Displacement detection method for modularized permanent magnet biased magnetic bearing without displacement sensor

Through the magnetic link observer and BP neural network of modular permanent magnet biased magnetic bearings, displacement detection without mechanical sensors is realized, solving the problem of limited accuracy of magnetic bearings in special environments, reducing design cost and control difficulty. Achieve high-precision nano-level detection effect.

CN118936288APending Publication Date: 2024-11-12NANJING TECH UNIV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410990201.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

Existing magnetic bearings need to install mechanical displacement sensors when obtaining displacement information, resulting in increased control difficulty and increased design cost, and limited measurement accuracy in special occasions such as high temperature, strong electromagnetic interference, and strong radiation.

Method used

Modular permanent magnet biased magnetic bearings without mechanical sensors are used to realize displacement detection through magnetic relay observers and BP neural networks, calculate the magnetic relay observation values ​​and input them into the trained BP neural network, output the rotor displacement values ​​to avoid The use of mechanical structures and corresponding complexity are achieved.

Benefits of technology

The displacement detection without mechanical sensors is realized, which reduces design costs and control difficulties, ensures detection accuracy in special environments, and can better reduce power consumption and extend service life, and achieves nano-level accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118936288A_ABST
    Figure CN118936288A_ABST
Patent Text Reader

Abstract

The invention discloses a displacement detection method for a modularized permanent magnet biased magnetic bearing without a displacement sensor, and the method comprises the steps: inputting an input current im and an input voltage um of the modularized permanent magnet biased magnetic bearing into a flux observer, and carrying out the integration of the input current im and the input voltage um through the flux observer, outputting a flux linkage observation value psi i of the flux linkage observer; and inputting the flux linkage observation value psi i into the trained BP neural network to output a displacement value of the rotor. According to the displacement detection method, a mechanical sensor does not need to be installed, so that equipment occupation is not needed, the economic cost can be greatly saved, the influence of a complex motion mode or a positioning problem is avoided, the maintenance is simple, the fault rate is low, serious interference in a specific working environment is avoided, the power consumption of the permanent magnet biased magnetic bearing can be better reduced, and the reliability of the permanent magnet biased magnetic bearing is improved. Therefore, the service life is prolonged, interference and influence of the environment of a working site are avoided, and the requirement of high-precision measurement can be met.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a displacement detection method for a modular permanent magnet biased magnetic bearing, in particular to a displacement detection method for a modular permanent magnet biased magnetic bearing without a displacement sensor. Background Art

[0002] With the introduction of the "dual carbon" policy in recent years, the country has vigorously promoted the establishment of a resource-saving and environmentally friendly society. Flywheel energy storage technology has gradually become an important development direction of future energy storage technology due to its high energy conversion rate, extremely low loss and long service life. Magnetic bearings are one of the key technologies for flywheel energy storage, and research in this field has gradually become a key topic in the current research of flywheel energy storage technology. Many research institutions at home and abroad have applied magnetic bearings to various scientific and technological research such as high-speed motors and satellite momentum flywheels.

[0003] As the most widely used magnetic bearing, hybrid magnetic bearing has the characteristics of no physical contact, no need for lubrication, high speed and low power loss. It is generally composed of a magnetic bearing stator body, a magnetic bearing rotor body, a sensor and a controller. Similar to traditional hybrid magnetic bearings, modular permanent magnet biased magnetic bearings replace the bias coils in conventional designs with permanent magnets to generate the required bias magnetic field, greatly reducing the energy loss of magnetic bearings. By changing the magnitude and direction of the current in the control winding, a control magnetic field is established at the air gap. The superposition of the bias magnetic field and the control magnetic field generates a controllable magnetic field attraction, thereby achieving stable suspension of the rotor. Modular permanent magnet biased magnetic bearings can overcome the problem that traditional magnetic bearings are not applicable when facing different force conditions through different combinations, and therefore have irreplaceable advantages in modern industrial technology.

[0004] Generally speaking, the acquisition of magnetic bearing displacement information needs to be achieved by installing a mechanical displacement sensor. However, the use of mechanical displacement sensors not only brings inconvenience to magnetic bearing control, but also increases the design cost. In addition, when magnetic bearings are in some special working situations, the use of mechanical displacement sensors greatly limits the measurement accuracy. Therefore, the research on magnetic bearing displacement sensor-free technology has become an important requirement for the development of magnetic bearings towards high reliability. Summary of the invention

[0005] The purpose of the invention is to provide a displacement detection method for a modular permanent magnet biased magnetic bearing without a displacement sensor, which can realize the displacement detection of the modular permanent magnet biased magnetic bearing without a mechanical sensor, reduce the control difficulty and design cost of the magnetic bearing, and ensure that the detection accuracy is not affected when dealing with special occasions such as high temperature, strong electromagnetic interference, and strong radiation.

[0006] Technical solution: The displacement detection method of the modular permanent magnet biased magnetic bearing without displacement sensor described in the present invention comprises the following steps:

[0007] Step 1: Set the input current i of the modular permanent magnet bias magnetic bearing to m And the input voltage u m Input flux observer, through the flux observer to input current i m And the input voltage u m Integrate and output the flux observation value ψ of the flux observer i ;

[0008] Step 2: Input the flux observation value ψi into the trained BP neural network to output the rotor displacement value.

[0009] Furthermore, in step 1, the calculation formula of the magnetic flux observation value ψi is:

[0010] ψ i =∫(u w -R w i w )dt

[0011] In the formula, u w is the equivalent voltage of the control winding, R w is the equivalent resistance of the control winding, i w is the equivalent current of the control winding, where i w Equivalent to the input current i m ,u w Equivalent to input voltage u m .

[0012] Furthermore, in step 2, the BP neural network needs to establish corresponding training data sets and test data sets during training. The specific steps are as follows:

[0013] Firstly, a simulation model is constructed according to the structural parameters of the modular permanent magnet biased magnetic bearing. Then, the magnetic flux φ of the rotor of the modular permanent magnet biased magnetic bearing simulation model when the displacement value x occurs is calculated according to the equivalent magnetic circuit method. xb ;

[0014] Then, according to the magnetic flux φ of the modular permanent magnet biased magnetic bearing xb Calculate the input current i for the input flux observer m ;

[0015] Then, according to the external input voltage u m 、Input current i m And the equivalent resistance R of the control winding of the modular permanent magnet biased magnetic bearing simulation model w Using the flux observation value ψ iThe calculation formula is used to calculate the magnetic flux observation value ψ i ;

[0016] Finally, using the flux observation value ψ i And the rotor displacement value x is used to construct a training data set and a test data set for training the BP neural network.

[0017] Furthermore, the magnetic flux φ when the rotor has a displacement value x is calculated xb The calculation formula is:

[0018]

[0019] In the formula, R x is the air gap reluctance when the rotor has a displacement value x, F m is the magnetomotive force of the permanent magnet, R m is the reluctance of the permanent magnet.

[0020] Furthermore, the air gap reluctance R x The calculation formula is:

[0021]

[0022] Where g0 is the air gap length when the rotor is in equilibrium position, μ0 is the air magnetic permeability, S z is the area of ​​the stator pole, and x is the offset of the magnetic bearing rotor.

[0023] Furthermore, the magnetomotive force F of the permanent magnet m The calculation formula is:

[0024] F m =H m L m

[0025] In the formula, H m is the magnetic field strength at the permanent magnet working point, L m is the thickness of the permanent magnet.

[0026] Furthermore, the reluctance R of the permanent magnet m The calculation formula is:

[0027]

[0028] In the formula, μ r is the relative magnetic permeability of the permanent magnet, S m is the cross-sectional area of ​​the permanent magnet.

[0029] Furthermore, according to the magnetic flux φ of the modular permanent magnet biased magnetic bearing xb Calculate the input current i for the input flux observer m The calculation formula is:

[0030]

[0031] Where: N is the number of turns of the control coil winding.

[0032] Furthermore, in step 1, the modular permanent magnet biased magnetic bearing includes a rotor and two stator structures; the rotor is located between the two stator structures, and the two stator structures are symmetrical about the rotor; the stator structure includes two stator units, each stator unit includes an outer stator, a permanent magnet, an inner stator, two control windings and two stator bosses; the permanent magnet is located between the outer stator and the inner stator; the two stator bosses are both located on the inner stator, and the ends of the stator bosses are close to the rotor on the corresponding side; the two control windings are respectively arranged around the two stator bosses.

[0033] Compared with the prior art, the present invention has the following beneficial effects: since there is no need to install mechanical sensors, no equipment space is required, which can greatly save economic costs; since there is no mechanical structure, it is not affected by complex motion patterns or positioning problems, maintenance is simple, and the failure rate is low; it will not be seriously disturbed under specific working environments (such as high temperature, strong electromagnetic field, strong radiation, etc.); displacement detection is achieved through a BP neural network detection program, so it can better reduce the power consumption of permanent magnet biased magnetic bearings, thereby extending the service life; since displacement detection is achieved through a flux observer and a BP neural network, it can more accurately detect the position and displacement of an object, and its accuracy is usually at the nanometer level, and it will not be affected by the interference of the workplace environment, and can meet the needs of high-precision measurement. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of the method of the present invention;

[0035] Figure 2 It is a front view structural schematic diagram of a modular permanent magnet biased magnetic bearing of the present invention;

[0036] Figure 3 It is a right side structural schematic diagram of the modular permanent magnet biased magnetic bearing of the present invention;

[0037] Figure 4 It is a schematic diagram of the flux observer integration algorithm of the present invention;

[0038] Figure 5 It is a schematic diagram of the BP neural network structure of the present invention. DETAILED DESCRIPTION

[0039] The technical solution of the present invention is described in detail below in conjunction with the accompanying drawings, but the protection scope of the present invention is not limited to the embodiments.

[0040] like Figure 1 As shown, the displacement detection method of the modular permanent magnet biased magnetic bearing without displacement sensor disclosed in the present invention comprises the following steps:

[0041] Step 1: Set the input current i of the modular permanent magnet bias magnetic bearing to m And the input voltage u m Input flux observer, through the flux observer to input current i m And the input voltage u m Perform integration, such as Figure 4 As shown, the flux observation value ψ of the output flux observer i ;

[0042] Step 2: Transform the magnetic flux observation value ψ i Input into the trained BP neural network to output the rotor displacement value.

[0043] Furthermore, in step 1, the flux observation value ψ i The calculation formula is:

[0044] ψ i =∫(u w -R w i w )dt

[0045] In the formula, u w is the equivalent voltage of the control winding, R w is the equivalent resistance of the control winding, i w is the equivalent current of the control winding, where i w Equivalent to the input current i m ,u w Equivalent to input voltage u m .

[0046] Furthermore, in step 2, the BP neural network needs to establish corresponding training data sets and test data sets during training. The specific steps are as follows:

[0047] Firstly, a simulation model is constructed according to the structural parameters of the modular permanent magnet biased magnetic bearing. Then, the magnetic flux φ of the rotor of the modular permanent magnet biased magnetic bearing simulation model when the displacement value x occurs is calculated according to the equivalent magnetic circuit method. xb ;

[0048] Then, according to the magnetic flux φ of the modular permanent magnet bias magnetic bearing xb Calculate the input current i for the input flux observer m ;

[0049] Then, according to the external input voltage u m 、Input current i mAnd the equivalent resistance R of the control winding of the modular permanent magnet biased magnetic bearing simulation model w Using the flux observation value ψ i The calculation formula is used to calculate the magnetic flux observation value ψ i ;

[0050] Finally, using the flux observation value ψ i And the rotor displacement value x is used to construct a training data set and a test data set for training the BP neural network.

[0051] Furthermore, the simulation model constructed according to the structural parameters of the modular permanent magnet biased magnetic bearing has the following restrictions:

[0052] (1) The back magnetomotive force reflects the rate of change of magnetic flux. When the rotor speed increases, the rate of change of magnetic flux will also increase, and the back magnetomotive force will increase accordingly. Therefore, it is necessary to limit the rotor speed range to less than 30,000 r / min.

[0053] (2) The total flux of the magnetic bearing is equal to the number of turns multiplied by the flux of its single turn. Therefore, the higher the number of turns, the greater the flux and the greater the back magnetomotive force. In the model of the present invention, the number of turns of the magnetic bearing coil is 170 turns.

[0054] (3) The magnetomotive force and magnetic resistance of a single-turn flux and magnetic bearing are related. Therefore, the greater the magnetomotive force, the smaller the magnetic resistance in the direction of the flux, and the greater the reverse magnetomotive force.

[0055] (4) The magnetic resistance of the magnetic bearing is related to the air gap. The larger the air gap, the larger the magnetic resistance, that is, the smaller the reverse magnetomotive force.

[0056] Furthermore, the magnetic flux φ when the rotor has a displacement value x is calculated xb The calculation formula is:

[0057]

[0058] In the formula, R x is the air gap reluctance when the rotor has a displacement value x, F m is the magnetomotive force of the permanent magnet, R m is the reluctance of the permanent magnet.

[0059] Furthermore, the air gap reluctance R x The calculation formula is:

[0060]

[0061] Where g0 is the air gap length when the rotor is in equilibrium position, μ0 is the air magnetic permeability, S z is the area of ​​the stator pole, and x is the offset of the magnetic bearing rotor.

[0062] Furthermore, the magnetomotive force F of the permanent magnetm The calculation formula is:

[0063] F m =H m L m

[0064] In the formula, H m is the magnetic field strength at the permanent magnet working point, L m is the thickness of the permanent magnet.

[0065] Furthermore, the reluctance R of the permanent magnet m The calculation formula is:

[0066]

[0067] In the formula, μ r is the relative magnetic permeability of the permanent magnet, S m is the cross-sectional area of ​​the permanent magnet.

[0068] Furthermore, according to the magnetic flux φ of the modular permanent magnet biased magnetic bearing xb Calculate the input current i for the input flux observer m The calculation formula is:

[0069]

[0070] Where: N is the number of turns of the control coil winding.

[0071] Further, in step 1, the modular permanent magnet biased magnetic bearing includes a rotor 1 and two stator structures, such as Figure 2 and 3 As shown; the rotor 1 is located between the two stator structures, and the two stator structures are symmetrical about the rotor from top to bottom; the stator structure includes two stator units, the stator units are fixedly installed by a fixed bracket, and each stator unit includes an outer stator 2, a permanent magnet 4, an inner stator 3, two control windings 6 and two stator bosses 5; the permanent magnet 4 is located between the outer stator 2 and the inner stator 3, and is in a suspended state; the two stator bosses 5 are both located on the inner stator 3, and the ends of the stator bosses 5 are close to the rotor 1 on the corresponding side, and the rotor 1 is suspended between the ends of the four stator bosses 5 on the upper and lower sides; the two control windings 6 are respectively arranged around the two stator bosses 5.

[0072] Furthermore, when the modular permanent magnet biased magnetic bearing is in operation, when the rotor 1 of the magnetic bearing moves, the magnetic bearing is in an unbalanced state, and the rotor 1 with the equilibrium position as the reference point generates a certain amount of displacement value x. Therefore, it is necessary to calculate the air gap magnetic resistance in the magnetic circuit through the equivalent magnetic circuit of the magnetic bearing. The magnetic flux of the magnetic bearing can be inferred based on the calculation result of the air gap magnetic resistance, and then the current can be inferred.

[0073] Furthermore, in step 4, when the BP neural network is trained, the specific steps are:

[0074] First, each magnetic flux observation value ψ i The data are respectively input into the simulation model of the modular permanent magnet biased magnetic bearing for simulation operation, and then the corresponding rotor displacement value is acquired by using an oscilloscope;

[0075] Then, using each magnetic flux observation value ψ i and the corresponding rotor displacement values ​​to construct training sets and test sets, and use the training sets to train the BP neural network;

[0076] Then the trained BP neural network is tested using the test set. If the error between the output displacement value and the corresponding test result value is within the set error range, the test is passed and the BP neural network training is completed. If the error between the output displacement value and the corresponding test result value is outside the set error range, the test is failed. The BP neural network is adjusted according to the error value until the error value is within the set error range.

[0077] Furthermore, the BP neural network of the present invention comprises an input layer, a hidden layer and an output layer; the input layer has two input nodes, the output layer has one output node, and the number of nodes in the hidden layer can be determined to be five according to an empirical formula, such as Figure 5 shown.

[0078] The BP neural network adopted in the present invention is an existing neural network structure, in which the input layer is the first step for data to enter the neural network, and its main function is to receive the original data and convert it into a format that the neural network can process; the hidden layer is the core part of the neural network, responsible for nonlinear transformation and feature extraction of the input signal, and the number of neurons in the hidden layer can be adjusted according to the complexity of the problem and the amount of data; the output layer is the last layer in the neural network, and its output is the prediction or classification result of the neural network for the input data.

[0079] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A displacement detection method for a modular permanent magnet biased magnetic bearing without a displacement sensor, characterized in that: The steps include: Step 1: Set the input current i of the modular permanent magnet bias magnetic bearing to m And the input voltage u m Input flux observer, through the flux observer to input current i m And the input voltage u m Integrate and output the flux observation value ψ of the flux observer i ; Step 2: Transform the magnetic flux observation value ψ i Input into the trained BP neural network to output the rotor displacement value.

2. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 1 is characterized in that: In step 1, the flux observation value ψ i The calculation formula is: ψ i =∫(u w -R w i w )dt In the formula, u w is the equivalent voltage of the control winding, R w is the equivalent resistance of the control winding, i w is the equivalent current of the control winding, where i w Equivalent to the input current i m ,u w Equivalent to input voltage u m .

3. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 2 is characterized in that: In step 2, the BP neural network needs to establish corresponding training data sets and test data sets during training. The specific steps are: Firstly, a simulation model is constructed according to the structural parameters of the modular permanent magnet biased magnetic bearing. Then, the magnetic flux φ of the rotor of the modular permanent magnet biased magnetic bearing simulation model when the displacement value x occurs is calculated according to the equivalent magnetic circuit method. xb ; Then, according to the magnetic flux φ of the modular permanent magnet bias magnetic bearing xb Calculate the input current i for the input flux observer m ; Then, according to the external input voltage u m 、Input current i m And the equivalent resistance R of the control winding of the modular permanent magnet biased magnetic bearing simulation model w Using the flux observation value ψ i The calculation formula is used to calculate the magnetic flux observation value ψ i ; Finally, using the flux observation value ψ i And the rotor displacement value x is used to construct a training data set and a test data set for training the BP neural network.

4. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 3 is characterized in that: Calculate the magnetic flux φ when the rotor has a displacement value x xb The calculation formula is: In the formula, R x is the air gap reluctance when the rotor has a displacement value x, F m is the magnetomotive force of the permanent magnet, R m is the reluctance of the permanent magnet.

5. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 4 is characterized in that: Air gap reluctance R x The calculation formula is: Where g0 is the air gap length when the rotor is in equilibrium position, μ0 is the air magnetic permeability, S z is the area of ​​the stator pole, and x is the offset of the magnetic bearing rotor.

6. The displacement detection method of modular permanent magnetic bias magnetic bearing without displacement sensor according to claim 4 is characterized in that: The magnetomotive force F of a permanent magnet m The calculation formula is: F m =H m L m In the formula, H m is the magnetic field strength at the permanent magnet working point, L m is the thickness of the permanent magnet.

7. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 4 is characterized in that: The reluctance R of the permanent magnet m The calculation formula is: In the formula, μ r is the relative magnetic permeability of the permanent magnet, S m is the cross-sectional area of ​​the permanent magnet.

8. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 3 is characterized in that: According to the magnetic flux φ of the modular permanent magnet biased magnetic bearing xb Calculate the input current i for the input flux observer m The calculation formula is: Where: N is the number of turns of the control coil winding.

9. The displacement detection method of modular permanent magnet biased magnetic bearing without displacement sensor according to claim 1, characterized in that: In step 1, the modular permanent magnet biased magnetic bearing includes a rotor and two stator structures; the rotor is located between the two stator structures, and the two stator structures are symmetrical about the rotor; the stator structure includes two stator units, each stator unit includes an outer stator, a permanent magnet, an inner stator, two control windings and two stator bosses; the permanent magnet is located between the outer stator and the inner stator; the two stator bosses are both located on the inner stator, and the ends of the stator bosses are close to the rotor on the corresponding side; the two control windings are respectively arranged around the two stator bosses.