1. A displacement determination method for a magnetic bearing, the magnetic bearing comprising: a magnetic bearing coil and a magnetic bearing rotor; the displacement determination method of the magnetic bearing comprises the following steps:

constructing a displacement measurement model based on a wavelet neural network, and recording the displacement measurement model as a displacement soft measurement model;

under the condition that the current displacement parameter of the magnetic bearing rotor needs to be determined, the current control current of the magnetic bearing coil is obtained;

and inputting the current control current of the magnetic bearing coil as an input parameter into the displacement soft measurement model, and determining an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor so as to determine the current displacement parameter of the magnetic bearing rotor.

2. The method of determining displacement of a magnetic bearing according to claim 1, wherein inputting the current control current of the magnetic bearing coils as an input parameter into the displacement soft measurement model and determining an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor comprises:

and arranging the displacement soft measurement model between a sampling end of the current control current of the magnetic bearing coil and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the displacement soft measurement model can output the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coil.

3. The method for determining displacement of magnetic bearing according to claim 1, wherein constructing a displacement measurement model based on wavelet neural network, denoted as displacement soft measurement model, comprises:

collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data;

and training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

4. The method for determining the displacement of the magnetic bearing according to claim 3, wherein the number of the magnetic bearing coils is one or more;

collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data, including:

selecting the set control current of more than one magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as characteristic variables;

and carrying out normalization processing on the characteristic variables to form sampling data which is used as sample data for training and testing.

5. The method of claim 3, wherein the training of the wavelet neural network based on the sample data with the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter to obtain the soft measurement model comprises:

setting a wavelet neural network model; under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one;

optimizing the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm to obtain the optimized wavelet neural network model;

and training the optimized wavelet neural network model by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

6. The displacement determination method for magnetic bearing according to any one of claims 1 to 5, further comprising:

after the determination of the current displacement parameter of the magnetic bearing rotor is realized, the current control current of the magnetic bearing coil is adjusted according to the current displacement parameter of the magnetic bearing rotor, and the adjustment value of the current control current of the magnetic bearing coil is obtained;

and controlling the magnetic bearing coil to operate according to the current control current adjustment value of the magnetic bearing coil.

7. A displacement determination device for a magnetic bearing, the magnetic bearing comprising: a magnetic bearing coil and a magnetic bearing rotor; a displacement determining apparatus of the magnetic bearing, comprising:

the modeling unit is configured to construct a displacement measurement model based on the wavelet neural network, and the displacement measurement model is recorded as a displacement soft measurement model;

an obtaining unit configured to obtain a current control current of the magnetic bearing coils in case a current displacement parameter of the magnetic bearing rotor needs to be determined;

a control unit configured to input the displacement soft measurement model with the current control current of the magnetic bearing coil as an input parameter, and determine an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor, thereby achieving determination of the current displacement parameter of the magnetic bearing rotor.

8. The displacement determination device for magnetic bearing according to claim 7, wherein the control unit inputs the current control current of the magnetic bearing coils as an input parameter into the displacement soft measurement model, and determines an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor, comprising:

and arranging the displacement soft measurement model between a sampling end of the current control current of the magnetic bearing coil and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the displacement soft measurement model can output the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coil.

9. The displacement determination device for magnetic bearings according to claim 7, wherein the modeling unit, which constructs a displacement measurement model based on a wavelet neural network, denoted as a displacement soft measurement model, comprises:

collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data;

and training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

10. The displacement determination device for magnetic bearing of claim 9, wherein the number of the magnetic bearing coils is one or more;

the modeling unit collects a set control current of the magnetic bearing coil and a set displacement parameter of the magnetic bearing rotor as sample data, and includes:

selecting the set control current of more than one magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as characteristic variables;

and carrying out normalization processing on the characteristic variables to form sampling data which is used as sample data for training and testing.

11. The device for determining displacement of magnetic bearing according to claim 9, wherein the modeling unit trains a wavelet neural network to obtain the soft measurement model of displacement based on the sample data with the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter, and comprises:

setting a wavelet neural network model; under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one;

optimizing the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm to obtain the optimized wavelet neural network model;

and training the optimized wavelet neural network model by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

12. The displacement determination device for magnetic bearing according to any one of claims 7 to 11, further comprising:

the control unit is further configured to adjust the current control current of the magnetic bearing coil according to the current displacement parameter of the magnetic bearing rotor after the determination of the current displacement parameter of the magnetic bearing rotor is achieved, so as to obtain an adjustment value of the current control current of the magnetic bearing coil;

the control unit is further configured to control the operation of the magnetic bearing coils according to the adjusted values of the current control currents of the magnetic bearing coils.

13. A magnetic bearing system, comprising: a displacement determination device for a magnetic bearing as claimed in any one of claims 7 to 12.

14. A storage medium, characterized in that the storage medium comprises a stored program, wherein the storage medium is controlled in a device in which the program is run to perform a displacement determination method of a magnetic bearing according to any one of claims 1 to 6.

15. A processor characterized in that the processor is configured to run a program, wherein the program is run to perform the displacement determination method of a magnetic bearing according to any one of claims 1 to 6.

Background

The magnetic suspension bearing stably suspends the magnetic bearing rotor in the air by utilizing magnetic field force, has the characteristics of no mechanical wear, no need of lubrication, no need of maintenance and the like, and can realize high-speed operation of a motor. In a related scheme, the magnetic suspension bearing system detects the displacement of the magnetic bearing rotor through a displacement sensor, compares the displacement with a given displacement value and sends the displacement to a suspension control system to generate suspension control current.

In the related scheme, an eddy current sensor is mostly adopted during the measurement of the radial displacement of the magnetic bearing rotor, and if the eddy current sensor is damaged, the magnetic bearing rotor is caused to have a falling accident.

The eddy current sensor occupies a large space, so that the volume of the magnetic suspension bearing control system is increased.

The above is only for the purpose of assisting understanding of the technical aspects of the present invention, and does not represent an admission that the above is prior art.

Disclosure of Invention

The invention aims to provide a displacement determination method and device of a magnetic bearing, a magnetic suspension bearing system, a storage medium and a processor, which aim to solve the problem that an eddy current sensor is mostly adopted when the radial displacement of a magnetic bearing rotor is measured, and a magnetic bearing rotor is dropped if the eddy current sensor is damaged, so that the radial displacement of the magnetic bearing rotor is measured by using a soft measurement mode, the dropping accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the eddy current sensor is used is avoided, and the suspension safety of the magnetic bearing rotor can be improved.

The present invention provides a method for determining the displacement of a magnetic bearing, the magnetic bearing comprising: a magnetic bearing coil and a magnetic bearing rotor; the displacement determination method of the magnetic bearing comprises the following steps: constructing a displacement measurement model based on a wavelet neural network, and recording the displacement measurement model as a displacement soft measurement model; under the condition that the current displacement parameter of the magnetic bearing rotor needs to be determined, the current control current of the magnetic bearing coil is obtained; and inputting the current control current of the magnetic bearing coil as an input parameter into the displacement soft measurement model, and determining an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor so as to determine the current displacement parameter of the magnetic bearing rotor.

In some embodiments, inputting the soft-measurement displacement model with the current control current of the magnetic bearing coils as an input parameter, and determining an output parameter of the soft-measurement displacement model as a current displacement parameter of the magnetic bearing rotor, comprises: and arranging the displacement soft measurement model between a sampling end of the current control current of the magnetic bearing coil and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the displacement soft measurement model can output the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coil.

In some embodiments, constructing a displacement measurement model based on a wavelet neural network, denoted as a displacement soft measurement model, includes: collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data; and training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

In some embodiments, the number of magnetic bearing coils is more than one; collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data, including: selecting the set control current of more than one magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as characteristic variables; and carrying out normalization processing on the characteristic variables to form sampling data which is used as sample data for training and testing.

In some embodiments, training a wavelet neural network to obtain the soft displacement measurement model by using the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data includes: setting a wavelet neural network model; under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one; optimizing the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm to obtain the optimized wavelet neural network model; and training the optimized wavelet neural network model by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

In some embodiments, further comprising: after the determination of the current displacement parameter of the magnetic bearing rotor is realized, the current control current of the magnetic bearing coil is adjusted according to the current displacement parameter of the magnetic bearing rotor, and the adjustment value of the current control current of the magnetic bearing coil is obtained; and controlling the magnetic bearing coil to operate according to the current control current adjustment value of the magnetic bearing coil.

In accordance with the above method, another aspect of the present invention provides a displacement determining apparatus for a magnetic bearing, the magnetic bearing including: a magnetic bearing coil and a magnetic bearing rotor; a displacement determining apparatus of the magnetic bearing, comprising: the modeling unit is configured to construct a displacement measurement model based on the wavelet neural network, and the displacement measurement model is recorded as a displacement soft measurement model; an obtaining unit configured to obtain a current control current of the magnetic bearing coils in case a current displacement parameter of the magnetic bearing rotor needs to be determined; a control unit configured to input the displacement soft measurement model with the current control current of the magnetic bearing coil as an input parameter, and determine an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor, thereby achieving determination of the current displacement parameter of the magnetic bearing rotor.

In some embodiments, the controlling unit inputting the displacement soft measurement model with the current control current of the magnetic bearing coils as an input parameter, and determining an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor, comprises: and arranging the displacement soft measurement model between a sampling end of the current control current of the magnetic bearing coil and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the displacement soft measurement model can output the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coil.

In some embodiments, the modeling unit, which constructs a displacement measurement model based on a wavelet neural network, is denoted as a displacement soft measurement model, and includes: collecting set control current of the magnetic bearing coil and set displacement parameters of the magnetic bearing rotor as sample data; and training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

In some embodiments, the number of magnetic bearing coils is more than one; the modeling unit collects a set control current of the magnetic bearing coil and a set displacement parameter of the magnetic bearing rotor as sample data, and includes: selecting the set control current of more than one magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as characteristic variables; and carrying out normalization processing on the characteristic variables to form sampling data which is used as sample data for training and testing.

In some embodiments, the modeling unit, based on the sample data, trains a wavelet neural network with a set control current of the magnetic bearing coil as an input parameter and a set displacement parameter of the magnetic bearing rotor as an output parameter, and obtains the soft displacement measurement model, and includes: setting a wavelet neural network model; under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one; optimizing the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm to obtain the optimized wavelet neural network model; and training the optimized wavelet neural network model by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

In some embodiments, further comprising: the control unit is further configured to adjust the current control current of the magnetic bearing coil according to the current displacement parameter of the magnetic bearing rotor after the determination of the current displacement parameter of the magnetic bearing rotor is achieved, so as to obtain an adjustment value of the current control current of the magnetic bearing coil; the control unit is further configured to control the operation of the magnetic bearing coils according to the adjusted values of the current control currents of the magnetic bearing coils.

In accordance with the above apparatus, a magnetic suspension bearing system according to another aspect of the present invention comprises: the displacement determining apparatus of the magnetic bearing described above.

In line with the above method, a further aspect of the present invention provides a storage medium comprising a stored program, wherein the program, when executed, controls a device in which the storage medium is located to perform the above-described method of determining displacement of a magnetic bearing.

In line with the above method, a further aspect of the invention provides a processor for running a program, wherein the program is run to perform the above described method of determining the displacement of a magnetic bearing.

Therefore, according to the scheme of the invention, the model is trained through the collected sample data, and the trained model can be directly connected in series to a magnetic suspension bearing system to form a radial displacement prediction module, so that the radial displacement of the magnetic suspension bearing rotor can be self-detected; therefore, the radial displacement of the magnetic bearing rotor is measured by using a soft measurement mode, the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the eddy current sensor is used is avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the scheme of the invention can also solve the problem that the volume of the magnetic suspension bearing control system is enlarged due to the fact that the eddy current sensor is mostly adopted when the radial displacement of the magnetic bearing rotor is measured, and the effect that the volume of the magnetic suspension bearing control system is enlarged due to the fact that the eddy current sensor is used for measuring the radial displacement of the magnetic bearing rotor in a soft measuring mode is achieved, and therefore the occupied space of the magnetic suspension bearing control system is saved.

Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention.

The technical solution of the present invention is further described in detail by the accompanying drawings and embodiments.

Drawings

FIG. 1 is a schematic flow chart diagram of one embodiment of a displacement determination method for a magnetic bearing of the present invention;

FIG. 2 is a schematic flow chart illustrating an embodiment of a method for constructing a displacement measurement model based on a wavelet neural network according to the present invention;

FIG. 3 is a schematic flow chart of one embodiment of collecting set control currents for the magnetic bearing coils and set displacement parameters for the magnetic bearing rotor in the method of the present invention;

FIG. 4 is a schematic flow chart illustrating an embodiment of training a wavelet neural network in the method of the present invention;

FIG. 5 is a schematic flow chart of one embodiment of controlling the current to the magnetic bearing coils based on the determined current displacement parameters of the magnetic bearing rotor in the method of the present invention;

FIG. 6 is a schematic structural view of an embodiment of a displacement determining apparatus of a magnetic bearing of the present invention;

FIG. 7 is a schematic flow chart diagram of one embodiment of constructing a wavelet neural network displacement soft measurement model;

FIG. 8 is a schematic diagram of an embodiment of a wavelet neural network architecture;

FIG. 9 is a schematic flow diagram of an embodiment of a process for an improved particle swarm optimization wavelet neural network;

fig. 10 is a schematic structural diagram of an embodiment of a displacement measurement circuit based on a wavelet neural network.

The reference numbers in the embodiments of the present invention are as follows, in combination with the accompanying drawings:

102-a modeling unit; 106-an obtaining unit; 104-control unit.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In consideration of the radial displacement measurement of the magnetic bearing rotor, an eddy current sensor is mostly adopted, and the eddy current sensor occupies a large space, so that the volume of the magnetic suspension bearing control system is increased, the cost of the magnetic suspension bearing control system is increased, and the magnetic suspension bearing control system is inconvenient to install and maintain. In a related scheme, a magnetic suspension bearing displacement self-detection method is provided, and the method mainly comprises a model reference self-adaption method and a support vector machine method. For example: in some schemes, an accurate mathematical model is not required to be relied on, the dynamic performance of the system is improved, but the stability, the learning capability and the generalization capability of the support vector machine mainly depend on the setting of parameters of the support vector machine, and the calculation amount of the algorithm is large, so that the network convergence speed is low and the local minimum value is easily caused. For another example: in other schemes, a soft measurement scheme is established based on an adaptive BP neural network (namely a multilayer feedforward neural network trained according to an error back propagation algorithm), and as a transfer excitation function of a hidden layer of the BP neural network is a tansig function (namely a neural network layer transfer function) which does not have good time-frequency localization property of a wavelet function, the BP neural network is easy to fall into local optimization, so that the prediction accuracy and robustness of the network are poor.

According to an embodiment of the present invention, a displacement determination method for a magnetic bearing is provided, as shown in fig. 1, which is a schematic flow chart of an embodiment of the method of the present invention. The magnetic bearing, comprising: a magnetic bearing coil and a magnetic bearing rotor; the displacement determination method of the magnetic bearing comprises the following steps: step S110 to step S130.

In step S110, a displacement measurement model based on the wavelet neural network is constructed and recorded as a displacement soft measurement model, i.e., a wavelet neural network displacement prediction model. The displacement soft measurement model is obtained by training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter.

In some embodiments, the wavelet neural network-based displacement measurement model is constructed in step S110, which is referred to as a specific process of displacement soft measurement model, see the following exemplary description.

The following further describes a specific process of constructing the wavelet neural network-based displacement measurement model in step S110, with reference to a flowchart of an embodiment of constructing the wavelet neural network-based displacement measurement model in the method of the present invention shown in fig. 2, including: step S210 and step S220.

Step S210, collecting the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as sample data. Among the set displacement parameters of the magnetic bearing rotor, the displacement parameters include: radial displacement of the magnetic bearing rotor.

In some embodiments, the number of magnetic bearing coils is more than one.

In step S210, the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor are collected as a specific process of sample data, which is described in the following exemplary description.

The specific process of collecting the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor in step S210 is further described with reference to the flowchart of an embodiment of collecting the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor in the method of the present invention shown in fig. 3, which includes: step S310 to step S320.

Step S310, selecting the set control current of more than one magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor as characteristic variables.

Step S320 normalizes the feature variables to form sample data, which is used as sample data for training and testing.

The scheme of the invention is to establish a nonlinear prediction model between the displacement and the current of the magnetic bearing by utilizing the algorithm of 'improved particle swarm optimization wavelet neural network', and can realize the displacement self-detection of the displacement-free sensor of the magnetic bearing. The model is established by adopting the formulas (1) to (6), so that the problem that the particle swarm falls into local optimum can be effectively avoided, the iteration times are reduced, and the prediction precision of the model is improved.

Fig. 7 is a schematic flow chart of an embodiment of constructing a wavelet neural network displacement soft measurement model. As shown in fig. 7, the soft measurement method for bearing displacement comprises the following steps:

the first step is as follows: and selecting characteristic variables. The characteristic variable of the displacement soft measurement model (namely the displacement measurement model based on the wavelet neural network) mainly comprises a model input variable and a model output variable and is measured by bearing radial displacement s_FX、s_FY、s_RX、s_RYThe suspension force control current i is a model output variable and can be known through model analysis_FX、i_FY、i_RX、i_RYIs correlated to bearing displacement, so it is determined to be the model input variable.

The second step is that: and preprocessing the sampled data. The input and output signals are normalized, and the normalization formula is as follows:

wherein A is_KIs normalized input data. x is the number of_KIs the original input data. x is the number of_max、x_minRespectively, the maximum and minimum values in the raw input data. And filtering the normalized data to obtain more accurate input and output data, and selecting N groups of representative sample data, wherein N is a positive integer. Wherein each set of sample data contains 4 input variables X_i＝[i_FX、i_FY、i_RX、i_RY]The corresponding output vector is Y_k＝[s_FX、s_FY、s_RX、s_RY]Sample data is formed for training and testing the model.

And S220, training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter based on the sample data to obtain the displacement soft measurement model.

Specifically, the scheme of the invention provides a method for magnetically suspended bearing rotor displacement soft measurement based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, a radial displacement prediction module (namely a displacement soft measurement model) is obtained through training, and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of a magnetically suspended bearing rotor can be self-detected without a mechanical eddy current sensor.

In some embodiments, in step S220, based on the sample data, the set control current of the magnetic bearing coil is used as an input parameter, and the set displacement parameter of the magnetic bearing rotor is used as an output parameter, and a wavelet neural network is trained to obtain a specific process of the displacement soft measurement model, which is described in the following exemplary description.

The following further describes a specific process of training the wavelet neural network in step S220 with reference to a flowchart of an embodiment of training the wavelet neural network in the method of the present invention shown in fig. 4, including: step S410 to step S430.

And step S410, setting a wavelet neural network model. And under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one.

And step S420, optimizing the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm to obtain the optimized wavelet neural network model.

As shown in fig. 7, the soft measurement method for bearing displacement further includes:

the third step: and designing a wavelet neural network. Fig. 8 is a schematic structural diagram of an embodiment of a wavelet neural network structure. The designed wavelet neural network model has three layers as shown in fig. 8, and is a multi-input multi-output model. The input layer variable of the network is X_i＝[i_FX、i_FY、i_RX、i_RY]Therefore, the number of input layer nodes is i (i is 1, 2, 3, 4). The output layer variable is Y_k＝[s_FX、s_FY、s_RX、s_RY]Therefore, the number of output layer nodes is k (k is 1, 2, 3, 4). The number of nodes in the hidden layer isAnd (4) respectively. As an example shown in fig. 8, the design uses Morlet mother wavelet function ψ (x) as the activation function of the hidden layer, which is expressed as follows:

the output layer excitation function is a Sigmoid function, and the formula is as follows:

if the ith input sample of the input layer is set to X_iThe k-th output value of the output layer is set to Y_kSetting the connection weight between the input layer node i and the hidden layer node j as omega_ijThe connection weight between the hidden layer node j and the output layer node k is set to be omega_jk. The scaling factor and the translation factor of the jth hidden layer node are respectively set as a_jAnd b_jThen, the mathematical model of the wavelet neural network is:

where, σ and u_kIs the excitation function of the output layer of the wavelet neural networkThe number of output layer nodes in (b) is K (K is 1, 2, 3, 4), and thus the characters in (b) are written as σ (u)_k) And represents the output value of the Kth node of the output layer.

As can be seen from the structure of the wavelet neural network shown in fig. 8, when the number of nodes in each layer is determined, the connection weight (ω) between each layer is determined_ij,ω_jk) Scale factor (a) with wavelet basis function_j) Translation factor (b)_j) And determining the quality of the performance of the wavelet neural network. Thus, pair (ω)_ij,ω_jk,a_j,b_j) It is important to optimize these parameters.

The fourth step: and optimizing parameters of the wavelet neural network. The main role of Particle Swarm Optimization (PSO) is to find a global optimal solution, which can be described as: assuming that a particle group is searched in an s-dimensional space, m particles form a group Z, which can be expressed as Z ═ Z (Z)₁，Z₂，Z₃，…，Z_m). The position of the ith particle is represented as Z_i＝(Z_i1，Z_i2，…，Z_is) The velocity of the ith particle is denoted V_i＝(V_i1，V_i2，…，V_is) The position of each particle represents a potential solution to the problem sought. Firstly, an adaptive value is obtained by using an objective function, and then the quality of the current particle performance is obtained according to the adaptive value of the particle. The particles continuously iterate to find new solutions according to the position updating formula, and the individual extreme value is marked as P_idAnd the total extreme value is denoted as P_gd. Each particle updates its own velocity according to equation (1) and its own position according to equation (2).

V_id(t+1)＝ωV_id(t)+η₁r₁(P_id-Z_id(t))+η₂r₂(P_gd-Z_id(t)) (1)。

Z_id(t+1)＝Z_id(t)+V_id(t+1) (2)。

Where ω is an inertial weight coefficient.η₁And η₂Is a non-negative learning factor. i is 1, 2, …, m. d is 1, 2, …, s. r1 and r2 are [0, 1 ]]The random number of (2).

Improving the inertia weight coefficient: the inertia weight coefficient omega is closely related to the global and local searching capability of the particle swarm. In general, the larger the value of ω, the stronger the global search ability of the particle group, and the smaller the value of ω, the stronger the local search ability of the particle group. Therefore, the global search capability in the early stage of particle swarm iteration is enhanced through the larger omega, and the local search in the later stage of particle swarm iteration is enhanced through the smaller omega, so that the swarm iteration speed and the iteration precision are improved.

Therefore, in the scheme of the invention, the particle swarm inertial weight coefficient is updated by adopting the formula (3).

Wherein, ω is_minAnd ω_maxRespectively the minimum value and the maximum value of the inertia weight change, t is the current iteration number, t_maxIs the set maximum number of iterations. For example: ω in the formula (3)_min、ω_max、t_maxThe values of (a) can be 0.2, 0.9, 500, respectively.

Improving the learning factor: due to η₁Determining the individual cognitive level contribution, η, of the population₂Determining the cognitive level contribution rate of the particle population. Therefore, in the early stage of iteration, the particle fitness is larger and can pass through larger eta₂To control the particles to develop in the optimal direction of the population. At the later stage of iteration, the particle fitness is gradually reduced and can pass through larger eta₁Individual cognition of the particles is released until an optimal position is found. In view of this, the scheme of the present invention adopts the following formula (4) and formula (5) to update the learning factor.

Wherein eta is_1min、η_2min、η_1maxAnd η_2maxThe minimum and maximum values of the learning factor are respectively indicated. Through improvement of the inertia weight and the learning factor of the particle swarm algorithm, the iteration times of the algorithm can be reduced, the calculation time is reduced, the algorithm precision can be improved, the problem that the particle swarm falls into local optimization can be effectively avoided through the improved model, and the prediction precision of the particle swarm algorithm is improved.

For example: eta in formula (4)_1min、η_1max、t_maxThe values of (A) are respectively 1.7, 2.3 and 500. Eta in formula (5)_2min、η_2max、t_maxThe values of (A) are also 1.7, 2.3, 500. The selection of the parameters can reduce the iteration times of the algorithm, and the calculation time is reduced by adjusting the development directions of particle swarms at different stages.

The particle swarm optimization wavelet neural network is to search parameters (omega) of the wavelet neural network in a search space of each particle through each particle in the network back propagation process_ij,ω_jk,a_j,b_j). The position of the particle in each dimension is the solution required, and the sum of the dimensions (ω) of the particles is established_ij,ω_jk,a_j,b_j) The weight (omega) of the neural network_ij,ω_jk) Scaling factor of number plus wavelet basis function (a)_j) Number and translation factor (b)_j) The sum of the numbers is equal to the dimension of the particle. For a three-layer network with a network structure of M-N-N, the number of network weights is M multiplied by N + N multiplied by N, and the number of scale factors and translation factors of the wavelet basis functions is N + N, so that the dimension of the particles is M multiplied by N + N multiplied by N + N + N. The mean square error function E is chosen as a function of the fitness value, as shown in equation (6):

wherein, M is the number of nodes of the input layer of the wavelet neural network, N is the number of nodes of the hidden layer of the wavelet neural network, and N is the number of nodes of the output layer of the wavelet neural network. And N is the number of sample data.

And step S430, based on the sample data, training the optimized wavelet neural network model by taking the set control current of the magnetic bearing coil as an input parameter and taking the set displacement parameter of the magnetic bearing rotor as an output parameter to obtain the displacement soft measurement model.

FIG. 9 is a flow diagram of an embodiment of a process for an improved particle swarm optimization wavelet neural network. The specific flow of the improved particle swarm optimization wavelet neural network training algorithm is shown in fig. 9. Assigning the value of each dimension of the optimized global optimal position of the particle to (omega)_ij,ω_jk,a_j,b_j) For a three-layer network with a network structure of M-N, the first mxn values are generally assigned to the weights between the input layer and the hidden layer, the values between the mxn +1 th and mxn + N are assigned to the scaling factors of the wavelet basis function, the values between the mxn + N +1 th and mxn +2N are assigned to the shifting factors of the wavelet basis function, and the remaining values are assigned to the weights between the hidden layer and the output layer.

As shown in fig. 7, the soft measurement method for bearing displacement further includes:

the fifth step: and establishing an improved particle swarm optimization-based wavelet neural network displacement soft measurement control system. And (3) connecting the wavelet neural network radial displacement prediction module which is optimally trained into a magnetic suspension bearing control system, and realizing the online real-time prediction of the radial displacement of the bearing rotor.

In step S120, in case that the current displacement parameter of the magnetic bearing rotor needs to be determined, the current control current of the magnetic bearing coil is obtained, for example, the current control current of the magnetic bearing coil sampled by a current sensor is obtained.

In step S130, the current control current of the magnetic bearing coil is used as an input parameter, the displacement soft measurement model is input, and an output parameter of the displacement soft measurement model is determined as a current displacement parameter of the magnetic bearing rotor, so as to determine the current displacement parameter of the magnetic bearing rotor. That is, the determination of the current displacement parameters of the magnetic bearing rotor is effected in accordance with the current control circuitry of the magnetic bearing coils. Among the current displacement parameters of the magnetic bearing rotor, displacement parameters include: radial displacement of the magnetic bearing rotor.

In specific implementation, a corresponding relationship between the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor may be established. And under the condition that the current displacement parameter of the magnetic bearing rotor needs to be determined, obtaining the current control current of the magnetic bearing coil. And determining the set displacement parameter corresponding to the set control current which is the same as the current control current of the magnetic bearing coil in the corresponding relation as the current displacement parameter of the magnetic bearing rotor corresponding to the current control current of the magnetic bearing coil, namely, determining the current displacement parameter of the magnetic bearing rotor according to the current control circuit of the magnetic bearing coil. In the setting displacement parameter of the magnetic bearing rotor and the current displacement parameter of the magnetic bearing rotor, the displacement parameter includes: radial displacement of the magnetic bearing rotor.

Wherein, constructing the corresponding relationship between the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor may include at least one of the following construction conditions:

the first construction case: it may be a numerical relationship or table that establishes a relationship between the set control currents of the magnetic bearing coils and the set displacement parameters of the magnetic bearing rotor.

The second construction case: the method can be used for constructing a displacement measurement model based on the wavelet neural network, and recording the displacement measurement model as a displacement soft measurement model, namely a wavelet neural network displacement prediction model. The displacement soft measurement model is a model which takes the control current of the magnetic bearing coil as an input parameter and takes the displacement parameter of the magnetic bearing rotor as an output parameter.

Thus, the scheme of the invention realizes the sensorless radial displacement self-detection technology by self-detecting the radial displacement of the magnetic suspension bearing rotor, namely, the radial displacement of the magnetic suspension bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, and the problems that the rotor suddenly falls off and the system is damaged by friction with a mechanical part when the mechanical sensor is damaged are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the system cost and the volume are reduced, the stability of the magnetic suspension bearing control system is improved, and the later maintenance is facilitated.

In some embodiments, the inputting the displacement soft measurement model with the current control current of the magnetic bearing coils as an input parameter and the determining the output parameter of the displacement soft measurement model as the current displacement parameter of the magnetic bearing rotor in step S130 includes: and arranging the displacement soft measurement model between a sampling end of the current control current of the magnetic bearing coil and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the displacement soft measurement model can output the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coil.

Therefore, the scheme of the invention provides a soft measurement method for the displacement of the magnetic bearing rotor based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, the trained model can be directly connected in series to a magnetic bearing system to form a radial displacement prediction module (namely a displacement soft measurement model), and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of the magnetic bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, the problems that the rotor suddenly falls off when the mechanical sensor is damaged and the system is damaged by friction with a mechanical part are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the problems of large volume and high cost of the magnetic bearing control system are solved, and the problem that the volume of the magnetic bearing control system is increased due to the fact that an eddy current sensor is adopted in the radial displacement measurement of the magnetic bearing rotor can be solved. In addition, the magnetic suspension bearing which is nonlinear and difficult to establish a rotor displacement calculation model realizes automatic measurement without a displacement sensor, does not need an additional complex circuit and is easy to realize.

In some embodiments, further comprising: a process of controlling the current of the magnetic bearing coils based on the determined current displacement parameters of the magnetic bearing rotor.

In the method of the present invention shown in fig. 5, a specific process of controlling the current of the magnetic bearing coil based on the determined current displacement parameter of the magnetic bearing rotor is further described, which includes: step S510 and step S520.

Step S510, after the determination of the current displacement parameter of the magnetic bearing rotor is achieved, the current control current of the magnetic bearing coil is adjusted according to the current displacement parameter of the magnetic bearing rotor, so as to obtain an adjustment value of the current control current of the magnetic bearing coil. Specifically, adjusting the current control current of the magnetic bearing coil according to the current displacement parameter of the magnetic bearing rotor comprises: and performing PID control according to the difference value between the current displacement parameter of the magnetic bearing rotor and the reference displacement parameter of the magnetic bearing rotor to obtain a first processing result. And performing PWM processing on the difference value of the first processing result and the current control current of the magnetic bearing coil to obtain a second processing result. And then carrying out power amplification processing on the second processing result to obtain an adjustment value of the current control current of the magnetic bearing coil.

And S520, controlling the magnetic bearing coil to operate according to the current control current adjustment value of the magnetic bearing coil.

Fig. 10 is a schematic structural diagram of an embodiment of a displacement measurement circuit based on a wavelet neural network. As shown in fig. 10, the circuit for measuring displacement based on wavelet neural network comprises: for each magnetic bearing coil, such as a radial forward X-direction coil (FX coil), a radial forward Y-direction coil (FY coil), a radial backward X-direction coil (RX coil), and a radial backward Y-direction coil (RY coil), a control branch is provided. The current signal output by each control branch, namely the control current of each magnetic bearing coil, passes through a Wavelet Neural Network (WNN) displacement measurement model (namely a displacement measurement model based on the wavelet neural network), and then outputs the predicted rotor displacement of the magnetic bearing rotor. And predicting the rotor displacement of each path and a given reference position signal of the path, obtaining the difference value of the rotor displacement and the given reference position signal of the path through a first comparator, outputting an adjusting value of the control current of the magnetic bearing coil of the path after passing through a control branch of the path, and controlling the magnetic bearing coil of the path according to the adjusting value of the control current of the magnetic bearing coil of the path. Wherein, every way control branch road includes: the PWM control module, the second comparator, the PWM module and the power amplifier. And the PWM control module is used for outputting a first processing result after PID processing is carried out on the difference value of each path of predicted rotor displacement and the given reference position signal of the path. And the difference value of the first processing result of each path and the predicted rotor displacement of the path is processed by the PWM module to obtain a second processing result. And the second processing result is processed by the power amplifier to obtain an adjusting value of the control current of the magnetic bearing coil.

In the example shown in fig. 10, a displacement measurement model based on a wavelet neural network is applied to a magnetic bearing control system. Wavelet neural network displacement prediction model based on current signal i of power amplifier (such as switching power amplifier)_FX、i_FY、i_RX、i_RYFor the input signal, a predicted rotor displacement s is output_FX、s_FY、s_RX、s_RYAnd compares it with a given reference position signal f_FX、f_FY、f_RX、f_RYAnd comparing, and finally realizing the stable suspension of the magnetic bearing by the difference value through a closed-loop controller and a switching power amplifier.

Therefore, the scheme of the invention provides a soft measurement method for the displacement of the magnetic bearing rotor based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, the trained model can be directly connected in series to a magnetic bearing system to form a radial displacement prediction module (namely a displacement soft measurement model), and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of the magnetic bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, the problems that the rotor suddenly falls off when the mechanical sensor is damaged and the system is damaged by friction with a mechanical part are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the problems of large volume and high cost of the magnetic bearing control system are solved, and the problem that the volume of the magnetic bearing control system is increased due to the fact that an eddy current sensor is adopted in the radial displacement measurement of the magnetic bearing rotor can be solved. In addition, the magnetic suspension bearing which is nonlinear and difficult to establish a rotor displacement calculation model realizes automatic measurement without a displacement sensor, does not need an additional complex circuit and is easy to realize.

It should be noted that the soft measurement method based on the wavelet neural network displacement in the scheme of the present invention can be used in a magnetic suspension bearing system, and is also suitable for other applications of eddy current sensors.

After a large number of tests, the technical scheme of the embodiment is adopted, the model is trained through the collected sample data, and the trained model can be directly connected in series to a magnetic suspension bearing system to form a radial displacement prediction module, so that the radial displacement of the magnetic suspension bearing rotor can be self-detected. Therefore, the radial displacement of the magnetic bearing rotor is measured by using a soft measurement mode, the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the eddy current sensor is used is avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the scheme of the invention can also solve the problem that the volume of the magnetic suspension bearing control system is enlarged due to the fact that the eddy current sensor is mostly adopted when the radial displacement of the magnetic bearing rotor is measured, and the effect that the volume of the magnetic suspension bearing control system is enlarged due to the fact that the eddy current sensor is used for measuring the radial displacement of the magnetic bearing rotor in a soft measuring mode is achieved, and therefore the occupied space of the magnetic suspension bearing control system is saved.

According to an embodiment of the present invention, there is also provided a displacement determining apparatus of a magnetic bearing corresponding to the displacement determining method of a magnetic bearing. Referring to fig. 6, a schematic diagram of an embodiment of the apparatus of the present invention is shown. The magnetic bearing, comprising: a magnetic bearing coil and a magnetic bearing rotor. A displacement determining apparatus of the magnetic bearing, comprising: a modeling unit 102, an acquisition unit 104, and a control unit 106.

The modeling unit 102 is configured to construct a displacement measurement model based on the wavelet neural network, and the displacement measurement model is recorded as a displacement soft measurement model, namely a wavelet neural network displacement prediction model. The displacement soft measurement model is obtained by training a wavelet neural network by taking the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter. The specific function and processing of the modeling unit 102 are referred to in step S110.

In some embodiments, the modeling unit 102, which constructs a displacement measurement model based on a wavelet neural network, referred to as a displacement soft measurement model, includes:

the modeling unit 102 is in particular further configured to collect as sample data set control currents of the magnetic bearing coils and set displacement parameters of the magnetic bearing rotor. Among the set displacement parameters of the magnetic bearing rotor, the displacement parameters include: radial displacement of the magnetic bearing rotor. The specific functions and processes of the modeling unit 102 are also referred to in step S210.

In some embodiments, the number of magnetic bearing coils is more than one.

The modeling unit 102 collects a set control current of the magnetic bearing coil and a set displacement parameter of the magnetic bearing rotor as sample data, and includes:

the modeling unit 102 is further configured to select one or more set control currents of the magnetic bearing coils and set displacement parameters of the magnetic bearing rotor as characteristic variables. The specific functions and processes of the modeling unit 102 are also referred to in step S310.

The modeling unit 102 is further configured to normalize the characteristic variables to form sample data, which is used as sample data for training and testing. The specific functions and processes of the modeling unit 102 are also referred to in step S320.

Fig. 7 is a schematic flow chart of an embodiment of constructing a wavelet neural network displacement soft measurement model. As shown in fig. 7, the bearing displacement soft measuring device includes:

the first step is as follows: and selecting characteristic variables. The characteristic variable of the displacement soft measurement model (namely the displacement measurement model based on the wavelet neural network) mainly comprises a model input variable and a model output variable and is measured by bearing radial displacement s_FX、s_FY、s_RX、s_RYThe suspension force control current i is a model output variable and can be known through model analysis_FX、i_FY、i_RX、i_RYIs correlated to bearing displacement, so it is determined to be the model input variable.

The second step is that: and preprocessing the sampled data. The input and output signals are normalized, and the normalization formula is as follows:

wherein A is_KIs normalized input data. x is the number of_KIs the original input data. x is the number of_max、x_minRespectively, the maximum and minimum values in the raw input data. And filtering the normalized data to obtain more accurate input and output data, and selecting N groups of representative sample data, wherein N is a positive integer. Wherein each set of sample data contains 4 input variables X_i＝[i_FX、i_FY、i_RX、i_RY]The corresponding output vector is Y_k＝[s_FX、s_FY、s_RX、s_RY]Sample data is formed for training and testing the model.

The modeling unit 102 is further configured to train a wavelet neural network to obtain the soft displacement measurement model based on the sample data, with the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter. The specific functions and processes of the modeling unit 102 are also referred to in step S220.

Specifically, the scheme of the invention provides a device for magnetically suspended bearing rotor displacement soft measurement based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, a radial displacement prediction module (namely a displacement soft measurement model) is obtained through training, and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of a magnetically suspended bearing rotor can be self-detected without a mechanical eddy current sensor.

In some embodiments, the modeling unit 102, based on the sample data, trains a wavelet neural network with the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter, and obtains the soft displacement measurement model, including:

the modeling unit 102 is further specifically configured to set a wavelet neural network model. And under the condition that the number of the magnetic bearing coils is more than one, the number of layers of the wavelet neural network model is more than one. The specific functions and processes of the modeling unit 102 are also referred to in step S410.

The modeling unit 102 is further specifically configured to optimize the set parameters of the wavelet neural network model based on a particle swarm optimization algorithm, so as to obtain the optimized wavelet neural network model. The specific function and processing of the modeling unit 102 are also referred to in step S420.

As shown in fig. 7, the bearing displacement soft measurement device further includes:

the third step: and designing a wavelet neural network. Fig. 8 is a schematic structural diagram of an embodiment of a wavelet neural network structure. The designed wavelet neural network model has three layers as shown in fig. 8, and is a multi-input multi-output model. The input layer variable of the network is X_i＝[i_FX、i_FY、i_RX、i_RY]Therefore, the number of input layer nodes is i (i is 1, 2, 3, 4). The output layer variable is Y_k＝[s_FX、s_FY、s_RX、s_RY]Therefore, the number of output layer nodes is k (k is 1, 2, 3, 4). The number of nodes in the hidden layer isAnd (4) respectively. As an example shown in fig. 8, the design uses Morlet mother wavelet function ψ (x) as the activation function of the hidden layer, which is expressed as follows:

the output layer excitation function is a Sigmoid function, and the formula is as follows:

if the ith input sample of the input layer is set to X_iThe k-th output value of the output layer is set to Y_kSetting the connection weight between the input layer node i and the hidden layer node j as omega_ijThe connection weight between the hidden layer node j and the output layer node k is set to be omega_jk. The scaling factor and the translation factor of the jth hidden layer node are respectively set as a_jAnd b_jThen, the mathematical model of the wavelet neural network is:

as can be seen from the structure of the wavelet neural network shown in fig. 8, when the number of nodes in each layer is determined, the connection weight (ω) between each layer is determined_ij,ω_jk) Scale factor (a) with wavelet basis function_j) Translation factor (b)_j) And determining the quality of the performance of the wavelet neural network. Thus, pair (ω)_ij,ω_jk,a_j,b_j) It is important to optimize these parameters.

The fourth step: and optimizing parameters of the wavelet neural network. The main role of Particle Swarm Optimization (PSO) is to find a global optimal solution, which can be described as: assuming the particle swarm searches in an s-dimensional space, there is mEach particle constitutes a population Z, which can be expressed as Z ═ Z (Z)₁，Z₂，Z₃，…，Z_m). The position of the ith particle is represented as Z_i＝(Z_i1，Z_i2，…，Z_is) The velocity of the ith particle is denoted V_i＝(V_i1，V_i2，…，V_is) The position of each particle represents a potential solution to the problem sought. Firstly, an adaptive value is obtained by using an objective function, and then the quality of the current particle performance is obtained according to the adaptive value of the particle. The particles continuously iterate to find new solutions according to the position updating formula, and the individual extreme value is marked as P_idAnd the total extreme value is denoted as P_gd. Each particle updates its own velocity according to equation (1) and its own position according to equation (2).

V_id(t+1)＝ωV_id(t)+η₁r₁(P_id-Z_id(t))+η₂r₂(P_gd-Z_id(t)) (1)。

Z_id(t+1)＝Z_id(t)+V_id(t+1) (2)。

Where ω is an inertial weight coefficient. Eta₁And η₂Is a non-negative learning factor. i is 1, 2, …, m. d is 1, 2, …, s. r1 and r2 are [0, 1 ]]The random number of (2).

Improving the inertia weight coefficient: the inertia weight coefficient omega is closely related to the global and local searching capability of the particle swarm. In general, the larger the value of ω, the stronger the global search ability of the particle group, and the smaller the value of ω, the stronger the local search ability of the particle group. Therefore, the global search capability in the early stage of particle swarm iteration is enhanced through the larger omega, and the local search in the later stage of particle swarm iteration is enhanced through the smaller omega, so that the swarm iteration speed and the iteration precision are improved.

Therefore, in the scheme of the invention, the particle swarm inertial weight coefficient is updated by adopting the formula (3).

Wherein, ω is_minAnd ω_maxRespectively the minimum value and the maximum value of the inertia weight change, t is the current iteration number, t_maxIs the set maximum number of iterations.

Improving the learning factor: due to η₁Determining the individual cognitive level contribution, η, of the population₂Determining the cognitive level contribution rate of the particle population. Therefore, in the early stage of iteration, the particle fitness is larger and can pass through larger eta₂To control the particles to develop in the optimal direction of the population. At the later stage of iteration, the particle fitness is gradually reduced and can pass through larger eta₁Individual cognition of the particles is released until an optimal position is found. In view of this, the scheme of the present invention adopts the following formula (4) and formula (5) to update the learning factor.

Wherein eta is_1min、η_2min、η_1maxAnd η_2maxThe minimum and maximum values of the learning factor are respectively indicated. Through improvement of the inertia weight and the learning factor of the particle swarm algorithm, the iteration times of the algorithm can be reduced, the calculation time is reduced, the algorithm precision can be improved, the problem that the particle swarm falls into local optimization can be effectively avoided through the improved model, and the prediction precision of the particle swarm algorithm is improved.

The particle swarm optimization wavelet neural network is to search parameters (omega) of the wavelet neural network in a search space of each particle through each particle in the network back propagation process_ij,ω_jk,a_j,b_j). The position of the particle in each dimension is the solution required, and the sum of the dimensions (ω) of the particles is established_ij,ω_jk,a_j,b_j) The weight (omega) of the neural network_ij,ω_jk) Scaling factor of number plus wavelet basis function (a)_j) Number andtranslation factor (b)_j) The sum of the numbers is equal to the dimension of the particle. For a three-layer network with a network structure of M-N-N, the number of network weights is M multiplied by N + N multiplied by N, and the number of scale factors and translation factors of the wavelet basis functions is N + N, so that the dimension of the particles is M multiplied by N + N multiplied by N + N + N. The mean square error function E is chosen as a function of the fitness value, as shown in equation (6):

the modeling unit 102 is further configured to train the optimized wavelet neural network model to obtain the soft displacement measurement model, based on the sample data, with the set control current of the magnetic bearing coil as an input parameter and the set displacement parameter of the magnetic bearing rotor as an output parameter. The specific functions and processes of the modeling unit 102 are also referred to in step S430.

FIG. 9 is a flow diagram of an embodiment of a process for an improved particle swarm optimization wavelet neural network. The specific flow of the improved particle swarm optimization wavelet neural network training algorithm is shown in fig. 9. Assigning the value of each dimension of the optimized global optimal position of the particle to (omega)_ij,ω_jk,a_j,b_j) For a three-layer network with a network structure of M-N, the first mxn values are generally assigned to the weights between the input layer and the hidden layer, the values between the mxn +1 th and mxn + N are assigned to the scaling factors of the wavelet basis function, the values between the mxn + N +1 th and mxn +2N are assigned to the shifting factors of the wavelet basis function, and the remaining values are assigned to the weights between the hidden layer and the output layer.

As shown in fig. 7, the bearing displacement soft measurement device further includes:

the fifth step: and establishing an improved particle swarm optimization-based wavelet neural network displacement soft measurement control system. And (3) connecting the wavelet neural network radial displacement prediction module which is optimally trained into a magnetic suspension bearing control system, and realizing the online real-time prediction of the radial displacement of the bearing rotor.

An obtaining unit 104 configured to obtain a current control current of the magnetic bearing coils, such as a current control current of the magnetic bearing coils sampled by a current sensor, in case a current displacement parameter of the magnetic bearing rotor needs to be determined. The specific function and processing of the acquisition unit 104 are referred to in step S120.

A control unit 106 configured to input the soft measurement model of displacement with the current control current of the magnetic bearing coils as an input parameter, and determine an output parameter of the soft measurement model of displacement as a current displacement parameter of the magnetic bearing rotor, enabling determination of the current displacement parameter of the magnetic bearing rotor. That is, the determination of the current displacement parameters of the magnetic bearing rotor is effected in accordance with the current control circuitry of the magnetic bearing coils. Among the current displacement parameters of the magnetic bearing rotor, displacement parameters include: radial displacement of the magnetic bearing rotor. The specific function and processing of the control unit 106 are shown in step S130.

In specific implementation, a corresponding relationship between the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor may be established. And under the condition that the current displacement parameter of the magnetic bearing rotor needs to be determined, obtaining the current control current of the magnetic bearing coil. And determining the set displacement parameter corresponding to the set control current which is the same as the current control current of the magnetic bearing coil in the corresponding relation as the current displacement parameter of the magnetic bearing rotor corresponding to the current control current of the magnetic bearing coil, namely, determining the current displacement parameter of the magnetic bearing rotor according to the current control circuit of the magnetic bearing coil. In the setting displacement parameter of the magnetic bearing rotor and the current displacement parameter of the magnetic bearing rotor, the displacement parameter includes: radial displacement of the magnetic bearing rotor.

Wherein, constructing the corresponding relationship between the set control current of the magnetic bearing coil and the set displacement parameter of the magnetic bearing rotor may include at least one of the following construction conditions:

the first construction case: it may be a numerical relationship or table that establishes a relationship between the set control currents of the magnetic bearing coils and the set displacement parameters of the magnetic bearing rotor.

The second construction case: the method can be used for constructing a displacement measurement model based on the wavelet neural network, and recording the displacement measurement model as a displacement soft measurement model, namely a wavelet neural network displacement prediction model. The displacement soft measurement model is a model which takes the control current of the magnetic bearing coil as an input parameter and takes the displacement parameter of the magnetic bearing rotor as an output parameter.

Thus, the scheme of the invention realizes the sensorless radial displacement self-detection technology by self-detecting the radial displacement of the magnetic suspension bearing rotor, namely, the radial displacement of the magnetic suspension bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, and the problems that the rotor suddenly falls off and the system is damaged by friction with a mechanical part when the mechanical sensor is damaged are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the system cost and the volume are reduced, the stability of the magnetic suspension bearing control system is improved, and the later maintenance is facilitated.

In some embodiments, the controlling unit 106, inputting the displacement soft measurement model with the current control current of the magnetic bearing coils as an input parameter, and determining an output parameter of the displacement soft measurement model as a current displacement parameter of the magnetic bearing rotor, comprises: the control unit 106, in particular, is further configured to arrange the soft measurement model of displacement between a sampling end of the current control current of the magnetic bearing coils and a feedback end of the current displacement parameter of the magnetic bearing rotor, so that the soft measurement model of displacement is capable of outputting the current displacement parameter of the magnetic bearing rotor based on the current control current of the magnetic bearing coils.

Therefore, the scheme of the invention provides a device for soft measurement of the displacement of the magnetic bearing rotor based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, the trained model can be directly connected in series to a magnetic bearing system to form a radial displacement prediction module (namely a displacement soft measurement model), and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of the magnetic bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, the problems that the rotor suddenly falls off when the mechanical sensor is damaged and the system is damaged by friction with a mechanical part are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the problems of large volume and high cost of the magnetic bearing control system are solved, and the problem that the volume of the magnetic bearing control system is increased due to the fact that an eddy current sensor is adopted in the radial displacement measurement of the magnetic bearing rotor can be solved. In addition, the magnetic suspension bearing which is nonlinear and difficult to establish a rotor displacement calculation model realizes automatic measurement without a displacement sensor, does not need an additional complex circuit and is easy to realize.

In some embodiments, further comprising: a process of controlling the current of the magnetic bearing coils based on the determined current displacement parameter of the magnetic bearing rotor, specifically as follows:

the control unit 106 is further configured to, after the determination of the current displacement parameter of the magnetic bearing rotor is achieved, adjust the current control current of the magnetic bearing coils according to the current displacement parameter of the magnetic bearing rotor, resulting in an adjustment value of the current control current of the magnetic bearing coils. The specific functions and processes of the control unit 106 are also referred to in step S510. Specifically, adjusting the current control current of the magnetic bearing coil according to the current displacement parameter of the magnetic bearing rotor comprises: and performing PID control according to the difference value between the current displacement parameter of the magnetic bearing rotor and the reference displacement parameter of the magnetic bearing rotor to obtain a first processing result. And performing PWM processing on the difference value of the first processing result and the current control current of the magnetic bearing coil to obtain a second processing result. And then carrying out power amplification processing on the second processing result to obtain an adjustment value of the current control current of the magnetic bearing coil.

The control unit 106 is further configured to control the operation of the magnetic bearing coils in accordance with the adjusted values of the current control currents of the magnetic bearing coils. The specific function and processing of the control unit 106 are also referred to in step S520.

Fig. 10 is a schematic structural diagram of an embodiment of a displacement measurement circuit based on a wavelet neural network. As shown in fig. 10, the circuit for measuring displacement based on wavelet neural network comprises: for each magnetic bearing coil, such as a radial forward X-direction coil (FX coil), a radial forward Y-direction coil (FY coil), a radial backward X-direction coil (RX coil), and a radial backward Y-direction coil (RY coil), a control branch is provided. The current signal output by each control branch, namely the control current of each magnetic bearing coil, passes through a Wavelet Neural Network (WNN) displacement measurement model (namely a displacement measurement model based on the wavelet neural network), and then outputs the predicted rotor displacement of the magnetic bearing rotor. And predicting the rotor displacement of each path and a given reference position signal of the path, obtaining the difference value of the rotor displacement and the given reference position signal of the path through a first comparator, outputting an adjusting value of the control current of the magnetic bearing coil of the path after passing through a control branch of the path, and controlling the magnetic bearing coil of the path according to the adjusting value of the control current of the magnetic bearing coil of the path. Wherein, every way control branch road includes: the PWM control module, the second comparator, the PWM module and the power amplifier. And the PWM control module is used for outputting a first processing result after PID processing is carried out on the difference value of each path of predicted rotor displacement and the given reference position signal of the path. And the difference value of the first processing result of each path and the predicted rotor displacement of the path is processed by the PWM module to obtain a second processing result. And the second processing result is processed by the power amplifier to obtain an adjusting value of the control current of the magnetic bearing coil.

In the example shown in fig. 10, a displacement measurement model based on a wavelet neural network is applied to a magnetic bearing control system. Wavelet neural network displacement prediction model based on current signal i of power amplifier (such as switching power amplifier)_FX、i_FY、i_RX、i_RYFor the input signal, a predicted rotor displacement s is output_FX、s_FY、s_RX、s_RYAnd compares it with a given reference position signal f_FX、f_FY、f_RX、f_RYAnd comparing, and finally realizing the stable suspension of the magnetic bearing by the difference value through a closed-loop controller and a switching power amplifier.

Therefore, the scheme of the invention provides a device for soft measurement of the displacement of the magnetic bearing rotor based on a particle swarm optimization wavelet neural network model, the wavelet neural network model is trained through collected sample data, the trained model can be directly connected in series to a magnetic bearing system to form a radial displacement prediction module (namely a displacement soft measurement model), and a sensorless radial displacement self-detection technology is realized, namely, the radial displacement of the magnetic bearing rotor can be self-detected, a mechanical eddy current sensor is not needed, the problems that the rotor suddenly falls off when the mechanical sensor is damaged and the system is damaged by friction with a mechanical part are solved, so that the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the electrical eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved. Meanwhile, the problems of large volume and high cost of the magnetic bearing control system are solved, and the problem that the volume of the magnetic bearing control system is increased due to the fact that an eddy current sensor is adopted in the radial displacement measurement of the magnetic bearing rotor can be solved. In addition, the magnetic suspension bearing which is nonlinear and difficult to establish a rotor displacement calculation model realizes automatic measurement without a displacement sensor, does not need an additional complex circuit and is easy to realize.

It should be noted that the displacement soft measurement device based on the wavelet neural network in the scheme of the present invention can be used in magnetic suspension bearing systems, and is also suitable for other applications of eddy current sensors.

Since the processes and functions implemented by the apparatus of this embodiment substantially correspond to the embodiments, principles and examples of the method shown in fig. 1 to 5, the description of this embodiment is not detailed, and reference may be made to the related descriptions in the foregoing embodiments, which are not repeated herein.

Through a large number of tests, the technical scheme of the invention is adopted, the model is trained through the collected sample data, the trained model can be directly connected in series to a magnetic suspension bearing system to form a radial displacement prediction module, the radial displacement of the magnetic suspension bearing rotor is self-detected, the problems that the rotor suddenly falls off and the system is damaged by friction with a mechanical part when a mechanical sensor is damaged are solved, the falling accident of the magnetic bearing rotor caused by the damage of the eddy current sensor when the eddy current sensor is used can be avoided, and the suspension safety of the magnetic bearing rotor can be improved.

There is also provided, in accordance with an embodiment of the present invention, a magnetic bearing system corresponding to a displacement determining apparatus of a magnetic bearing. The magnetic bearing system may include: the displacement determining apparatus of the magnetic bearing described above.

Since the processes and functions of the magnetic suspension bearing system of this embodiment are basically corresponding to the embodiments, principles and examples of the apparatus shown in fig. 6, the description of this embodiment is not given in detail, and reference may be made to the related descriptions in the embodiments, which are not described herein again.

Through a large number of tests, the technical scheme of the invention is adopted, the model is trained through the collected sample data, and the trained model can be directly connected in series to the magnetic suspension bearing system to form a radial displacement prediction module, so that the radial displacement of the magnetic suspension bearing rotor is self-detected, the problems of large volume and high cost of the magnetic suspension bearing control system are solved, and the problem that the volume of the magnetic suspension bearing control system is increased due to the fact that an eddy current sensor is additionally adopted during the measurement of the radial displacement of the magnetic bearing rotor can be avoided.

According to an embodiment of the present invention, there is also provided a storage medium corresponding to a displacement determination method of a magnetic bearing, the storage medium including a stored program, wherein a device on which the storage medium is located is controlled to execute the displacement determination method of a magnetic bearing described above when the program is run.

Since the processing and functions implemented by the storage medium of this embodiment substantially correspond to the embodiments, principles, and examples of the methods shown in fig. 1 to fig. 5, details are not described in the description of this embodiment, and reference may be made to the related descriptions in the foregoing embodiments, which are not described herein again.

Through a large number of tests, the technical scheme of the invention is adopted, the model is trained through the collected sample data, the trained model can be directly connected in series to a magnetic suspension bearing system to form a radial displacement prediction module, the radial displacement of the rotor of the magnetic suspension bearing is self-detected, the automatic measurement without a displacement sensor is realized for the magnetic suspension bearing which is nonlinear and is difficult to establish a rotor displacement calculation model, an additional complex circuit is not needed, and the method is easy to realize.

According to an embodiment of the present invention, there is also provided a processor corresponding to the displacement determination method of the magnetic bearing, the processor being configured to run a program, wherein the program is run to execute the displacement determination method of the magnetic bearing described above.

Since the processing and functions implemented by the processor of this embodiment substantially correspond to the embodiments, principles, and examples of the methods shown in fig. 1 to fig. 5, details are not described in the description of this embodiment, and reference may be made to the related descriptions in the foregoing embodiments, which are not described herein again.

Through a large number of tests, the technical scheme of the invention is adopted, the model is trained through the collected sample data, and the trained model can be directly connected in series to the magnetic suspension bearing system to form a radial displacement prediction module, so that the radial displacement of the rotor of the magnetic suspension bearing can be self-detected, the cost and the volume of the system are reduced, the stability of the magnetic suspension bearing control system is improved, and the later maintenance is facilitated.

In summary, it is readily understood by those skilled in the art that the advantageous modes described above can be freely combined and superimposed without conflict.

The above description is only an example of the present invention, and is not intended to limit the present invention, and it is obvious to those skilled in the art that various modifications and variations can be made in the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the scope of the claims of the present invention.