SSV model parameter identification method and device based on intelligent optimization algorithm and application

Through the step-by-step identification method of the intelligent optimization algorithm, combined with the least squares method and the PSO algorithm, the problem of low parameter identification accuracy of the existing friction model is solved, the identification accuracy of the friction model is improved, and the control capability of the CNC system is enhanced.

CN119065244APending Publication Date: 2024-12-03HUAZHONG UNIV OF SCI & TECH +1
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202411151922.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-21
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The existing friction model parameter identification methods have low accuracy, resulting in inaccurate friction model and affecting the design, analysis and control of mechanical systems.

Method used

Using a step-by-step identification method based on intelligent optimization algorithm, data is collected through sine-like excitation signals, first identify static friction parameters using the least squares method, determine the value range of dynamic friction parameters, and then use the PSO algorithm to identify the remaining parameters to improve the identification accuracy .

Benefits of technology

It improves the accuracy of friction model parameter identification, enhances the accuracy of the system dynamic model, reflects the dynamic characteristics of the machine tool more realistic, and improves the control accuracy of the CNC system.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119065244A_ABST
    Figure CN119065244A_ABST
Patent Text Reader

Abstract

The invention belongs to the related technical field of the numerical control technology, and discloses an SSV model parameter identification method and device based on an intelligent optimization algorithm and application, and the method comprises the following steps: (1) inputting a sine-like speed excitation signal to a numerical control system, collecting the data of a motor encoder and a current sensor of a feed shaft, speed friction force data are obtained through data processing; (2) intercepting data of a high-speed section in the speed friction force data, and identifying a forward coulomb friction force Fc and a forward viscous friction coefficient B by adopting a least square method; (3) determining the forward maximum static friction force Fs based on the forward coulomb friction force Fc and the forward viscous friction coefficient B, and controlling the value ranges of hysteretic curve shape parameters sigma and n; and (4) intercepting data of a low-speed section in the speed friction force data, and further using an intelligent optimization algorithm to identify the remaining parameters needing to be identified of the SSV model. The dynamic friction parameter identification accuracy of the identification method is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field related to numerical control technology, and more specifically, relates to a SSV model parameter identification method, device and application based on an intelligent optimization algorithm. Background Art

[0002] As an intelligent "industrial mother machine", CNC machine tools are an important foundation for intelligent manufacturing equipment, and high-end CNC machine tools are the core of intelligent manufacturing in the high-end equipment manufacturing industry. High-precision CNC systems have nonlinearity and uncertainty, and the friction link in the system will cause adverse factors such as stick-slip motion, limit cycle oscillation, and tracking error. Friction exists between the contact surfaces that move against each other and is the tangential force generated between the two contact surfaces. Friction is related to many factors, such as relative sliding speed, acceleration, displacement, lubrication conditions, and the condition of the contact surface. The role of friction in mechanical systems is becoming increasingly prominent, and it has become the focus of current research to reasonably solve the constraints of friction links in mechanical systems, especially nonlinear friction links.

[0003] Due to the complexity of friction, it is difficult to obtain an accurate and unique mathematical model from the mechanism. So far, dozens of friction models have been proposed. An accurate friction model not only helps to correctly understand the mechanism of friction and effectively predict friction behavior, but also plays a positive and key role in the design, analysis, control and compensation of mechanical systems containing friction. Therefore, it is extremely important to identify the parameters of the friction model and obtain an accurate friction model. Accurate identification of friction model parameters can improve the accuracy of the system dynamics model, thereby more realistically reflecting the dynamic characteristics of the machine tool feed system.

[0004] The single state variable friction model (SSV) based on Bouc-Wen is more flexible and effective than other friction models in describing pre-slip and friction overshoot. The model has a simple structure, is easy to identify, has high computational efficiency, and is suitable for control and other related application fields. The friction calculation formula predicted by the SSV model is as follows:

[0005]

[0006] Among them, F sc (k) is the friction force that characterizes the Stribeck characteristic, F bw (k) is the friction force that characterizes the friction hysteresis part.

[0007] Due to the existence of differential equations and the absence of analytical solutions, the SSV model can only be identified using a discrete model. The parameters required for identification of the SSV model are shown in Table 1:

[0008] Table 1 Parameters to be identified for the SSV model

[0009]

[0010] Most existing methods for identifying friction model parameters use an overall identification method, which usually involves filtering and denoising the velocity friction data and then using a nonlinear least squares method to identify the parameters. However, such identification results are often inaccurate and the identification accuracy is not high. Summary of the invention

[0011] In view of the above defects or improvement needs of the prior art, the present invention provides a SSV model parameter identification method, device and application based on an intelligent optimization algorithm, which aims to solve the problem of low accuracy of existing parameter identification.

[0012] To achieve the above object, according to one aspect of the present invention, a method for identifying SSV model parameters based on an intelligent optimization algorithm is provided, the method comprising the following steps:

[0013] (1) Inputting a quasi-sinusoidal speed excitation signal into the numerical control system to drive the feed axis to run, and collecting data from the motor encoder and current sensor of the feed axis to obtain speed friction data through data processing;

[0014] (2) extracting the high-speed segment data from the speed friction data, and using the least square method based on the high-speed segment speed friction data to identify the positive Coulomb friction Fc and the positive viscous friction coefficient B;

[0015] (3) Determine the value range of the positive maximum static friction force Fs, the control hysteresis curve shape parameter σ, and the control hysteresis curve shape parameter n based on the identified positive Coulomb friction force Fc and the positive viscous friction coefficient B;

[0016] (4) The low-speed segment data in the speed friction data is intercepted, and then based on the value range of the identified parameters, the intelligent optimization algorithm is used to identify the remaining parameters that need to be identified for the SSV model.

[0017] Furthermore, the high-speed section is a speed section with a speed value greater than 0.01 m / s, i.e., 600 mm / min; the low-speed section is a speed section with a speed value less than 0.004 m / s, i.e., 240 mm / min.

[0018] Furthermore, the PSO algorithm is used to identify the remaining parameters of the SSV model.

[0019] Furthermore, different amplitudes v are designed according to the machine tool working conditions. max , the quasi-sinusoidal velocity excitation signal v(t) of frequency f, the corresponding formula is:

[0020]

[0021] Furthermore, the speed data vact (t) Position data S collected by the motor encoder act (t) is differentiated to obtain the actual friction force data F fact (t) is obtained by converting the current data I(t) collected by the current sensor and removing the inertia moment, and finally obtaining the friction force data corresponding to different speeds at different times.

[0022] Furthermore, the speed data v act (t), angular acceleration a of the motor act And the actual friction data F fact The calculation formulas for (t) are:

[0023]

[0024] F fact (t) = k t I(t)-Ja act (t) (5)

[0025] Where ΔT is the period of data acquisition, k t is the torque constant and J is the load inertia.

[0026] Furthermore, F s The value range is 1.1F c ~2F c , σ ranges from 0 to 1, n represents the smoothness of pre-slip, and its range is similar to that of the viscous damping coefficient, which is 0.1 to 2;

[0027] The friction force F predicted by the SSV friction model fpred The actual friction force F fact The obtained error is used as the loss function of the PSO algorithm, and the corresponding formula is:

[0028] loss=(F fpred -F fact ) 2 (6).

[0029] The present invention also provides an SSV model parameter identification system based on an intelligent optimization algorithm, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing the SSV model parameter identification method based on the intelligent optimization algorithm as described above when executing the computer program.

[0030] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the SSV model parameter identification method based on the intelligent optimization algorithm as described above.

[0031] The present invention also provides an application of the SSV model parameter identification method based on the intelligent optimization algorithm as described above in numerical control system control.

[0032] In general, compared with the prior art, the SSV model parameter identification method, device and application based on the intelligent optimization algorithm provided by the present invention mainly have the following beneficial effects:

[0033] 1. The present invention adopts a quasi-sinusoidal velocity excitation signal with smooth transition, collects data, and smoothly transitions the reverse segment, which better reflects the friction characteristics of the reverse segment and increases the accuracy of the identification method in identifying dynamic friction parameters.

[0034] 2. The present invention adopts a step-by-step identification strategy, first using the least squares method to fit the static friction parameters, and then determining the value range of the dynamic friction model parameters based on the static friction parameters, which also has guiding significance for parameter identification of other mechanical models.

[0035] 3. The parameter range of the present invention is determined by physical laws and does not require manual adjustment.

[0036] 4. The present invention selects the SSV friction model as the model for parameter identification. Compared with the traditional Stribeck friction model, it can reflect the pre-slip characteristics and has higher modeling accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of a SSV model parameter identification method based on an intelligent optimization algorithm provided by the present invention;

[0038] Figure 2 is a schematic diagram of an algorithm flow for identifying the remaining parameters of the SSV model using the PSO algorithm in an embodiment of the present invention;

[0039] Figure 3 is a schematic diagram of a quasi-sinusoidal speed excitation signal used in an embodiment of the present invention;

[0040] Figure 4 It is a speed friction force data diagram obtained by calculating and processing the data collected from the feed axis of a certain model of a small five-axis machine tool in an embodiment of the present invention;

[0041] Figure 5 (a) and (b) are respectively a comparison diagram and an error diagram of the predicted friction force and the actual friction force of a feed axis of a certain model of a small five-axis machine tool in an embodiment of the present invention after using the PSO algorithm to identify the SSV model parameters. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0043] The present invention provides a method for identifying SSV model parameters based on an intelligent optimization algorithm. The method only needs to use a quasi-sinusoidal speed excitation signal and a motor encoder and a current sensor built into a CNC machine tool to collect speed friction data. Using a step-by-step identification method, firstly, the high-speed segment data is intercepted and the model parameters such as Fc and B of the SSV model are identified using the least squares method. On this basis, the low-speed segment data is intercepted, and the remaining model parameters of the SSV model are identified using an intelligent optimization algorithm.

[0044] The identification method mainly comprises the following steps:

[0045] (1) Set quasi-sinusoidal velocity excitation signals with different amplitudes and frequencies.

[0046] The speed excitation signal is different under different machine tool working conditions, and using a quasi-sinusoidal signal instead of a sine signal can improve the applicability of the speed excitation signal to various machine tools.

[0047] (2) A quasi-sinusoidal speed excitation signal is input into the CNC system to drive the feed axis to run, and the data of the motor encoder and current sensor of the feed axis are collected to obtain the speed friction data through data processing.

[0048] Inertia identification is performed first to obtain the inertia moment. After the current data is converted, the inertia moment is removed to obtain the friction data.

[0049] (3) The high-speed segment data in the speed friction data is intercepted, and the positive Coulomb friction Fc and the positive viscous friction coefficient B are identified by using the least square method based on the speed friction data in the high-speed segment.

[0050] The high speed section is determined by V>4Vs. At this time, the Stribeck speed Vs has not been identified yet, so the value can be predicted by experience. In this embodiment, the high speed section is a speed section with a speed value greater than 0.01m / s, that is, 600mm / min.

[0051] (4) Based on the identified positive Coulomb friction force Fc and positive viscous friction coefficient B, the value range of the positive maximum static friction force Fs, the control hysteresis curve shape parameter σ, and the control hysteresis curve shape parameter n are determined.

[0052] The range of model parameters should be different under different machine tool conditions. For machine tool conditions with heavy load and lubrication, the range of model parameters should be larger.

[0053] (5) The low-speed segment data in the speed friction data is intercepted, and then based on the value range of the identified parameters, the intelligent optimization algorithm is used to identify the remaining parameters that need to be identified for the SSV model.

[0054] The low speed section refers to a speed section where the speed value is less than 0.004 m / s, ie, 240 mm / min. The intelligent optimization algorithm used in this embodiment is the PSO algorithm, and other optimization algorithms may also be used.

[0055] The present invention also provides an SSV model parameter identification system based on an intelligent optimization algorithm, the system comprising a memory and a processor, the memory storing a computer program, and the processor executing the SSV model parameter identification method based on the intelligent optimization algorithm as described above when executing the computer program.

[0056] The present invention also provides a computer-readable storage medium, which stores machine-executable instructions. When the machine-executable instructions are called and executed by a processor, the machine-executable instructions prompt the processor to implement the SSV model parameter identification method based on the intelligent optimization algorithm as described above.

[0057] The present invention also provides an application of the SSV model parameter identification method based on the intelligent optimization algorithm as described above in numerical control system control.

[0058] The present invention is further described in detail below with reference to specific embodiments.

[0059] The identification process of the SSV model parameter identification method based on the intelligent optimization algorithm provided by the embodiment of the present invention is as follows: Figure 1 The experimental steps of this embodiment are as follows:

[0060] (1) Warm up the machine for 1 hour.

[0061] (2) Design different amplitudes v according to the machine tool working conditions max , a quasi-sinusoidal velocity excitation signal v(t) with a frequency of f. Formula (2) is the calculation formula, and the signal curve is as follows: Figure 2 shown.

[0062]

[0063] (3) The quasi-sinusoidal speed excitation signal v(t) of the specified frequency and amplitude is input into the motor through the machine tool servo system, the machine tool is operated, and the motor encoder data S is collected. act (t) and current sensor data I(t).

[0064] (4) Perform data processing, speed data v act (t) Position data S collected by the motor encoder act (t) is differentiated to obtain the actual friction force data F fact (t) is obtained by converting the current data I(t) collected by the current sensor and removing the inertia moment, and finally obtaining the friction data corresponding to different speeds at different times. The specific formulas are formulas (3), (4), and (5). The final speed-friction relationship curve is shown in Figure 3 shown.

[0065]

[0066] F fact (t) = k t I(t)-Ja act (t) (5)

[0067] Where ΔT is the period of data acquisition, k t is the moment constant, J is the load inertia, a act is the angular acceleration of the motor.

[0068] (5) Intercept the speed friction data of the high-speed section (speed value greater than 0.01m / s, i.e. 600mm / min). In the high-speed section, the friction model formula can be simplified to F f =F c +Bv, using the least squares method to fit the data, the model parameters such as Fc and B can be identified first.

[0069] (6) According to relevant data and experience, set the adjustment range of the remaining parameters to be identified in the SSV model. s The value range is 1.1F c ~2F c According to formula (1), the value range of σ is 0-1, and n represents the smoothness of pre-slip, and its value range is similar to that of the viscous damping coefficient, generally 0.1-2. The range of other parameters is determined by the size of the parameters in the simulation model.

[0070] (7) Intercept the speed friction data of the low-speed section (speed value less than 0.004m / s, i.e. 240mm / min), determine the value range of the remaining parameters to be identified, and convert the friction force F predicted by the simulated SSV friction model into fpred The actual friction force F fact The obtained error is used as the loss function of the PSO algorithm, as shown in formula (6).

[0071] loss=(F fpred -F fact ) 2(6)

[0072] (8) Set the number of iterations of the PSO algorithm, update the step size and other parameters, and optimize the remaining parameters to be identified. Figure 4 , the specific optimization steps are:

[0073] S1, initialize the coordinates and speed of the particles at any time according to the set range of parameters to be identified.

[0074] S2, determine the initial fitness value of each particle (predict friction) based on the initial state parameters.

[0075] S3, the initial global optimal solution is the location of the particle whose initial fitness (predicted friction) is closest to the actual friction.

[0076] S4, update particle position (parameters) and movement speed.

[0077] S5, determine whether the maximum number of iterations has been reached, if yes, the optimal solution is obtained and the optimization ends; otherwise, go to step S3.

[0078] (9) After the optimization is completed, a comparison chart between the predicted friction force and the actual friction force is drawn, such as Figure 5 As shown, a comparative analysis is performed and the value of the parameter to be identified is observed. If it is very close to the boundary of the parameter adjustment range set in step (6), the adjustment range is expanded and step (8) is repeated for optimization until the result of the parameter to be identified meets expectations.

[0079] Specifically, the implementation method of the PSO particle swarm optimization algorithm in step (8) is as follows: within the adjustment range of the set parameter to be identified, the parameter value is changed, the predicted friction force is calculated, and the parameter value of the next iteration is adjusted according to the error change with the actual friction force, and the process is repeated until the specified number of iterations is reached.

[0080] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for identifying SSV model parameters based on intelligent optimization algorithm, characterized in that: The method comprises the following steps: (1) Inputting a quasi-sinusoidal speed excitation signal into the numerical control system to drive the feed axis to run, and collecting data from the motor encoder and current sensor of the feed axis to obtain speed friction data through data processing; (2) extracting the high-speed segment data from the speed friction data, and using the least square method based on the high-speed segment speed friction data to identify the positive Coulomb friction Fc and the positive viscous friction coefficient B; (3) Determine the value range of the positive maximum static friction force Fs, the control hysteresis curve shape parameter σ, and the control hysteresis curve shape parameter n based on the identified positive Coulomb friction force Fc and the positive viscous friction coefficient B; (4) The low-speed segment data in the speed friction data is intercepted, and then based on the value range of the identified parameters, the intelligent optimization algorithm is used to identify the remaining parameters that need to be identified for the SSV model.

2. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 1, characterized in that: The high-speed section is the speed section with a speed value greater than 0.01m / s, i.e. 600mm / min; the low-speed section is the speed section with a speed value less than 0.004m / s, i.e. 240mm / min.

3. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 1, characterized in that: The PSO algorithm is used to identify the remaining parameters of the SSV model.

4. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 1, characterized in that: Design different amplitudes v according to machine tool working conditions max , the quasi-sinusoidal velocity excitation signal v(t) of frequency f, the corresponding formula is:

5. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 1, characterized in that: Speed ​​data act (t) Position data S collected by the motor encoder act (t) is differentiated to obtain the actual friction force data F fact (t) is obtained by converting the current data I(t) collected by the current sensor and removing the inertia moment, and finally obtaining the friction force data corresponding to different speeds at different times.

6. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 5, characterized in that: Speed ​​data act (t), angular acceleration a of the motor act And the actual friction data F fact The calculation formulas for (t) are: Where ΔT is the period of data acquisition, k t is the torque constant and J is the load inertia.

7. The SSV model parameter identification method based on intelligent optimization algorithm according to claim 1, characterized in that: F s The value range is 1.1F c ~2F c , σ ranges from 0 to 1, n represents the smoothness of pre-slip, and its range is similar to that of the viscous damping coefficient, which is 0.1 to 2; The friction force F predicted by the SSV friction model fpred The actual friction force F fact The obtained error is used as the loss function of the PSO algorithm, and the corresponding formula is: loss=(F fpred -F fact ) 2 (6)。 8. A SSV model parameter identification system based on intelligent optimization algorithm, characterized by: The system includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the SSV model parameter identification method based on the intelligent optimization algorithm described in any one of claims 1 to 7 is executed.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores machine-executable instructions. When the machine-executable instructions are called and executed by the processor, the machine-executable instructions prompt the processor to implement the SSV model parameter identification method based on the intelligent optimization algorithm described in any one of claims 1 to 7.

10. Application of the SSV model parameter identification method based on intelligent optimization algorithm as described in any one of claims 1 to 7 in numerical control system control.

Citation Information

Cited By

  • Robot friction parameter identification method and device, terminal and medium

    CN120921401A

  • A robot friction parameter identification method and device, a terminal and a medium

    CN120921401B