Parameter setting method of magnetic suspension bearing

Abstract

The application provides a parameter setting method of a magnetic suspension bearing, which comprises the following steps: calculating electromagnetic force and acceleration according to current and position in a predetermined time interval; calculating displacement and speed increment according to the acceleration and instantaneous speed; repeating the above steps according to different current and position in a predetermined update number; and finally outputting the displacement to realize parameter setting. The displacement is further calculated according to the electromagnetic force and the acceleration in the predetermined time interval, and the final displacement output is corrected by continuously updating the current and position in different time periods, so that the technical problem of lacking a displacement output method with small result error in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of parameter setting of magnetic suspension bearing, and particularly to a parameter setting method of magnetic suspension bearing. BACKGROUND

[0002] Since the invention of magnetic suspension technology, it has been widely studied due to its advantages of no contact, no lubrication, and no wear, and has been widely used in various fields. With the development of industry, magnetic suspension bearing has also entered the field of view of people.

[0003] Compared with traditional bearings, the stiffness and damping of magnetic bearings are variable, and the performance is mainly determined by the controller. Therefore, the performance of the controller directly determines the stability and anti-interference of the stator suspension. The most commonly used control strategy for magnetic suspension is PID control. However, the PID control system of the magnetic suspension bearing is a high-order system, which cannot be directly calculated to optimize the control performance. Usually, it needs to be converted into a low-order system or linearized in a small working range for equivalent calculation. Due to system errors, equivalent errors and other factors, there is a large difference between the output performance and the calculated performance, which causes a certain error between the simulation results and the actual situation.

[0004] Therefore, there is an urgent need for a parameter setting method of magnetic suspension bearing to solve the technical problem of lack of a displacement output method with small result error in the prior art. SUMMARY

[0005] In an embodiment, the present application provides a parameter setting method of magnetic suspension bearing, which further calculates displacement according to electromagnetic force and acceleration in a predetermined time interval, and continuously updates the current displacement and position to correct the final displacement output, thereby solving the technical problem of lack of a displacement output method with small result error in the prior art.

[0006] The parameter setting method comprises:

[0007] calculating electromagnetic force and acceleration according to current displacement and position in a predetermined time interval;

[0008] calculating displacement and velocity increment according to the acceleration and the instantaneous velocity;

[0009] repeating the above steps according to different current displacement and position updated for a predetermined number of times;

[0010] Finally, the displacement output is realized for subsequent parameter setting.

[0011] In an embodiment, before the step of calculating electromagnetic force and acceleration according to current displacement and position in a predetermined time interval, the method further comprises:

[0012] Setting initial parameters, wherein the initial parameters include control time and initial measurement current time, and position measurement current time;

[0013] Equally dividing the control time into multiple segments, wherein each segment of the control time is the predetermined time interval and is divided into m segments.

[0014] In an embodiment, after the step of updating different current and position according to the predetermined number of updates is repeated, the method further comprises:

[0015] Determining whether the number of repetitions is m times, if the number of repetitions is m times, updating the control time, if the control time is less than the simulation time, repeating the calculation.

[0016] In an embodiment, after the step of setting the displacement output implementation parameter is completed, the method further comprises:

[0017] Building a neural network prediction model with the displacement output as the input layer and the control parameter as the output layer.

[0018] In an embodiment, the displacement output is divided into a training group and a validation group.

[0019] In an embodiment, after the step of building a neural network prediction model with the displacement output as the input layer and the control parameter as the output layer is completed, the method further comprises:

[0020] If the output result of the validation group meets the predetermined expected value, input the required performance index value.

[0021] In an embodiment, after the step of building a neural network prediction model with the displacement output as the input layer and the control parameter as the output layer is completed, the method further comprises:

[0022] If the output result of the validation group does not meet the predetermined expected value, perform error analysis, and after the control parameter step size is traversed and calculated, the neural network prediction model is built again.

[0023] In an embodiment, the mechanical expression of the acceleration is calculated as:

[0024]

[0025] μ0, vacuum permeability A0, air gap area, N1 coil turns, I1 bias current, ΔI, control current, Δx1 position deviation, m mass.

[0026] In an embodiment, the control parameters include initial position, initial speed, target position value, and simulation time.

[0027] In an embodiment, the displacement and the velocity calculation formula are as follows:

[0028]

[0029] wherein x f the displacement of the first calculation, v i : the velocity of the first calculation, Δt2: the time interval of the simulation calculation; a i : the acceleration of the first calculation. BRIEF DESCRIPTION OF DRAWINGS

[0030] Figure 1 is a flowchart of a parameter setting method of a magnetic suspension bearing in an embodiment of the present application;

[0031] Figure 2 is a flowchart of a PID control parameter setting method of a magnetic suspension bearing in another embodiment of the present application;

[0032] Figure 3 is a flowchart of a simulation model calculation in another embodiment of the present application. DETAILED DESCRIPTION

[0033] In order for those skilled in the art to better understand the technical solutions of the present application, the present application will be described in detail below in combination with the drawings and specific embodiments.

[0034] The various aspects and features of the present application are described herein with reference to the accompanying drawings.

[0035] These and other characteristics of the present application will become more apparent from the following description of the preferred forms given, by way of non-limiting example, with reference to the attached drawings.

[0036] It should also be understood that, although the present application has been described with reference to some specific examples, a person skilled in the art can certainly determine many other equivalent forms of the present application, which have the features as claimed and thus all fall within the protection scope defined thereby.

[0037] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings, in which:

[0038] application is made to the best mode contemplated, and any specific embodiment described herein is intended to be illustrative only and not as limiting of the application as a whole. Various modifications and changes can be made thereto by those skilled in the art which fall within the scope and spirit of the application as described.

[0039] The description herein can use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which can refer to one or more of the same or different embodiments of the application.

[0040] For a better understanding of the technical solution of the present application, the following will describe the present application in detail in connection with the drawings and specific embodiments.

[0041] The various aspects and features of the present application are described herein with reference to the accompanying drawings.

[0042] These and other characteristics of the present application will become apparent from the following description of the preferred forms given, by way of non-limiting example, with reference to the attached drawings.

[0043] It is also to be understood that even though a number of embodiments of the application have been described herein, the application covers all possible combinations and sub-combinations of the various elements and features described herein.

[0044] The above and other aspects, features and advantages of the present application will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings, in which:

[0045] The specific embodiments of the present application are described herein with reference to the accompanying drawings. However, it is to be understood that the embodiments are merely exemplary of the application and are not intended to limit the scope of the application. The application can be embodied in many different forms and should not be construed as limited to the specific embodiments set forth herein. Rather, the specific structural and functional details disclosed herein are merely representative examples of the application and do not limit the scope of the application as stated in the claims.

[0046] The description herein can use the phrases "in one embodiment," "in another embodiment," "in yet another embodiment," or "in other embodiments," which can refer to one or more of the same or different embodiments of the application.

[0047] Noun explanation: system identification is to determine the mathematical model describing the behavior of the system according to the input and output time function of the system.

[0048] The commonly used method for PID control parameter setting is trial and error method, which is more tedious to adjust parameters, a large number of experiments need to be verified, and only good output performance can be obtained. Therefore, some people use online neural network to adjust the control parameters in real time through the control error, which can optimize the parameters in real time through the error, but there are some problems, not only the control chip computing performance is required, the control frequency cannot be obtained, and only the steady-state error value can be optimized in the control process, other performance parameters cannot be optimized, the stiffness and damping cannot be controlled artificially, and the parameter adjustment in the dynamic state has a certain danger, and collision danger is easy to occur.

[0049] In view of the above problems, the model backstepping method can be used for system identification of the magnetic suspension model coefficient, the output response under different parameters is recorded through certain bearing magnetic suspension test. Meanwhile, a model establishing method is proposed, which greatly reduces the error between the model and the actual situation, and the uncertain parameters and experimental control parameters are iterated and output in the error range by using the model, and the model and the experimental test are matched by using the minimum error principle, so that the real model of the magnetic suspension is obtained. And in the stable domain of the control parameter, the different parameters are iterated and output in small steps, the output response performance under different control parameters is obtained, and part of the data is taken as the input layer and the control parameter is taken as the output layer to train the model. The remaining data is used for model verification until the model accuracy reaches the standard. Given the target performance index, the trained model is used for control parameter prediction, so as to solve the problem of difficult parameter adjustment,

[0050] The purpose of the present application is to provide a kind of magnetic suspension bearing PID control parameter setting method based on comprehensive index, the present application is established by proposing a kind of mathematical model using mechanical expression, the system identification of magnetic suspension control can determine unknown or parameters with large measurement error. At the same time, the method combining actual test and model is used for parameter debugging, which not only solves the problem of difficult control parameter adjustment, but also optimizes other parameters under the premise of ensuring system stability.

[0051] Figure 1 It is a flowchart of a parameter setting method of a magnetic suspension bearing in an embodiment of the present application; Figure 2 It is a flowchart of a magnetic suspension bearing PID control parameter setting in another embodiment of the present application; Figure 3 It is a flowchart of a simulation model calculation in another embodiment of the present application.

[0052] As Figures 1 to 3As shown in an embodiment, the present application provides a parameter setting method of a magnetic suspension bearing, which comprises:

[0053] S101, calculating electromagnetic force and acceleration according to current and position in a predetermined time interval.

[0054] In this step, a specific step of calculating electromagnetic force and acceleration according to current and position in a predetermined time interval is provided.

[0055] S102, calculating displacement and speed increment according to acceleration and instantaneous speed.

[0056] In this step, a specific step of calculating displacement and speed increment according to acceleration and instantaneous speed is provided.

[0057] S103, repeating the above steps according to different current and position with a predetermined number of updates.

[0058] In this step, a specific step of repeating the above steps according to different current and position with a predetermined number of updates is provided.

[0059] S104, finally outputting the displacement to realize subsequent parameter setting.

[0060] In this step, a specific step of finally outputting the displacement to realize parameter setting is provided. In addition to the displacement, the speed increment can also be included.

[0061] In this embodiment, a specific implementation of a parameter setting method of a magnetic suspension bearing is provided. First, this method is a displacement simulation data forming method based on simulation calculation, which is used as the basis and important foundation for subsequent control parameter setting. Data generation is performed once every predetermined time interval, and the above process is repeatedly performed to obtain a discrete displacement output curve in the simulation time. According to the displacement output curve, the response performance such as steady-state error, adjustment time and overshoot under the parameter control is calculated. It should be noted that the simulation time is the total time, and the predetermined time interval is the time period after the simulation time is divided multiple times. The number of calculations is the number of times the simulation time is divided by the predetermined time interval. Finally, we can use the displacement and the speed increment as the data source and data basis for subsequent establishment of control parameter setting. The setting range will be described in detail in the following text, and will not be described here in detail, which is helpful to solve the technical problem of lack of a displacement output method with small result error in the prior art.

[0062] In an embodiment, before the step of calculating electromagnetic force and acceleration according to current and position in a predetermined time interval, the method further comprises:

[0063] S201, setting initial parameters, wherein the initial parameters include position and initial speed.

[0064] In this step, an embodiment of setting the initial parameters before simulation is provided. It should be noted that the position and initial speed refer to the offset position and displacement of the magnetic suspension bearing axis and the axis, and the instantaneous speed.

[0065] S202, dividing the control time into multiple segments, wherein each segment of the control time is the predetermined time interval and is divided into m segments.

[0066] In this step, a specific time segmentation step is provided. It should be noted that the simulation time is the total time, the control time is the second layer, and the predetermined time interval is the time segment formed by dividing the control time into m segments again.

[0067] In this embodiment, a specific embodiment of how to construct multiple predetermined time intervals is provided. Here, the division of the control time and the simulation time is involved, for example, the simulation time is 1 second, the control time is 0.001 second, and it is divided into m segments. When m is 10, the control time of each segment is 0.0001 second, so in the simulation time, a total of 10,000 times, i.e. 10,000 times, are calculated.

[0068] In an embodiment, after the above steps of updating different current and position according to the predetermined update times are repeated, the method further comprises:

[0069] determining whether the repetition number is m times, if the repetition number is m times, updating the control time, if the control time is less than the simulation time, repeating the above calculation.

[0070] In this embodiment, a specific embodiment of determining whether the simulation is completed after m times of calculation is provided. For example, when the above displacement has been output m times, one control time should be completed, that is, according to the above example, 0.001 second simulation should be completed. Since 0.001 second is less than the simulation time, i.e. 1 second, m times of calculation are performed again, the control time is accumulated to 0.002 second, and 0.002 second is still less than the simulation time. Therefore, m times of calculation are performed again until the accumulated control time is greater than 1 second, and all displacements are output, indicating that the simulation is completed.

[0071] As shown in Figure 2 In an embodiment, after the final displacement output parameter setting step, the method further comprises:

[0072] A neural network prediction model is constructed with the displacement output as the input layer and the control parameter as the output layer.

[0073] In the embodiment, a specific implementation of constructing a neural network prediction model according to an artificial intelligence neural network is provided. The neural network prediction model is constructed with an output performance index as an input layer and a control parameter as an output layer. The data in step seven is divided into a training group and a verification group, and the training group is used to train the model.

[0074] Before constructing the neural network prediction model, a simulation model is constructed by a traditional displacement current mechanics equation. In the simulation model, in addition to the given measurement parameters, part of the control parameters are involved. The measurement parameters include electromagnetic force and instantaneous speed, which are measurable and controllable parameters and accurate values. However, the control parameters need to be selected, which are existing technologies that can be implemented by those skilled in the art. The control parameter is a parameter range. Thereafter, the displacement obtained under the same external conditions by the magnetic levitation test and the simulation calculation mentioned above is compared. It is found that there is a certain deviation between the displacement of the magnetic levitation test and the displacement of the simulation calculation. This is the significance of system identification. Figure 2 The control parameter is adjusted, and the control parameter is calculated in a loop to make the control parameter of the simulation calculation close to the displacement parameter of the magnetic levitation test. When a predetermined fitting degree is reached, the current control parameter can be selected to participate in the construction of the subsequent neural network prediction model. This process is called setting.

[0075] Specifically, the closed-loop transfer function of the current and displacement is calculated according to the mechanics expression and the control flowchart. The characteristic equation is calculated according to the transfer function, and the stable domain parameter range of the system is calculated by the Routh criterion as a subsequent range reference. A group of control parameters is selected, the magnetic levitation test is performed, the data of the levitation segment is recorded, and the PID discrete control simulation model is established by the mechanics expression. According to the actual control period, the time interval Δt1 is set. In the time interval Δt1, the PID calculation is first performed according to the current position error value, the error integral, and the error differential value to obtain the control current value. That is, the target current calculated by the PID is unchanged in the time interval Δt1. Δt1 is again divided into m parts with equal steps, and each time interval is Δt2. When m is large enough, the stator can be regarded as uniformly accelerated motion in Δt2. The displacement stiffness coefficient and the current stiffness coefficient can be calculated from the current position and the current size, so that the electromagnetic force can be calculated. The rotor acceleration can be calculated from the electromagnetic force. The speed increment and the displacement increment can be calculated according to the acceleration and the current speed, and the current position and speed values are updated. Before the next discrete time calculation, the electromagnetic force is updated by the displacement, and the final displacement and speed in one control period are obtained by iterative calculation, that is, the displacement and speed after Δt1 are calculated.

[0076] And the unknown parameters in the error range of small step size loop calculation, obtain the simulation calculation output response. Error analysis of experimental data and simulation data, through the minimum error system identification, determine the unknown parameters, to improve the fourth step simulation calculation model, the control parameters with equal step size traversal calculation, obtain a large number of control parameters corresponding to the output response performance.

[0077] In an embodiment, the displacement output is divided into a training group and a validation group.

[0078] In this embodiment, the displacement output is divided into two parts, which are used for training the model and verifying the model, respectively.

[0079] In an embodiment, after the step of constructing a neural network prediction model with the displacement output as the input layer and the control parameters as the output layer, the method further comprises:

[0080] If the output result of the validation group meets the predetermined expected value, the required performance index value is input, and the corresponding predicted control parameter value output is obtained.

[0081] In an embodiment, after the step of constructing a neural network prediction model with the displacement output as the input layer and the control parameters as the output layer, the method further comprises:

[0082] If the output result of the validation group does not meet the predetermined expected value, error analysis is performed, and the control parameters are calculated with equal step size, and the neural network prediction model is constructed again.

[0083] In addition, according to the target performance index, the predicted parameters are experimentally verified, and if the performance meets the standard, the predicted parameters are the final control parameters, otherwise, the process is repeated.

[0084] In an embodiment, the mechanical expression of the acceleration is calculated as:

[0085]

[0086] Where, μ0 is the vacuum permeability, A0 is the air gap area, N1 is the number of turns of the coil, I is the set current, ΔI is the control current, Δx is the position deviation, and m is the mass.

[0087] In an embodiment, the control parameters include an initial position, an initial speed, a target position value, and the simulation time.

[0088] In an embodiment, the displacement and the speed calculation formula are as follows:

[0089]

[0090] Where, x jthe displacement of the nth calculation, v i the velocity of the nth calculation, Δt2: the time interval of the simulation calculation i the acceleration of the nth calculation.

[0091] The above embodiments are only exemplary embodiments of the present application, and are not intended to limit the present application. The scope of protection of the present application is defined by the claims. Those skilled in the art can make various modifications or equivalent replacements to the present application within the spirit and protection scope of the present application, and such modifications or equivalent replacements should also be considered to fall within the protection scope of the present application.

Claims

1. A method of parameter setting for a magnetic bearing, characterized by, The parameter setting method comprises: calculating electromagnetic force and acceleration according to current and position in a predetermined time interval, wherein the predetermined time interval is a time period after dividing simulation time for multiple times; calculating displacement and velocity increment according to the acceleration and instantaneous velocity; repeating the above steps according to different current and position with a predetermined update number; finally outputting the displacement to realize subsequent parameter setting.

2. The parameter tuning method of a magnetic bearing according to claim 1, characterized in that, Before the step of calculating electromagnetic force and acceleration according to current and position in a predetermined time interval, the method further comprises: setting initial parameters, wherein the initial parameters comprise position and initial velocity; equally dividing the control time into multiple segments, wherein each segment of the control time is the predetermined time interval and is divided into m segments.

3. The parameter tuning method of a magnetic bearing according to claim 2, wherein After the step of repeating the above steps according to different current and position with a predetermined update number, the method further comprises: judging whether the repetition number is m times, if the repetition number is m times, updating the control time, and if the control time is less than the simulation time, cyclically calculating.

4. The parameter tuning method of a magnetic bearing according to claim 3, wherein After the step of finally outputting the displacement to realize parameter setting, the method further comprises: constructing a neural network prediction model with the displacement output as an input layer and control parameters as an output layer.

5. The parameter tuning method of a magnetic bearing according to claim 4, wherein The displacement output is divided into a training group and a verification group.

6. The parameter tuning method of a magnetic bearing according to claim 5, wherein After the step of constructing a neural network prediction model with the displacement output as an input layer and control parameters as an output layer, the method further comprises: if the output result of the verification group meets a predetermined expected value, inputting a required performance index value.

7. The parameter tuning method of a magnetic bearing according to claim 6, wherein After the step of constructing a neural network prediction model with the displacement output as an input layer and control parameters as an output layer, the method further comprises: if the output result of the verification group does not meet the predetermined expected value, performing error analysis, and after stepwise calculation of the control parameters, constructing a neural network prediction model again.

8. The parameter tuning method of a magnetic bearing according to claim 7, wherein The mechanical expression for calculating the acceleration is: wherein μ0 is vacuum permeability, A0 is air gap area, N is coil turns, I is bias current, ΔI is control current, Δx is position deviation, and m is mass.

9. The parameter tuning method of a magnetic bearing according to claim 8, wherein, The control parameters comprise initial position, initial velocity, target position value, and the simulation time.

10. The parameter tuning method of a magnetic bearing according to claim 9, wherein, The displacement and the velocity calculation formula are as follows: where x j : displacement of the i-th calculation; v i : velocity of the i-th calculation; Δt2: time interval of the simulation calculation; α i : acceleration of the i-th calculation.

Images