g(k)*g(k-1))；
Where lambda reflects the gradient direction. Therefore, the learning rate gradient directions for the output layer and the hidden layer are respectively:
the learning rate may be updated according to the above.
The weights may then be updated based on the collected data.The output value of the neural network output layer is k of the pid parameter p 、k i 、k d Then, the controller output is obtained by using an incremental pid algorithm:
u(k)＝u(k-1)+k p (e(k)-e(k-1))+k i e(k)+k d (e(k)-2e(k-1)+e(k-2))
the output u (k) of the controller is the q-axis reference current i q * The q-axis reference current is input into a current loop feedforward pi controller to control the motor to operate.
Simulation result analysis: in order to verify the feasibility of the proposed method, a simulation verification was performed in a Matlab/Simulink environment and compared with a PI regulator.
FIG. 5 is a graph of the speed of the present invention versus a conventional pid algorithm when the motor is idle, and it can be seen from FIG. 5 that the overshoot of the speed curve of the present invention is much less than that of the conventional pid algorithm, and the settling time to reach steady state is much less than that of the conventional pid algorithm; FIG. 6 is a graph of the speed of the present invention versus a conventional pid algorithm when the motor is directly loaded, and it can be seen from FIG. 6 that the overshoot of the speed curve of the present invention is much less than that of the conventional pid algorithm, and the settling time to reach steady state is much less than that of the conventional pid algorithm; FIG. 7 is a graph of the speed of the invention compared with the conventional pid algorithm when the load is changed, and FIG. 7 is that the motor runs idle at 0-0.2s, then suddenly loads at 0.2s, and loads at 0.2-0.4s, then the time for recovering the stable rotating speed after suddenly loading is less than that of the conventional pid algorithm; from the above figures, it can be seen that the overshoot and the adjustment time of the algorithm of the present invention are superior to the conventional pid algorithm. In summary, the algorithm provided herein can achieve the expected control accuracy, and can effectively improve the running performance of the motor.
Finally, it should be noted that: the foregoing description is only a preferred embodiment of the present invention, and the present invention is not limited thereto, but it is to be understood that modifications and equivalents of some of the technical features described in the foregoing embodiments may be made by those skilled in the art, although the present invention has been described in detail with reference to the foregoing embodiments. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (7)
1. The method for self-setting the pid parameters of the permanent magnet synchronous motor control system based on the neural network self-adaptive control is characterized by comprising the following steps:
1) Designing a neural network structure of a self-adaptive controller of a rotating speed ring, determining the output of a neural network input layer according to the structure of a motor control system, determining the output of a neural network hidden layer according to the output of the input layer and a hidden layer activation function, and determining the output of a neural network output layer according to the output of the neural network hidden layer and the activation function;
2) Acquiring motor rotating speed and rotor position information from a motor sensor, taking the motor reference rotating speed, the motor actual rotating speed and the rotating speed difference of the motor reference rotating speed and the motor actual rotating speed as the input of a rotating speed loop neural network self-adaptive controller, and utilizing a gradient descent algorithm to adjust the neural network weight of the neural network self-adaptive controller on line; the output value of the neural network output layer is the gain link parameter k of the rotating speed ring pid p Integral link parameter k i Differential link parameter k d And then obtaining the output of the neural network self-adaptive controller by using an incremental pid algorithm:
u(k)＝u(k-1)+k p (e(k)-e(k-1))+k i e(k)
+k d (e(k)-2e(k-1)+e(k-2))
where k represents the discrete amount of time, the controller output u (k) is the q-axis reference current i q * ，
3) Inputting the q-axis reference current and the actual q-axis current value into a current loop feedforward pi controller, thereby obtaining q-axis reference voltages and d-axis reference voltages; and according to the q-axis reference voltage, the d-axis reference voltage and the rotor position information, the motor rotating speed is controlled by adjusting the switch of the inverter bridge of the permanent magnet synchronous motor control system in a svpwm modulation mode through coordinate transformation, and the feedback control of the motor rotating speed is realized.
2. The method according to claim 1, wherein the neural network structure of the adaptive controller for designing the rotation speed loop in step 1) is specifically:
selecting a three-layer neural network, wherein the three-layer neural network comprises an input layer, a hidden layer and an output layer, and the input nodes correspond to the motor reference rotating speed, the actual rotating speed and the rotating speed difference of the motor which are operated by a motor control system; the output nodes correspond to three parameters kp, ki, kd of the rotating speed ring pid; the number j of the input layer nodes is 3, the number i of the output layer nodes is 3, and the number of the hidden layer nodes is 3
3. The method according to claim 1, wherein the determining the input layer output in step 1) according to the motor control system structure is specifically:
for a motor control system, the output of the input layer of the selective neural network is:
wherein x (j) represents the input quantity of the neural network, namely the reference rotating speed of the motor, the actual rotating speed of the motor and the rotating speed difference, and j represents a neuron node; the corner mark (1) represents the first layer of the neural network.
4. The method according to claim 1, wherein the determining the hidden layer output according to the input layer output and the hidden layer activation function in step 1) is specifically:
the hidden layer input obtained by the input layer output is:
wherein omega ij Representing connection weights between corresponding neurons; o (O) j Output for the input layer; the corner mark (2) represents a hidden layer;
the output of the hidden layer is:
wherein g (x) represents the activation function of the hidden layer,
5. the method according to claim 1, wherein the determining the output layer output according to the hidden layer output and the activation function in step 1) is specifically:
the input of the output layer is as follows:
the output of the output layer is:
wherein omega li Representing connection weights between corresponding neurons of the hidden layer and the output layer; o (O) i Is the output of the hidden layer; the corner mark (3) represents an output layer; f (x) represents the activation function of the hidden layer, since the output layer output corresponds to the pid parameter k p 、k i 、k d Thus, f (x) is a non-negative activation function, f (x) =relu (x) =max (0, x);
6. the method according to claim 1, wherein the on-line adjusting the neural network weight of the neural network adaptive controller in step 2) specifically comprises:
the performance index function is taken as follows:
wherein r (k) represents a reference rotational speed, and y (k) represents an actual rotational speed;
correcting the connection weight function of the neural network by using a gradient descent algorithm, and adding an inertia term to enable the search to be quickly converged to the global minimum, so that the method is obtained:
wherein eta is learning rate, alpha is inertia coefficient,solving a first term of an expression for the variable quantity of the connection weight between the hidden layer and the output layer:
the actual motor control pid algorithm employs a discrete incremental algorithm that, for the incremental algorithm,
then, it is possible to obtain:
the input-output functions of the previous layers can then be obtained:
order theThe above can be written as:
similarly, the hidden layer weight calculation formula is as follows:
wherein the method comprises the steps of
The weights may then be updated based on the collected data.
7. The method according to claim 6, wherein in the step 2), the learning rate η is updated, specifically:
learning rate η (k) =σ×η (k-1)
Wherein the change factor σ=2 λ ，λ＝sign(g(k)*g(k-1))；
Wherein λ reflects the gradient direction; therefore, the learning rate gradient directions for the output layer and the hidden layer are respectively:
and updating the learning rate according to the above.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310176216.2A CN116455278A (en) | 2023-02-28 | 2023-02-28 | Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310176216.2A CN116455278A (en) | 2023-02-28 | 2023-02-28 | Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116455278A true CN116455278A (en) | 2023-07-18 |
Family
ID=87120971
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310176216.2A Pending CN116455278A (en) | 2023-02-28 | 2023-02-28 | Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116455278A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117879408A (en) * | 2024-03-11 | 2024-04-12 | 深圳市昱森机电有限公司 | Self-adaptive intelligent control method of linear motor and related equipment |
-
2023
- 2023-02-28 CN CN202310176216.2A patent/CN116455278A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117879408A (en) * | 2024-03-11 | 2024-04-12 | 深圳市昱森机电有限公司 | Self-adaptive intelligent control method of linear motor and related equipment |
| CN117879408B (en) * | 2024-03-11 | 2024-05-31 | 深圳市昱森机电有限公司 | Self-adaptive intelligent control method of linear motor and related equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Xu et al. | Model-free cooperative adaptive sliding-mode-constrained-control for multiple linear induction traction systems | |
| Jie et al. | Speed regulation based on adaptive control and RBFNN for PMSM considering parametric uncertainty and load fluctuation | |
| CN116455278A (en) | Self-tuning method for pid parameters of permanent magnet synchronous motor control system based on neural network self-adaptive control | |
| Jamoussi et al. | Robust sliding mode control using adaptive switching gain for induction motors | |
| Fei et al. | Adaptive super-twisting sliding mode control for micro gyroscope based on double loop fuzzy neural network structure | |
| CN112859618B (en) | Self-adaptive learning sliding mode control method for multi-degree-of-freedom magnetic suspension planar motor | |
| Zhong et al. | An ADRC‐based PID tuning rule | |
| Wang et al. | Model predictive position control of permanent magnet synchronous motor servo system with sliding mode observer | |
| Yu et al. | Finite-time composite adaptive fuzzy control of permanent magnet synchronous motors | |
| Meng et al. | Adaptive EPCH strategy for nonlinear systems with parameters uncertainty and disturbances | |
| CN117691895A (en) | Double-shaft linkage system speed cooperative control method based on explicit model predictive control | |
| Lin et al. | Observer-Based Prescribed Performance Speed Control for PMSMs: A Data-Driven RBF Neural Network Approach | |
| Jin et al. | High precision tracking control for linear servo system based on intelligent second-order complementary sliding mode | |
| CN113343592B (en) | Intelligent control method for DQN (direct current) of permanent magnet synchronous motor of new energy aircraft | |
| CN113691179B (en) | Permanent magnet synchronous motor control method based on fixed time power-varying exponent approach law | |
| Dutta et al. | An adaptive feedback linearized model predictive controller design for a nonlinear multi‐input multi‐output system | |
| Zhao et al. | Design of model free adaptive controller for PMSM speed control | |
| Zhao et al. | Design of MRAC and Modified MRAC for the Turntable | |
| CN112596390A (en) | Method for determining parameters of PWPF modulator of spacecraft attitude control system | |
| Mahmoudi et al. | Neuro-Genetic sensorless sliding mode control of a permanent magnet synchronous motor using Luenberger observer | |
| Deepika et al. | Globally robust adaptive critic based neuro-integral terminal sliding mode technique with UDE for nonlinear systems | |
| Téczely et al. | LPV Control of Systems via Nonlinear Actuator Dynamics Mapping | |
| CN117879419A (en) | Improved sensorless control method for permanent magnet synchronous motor | |
| Khan et al. | Fuzzy logic based MRAS speed observer for control of induction motor drive | |
| CN113659894B (en) | Asynchronous motor random limited time fuzzy self-adaptive control method based on instruction filtering |
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 |