CN114844430B - Fuzzy neural network control method for magnetic suspension switch reluctance motor - Google Patents
Fuzzy neural network control method for magnetic suspension switch reluctance motor Download PDFInfo
- Publication number
- CN114844430B CN114844430B CN202210631171.9A CN202210631171A CN114844430B CN 114844430 B CN114844430 B CN 114844430B CN 202210631171 A CN202210631171 A CN 202210631171A CN 114844430 B CN114844430 B CN 114844430B
- Authority
- CN
- China
- Prior art keywords
- reluctance motor
- neural network
- fuzzy neural
- switch reluctance
- fuzzy
- 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
- 238000013528 artificial neural network Methods 0.000 title claims abstract description 72
- 239000000725 suspension Substances 0.000 title claims abstract description 60
- 238000000034 method Methods 0.000 title claims abstract description 41
- 238000005339 levitation Methods 0.000 claims abstract description 39
- 230000008569 process Effects 0.000 claims abstract description 10
- 238000013459 approach Methods 0.000 claims abstract description 7
- 230000006870 function Effects 0.000 claims description 32
- 239000013598 vector Substances 0.000 claims description 18
- 238000009795 derivation Methods 0.000 claims description 7
- 238000013461 design Methods 0.000 claims description 7
- 230000003044 adaptive effect Effects 0.000 claims description 6
- 238000006467 substitution reaction Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 238000012217 deletion Methods 0.000 claims description 3
- 230000037430 deletion Effects 0.000 claims description 3
- 239000011159 matrix material Substances 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000005381 potential energy Methods 0.000 claims description 3
- 230000007547 defect Effects 0.000 abstract description 4
- 238000011217 control strategy Methods 0.000 abstract description 2
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002040 relaxant effect Effects 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
- H02P23/00—Arrangements or methods for the control of AC motors characterised by a control method other than vector control
- H02P23/0004—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control
- H02P23/0018—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
- H02P23/00—Arrangements or methods for the control of AC motors characterised by a control method other than vector control
- H02P23/0004—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control
- H02P23/0013—Control strategies in general, e.g. linear type, e.g. P, PI, PID, using robust control using fuzzy 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
- H02P23/00—Arrangements or methods for the control of AC motors characterised by a control method other than vector control
- H02P23/14—Estimation or adaptation of motor parameters, e.g. rotor time constant, flux, speed, current or voltage
-
- 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
- H02P25/08—Reluctance motors
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/60—Other road transportation technologies with climate change mitigation effect
- Y02T10/72—Electric energy management in electromobility
Landscapes
- Engineering & Computer Science (AREA)
- Power Engineering (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Automation & Control Theory (AREA)
- Fuzzy Systems (AREA)
- Control Of Electric Motors In General (AREA)
- Feedback Control In General (AREA)
Abstract
The invention discloses a fuzzy neural network control method of a magnetic suspension switch reluctance motor, which comprises the steps of establishing a dynamic model of the magnetic suspension switch reluctance motor; establishing a magnetic levitation switched reluctance motor rotating speed inversion controller by using an inversion method, so that the magnetic levitation switched reluctance motor operates according to expectations; the structure and parameter full-adjustment fuzzy neural network structure is adopted, and the self-adaptive magnetic suspension switch reluctance motor fuzzy neural network controller is designed, so that the defect that inversion control strategy needs accurate system information is overcome, and the system robustness is further improved. The problem that the uncertainty is difficult to effectively process in the control of the general magnetic suspension switch reluctance motor can be effectively solved, and the limitation that an uncertainty function needs to be predicted and an upper interference bound is relaxed by adopting the fuzzy neural network to approach the inversion controller.
Description
Technical Field
The invention relates to a magnetic suspension switch reluctance motor control technology, in particular to a fuzzy neural network control method for a magnetic suspension switch reluctance motor.
Background
The magnetic suspension switch reluctance motor inherits the advantages of simple structure, flexible control, high mechanical strength and strong fault tolerance of the switch reluctance motor, and has wide application prospect due to the fact that the magnetic bearing structure is adopted, and the magnetic suspension switch reluctance motor has no mechanical friction and two-degree-of-freedom suspension.
The magnetic suspension switch reluctance motor can cause internal parameter change when the temperature rises in the operation process, so that uncertainty exists in dynamic characteristics, and a certain difficulty exists in controller design. At present, the problem that the control precision of the magnetic suspension switch reluctance motor is affected by uncertainty factors such as model uncertainty, external interference and the like is mostly solved by designing a sliding mode controller, however, the method needs known moment of inertia, which is difficult to achieve in a practical system.
Disclosure of Invention
In order to overcome the defects, the invention aims to provide the fuzzy neural network control method for the magnetic levitation switch reluctance motor, which not only can effectively solve the problem that the control of the general magnetic levitation switch reluctance motor is difficult to effectively treat uncertainty, but also adopts the full-adjustment fuzzy neural network to approach and establish the magnetic levitation switch reluctance motor rotating speed inversion controller, thereby relaxing the limit of needing to predict an uncertainty function and an interference upper bound.
In order to achieve the above purpose, the invention adopts the following technical scheme: a fuzzy neural network control method of a magnetic suspension switch reluctance motor comprises the steps of firstly, establishing a mechanical model of the magnetic suspension switch reluctance motor; step two, establishing a magnetic levitation switched reluctance motor rotating speed inversion controller, wherein the magnetic levitation switched reluctance motor rotating speed inversion controller calculates to obtain the actual rotating speed omega and the expected rotating speed omega of the magnetic levitation switched reluctance motor d Is the tracking error e 1 The method comprises the steps of carrying out a first treatment on the surface of the Step three, establishing a magnetic suspension switch reluctance motor fuzzy neural network controller, wherein the magnetic suspension switch reluctance motor fuzzy neural network controller approaches the magnetic suspension switch reluctance motor rotating speed inversion controller by adopting a fuzzy neural network, eliminates an uncertainty upper bound D in the magnetic suspension switch reluctance motor rotating speed inversion controller, and is based on the tracking error e 1 Obtaining control force tau through built-in fuzzy neural network FNN The method comprises the steps of carrying out a first treatment on the surface of the Step four, the control force tau is controlled FNN The magnetic suspension is input into a magnetic suspension switch reluctance motor to achieve the rotation speed control of the magnetic suspension switch reluctance motor.
Further, the first step specifically comprises,
the magnetic suspension switch reluctance motor is 2-phase 12/14-pole, and the kinetic equation is as follows:
wherein T is e Is electromagnetic torque, T L The load torque, ω is the angular velocity, J is the moment of inertia, and I is the coefficient of friction;
the derivation of formula (1) can be obtained:
wherein τ= -IT e For the control law to be designed,is a lumped term, whose upper bound is D.
In the second step, the magnetic levitation switched reluctance motor rotation speed inversion controller is designed by using an inversion method, and the control objective of the magnetic levitation switched reluctance motor rotation speed inversion controller is to design a proper control law τ to enable the actual rotation speed ω of the magnetic levitation switched reluctance motor to track the expected rotation speed ω d The rotating speed inversion controller of the magnetic suspension switch reluctance motor is designed as follows
Wherein c 1 、c 2 Is a non-zero positive constant, e 1 ＝ω-ω d ，sgn (e) is a sign function, where
Furthermore, the fuzzy neural network adopts a four-layer network structure, which is respectively an input layer, a fuzzification layer, a fuzzy reasoning layer and an output layer.
Further, in the third step, the tracking error e is used 1 Obtaining control force tau through built-in fuzzy neural network FNN The method comprises the following specific steps:
a first layer: input layer
The input layer comprises a plurality of nodes, each node is directly connected with the input tracking error e 1 Is connected to input tracking error e 1 Passes the components of (a) to the next layer;
a second layer: blurring layer
Using Gaussian functions as membership functions Representing tracking error e 1 Element of (a)>And->The central vector and base width of membership functions of the jth fuzzy set of the ith input variable, respectively, i.e
For ease of calculation, N is employed pi Representing individual numbers of membership functions and defining adaptive parameter vectors b and c representing sets of all basis widths and center vectors of gaussian membership functions, respectively, i.e，Wherein->Representing the total number of membership functions;
third layer: fuzzy inference layer
The fuzzy inference layer adopts a fuzzy inference mechanism, and the output of each node of the fuzzy inference layer is the product of all input signals of the node, namely
Wherein l k (k＝1,...,N y ) Representing the kth output of the fuzzy inference layer,representing the connection weight matrix between the fuzzification layer and the fuzzy inference layer, taken here as a unit vector, N y Is the total number of rules;
fourth layer: output layer
The nodes of the output layer represent output variables, each node y of the output layer o (o＝1,...,,N o ) The output of (a) is the sum of all input signals of the node, i.e
In the whole fuzzy neural network, the input-output relationship is as follows:
wherein, obtaining tau FNN 。
Furthermore, the fuzzy neural network comprises two self-regulation strategies, namely data learning and data deleting, wherein the data learning is used for online evolution and parameter updating of rules of current input data in the fuzzy neural network, and the data deleting is that network parameters do not need to be updated when the current tracking error is close to the error of the previous fuzzy neural network iterative calculation process.
Further, the data learning is embodied as
Determining fuzzy rules, i.e.And psi is<E s Where ψ is the spherical potential energy, representing the novelty of the input data,
wherein E is s And E is a Is novel and the addition of a threshold value,E a can be adjusted in a self-adjusting manner,
wherein the method comprises the steps ofRepresents the tracking error, delta represents the slope factor,
when a new fuzzy rule needs to be added, the parameters of the fuzzy rule are initialized,
where κ is the overlap parameter of the predefined fuzzy rule,
when (when)When the rule parameters and the threshold E are adjusted l Also based on tracking error e 1 To perform self-operationRegulated by
Data deletion is specifically
In the fuzzy neural network learning process, the contribution degree of the q-th rule
Wherein,n represents the dimension of the input and,
if the degree of contribution of the q-th rule to the input is below the threshold E p The rule is deleted.
In the third step, the fuzzy neural network controller of the magnetic levitation switch reluctance motor approximates the rotational speed inversion controller of the magnetic levitation switch reluctance motor by adopting a fuzzy neural network, the fuzzy neural network approximates the rotational speed inversion controller of the magnetic levitation switch reluctance motor by a parameter self-adaptive law, and whether the designed control method of the magnetic levitation switch reluctance motor is stable or not is verified.
The fuzzy neural network has optimal control force according to the universal approximation theoryOptimal control forceSatisfy the following requirements
Wherein epsilon is the minimum reconstruction error vector, W * 、b * 、c * And l * W, b, c and l, respectivelyAn optimal value;
assume the output control force of a fuzzy neural networkThe following form:
wherein,and->Respectively W * 、b * 、c * And l * Is a function of the estimated value of (2);
defining an approximation error:
by Taylor series expansion, can be obtained
Wherein,b * and c * Is the optimal value of b and c, +.>And->Is b * And c * Estimated value of (2), O nv Is a higher term,/->
Then (23) substitution (22) is available
Wherein,
the designed adaptive law of the weight, base width and center vector of the fuzzy neural network can be designed as follows:
wherein,e is 2 Element sigma of (a) ω ,σ b ,σ c Is of normal number>Is->Estimated value of ∈10->Is omega i Is the optimal value of r 1 ，r 2 ，r 3 Is a normal number of times, and the number of times is equal to the normal number,
defining a lyapunov function as
Derivation of formula (28) and substitution of (25-27) into
If the D is more than or equal to D plus E, the designed control method of the magnetic suspension switch reluctance motor fuzzy neural network is stable.
The invention has the beneficial effects that: the structure and parameter full-adjustment fuzzy neural network structure is adopted, and the self-adaptive magnetic suspension switch reluctance motor fuzzy neural network controller is designed, so that the defect that inversion control strategy needs accurate system information is overcome, and the system robustness is further improved. The problem that the uncertainty is difficult to effectively process in the control of the general magnetic suspension switch reluctance motor can be effectively solved, and the limitation that an uncertainty function needs to be predicted and an upper interference bound is relaxed by adopting the fuzzy neural network to approach the inversion controller.
Drawings
Fig. 1 is a block diagram of an embodiment of the present invention.
Detailed Description
The preferred embodiments of the present invention will be described in detail below with reference to the accompanying drawings so that the advantages and features of the present invention can be more easily understood by those skilled in the art, thereby making clear and defining the scope of the present invention.
Referring to fig. 1, the invention discloses a fuzzy neural network control method for a magnetic suspension switch reluctance motor, which comprises the steps of firstly establishing a mechanical model of the magnetic suspension switch reluctance motor. Then, a magnetic levitation switch reluctance motor rotating speed inversion controller is established, and the magnetic levitation switch reluctance motor rotating speed inversion controller is designed by using an inversion method, so that the magnetic levitation switch reluctance motor rotates according to the expected rotating speed omega d And (3) running. Wherein the magnet isThe rotation speed inversion controller of the suspension switch reluctance motor calculates to obtain the actual rotation speed omega and the expected rotation speed omega of the suspension switch reluctance motor d Is the tracking error e 1 . Then establishing a fuzzy neural network controller of the magnetic suspension switch reluctance motor, wherein the fuzzy neural network controller of the magnetic suspension switch reluctance motor approaches a rotational speed inversion controller of the magnetic suspension switch reluctance motor by adopting the fuzzy neural network, and an uncertainty upper bound D in the rotational speed inversion controller of the magnetic suspension switch reluctance motor is eliminated to obtain a control force tau FNN . Receiving tracking error e by magnetic suspension switch reluctance motor fuzzy neural network controller 1 And according to tracking error e 1 Obtaining control force tau through built-in fuzzy neural network FNN . Finally, the force tau is controlled FNN The magnetic suspension is input into a magnetic suspension switch reluctance motor to achieve the rotation speed control of the magnetic suspension switch reluctance motor.
The method comprises the following specific steps:
step 1, establishing a mechanical model of a magnetic suspension switch reluctance motor
Consider a 2-phase 12/14 pole magnetic levitation switched reluctance motor whose kinetic equation is:
wherein T is e Is electromagnetic torque, T L For load torque, ω is angular velocity, J is moment of inertia, and I is coefficient of friction.
The derivation of formula (1) can be obtained:
wherein τ= -IT e For the control law to be designed,is a lumped term, whose upper bound is D.
Step 2, establishing a magnetic suspension switch reluctance motor rotating speed inversion controller
The inversion method is utilized to design a magnetic levitation switch reluctance motor rotating speed inversion controller, and the control aim of the magnetic levitation switch reluctance motor rotating speed inversion controller is to design a proper control law tau so as to enable the actual rotating speed omega of the magnetic levitation switch reluctance motor to track the expected rotating speed omega on the magnetic levitation switch reluctance motor d . The specific establishment steps are as follows:
step 21:
definition of tracking error e 1 Is that
e 1 ＝ω-ω d (3)
Deriving (3)
Selecting virtual control quantity
Wherein c 1 Is a non-zero positive constant.
Taking Lyapunov function (Lyapunov function) as
Defining the amount of deviationDeriving (6)
If e 2 =0, thenThe stability requirement is met at this point. Therefore, it is necessary to continue the design to let the deviation e 2 ＝0。
Step 22: for deviation e 2 Conduct derivation, e 2 The derivative of (2) can be expressed as
Wherein c 2 Is a non-zero positive constant.
The rotating speed inversion controller of the magnetic suspension switch reluctance motor is designed as follows
Definition of Lyapunov function (Lyapunov function)
Deriving (10) and bringing (9) into availability
From the above formula (11), the designed magnetic levitation switched reluctance motor inversion controller is stable. However, the controller needs detailed system information, and the uncertainty upper bound D is difficult to determine in a practical system, which problems indicate that the magnetic levitation switched reluctance motor inversion controller is difficult to realize in practical application. Therefore, in order to overcome these problems, a magnetic levitation switched reluctance motor fuzzy neural network controller is proposed.
And thirdly, establishing a fuzzy neural network controller of the magnetic suspension switch reluctance motor.
The designed magnetic levitation switch reluctance motor fuzzy neural network controller adopts the fuzzy neural network to approach the establishment of the magnetic levitation switch reluctance motor rotating speed inversion controller in the second step, thereby overcoming the defect that the establishment of the magnetic levitation switch reluctance motor rotating speed inversion controller needs system detailed information. Magnetic suspension switch magnetThe fuzzy neural network adopted by the fuzzy neural network controller of the resistance motor consists of an input layer, a fuzzification layer, a fuzzy reasoning layer and an output layer, wherein the input of the fuzzy neural network controller of the magnetic suspension switch reluctance motor is tracking error e 1 The output is the control force tau FNN . The functions of the signal propagation and layers in the fuzzy neural network are expressed as follows:
a first layer: input layer
The input layer comprises a plurality of nodes, each node is directly connected with the input tracking error e 1 Is connected to input tracking error e 1 Is passed on to the next layer.
Second time: blurring layer
Using Gaussian functions as membership functions Representing tracking error e 1 Element of (a)>And->The central vector and base width of membership functions of the jth fuzzy set of the ith input variable, respectively, i.e
For ease of calculation, N is employed pi Representing individual numbers of membership functions and defining adaptive parameter vectors b and c representing sets of all basis widths and center vectors of gaussian membership functions, respectively, i.e，Wherein->Representing the total number of membership functions.
Third layer: fuzzy inference layer
The fuzzy inference layer adopts a fuzzy inference mechanism, and the output of each node of the fuzzy inference layer is the product of all input signals of the node, namely
Wherein l k (k＝1,...,N y ) Representing the kth output of the fuzzy inference layer,representing the connection weight matrix between the fuzzification layer and the fuzzy inference layer, taken here as a unit vector, N y Is the total number of rules.
Fourth layer: and an output layer.
The nodes of the output layer represent output variables. Each node y of the output layer o (o＝1,...,,N o ) The output of (a) is the sum of all input signals of the node, i.e
In the whole fuzzy neural network, the input-output relationship is as follows:
wherein, obtaining tau FNN 。
In particular, in one embodiment, the fuzzy neural network controller of the magnetic levitation switch reluctance motor in the third step considers two self-regulating strategies of data learning and data deleting, thereby being beneficial to efficiently executing real-time control tasks.
First is a data learning strategy. The data learning process of the fuzzy neural network involves online evolution of rules and parameter updating that are closest to the current input data.
The fuzzy rule is determined stepwise according to the condition thatAnd psi is<E s . Where ψ is spherical potential energy, representing the novelty of the input data, is given by:
wherein E is s And E is a Is novel and the addition of a threshold value,E a the self-adjustment can be performed according to the following formula,
wherein the method comprises the steps ofRepresenting tracking errors. Delta represents a slope factor.
When a new fuzzy rule (l+1) needs to be added, its parameters are initialized,
where κ is the overlap parameter of the predefined fuzzy rule.
When (when)And adjusting rule parameters. Threshold E l Is also self-adjusting in terms of tracking error, given by,
then a data deletion policy. Specifically, when the current tracking error is close to the error in the previous fuzzy neural network iterative computation process, network parameters do not need to be updated, excessive learning is avoided, and the computation burden is reduced.
During the learning process, the contribution of a rule to the output may decrease. In this case, insignificant rules should be deleted from the rule base to avoid over-computation. The contribution of the q-th rule is given by:
wherein,n represents the dimension of the input.
If the degree of contribution of the q-th rule to the input is below the threshold E p The rule is deleted.
The fuzzy neural network has optimal control force according to the universal approximation theorySatisfy the following requirements
Wherein epsilon is the minimum reconstruction error vector, W * 、b * 、c * And l * The optimum values for W, b, c and l, respectively.
Assume the output control force of a fuzzy neural networkThe following form:
wherein,and->Respectively W * 、b * 、c * And l * Is used for the estimation of the estimated value of (a).
Defining an approximation error:
by Taylor series expansion, can be obtained
Wherein,b * and c * Is the optimal value of b and c, +.>And->Is b * And c * Estimated value of (2), O nv Is a higher term,/->
Then (23) substitution (22) is available
Wherein,
the designed adaptive law of the weight, base width and center vector of the fuzzy neural network can be designed as follows:
wherein,e is 2 Element sigma of (a) ω ,σ b ,σ c Is of normal number>Is omega i * Estimated value of ∈10->Is omega i Is the optimal value of r 1 ，r 2 ，r 3 Is a positive constant.
And (3) proving: defining a lyapunov function as
Derivation of formula (28) and substitution of (25-27) into
If the D is more than or equal to D plus E, the designed control method of the magnetic suspension switch reluctance motor fuzzy neural network is stable.
The above embodiments are only for illustrating the technical concept and features of the present invention, and are intended to enable those skilled in the art to understand the content of the present invention and to implement the same, but are not intended to limit the scope of the present invention, and all equivalent changes or modifications made according to the spirit of the present invention should be included in the scope of the present invention.
Claims (4)
1. A fuzzy neural network control method for a magnetic suspension switch reluctance motor is characterized by comprising the following steps of: comprising
Step one, establishing a mechanical model of a magnetic suspension switch reluctance motor, which specifically comprises the following steps:
the magnetic suspension switch reluctance motor is 2-phase 12/14-pole, and the kinetic equation is as follows:
wherein T is e Is electromagnetic torque, T L The load torque, ω is the angular velocity, J is the moment of inertia, and I is the coefficient of friction;
the derivation of formula (1) can be obtained:
wherein τ= -IT e For the control law to be designed,is a lumped item, and the upper bound is D;
step two, establishing a magnetic levitation switched reluctance motor rotating speed inversion controller, wherein the magnetic levitation switched reluctance motor rotating speed inversion controller calculates to obtain the actual rotating speed omega and the expected rotating speed omega of the magnetic levitation switched reluctance motor d Is the tracking error e 1 ；
In the second step, the magnetic levitation switched reluctance motor rotating speed inversion controller is designed by using an inversion method, and the control objective of the magnetic levitation switched reluctance motor rotating speed inversion controller is to design a proper control law tau so as to enable the actual rotating speed omega of the magnetic levitation switched reluctance motor to track the expected rotating speed omega d The design of the magnetic suspension switch reluctance motor rotating speed inversion controller is as follows:
wherein c 1 、c 2 Is a non-zero positive constant, e 1 ＝ω-ω d ，sgn (e) is a sign function, where
Step three, establishing a fuzzy neural network controller of the magnetic suspension switch reluctance motorThe fuzzy neural network controller of the magnetic levitation switch reluctance motor approaches the magnetic levitation switch reluctance motor rotating speed inversion controller by adopting the fuzzy neural network, an uncertainty upper bound D in the magnetic levitation switch reluctance motor rotating speed inversion controller is eliminated, and the tracking error e is used 1 Obtaining control force tau through built-in fuzzy neural network FNN The method specifically comprises the following steps:
a first layer: an input layer;
the input layer comprises a plurality of nodes, each node is directly connected with the input tracking error e 1 Is connected to input tracking error e 1 Passes the components of (a) to the next layer;
a second layer: a blurring layer;
using Gaussian functions as membership functions Representing tracking error e 1 Element of (a)>And->The central vector and base width of membership functions of the jth fuzzy set of the ith input variable, respectively, i.e
For ease of calculation, N is employed pi Representing individual numbers of membership functions and defining adaptive parameter vectors b and c representing sets of all basis widths and center vectors of gaussian membership functions, respectively, i.e，Wherein->Representing the total number of membership functions;
third layer: a fuzzy reasoning layer;
the fuzzy inference layer adopts a fuzzy inference mechanism, and the output of each node of the fuzzy inference layer is the product of all input signals of the node, namely
Wherein l k (k＝1,...,N y ) Representing the kth output of the fuzzy inference layer,representing the connection weight matrix between the fuzzification layer and the fuzzy inference layer, taken here as a unit vector, N y Is the total number of rules;
fourth layer: an output layer;
the nodes of the output layer represent output variables, each node y of the output layer o (o＝1,...,,N o ) The output of (a) is the sum of all input signals of the node, i.e
In the whole fuzzy neural network, the input-output relationship is as follows:
wherein, obtaining tau FNN ；
Step four, the control force tau is controlled FNN The magnetic suspension is input into a magnetic suspension switch reluctance motor to achieve the rotation speed control of the magnetic suspension switch reluctance motor.
2. The method for controlling the fuzzy neural network of the magnetic suspension switch reluctance motor according to claim 1, wherein the method comprises the following steps: the fuzzy neural network comprises two self-regulation strategies of data learning and data deleting, wherein the data learning is used for online evolution and parameter updating of rules of current input data in the fuzzy neural network, and the data deleting is that network parameters do not need to be updated when the current tracking error is close to the error of the previous iterative calculation process of the fuzzy neural network.
3. The method for controlling the fuzzy neural network of the magnetic suspension switch reluctance motor according to claim 2, which is characterized in that: the data learning is specifically
Determining fuzzy rules, i.e.And psi is<E s Where ψ is the spherical potential energy, representing the novelty of the input data,
wherein E is s And E is a Is novel and the addition of a threshold value,E a can be adjusted in a self-adjusting manner,
wherein the method comprises the steps ofRepresents the tracking error, delta represents the slope factor,
when a new fuzzy rule needs to be added, the parameters of the fuzzy rule are initialized,
where κ is the overlap parameter of the predefined fuzzy rule,
when (when)When the rule parameters and the threshold E are adjusted l Also based on tracking error e 1 Self-regulating
Data deletion is specifically
In the fuzzy neural network learning process, the contribution degree of the q-th rule
Wherein,n represents the dimension of the input and,
if the degree of contribution of the q-th rule to the input is below the threshold E p The rule is deleted.
4. According to claim 1The fuzzy neural network control method of the magnetic suspension switch reluctance motor is characterized by comprising the following steps of: in the third step, the fuzzy neural network controller of the magnetic suspension switch reluctance motor approximates the rotational speed inversion controller of the magnetic suspension switch reluctance motor by adopting a fuzzy neural network, and the method specifically comprises the step that the fuzzy neural network has optimal control force according to a universal approximation theoryOptimal control force->Satisfy the following requirements
Wherein epsilon is the minimum reconstruction error vector, W * 、b * 、c * And l * The optimal values of W, b, c and l, respectively;
assume the output control force of a fuzzy neural networkThe following form:
wherein,and->Respectively W * 、b * 、c * And l * Is a function of the estimated value of (2);
defining an approximation error:
by Taylor series expansion, can be obtained
Wherein,b * and c * Is the optimal value of b and c, +.>And->Is b * And c * Estimated value of (2), O nv Is a higher term,/->
Then (23) substitution (22) is available
Wherein,
the designed adaptive law of the weight, base width and center vector of the fuzzy neural network can be designed as follows:
wherein,e is 2 Element sigma of (a) ω ,σ b ,σ c Is of normal number>Is->Estimated value of ∈10->Is omega i Is the optimal value of r 1 ，r 2 ，r 3 Is a normal number of times, and the number of times is equal to the normal number,
defining a lyapunov function as
Derivation of formula (28) and substitution of (25-27) into
If the D is more than or equal to D plus E, the designed control method of the magnetic suspension switch reluctance motor fuzzy neural network is stable.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210631171.9A CN114844430B (en) | 2022-06-06 | 2022-06-06 | Fuzzy neural network control method for magnetic suspension switch reluctance motor |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210631171.9A CN114844430B (en) | 2022-06-06 | 2022-06-06 | Fuzzy neural network control method for magnetic suspension switch reluctance motor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114844430A CN114844430A (en) | 2022-08-02 |
| CN114844430B true CN114844430B (en) | 2024-03-01 |
Family
ID=82574449
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210631171.9A Active CN114844430B (en) | 2022-06-06 | 2022-06-06 | Fuzzy neural network control method for magnetic suspension switch reluctance motor |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114844430B (en) |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109103884A (en) * | 2018-09-18 | 2018-12-28 | 河海大学常州校区 | Active Power Filter-APF back stepping control method based on metacognition fuzzy neural network |
| CN112757298A (en) * | 2020-12-29 | 2021-05-07 | 苏州连恺自动化有限公司 | Intelligent inversion control method for manipulator |
| CN112947066A (en) * | 2021-01-26 | 2021-06-11 | 苏州连恺自动化有限公司 | Manipulator improved finite time inversion control method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6701236B2 (en) * | 2001-10-19 | 2004-03-02 | Yamaha Hatsudoki Kabushiki Kaisha | Intelligent mechatronic control suspension system based on soft computing |
-
2022
- 2022-06-06 CN CN202210631171.9A patent/CN114844430B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109103884A (en) * | 2018-09-18 | 2018-12-28 | 河海大学常州校区 | Active Power Filter-APF back stepping control method based on metacognition fuzzy neural network |
| CN112757298A (en) * | 2020-12-29 | 2021-05-07 | 苏州连恺自动化有限公司 | Intelligent inversion control method for manipulator |
| CN112947066A (en) * | 2021-01-26 | 2021-06-11 | 苏州连恺自动化有限公司 | Manipulator improved finite time inversion control method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114844430A (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109946975B (en) | Reinforced learning optimal tracking control method of unknown servo system | |
| Lin et al. | Data-based fault tolerant control for affine nonlinear systems through particle swarm optimized neural networks | |
| Kayacan et al. | A servo system control with time-varying and nonlinear load conditions using type-2 TSK fuzzy neural system | |
| CN108984815B (en) | Control method of vehicle-mounted flywheel battery radial suspension supporting system based on working condition | |
| Keighobadi et al. | From nonlinear to fuzzy approaches in trajectory tracking control of wheeled mobile robots | |
| Khater et al. | A novel structure of actor-critic learning based on an interval type-2 TSK fuzzy neural network | |
| Reyes-Reyes et al. | Bounded neuro-control position regulation for a geared DC motor | |
| CN112564557B (en) | Control method, device and equipment of permanent magnet synchronous motor and storage medium | |
| CN111007716A (en) | Alternating current servo motor variable discourse domain fuzzy PI control method based on prediction function | |
| CN110401391A (en) | Asynchronous motor stochastic system fuzzy self-adaption dynamic surface control method | |
| Hsu et al. | Microcontroller-based B-spline neural position control for voice coil motors | |
| CN113093526A (en) | Overshoot-free PID controller parameter setting method based on reinforcement learning | |
| Su et al. | Deep reinforcement learning method based on DDPG with simulated annealing for satellite attitude control system | |
| CN114844430B (en) | Fuzzy neural network control method for magnetic suspension switch reluctance motor | |
| Liu et al. | Her-pdqn: A reinforcement learning approach for uav navigation with hybrid action spaces and sparse rewards | |
| Lee et al. | Speed estimation and control of induction motor drive using hybrid intelligent control | |
| CN113009834A (en) | Fuzzy PID control optimization method for magnetic suspension flywheel motor | |
| Kumar et al. | Neural/fuzzy self learning Lyapunov control for non linear systems | |
| Neema et al. | Speed control of induction motor using fuzzy rule base | |
| CN113343592B (en) | Intelligent control method for DQN (direct current) of permanent magnet synchronous motor of new energy aircraft | |
| CN115102444A (en) | Self-adaptive integral sliding mode prediction control method for permanent magnet synchronous motor | |
| Iwasaki et al. | Development of a reference signal self-organizing control system based on deep reinforcement learning | |
| Prakash et al. | Neuro-PI controller based model reference adaptive control for nonlinear systems | |
| Brasch et al. | Lateral control of a vehicle using reinforcement learning | |
| Cigdem et al. | A novel training method based on variable structure systems theory for fuzzy 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |