An improved method of active disturbance rejection controller based on electro-hydraulic servo system

By improving the algorithms of tracking differentializer, expansion state observer and feedback control law in ADRC, the problems of insufficient estimation capabilities and difficult parameters in electro-hydraulic servo systems are solved, and the anti-interference control accuracy and stability of the system are improved sex.

CN114594688BActive Publication Date: 2024-12-03HARBIN UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210222549.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-03-07
Publication Date
2024-12-03
Estimated Expiration
2042-03-07

AI Technical Summary

Technical Problem

Traditional self-immune interference controllers (ADRCs) have problems in electro-hydraulic servo systems with insufficient estimation capabilities, difficult parameters to adjust, and poor response to fast time-varying interference, which affects the system's anti-interference control accuracy.

Method used

By improving the tracking differentializer (TD) algorithm, adjusting the parameter values ​​in the expansion state observer (ESO), and modifying the fal function in the feedback control law (NLSEF) through a nonlinear differential combination method, the ADRC control algorithm is optimized.

Benefits of technology

It improves the system's tracking accuracy in a steady-state interference environment, reduces the estimation deviation of the expansion state observer, and enhances the anti-interference control ability for external interference and internal parameter perturbation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN114594688B_ABST
    Figure CN114594688B_ABST
Patent Text Reader

Abstract

The present invention discloses an improved method for an active disturbance rejection controller based on an electro-hydraulic servo system. Firstly, a tracking differentiator (TD) in an active disturbance rejection controller (ADRC) is improved, and the algorithm of the tracking differentiator itself is optimized and improved; then, an extended state observer (ESO) is improved, and the β 01 , β 02 , β 03 and b parameter values ​​to improve the performance of ESO; finally, the feedback control law (NLSEF) is improved, and the fal function is transformed by the nonlinear difference combination method. By improving the anti-disturbance controller, the anti-disturbance control accuracy of the electro-hydraulic servo system to external disturbances and internal parameter perturbations is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the field of servo technology control, and in particular relates to an improved method for an auto-disturbance rejection controller based on an electro-hydraulic servo system. Background Art

[0002] The electro-hydraulic servo system has many complex disturbance factors such as model uncertainty, parameter perturbation, and variable load force, which have a great impact on the anti-disturbance accuracy of the electro-hydraulic position servo system, and will also greatly affect the tracking and control accuracy of the system.

[0003] The active disturbance rejection controller (ADRC) is mainly composed of three parts: tracking differentiator (TD), extended state observer (ESO), and feedback control law (NLSEF). This controller integrates the essence of modern control theory and classical PID theory. By real-time observation of the various order state variables of the electro-hydraulic position servo system, it estimates and compensates for the unknown disturbances in the system in real time. The active disturbance rejection controller can effectively suppress the unknown disturbances in the system, such as the nonlinear and coupled disturbances of the internal and external environment of the electro-hydraulic system. For traditional PID control, the adverse consequences of these interference sources can only be effectively suppressed by establishing an accurate mathematical model, which poses a huge challenge to controllers. However, the feature that ADRC does not rely on the specific mathematical model of the control object can improve the stability and robustness of the system and improve the anti-interference control accuracy of the control system.

[0004] However, the traditional ADRC control algorithm also has certain defects, such as: generally only requiring ESO to be asymptotically stable, there is room for improvement in convergence speed, insufficient estimation ability for fast time-varying disturbances, limited accuracy of disturbance estimation, many parameters to be selected and difficult to adjust, etc. In response to the problems of the traditional ADRC method, the mainstream improvement method is to linearize and parameterize the nonlinear form of ESO, and adjust the parameters to facilitate practical application; or introduce the theory of variable structure control into the design of ESO in the anti-disturbance algorithm, while ensuring the advantages of the original controller and reducing the adjustable parameters. Summary of the invention

[0005] In view of this, in order to solve the above problems, the present invention provides an improved method for an active disturbance rejection controller based on an electro-hydraulic servo system. The present invention optimizes and improves the tracking differentiator (TD) algorithm in the active disturbance rejection controller (ADRC); 01 , β 02 , β 03 and b parameter values ​​to improve the performance of ESO; through the method of nonlinear difference combination, the fal function is transformed to improve the feedback control law (NLSEF) and the improvement of the ADRC control algorithm is completed, which specifically includes the following steps:

[0006] Step 1: Improve the tracking differentiator (TD) in the active disturbance rejection controller (ADRC) and optimize the algorithm of the tracking differentiator itself;

[0007] Step 2, improvement of the extended state observer (ESO), by adjusting the parameter values ​​of β01, β02, β03 and b to improve the performance of ESO;

[0008] Step 3, improvement of the feedback control law (NLSEF), by transforming the fal function through the method of nonlinear difference combination.

[0009] Furthermore, in step 1, the tracking differentiator (TD) inputs a given signal and arranges the transition process:

[0010]

[0011] Furthermore, a smooth and transitional input signal v1 and its differential signal v2 are arranged according to the given signal v. The second-order continuous type TD is as follows:

[0012]

[0013] Among them, r and δ are parameters to be adjusted. Parameter r is the convergence factor, which determines the speed of the transition process tracking the input signal. The larger the r value, the faster the tracking speed. However, when r is large to a certain extent, the tracking speed will reach saturation and no longer increase. If r is too large, the extracted tracking differential signal will oscillate. δ is the linear interval factor, which determines the size of the linear interval. The value range is generally between (0.01 and 1).

[0014] The improved transition process algorithm is as follows:

[0015]

[0016] Among them, h is the sampling step, y is the output of the system, and e is the deviation between the two.

[0017] Furthermore, the fhan function is a function of nonlinear combination. In TD, the nonlinear combination of fhan functions is optimized.

[0018] Furthermore, in step 2, the system output y and input u are used to track and estimate the system state and disturbance:

[0019]

[0020] Among them, β 01 , β 02 , β 03 is a set of parameters. 01 , β 02 , β03 are three important parameters of the linear extended state observer. 01 , β 01 The size of β is roughly the same order of magnitude as 1 / h. 02 , β 03 It needs to be within a certain range of values.

[0021] Furthermore, the improved ESO algorithm is as follows:

[0022]

[0023] Among them, Z2 is the real second-order differential signal output by the system, Z3 is the disturbance state quantity of ESO expansion, which is the core variable of ESO. Parameter b0 is the "compensation factor" that determines the size of disturbance compensation, that is, the estimated value of the amplification coefficient b when the control quantity u acts on the system.

[0024] Furthermore, in step 3, NLSEF improves the fal function in u0=β1fal(e1,a1,δ)+β2fal(e2,a2,δ) by a nonlinear difference combination method, and the second-order nonlinear state error feedback control law formula is as follows:

[0025]

[0026] Among them, a1, a2, b1, b2, b3, δ are the parameter variables to be adjusted. b1 is similar to the proportional gain coefficient of the PID controller. Increasing b1 will increase the system running speed and reduce the static error, thereby improving the tracking accuracy. However, too large b1 will reduce the dynamic performance of the system and even cause oscillation instability. b2 is similar to the differential gain coefficient in PID. Increasing it within a certain degree will improve the tracking accuracy of the system and speed up the dynamic response speed of the system.

[0027] Furthermore, considering that the electro-hydraulic position servo system should achieve the goal of "large error and small gain, small error and large gain", the range of error gain coefficients a1 and a2 can be set as:

[0028] 0 <a1<1,1<a2

[0029] In order to avoid high-frequency vibration during numerical simulation, the fal function is modified to become a power function with continuous linear segments near the origin. The improved function fal is in the form of:

[0030]

[0031] Where: δ is the interval length of the linear segment.

[0032] In general, the above technical solutions conceived by the present invention can achieve the following beneficial results compared with the prior art:

[0033] 1. The present invention improves the tracking accuracy of the fhan function in TD under steady-state interference environment by optimizing the tracking differentiator algorithm itself.

[0034] 2. The present invention improves the performance of the ESO by adjusting the parameter values ​​of β01, β02, β03 and b, thereby reducing the estimation deviation of the ESO and improving the ability of the ESO to eliminate disturbances.

[0035] 3. The present invention improves the anti-interference control accuracy of the electro-hydraulic position servo system against external interference and internal parameter perturbations. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 The structural block diagram of the improved ADRC of the present invention;

[0037] Figure 2 It is a nonlinear combination optimization graph of fhan function in TD in the embodiment;

[0038] Figure 3 It is a simulation diagram of extracting the sinusoidal position signal of the fhan function in TD in the embodiment;

[0039] Figure 4 It is a simulation diagram of extracting the step position signal of the fhan function in TD in the embodiment;

[0040] Figure 5 It is a sinusoidal signal tracking curve when there is no noise in the embodiment;

[0041] Figure 6 is a step signal tracking curve when there is no noise in the embodiment;

[0042] Figure 7 It is a sine signal tracking curve under white noise in the embodiment;

[0043] Figure 8 is a step signal tracking curve under white noise in the embodiment;

[0044] Fig. 9 It is a sine signal tracking curve under square wave interference in the embodiment;

[0045] Fig.10 It is a step signal tracking curve under square wave interference in the embodiment. DETAILED DESCRIPTION

[0046] The following will be combined with the accompanying drawings in the present invention to clearly and completely describe the purpose, technical solutions and advantages of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] The purpose of the present invention is to provide an improved method for an active disturbance rejection controller based on an electro-hydraulic servo system. The present invention optimizes and improves the tracking differentiator (TD) algorithm in an active disturbance rejection controller (ADRC); 01 , β 02 , β 03 and b parameter values ​​to improve the performance of ESO; through the method of nonlinear difference combination, the fal function is transformed to improve the feedback control law (NLSEF), complete the improvement of ADRC, and improve the anti-interference control accuracy of the electro-hydraulic position servo system to external interference and internal parameter perturbations. The specific steps are as follows:

[0048] Step 1: Improve the tracking differentiator (TD) in the active disturbance rejection controller (ADRC) and optimize the algorithm of the tracking differentiator itself.

[0049] According to the given signal v, a smooth and transitional input signal v1 and its differential signal v2 are arranged. The second-order continuous type TD is as follows:

[0050]

[0051] In the formula: r and δ are parameters to be adjusted. Parameter r is the convergence factor, which determines the speed of the transition process tracking the input signal. The larger the r value, the faster the tracking speed. However, when r is large to a certain extent, the tracking speed will reach saturation and no longer increase. If r is too large, the extracted tracking differential signal will oscillate. δ is the linear interval factor, which determines the size of the linear interval. The value range is generally between (0.01 and 1).

[0052] The improved transition process algorithm is as follows:

[0053]

[0054] As shown in the above formula, h is the sampling step, y is the output of the system, and e is the deviation between the two. The fhan function is a nonlinear combination function. The above is the most basic algorithm for the transition process, which plays a role in extracting differential signals and arranging the transition process for the control system. In TD, the fhan function plays a vital role in improving tracking accuracy. Therefore, the algorithm itself is optimized to obtain a better transition effect. The nonlinear combination optimization diagram is shown in the figure below. Figure 2 shown.

[0055] After improving the fhan function and optimizing the nonlinear combination, in order to verify the changes of the first-order differential signal extracted by the fhan function before and after the improvement, the sine signal and the step signal are used as the excitation signal to perform simulation comparisons. The results are as follows Figure 3 and Figure 4 shown.

[0056] Depend on Figure 3 It can be seen that when a sinusoidal excitation signal is given to the system, the blue line is the target curve, the red line is the tracking and extraction signal of the fhan function before improvement, and the black line is the tracking and extraction signal of the fhan function after improvement. The amplitude of the position signal of the original fhan function is attenuated to 0.05mm, the phase lag is about 4%, and the tracking error accuracy is about 0.02mm. The amplitude of the position signal of the improved fhan function is attenuated to 0.04mm, the phase lag is about 3%, and the tracking error accuracy is about 0.01mm. This shows that the fhan function after the improved design has been greatly improved compared with the one before improvement.

[0057] Give the system a step excitation signal, Figure 4 It can be seen that before the improvement, the overshoot of the tracking curve is about 0.01mm, the response time is about 0.03s, the time to reach the steady state is about 0.5s, and the steady-state error is about 0.005mm; after the improvement, the overshoot of the tracking curve is about 0.01mm, the response time is about 0.01s, the time to reach the steady state is about 0.25s, and the steady-state error is about 0.005mm. Compared with before the improvement, the response time is about 0.02s faster, and the time to reach the steady state is about 0.25s faster, indicating that the fhan function has better tracking accuracy for the system under the steady-state interference environment than before the improvement.

[0058] Step 2, the improvement of the extended state observer (ESO), improves the performance of the ESO by adjusting the parameter values ​​of β01, β02, β03 and b.

[0059] ADRC extracts the dynamic information of the system through ESO, modifies the system characteristics according to the dynamic information, offsets the disturbance, and ensures the structure and physical meaning of the system.

[0060] β 01 , β 02 , β 03 are three important parameters of the linear extended state observer, which play a vital role in steady-state regulation. 01 It can improve the observation speed of the system, but too large a value will cause violent oscillation. 01 The size of is roughly the same order of magnitude as 1 / h. 02 If the value is too small, the system response time will be delayed and a large degree of oscillation will occur. If it is too large, although the speed can be significantly improved, the control quality will be deteriorated. 02It needs to be within a certain range. 03 If β is too small, the observation speed of the system will be too small. Too small observation speed will cause the system to have phase lag and even fail to track the state variables in real time. 03 If β is too large, the observation speed of the system will be too large, which will cause violent oscillation. 02 , β 03 It needs to be within a certain range of values. After considering the characteristics of various system parameters, the improved ESO algorithm is as follows:

[0061]

[0062] Z2 is the real second-order differential signal output by the system, and Z3 is the disturbance state quantity of ESO expansion, which is the core variable of ESO. Parameter b0 is the "compensation factor" that determines the size of disturbance compensation, that is, the estimated value of the amplification factor b when the control quantity u acts on the system. By adjusting β 01 , β 02 , β 03 The performance of ESO can be improved by adjusting the parameter values ​​of α and b, reducing the estimation bias of ESO and improving the ability of ESO to eliminate disturbances.

[0063] Step 3, improvement of the feedback control law (NLSEF), by transforming the fal function through the method of nonlinear difference combination.

[0064] The nonlinear state error feedback control law (NLSEF) is a process of combining the first two parts of ADRC with nonlinear difference.

[0065] The input signal v can be extracted into differential states v1, v2, ..., v after the transition process. n After the input and output of the system pass through ESO, the state variables of each order of the controlled object are obtained: z1, z2, ..., z n The errors formed by these two sets of variables before entering NLSEF are:

[0066]

[0067] Therefore, NLSEF is processed by a nonlinear difference combination method. Several common nonlinear functions are as follows:

[0068]

[0069] Using the first nonlinear function and improving the FAL function, the second-order nonlinear state error feedback control law formula is as follows:

[0070]

[0071] Among them, a1, a2, b1, b2, b3, δ are parameter variables to be adjusted. b1 is similar to the proportional gain coefficient of the PID controller. Increasing b1 will increase the system speed and reduce the static error, thereby improving the tracking accuracy. However, too large b1 will reduce the dynamic performance of the system and even cause oscillation instability. b2 is similar to the differential gain coefficient in PID. Increasing it to a certain extent will improve the tracking accuracy of the system and speed up the dynamic response speed of the system. a1 and a2 are error gain coefficients. Considering that the electro-hydraulic position servo system must achieve the goal of "large error and small gain, small error and large gain", the range of a1 and a2 can be

[0072] 0 <a1<1,1<a2

[0073] In order to avoid high-frequency vibration during numerical simulation, the fal function is modified to become a power function with continuous linear segments near the origin. The improved function fal is in the form of:

[0074]

[0075] Where: δ is the interval length of the linear segment.

[0076] Step 4: After adjusting the parameters, perform simulation analysis to verify the anti-interference effect of the improved ADRC control algorithm.

[0077] After adjusting the parameters, the simulation results without disturbance are shown in the figure below: Figure 5 and Figure 6 shown.

[0078] like Figure 5 It can be seen that in the absence of disturbance, the amplitude attenuation of the original ADRC sinusoidal curve is about 1mm, the phase lag is 6%, while the amplitude attenuation of the improved active disturbance rejection controller sinusoidal curve is about 0.4mm, and the phase lag is 4%. It can be seen that the tracking performance of the improved active disturbance rejection controller is better than that of the ADRC controller.

[0079] Depend on Figure 6It can be seen that the overshoot of the original ADRC step curve is about 6mm, the response time is about 0.02s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.6mm; the overshoot of the improved ADRC step curve is about 4mm, the response time is about 0.015s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.4mm. At the same time, the original ADRC has a parameter perturbation problem due to the large number of internal parameters of the controller. At the beginning, there will be a certain perturbation deviation, and the improved ADRC will have a certain inhibitory effect on this perturbation deviation through the improved algorithm designed in Section 3.2. It can be seen from the figure that the perturbation deviation of the original ADRC is 0.4mm, and the perturbation deviation of the improved ADRC is only 0.2mm. This shows that in the absence of disturbance, the stability of the designed improved ADRC is stronger than that of the original ADRC.

[0080] A disturbance source signal of the electro-hydraulic position servo system is simulated with white noise. A white noise with a noise power of 0.1 is given. Under the interference of white noise, a 1 rad / s sinusoidal excitation signal and a step excitation signal with an amplitude of 20 mm are also given. The simulation results are as follows: Figure 7 and Figure 8 shown.

[0081] Depend on Figure 7 It can be seen that under the influence of white noise, the sinusoidal following curve of the original ADRC has a certain jitter, the sinusoidal following curve of the improved anti-disturbance controller has only a slight jitter, the amplitude attenuation is about 0.4mm, the phase lag is 4%, and the amplitude and phase have not changed significantly. The jitter wave of the following curve of the original ADRC is larger, the amplitude attenuation is about 1.2mm, and the phase lag is 6.5%. Compared with the noise-free state, the amplitude has increased by 0.2mm and the phase lag has increased by 0.5%.

[0082] Depend on Figure 8It can be seen that in the white noise environment, the overshoot of the original ADRC step curve is about 6mm, the response time is about 0.02s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.6mm; the overshoot of the improved ADRC step curve is about 4mm, the response time is about 0.015s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.4mm. Its basic step performance indicators are roughly similar to those in the noise-free state. However, the ADRC parameter perturbation problem is further strengthened. The maximum perturbation deviation of the original ADRC is about 0.8mm at this time, and the maximum perturbation deviation of the improved ADRC is about 0.7mm at this time. It can be calculated that the average perturbation deviation of the original ADRC is about 0.4mm, and the average perturbation deviation of the improved ADRC is about 0.16mm. The perturbation deviation of the improved ADRC under white noise interference is significantly lower than that of ADRC. This shows that the anti-disturbance effect of the improved ADRC is stronger than that of ADRC under white noise.

[0083] Because the interference source of the electro-hydraulic position servo is unknown, in order to further verify its anti-interference control effect, in addition to white noise, the present invention also uses a square wave interference signal to simulate an interference source signal of the electro-hydraulic position servo system. Similarly, a 1rad / s sinusoidal excitation signal and a step excitation signal with an amplitude of 20mm are given. Under the square wave interference signal with an amplitude of 1, the simulation results are as follows: Fig. 9 and Fig.10 shown.

[0084] Depend on Fig. 9 It can be seen that under the interference of square wave signal, the amplitude attenuation of the original ADRC sinusoidal curve is about 1mm, the phase lag is 6%, while the amplitude attenuation of the improved sinusoidal curve is about 0.4mm, and the phase lag is 4%. This is not much different from the noise-free state, but under the condition of square wave interference, the system will have a following deviation at the beginning. From the figure, it can be seen that the maximum following deviation of the improved ADRC is about 0.8mm, and the maximum following deviation of the original ADRC is about 1.8mm.

[0085] Depend on Fig.10It can be seen that the overshoot of the original ADRC step curve is about 6mm, the response time is about 0.02s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.6mm; the overshoot of the improved ADRC step curve is about 4mm, the response time is about 0.015s, the time to reach steady state is about 0.5s, and the steady-state error is about 0.4mm. However, under the action of the square wave interference source with an amplitude of 1, there are two large jitter deviations in the system. After local amplification, it can be seen that the two jitter deviations are mainly concentrated in the two time periods of 0-0.2s and 0.5-0.65s. In the time period of 0-0.2s, the maximum jitter deviation of the improved ADRC is about 0.15mm, and the maximum jitter deviation of the original ADRC is about 0.2mm. After calculation, it can be seen that the average jitter deviation of the improved ADRC is about 0.08mm, and the average jitter deviation of the original ADRC is about 0.12mm. In the time period of 0.5 to 0.65s, the maximum jitter deviation of the improved ADRC is about 0.12mm, and the maximum jitter deviation of the original ADRC is about 0.2mm. After calculation, it can be seen that the average jitter deviation of the improved ADRC is about 0.04mm, and the average jitter deviation of the original ADRC is about 0.1mm. This shows that under the action of square wave interference, the anti-disturbance effect of the improved ADRC is stronger than that of the original ADRC.

[0086] The above implementation modes are only used to illustrate the present invention, but not to limit the present invention. Ordinary technicians in the relevant technical field may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, all equivalent technical solutions also belong to the scope of the present invention. The protection scope of the present invention should be defined by the claims.

Claims

1. An improved method for an active disturbance rejection controller based on an electro-hydraulic servo system, characterized in that: The following steps are involved: Step 1: Improve the tracking differentiator (TD) in the active disturbance rejection control (ADRC) and optimize the algorithm of the tracking differentiator itself. The improved algorithm is as follows: In the formula, h is the sampling step, y is the output of the system, and e is the deviation between the two; Step 2, the improvement of the extended state observer (ESO), by adjusting β 01 , β 02 , β 03 The performance of ESO is improved by adjusting the parameter values ​​of and b. The algorithm of the improved ESO is as follows: In the formula, Z2 is the real second-order differential signal output by the system, Z3 is the disturbance state quantity of ESO expansion, which is the core variable of ESO, and parameter b0 is the "compensation factor" that determines the size of disturbance compensation, that is, the estimated value of the amplification coefficient b when the control quantity u acts on the system; Step 3, improvement of the feedback control law (NLSEF), through the method of nonlinear difference combination, the fal function in u0 = β1fal (e1, a1, δ) + β2fal (e2, a2, δ) is improved, and the formula of the second-order nonlinear state error feedback control law is as follows: In the formula, a1, a2, b1, b2, b3, δ are parameter variables to be adjusted; b1 is similar to the proportional gain coefficient of the PID controller. Increasing b1 will increase the system running speed and reduce the static error, thereby improving the tracking accuracy. However, too large b1 will reduce the dynamic performance of the system and even cause oscillation instability; b2 is similar to the differential gain coefficient in PID. Increasing it to a certain extent will improve the tracking accuracy of the system and speed up the dynamic response speed of the system; a1 and a2 are error gain coefficients. Considering that the electro-hydraulic position servo system should achieve the goal of "large error and small gain, small error and large gain", the range of a1 and a2 is set to 0. <a1<1,1<a2。 2. The improved method of the active disturbance rejection controller based on the electro-hydraulic servo system according to claim 1 is characterized in that: In step 1, the fhan function is a nonlinear combination function, and the fhan function is optimized by nonlinear combination in TD in order to obtain a better transition effect.

3. The improved method of the active disturbance rejection controller based on the electro-hydraulic servo system according to claim 1 is characterized in that: In numerical simulation, in order to avoid high-frequency chattering, the fal function is transformed into a power function with continuous linear segments near the origin. The improved function fal is in the form of: Where δ is the interval length of the linear segment.