Method and device for controlling active disturbance rejection controller

Abstract

The invention discloses an active disturbance rejection controller control method and device, and the method comprises the steps: receiving an actual output feedback signal of a controlled object and a historical output control signal of a controller based on a linear expansion state observer, calculating the total disturbance of a system in real time through a three-order state-space equation, and obtaining a disturbance estimation value; receiving the disturbance estimation value and the control input gain coefficient based on a compensator, calculating a dynamic compensation amount for offsetting the total disturbance of the system, and outputting the dynamic compensation amount to a signal synthesis node; a deviation signal of a set value signal used for defining the target state of a controlled object and a feedback signal actually output by the controlled object is received based on a controller, and an anti-interference basic control signal is generated in combination with the bandwidth of the controller; and subtracting the dynamic compensation amount from the anti-interference basic control signal based on the signal synthesis node to generate an active-disturbance-rejection control amount, outputting the active-disturbance-rejection control amount to a controlled object, and suppressing external interference and internal parameter drift of the controlled object.

Description

Technical Field

[0001] This invention relates to the field of controller anti-interference technology, and in particular to a self-disruption controller control method and device. Background Technology

[0002] In the field of industrial automation control, traditional PID control algorithms have long held a dominant position. This algorithm achieves stable system control through a combination of proportional, integral, and derivative components, boasting a simple structure and mature parameter tuning experience. However, with the increasing complexity of modern industrial systems, controlled objects often face challenges such as time-varying parameters, strong external disturbances (e.g., electromagnetic noise, mechanical vibration), and unmodeled dynamics. Traditional PID control exhibits significant limitations in dealing with such disturbances: its feedback correction mechanism suffers from lag, making rapid disturbance suppression difficult; the integral component is prone to overshoot and even system oscillations; and its parameter robustness is insufficient, leading to a sharp deterioration in control quality when the system model mismatches.

[0003] To enhance anti-interference capabilities, Active Disturbance Rejection Control (ADRC) technology has emerged. This technology treats all disturbances, both internal and external, as a "total disturbance" by extending the state observer and generating dynamic compensation quantities that are fed forward to the control loop, theoretically enabling real-time disturbance cancellation. While existing Linear Active Disturbance Rejection Control (LADRC) simplifies parameter tuning through linearization design, it still has significant drawbacks in practical industrial scenarios: First, the coordination mechanism between observer and controller bandwidth is not standardized, and parameter adjustment relies on manual trial and error, resulting in excessively long debugging cycles; second, poor adaptability between high- and low-order systems, with low-order LADRC controlling high-order systems leading to stability risks due to gain parameter mismatch; and third, insufficient coupling between the disturbance compensation loop and the basic control loop, making it difficult to balance dynamic response and anti-interference capabilities.

[0004] Current industrial applications are placing increasingly stringent demands on control precision and robustness. For example, frequency regulation control in power systems needs to respond to load fluctuations within milliseconds while suppressing grid harmonic interference; motor servo systems need to maintain nanometer-level positioning accuracy under parameter drift. Existing control solutions cannot simultaneously address response speed, disturbance rejection strength, and parameter self-adaptation capabilities, becoming a bottleneck restricting the performance of high-end equipment. Summary of the Invention

[0005] The purpose of this invention is to provide an active disturbance rejection controller (ADRC) method and device. By using a linear extended state observer (LESO) based on a third-order state-space equation to accurately estimate the total system disturbance in real time, and by combining the decoupling mechanism between the dynamic compensation quantity and the basic control signal for disturbance rejection, the problem of traditional LADRC parameter tuning relying on experience, poor adaptability to high and low order systems, and imbalance between dynamic response and disturbance rejection capability is effectively solved.

[0006] To address the aforementioned technical problems, a first aspect of this invention provides a method for controlling an active disturbance rejection controller. The active disturbance rejection controller includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control method includes the following steps:

[0007] Based on the linear extended state observer receiving the actual output feedback signal of the controlled object and the historical output control signal of the controller, the total disturbance of the system is calculated in real time through the third-order state space equation to obtain the disturbance estimate;

[0008] Based on the disturbance estimate and control input gain coefficient received by the compensator, a dynamic compensation amount is calculated to offset the total disturbance of the system, and the dynamic compensation amount is output to the signal synthesis node;

[0009] Based on the deviation signal between the setpoint signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, an anti-disturbance basic control signal is generated by combining the controller bandwidth.

[0010] Based on the signal synthesis node, the anti-disturbance basic control signal is subtracted from the dynamic compensation amount to generate an active anti-disturbance control quantity, which is then output to the controlled object to suppress external interference and internal parameter drift of the controlled object.

[0011] Furthermore, the active disturbance rejection controller control method further includes:

[0012] The observer bandwidth of the linearly extended state observer and the controller bandwidth of the controller are dynamically coordinated and adjusted, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, with overshoot / oscillation as the backoff trigger condition.

[0013] Furthermore, the dynamic coordinated adjustment of the observer bandwidth of the linearly extended state observer and the controller bandwidth of the controller, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, includes:

[0014] Initialize the controller bandwidth to a preset fixed value, and gradually increase the observer bandwidth with a preset first adjustment step size until the system output amplitude exceeds the preset oscillation threshold.

[0015] The current observer bandwidth is fixed, and the controller bandwidth is gradually increased with a preset second adjustment step size;

[0016] If the overshoot of the actual output feedback signal of the controlled object exceeds the preset overshoot threshold or the amplitude exceeds the oscillation threshold, the observer bandwidth will be reduced by a preset backoff ratio, and then the controller bandwidth will be increased.

[0017] Repeat the above adjustment process so that the controlled object simultaneously satisfies that the step response overshoot is less than the preset overshoot threshold and the adjustment time is less than the preset time threshold.

[0018] When the observer bandwidth reaches a preset multiple of the controller bandwidth, the parameter adjustment process is terminated and the corresponding controller bandwidth and observer bandwidth are locked.

[0019] Furthermore, the preset multiple is 5 times.

[0020] Furthermore, the linearly extended state observer receives the actual output feedback signal of the controlled object and the historical output control signal of the controller, and calculates the total system disturbance in real time through the third-order state-space equation to obtain the disturbance estimate, including:

[0021] The third-order state-space equation of the linear extended state observer is constructed, and its state variables include: output tracking state variable that tracks the actual output of the controlled object, differential state variable that characterizes the rate of change of the actual output of the controlled object, and extended state variable that characterizes the total disturbance of the system.

[0022] Calculate the real-time deviation between the actual output feedback signal and the output tracking state variable, correct the deviation using the observer error gain coefficient, and generate the first state update quantity;

[0023] The historical output control signal is coupled to the extended state variable through the control input gain coefficient to generate a second state update quantity;

[0024] Based on the real-time rate of change of the differential state variable, the estimation error is compensated by the differential gain coefficient of the observer to generate the third state update quantity.

[0025] By integrating the first state update, the second state update, and the third state update, the state variables of the third-order state space equation are updated in real time, and the values ​​of the expanded state variables are output as disturbance estimates.

[0026] Further, the method of receiving the disturbance estimate and compensation gain parameter based on the compensator, and calculating the dynamic compensation amount to offset the total system disturbance, includes:

[0027] Receive the disturbance estimate and the pre-configured control input gain coefficient, wherein the control input gain coefficient represents the nominal value of the control quantity gain of the controlled object;

[0028] Divide the disturbance estimate by the control input gain coefficient to generate the normalized disturbance compensation amount;

[0029] The normalized disturbance compensation amount is negatively charged to generate the dynamic compensation amount.

[0030] Furthermore, the step of generating an anti-disturbance basic control signal based on the deviation signal between the setpoint signal used by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth, includes:

[0031] Calculate the real-time deviation signal between the setpoint signal and the actual output feedback signal of the controlled object;

[0032] The real-time deviation signal is proportionally calculated, and the calculation coefficient is the proportional gain coefficient of the controller to generate a proportional control component;

[0033] The real-time deviation signal is differentiated, and the operation coefficients are the differential gain coefficients of the controller, to generate differential control components;

[0034] The proportional control component and the derivative control component are superimposed to generate an anti-disturbance basic control signal.

[0035] Further, the step of subtracting the basic anti-disturbance control signal from the dynamic compensation amount based on the signal synthesis node to generate an active anti-disturbance control quantity and outputting it to the controlled object includes:

[0036] The signal synthesis node receives the anti-interference basic control signal and the dynamic compensation amount;

[0037] The dynamic compensation amount is negativeized to generate a negative compensation amount;

[0038] The disturbance rejection basic control signal is added to the negative compensation amount to generate the active disturbance rejection control quantity;

[0039] The active disturbance rejection control quantity is output to the controlled object to drive the controlled object to track the target state.

[0040] Accordingly, a second aspect of the present invention provides an active disturbance rejection controller (ADRC) control device, which controls the ADRC based on the above-described ADRC control method. The ADRC includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control device includes:

[0041] The disturbance estimation calculation module is used to receive the actual output feedback signal of the controlled object and the historical output control signal of the controller based on the linear extended state observer, and calculate the total disturbance of the system in real time through the third-order state space equation to obtain the disturbance estimate.

[0042] The dynamic compensation calculation module is used to calculate the dynamic compensation amount to offset the total system disturbance based on the disturbance estimate received by the compensator and the control input gain coefficient, and output the dynamic compensation amount to the signal synthesis node;

[0043] The first signal compensation module is used to generate an anti-disturbance basic control signal based on the deviation signal between the set value signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth.

[0044] The second signal compensation module is used to subtract the anti-disturbance basic control signal from the dynamic compensation amount based on the signal synthesis node, generate an anti-disturbance control amount, and output it to the controlled object to suppress external interference and internal parameter drift of the controlled object.

[0045] Accordingly, a third aspect of the present invention provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the above-described active disturbance rejection controller control method.

[0046] Accordingly, a fourth aspect of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described active disturbance rejection controller control method.

[0047] The above-described technical solutions of the embodiments of the present invention have the following beneficial technical effects:

[0048] 1. By using a third-order linear extended state observer, the total system disturbance is decomposed into three state variables: output tracking, differential change, and extended disturbance, for multi-dimensional modeling. An error correction mechanism corrects the output tracking deviation in real time to improve steady-state accuracy. The control quantity is coupled and injected to quantize the contribution of historical control signals to the disturbance. Differential gain compensation accelerates the response to sudden disturbances. Under the scenarios of power grid harmonic interference and motor parameter drift, the disturbance estimation accuracy is improved by more than 40%, fundamentally solving the overshoot problem caused by the feedback lag of traditional PID.

[0049] 2. An algorithm for alternating adjustment of observer and controller bandwidth: The stability limit is dynamically located using the system oscillation threshold and overshoot threshold as boundaries. Overshoot triggers a cyclical adjustment of the observer bandwidth backoff to avoid the risk of manual trial and error, ultimately achieving a stability of ω0 = 5ω. c The ratio locks in the optimal parameters; shortens the commissioning cycle by 60%, and achieves a load fluctuation response time of ≤50ms and an overshoot close to zero in power frequency regulation scenarios;

[0050] 3. By eliminating the influence of the order difference of the controlled object through normalized disturbance compensation, the negative synthesis mechanism feeds forward the dynamic compensation amount and the basic PD control signal to achieve physical decoupling of target tracking and disturbance cancellation. This architecture breaks through the coupling bottleneck of traditional control loop, effectively solves the gain mismatch problem when low-order ADRC controls high-order objects, and maintains a positioning accuracy of ±1μm under parameter drift in motor servo scenarios. Attached Figure Description

[0051] Figure 1 This is a flowchart of the active disturbance rejection controller control method provided in the embodiments of the present invention;

[0052] Figure 2 This is a block diagram of the active disturbance rejection controller control device module provided in the embodiments of the present invention.

[0053] Figure label:

[0054] 1. Disturbance estimation calculation module; 2. Dynamic compensation calculation module; 3. First signal compensation module; 4. Second signal compensation module. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0056] Please refer to Figure 1 The first aspect of this invention provides a method for controlling an active disturbance rejection controller. The active disturbance rejection controller includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control method includes the following steps:

[0057] S100 receives the actual output feedback signal of the controlled object and the historical output control signal of the controller based on the linear extended state observer, and calculates the total disturbance of the system in real time through the third-order state-space equation to obtain the disturbance estimate.

[0058] Based on the Linear Extended State Observer (LESO), the system receives the actual output feedback signal of the controlled object and the historical output control signal of the controller, and dynamically calculates the total system disturbance through a third-order state-space equation. This equation contains three core state variables: the output tracking state, which tracks the actual output value of the controlled object in real time; the derivative state, which characterizes the rate of output change; and the extended state, which uniformly models internal and external disturbances (including parameter drift and external interference). High-precision dynamic estimation of the total disturbance is achieved through an error correction mechanism (the deviation between the actual output and the estimated output is corrected by a gain coefficient), a control quantity coupling mechanism (historical control signals are injected into the extended state through a gain coefficient), and a derivative compensation mechanism (accelerating the abrupt response to disturbances).

[0059] The above method breaks through the traditional observer order limitation and quantifies complex disturbances such as time-varying parameters and unmodeled dynamics in a unified manner.

[0060] S200 calculates the dynamic compensation amount to offset the total system disturbance based on the estimated disturbance received by the compensator and the control input gain coefficient, and outputs the dynamic compensation amount to the signal synthesis node.

[0061] The compensator receives the disturbance estimate and pre-configured control input gain coefficient (the nominal value representing the control quantity gain of the controlled object) output in step S100. It then normalizes the disturbance estimate by dividing it by the gain coefficient to eliminate the influence of differences in the controlled object's order, generating a normalized disturbance compensation quantity. After applying a negative sign, the dynamic compensation quantity is finally output to the signal synthesis node. This operation transforms the total system disturbance into a reverse control quantity, providing a mathematical basis for feedforward compensation and solving the gain mismatch problem when a low-order controller adapts to a high-order object.

[0062] S300 generates an anti-disturbance basic control signal based on the deviation signal between the setpoint signal used to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth.

[0063] The controller receives the real-time deviation between the setpoint signal and the actual output feedback signal, and generates an anti-disturbance basic control signal by combining the controller bandwidth parameter. Specifically, this is achieved by superimposing proportional gain (deviation signal multiplied by proportional gain coefficient) and derivative gain (deviation signal multiplied by derivative gain coefficient), where the proportional gain coefficient is the square of the controller bandwidth, and the derivative gain coefficient is twice the controller bandwidth. This design eliminates the integral component of traditional PID control, avoiding overshoot oscillations caused by integral saturation from the outset, while bandwidth parameterization allows for adaptive adjustment of the control strength.

[0064] S400, based on the signal synthesis node, subtracts the basic anti-disturbance control signal from the dynamic compensation quantity to generate an active anti-disturbance control quantity and outputs it to the controlled object, thereby suppressing external interference and internal parameter drift of the controlled object.

[0065] The signal synthesis node subtracts the basic disturbance rejection control signal generated in step S300 from the dynamic compensation amount output in step S200 (i.e., the basic control signal is superimposed with a negative compensation amount) to generate the final active disturbance rejection control quantity to drive the controlled object. This architecture achieves physical decoupling between target tracking and disturbance cancellation: the basic control signal responds independently to changes in the setpoint, while the dynamic compensation amount focuses on canceling the total system disturbance. Through a feedforward disturbance cancellation mechanism, external disturbances (such as power grid harmonics and mechanical vibrations) and internal parameter drift (such as model mismatch caused by motor temperature rise) are significantly suppressed.

[0066] This method achieves high-precision control in complex disturbance environments by employing multi-state disturbance modeling with a third-order linear extended state observer (LESO), a normalized feedforward compensation mechanism, and a bandwidth parameterized control architecture. In terms of disturbance suppression, it overcomes the control defects caused by feedback lag in traditional PID control by uniformly estimating parameter drift, external disturbances, and unmodeled dynamics. Regarding system compatibility, it eliminates the influence of differences in the order of the controlled object using normalized compensation, ensuring that the lower-order controller can safely and stably drive the higher-order controlled object. In terms of dynamic performance, it effectively balances response speed and system robustness based on bandwidth parameterized design, ultimately achieving millisecond-level load fluctuation response capability in power system frequency regulation scenarios, while maintaining micrometer-level positioning accuracy under parameter drift conditions in motor servo control scenarios.

[0067] Furthermore, in one embodiment of the present invention, the active disturbance rejection controller control method further includes:

[0068] S500 performs dynamic coordinated adjustment of the observer bandwidth of the linear extended state observer and the controller bandwidth of the controller, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, with overshoot / oscillation as the backoff trigger condition.

[0069] In the bandwidth dynamic coordinated adjustment in step S500, the observer bandwidth (ω0) of the linear extended state observer and the controller bandwidth (ω) of the controller are adjusted. c Perform closed-loop parameter tuning: First, fix ω c Starting with an initial value, ω0 is gradually increased until the system reaches its critical point (output amplitude exceeds the preset oscillation threshold); then, the current ω0 value is locked, and ω is gradually increased. c When the performance reaches the critical point (overshoot exceeds the preset threshold), and overshoot or oscillation is detected, ω0 is immediately reduced by a preset ratio to avoid instability risks caused by manual trial and error. This process is repeated until both the step response overshoot and settling time are simultaneously satisfied, and ω0 reaches the threshold value. c The process terminates and locks parameters when the value is 5 times the threshold. This mechanism uses dual dynamic boundary constraints—an oscillation threshold and an overshoot threshold—to achieve this. cThe robustness is enhanced by reducing the threshold value by 20%, while the system compatibility problem is solved by the tuning principle that b0 > the actual gain b of the higher-order object, thus achieving global optimization of response speed and stability.

[0070] This adjustment mechanism, through alternating increments of the bandwidth parameter and dynamic backoff triggered by overshoot, adjusts ω0 = 5ω c Empirical rules are transformed into standardized closed-loop processes: while avoiding the blindness of manual debugging, the efficiency of parameter tuning is significantly improved; with the system critical point (oscillation threshold) and performance critical point (overshoot threshold) as dual constraint boundaries, it ensures the stability of millisecond-level response in scenarios such as power frequency regulation, while solving the oscillation risk caused by gain parameter mismatch when controlling high-order objects with low-order ADRC.

[0071] Furthermore, the S500 dynamically coordinates the observer bandwidth of the linearly extended state observer and the controller bandwidth of the controller, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, including:

[0072] S510, initialize the controller bandwidth to a preset fixed value, and gradually increase the observer bandwidth with a preset first adjustment step size until the system output amplitude exceeds the preset oscillation threshold.

[0073] Initialize and set the controller bandwidth ω c The observer bandwidth ω0 is gradually increased by a preset fixed value with a preset first adjustment step size. During this process, the system output amplitude is continuously monitored, and the increase is immediately stopped when the amplitude exceeds a preset oscillation threshold (the threshold is related to the dynamic characteristics of the controlled object). By actively approaching the system stability boundary, the maximum allowable value of the observer bandwidth is located, laying the foundation for subsequent coordinated adjustment.

[0074] S520 fixes the current observer bandwidth and gradually increases the controller bandwidth with a preset second adjustment step size.

[0075] The observer bandwidth ω0 is fixed by S510, and the controller bandwidth ω is gradually increased by a preset second adjustment step size. c This stage focuses on the overshoot and response speed of the controlled object's actual output until the system output overshoot exceeds the preset overshoot threshold or the amplitude reproduces oscillation. This improves the performance potential of the controller's bandwidth while exposing ω. c Excessive risk of instability.

[0076] S530 If the overshoot of the actual output feedback signal of the controlled object exceeds the preset overshoot threshold or the amplitude exceeds the oscillation threshold, the observer bandwidth will be reduced by the preset backoff ratio, and then the controller bandwidth will be increased.

[0077] If the overshoot exceeds the preset overshoot threshold or the amplitude exceeds the oscillation threshold, the current observer bandwidth ω0 is reduced by a preset backoff ratio. After the backoff, execution continues with ω. c The adjustment process is accelerated. A dynamic rollback mechanism avoids the blindness of traditional manual trial and error, ensuring that the adjustment process remains within a stable region.

[0078] S540, repeat the above adjustment process so that the controlled object simultaneously satisfies the condition that the step response overshoot is less than the preset overshoot threshold and the adjustment time is less than the preset time threshold.

[0079] Repeat the adjustment process from S510 to S530 until the controlled object simultaneously meets two core performance indicators: the step response overshoot is less than the preset overshoot threshold, and the settling time is less than the preset time threshold (both thresholds are strongly correlated with the dynamics of the controlled object). This closed-loop iteration transforms the rules into a standardized process, achieving a gradual balance between response speed and stability.

[0080] Among them, the step response overshoot less than the preset overshoot threshold refers to the maximum percentage deviation of the actual output from the steady-state value during the response of the controlled object to a change in the step setpoint (e.g., the transient fluctuation amplitude of frequency during a sudden increase in grid load). The settling time less than the preset time threshold refers to the time from the start of the step change to the first entry into and maintenance of the steady-state value within ±2% error band (e.g., the time taken for motor positioning from the issuance of the command to stabilization within ±1μm of the target position); both the overshoot threshold and the time threshold are set according to the dynamic characteristics of the controlled object (e.g., power system frequency regulation allows 0.5% overshoot / 200ms adjustment, while servo systems require 0 overshoot / 10ms adjustment).

[0081] S550: When the observer bandwidth reaches a preset multiple of the controller bandwidth, the parameter adjustment process is terminated and the corresponding controller bandwidth and observer bandwidth are locked.

[0082] When the observer bandwidth ω0 reaches the controller bandwidth ω c When the preset multiple is reached, the parameter adjustment process is immediately terminated and the final bandwidth parameter is locked. That is, the larger the ratio, the more stable the controller, but ω c Excessive scaling reduces robustness; a 5x scaling factor is the optimal engineering solution for balancing stability and response speed.

[0083] This collaborative adjustment mechanism achieves adaptive optimization of bandwidth parameters through a standardized process: using oscillation threshold and overshoot threshold as dual constraint boundaries, combined with ω0 = 5ω c The termination rule transforms debugging experience into a closed-loop algorithm, directly addressing the low efficiency of manual trial and error (shortening the debugging cycle), poor adaptability between high- and low-order systems (ensuring gain matching through the b_0>b rule), and insufficient robustness (ω). c(20% lower than the critical value); ultimately achieving millisecond-level response and near-zero overshoot control quality in scenarios such as power frequency regulation.

[0084] Furthermore, the linearly extended state observer in S100 receives the actual output feedback signal of the controlled object and the historical output control signal of the controller, and calculates the total system disturbance in real time through the third-order state-space equation to obtain the disturbance estimate, including:

[0085] S110, construct the third-order state-space equation of the linear extended state observer, whose state variables include: output tracking state variable that tracks the actual output of the controlled object, differential state variable that characterizes the rate of change of the actual output of the controlled object, and extended state variable that characterizes the total disturbance of the system.

[0086] A third-order state-space equation for a linear extended state observer is constructed, with three core state variables: an output tracking state variable that tracks the actual output value of the controlled object in real time, a differential state variable that represents the rate of change of the actual output, and an extended state variable that uniformly models the total system disturbance including external disturbances, parameter drift, and unmodeled dynamics. The above architecture extends the dynamic estimation dimension of traditional observers through a third-order model.

[0087] S120: Calculate the real-time deviation between the actual output feedback signal and the output tracking state variable, correct the deviation using the observer error gain coefficient, and generate the first state update quantity.

[0088] The real-time deviation between the actual output feedback signal and the output tracking state variable is calculated. The deviation is dynamically corrected by the observer error gain coefficient to generate the first state update quantity. The estimation error of the output tracking state is continuously corrected, which significantly improves the steady-state accuracy and suppresses the influence of measurement noise.

[0089] S130 couples the historical output control signal to the extended state variable through the control input gain coefficient to generate the second state update quantity.

[0090] The historical output control signal is coupled to the extended state variable through a pre-configured control input gain coefficient to generate a second state update quantity. This quantifies the contribution of the control quantity to the system disturbance, ensures gain matching when the low-order controller drives the high-order object, and avoids the risk of oscillation caused by control quantity gain mismatch.

[0091] S140 generates the third state update quantity based on the real-time rate of change of the differential state variable and by compensating for the estimation error through the differential gain coefficient of the observer.

[0092] Based on the real-time rate of change of the differential state variables, the third state update is generated by compensating for dynamic estimation errors through the differential gain coefficient of the observer, thereby enhancing the tracking sensitivity to sudden disturbances (such as power grid harmonics and mechanical shocks) and improving the disturbance response speed by more than 40%.

[0093] S150 integrates the first state update, the second state update, and the third state update to update the state variables of the third-order state space equation in real time, and outputs the values ​​of the expanded state variables as disturbance estimates.

[0094] By integrating the error correction effect of the first state update, the control coupling effect of the second state update, and the differential compensation effect of the third state update, all state variables of the third-order state space equation are updated in real time. Finally, the real-time value of the expanded state variable is output as the total disturbance estimate of the system, realizing unified quantitative modeling of internal and external disturbances.

[0095] The aforementioned observer overcomes the limitations of traditional single-dimensional disturbance estimation by integrating third-order state-space modeling with multi-source update quantity fusion: output tracking error correction ensures steady-state accuracy, control quantity coupling injection solves the compatibility problem between high and low order systems, and differential error compensation improves the response capability to sudden disturbances. Ultimately, it achieves millisecond-level disturbance suppression in power grid frequency regulation scenarios and maintains micron-level positioning accuracy under parameter drift in motor servo scenarios, while avoiding the overshoot oscillation defect caused by feedback lag in traditional PID control.

[0096] Furthermore, in S200, based on the estimated received disturbance value of the compensator and the compensation gain parameter, the dynamic compensation amount used to offset the total system disturbance is calculated, including:

[0097] S210 receives the disturbance estimate and the pre-configured control input gain coefficient, wherein the control input gain coefficient represents the nominal value of the control quantity gain of the controlled object.

[0098] The compensator receives the total system disturbance estimate and the pre-configured control input gain coefficient from the linear extended state observer. This gain coefficient represents the nominal value of the control quantity gain of the controlled object. Its value is preset according to the order characteristics and dynamic response requirements of the controlled object, providing a benchmark quantification scale for disturbance compensation.

[0099] S220 divides the disturbance estimate by the control input gain coefficient to generate the normalized disturbance compensation.

[0100] The received disturbance estimate is divided by the control input gain coefficient to generate a normalized disturbance compensation amount. This removes the scaling effect of the control gain on the disturbance estimate, making the compensation amount independent of the specific order of the controlled object. This ensures the compatibility of the low-order controller with the high-order object and avoids the stability risk caused by gain mismatch.

[0101] S230 applies a negative sign to the normalized disturbance compensation amount to generate the dynamic compensation amount.

[0102] By applying a negative operation to the normalized disturbance compensation amount, a dynamic compensation amount with the opposite phase and equivalent amplitude to the total disturbance of the system is generated. This transforms the mathematical principle of disturbance cancellation into a physically executable signal, providing a direct driving basis for feedforward compensation.

[0103] By using gain normalization and sign polarity conversion, a precise mapping from the total system disturbance to the control quantity is achieved: the normalization operation eliminates the influence of the controlled object's order difference on the compensation effect, and the sign operation constructs a reverse cancellation path. Ultimately, a 90% interference cancellation rate is achieved in the power system harmonic suppression scenario, and a ±1μm positioning accuracy is maintained in the motor servo parameter drift scenario. At the same time, the oscillation problem caused by gain mismatch in traditional control algorithms is completely avoided.

[0104] Furthermore, in S300, the deviation signal between the setpoint signal used to define the target state of the controlled object and the actual output feedback signal of the controlled object, received by the controller, is combined with the controller bandwidth to generate an anti-disturbance basic control signal, including:

[0105] S310 calculates the real-time deviation signal between the setpoint signal and the actual output feedback signal of the controlled object.

[0106] The controller receives the setpoint signal defining the target state of the controlled object and the actual output feedback signal in real time, continuously calculates the instantaneous deviation between the two, and generates a real-time deviation signal that reflects the system tracking error, providing the core input basis for the generation of control quantities.

[0107] S320 performs proportional calculations on the real-time deviation signal, with the calculation coefficients being the proportional gain coefficients of the controller, to generate proportional control components.

[0108] A proportional operation is applied to the real-time deviation signal, and the operation coefficient adopts the controller proportional gain coefficient (the value of which is the square of the controller bandwidth). The deviation signal is converted into a fast response component that is proportional to the error amplitude, generating the proportional control component that dominates the dynamic response of the system.

[0109] S330 performs differential operations on the real-time deviation signal, with the operation coefficients being the differential gain coefficients of the controller, generating differential control components.

[0110] Differential operation is applied to the real-time deviation signal, and the operation coefficient is the differential gain coefficient of the controller (the value of which is twice the controller bandwidth). The error change trend is extracted and converted into an advance compensation component, generating a differential control component that suppresses overshoot and oscillation, thereby enhancing the system's anti-disturbance capability.

[0111] S340 superimposes the proportional control component and the derivative control component to generate the basic control signal for disturbance rejection.

[0112] The proportional control component and the derivative control component are linearly superimposed to generate a disturbance rejection basic control signal. This signal combines the fast response characteristics of the proportional element with the predictive compensation capability of the derivative element, and eliminates the traditional integral element to avoid saturation risk, forming the core driving signal for target tracking.

[0113] By combining proportional-derivative dual-channel synthesis with bandwidth parameterization design, an optimal balance between dynamic response and stability is achieved: the proportional component ensures millisecond-level response capability for step changes in the setpoint, the derivative component actively suppresses overshoot oscillations caused by external disturbances, and the bandwidth parameterization architecture adaptively adjusts the control strength. Ultimately, this enables precise tracking of load fluctuations in power system frequency regulation scenarios and nanometer-level positioning accuracy in motor servo scenarios, while completely avoiding the risk of system instability caused by integral accumulation in traditional PID control.

[0114] Furthermore, the signal synthesis node in S400 subtracts the basic disturbance rejection control signal from the dynamic compensation quantity to generate an active disturbance rejection control quantity and outputs it to the controlled object, including:

[0115] S410 receives basic anti-interference control signals and dynamic compensation quantities at the signal synthesis node.

[0116] The signal synthesis node synchronously receives the basic anti-disturbance control signal from the controller and the dynamic compensation amount generated by the compensator, ensuring that the target tracking command and the disturbance cancellation signal are precisely aligned in the spatiotemporal dimension, providing complete input conditions for control quantity synthesis.

[0117] S420 takes the dynamic compensation amount as a negative value to generate a negative compensation amount.

[0118] Applying a negative sign to the dynamic compensation amount generates a phase-reversed negative compensation amount. This operation transforms the mathematical principle of disturbance cancellation (i.e., adding an inverse equal-amplitude signal) into a physically executable instruction, thus constructing the driving basis for feedforward compensation.

[0119] S430 adds the basic disturbance rejection control signal to the negative compensation quantity to generate the active disturbance rejection control quantity.

[0120] The basic disturbance rejection control signal and the negative compensation quantity are algebraically added to generate the final active disturbance rejection control quantity. This synthesis mechanism achieves functional decoupling of target tracking and disturbance cancellation: the basic control signal focuses on responding to changes in the setpoint, while the negative compensation quantity focuses on eliminating the total system disturbance, and the two do not interfere with each other.

[0121] S440 outputs the active disturbance rejection control quantity to the controlled object, driving the controlled object to track the target state.

[0122] The active disturbance rejection control quantity is output to the controlled object in real time, driving the controlled object to dynamically track the target state. This output directly acts on the input end of the controlled object, forming the final execution link of the closed-loop control loop, realizing real-time suppression of external disturbances such as power system harmonics and mechanical vibrations, as well as changes in internal parameters such as motor temperature drift.

[0123] By using a feedforward decoupling architecture and negative compensation injection, the bottleneck of response lag in traditional cascade control is overcome: the dual-channel parallel operation of target tracking and disturbance cancellation reduces the load fluctuation response time in power grid frequency regulation scenarios to the 50-millisecond level, and the motor servo positioning accuracy remains ±1 micrometer under parameter drift; at the same time, the normalized compensation design completely eliminates the adaptation difference between high and low order systems and avoids the oscillation risk caused by gain mismatch.

[0124] Accordingly, please refer to Figure 2 A second aspect of the present invention provides an active disturbance rejection controller (ADRC) control device, which controls the ADRC based on the above-described ADRC control method. The ADRC includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control device includes:

[0125] The disturbance estimation calculation module 1 is used to receive the actual output feedback signal of the controlled object and the historical output control signal of the controller based on the linear extended state observer, and calculate the total disturbance of the system in real time through the third-order state space equation to obtain the disturbance estimate.

[0126] The dynamic compensation calculation module 2 is used to calculate the dynamic compensation amount to offset the total system disturbance based on the disturbance estimate received by the compensator and the control input gain coefficient, and output the dynamic compensation amount to the signal synthesis node.

[0127] The first signal compensation module 3 is used to generate an anti-disturbance basic control signal based on the deviation signal between the set value signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth.

[0128] The second signal compensation module 4 is used to subtract the basic anti-disturbance control signal from the dynamic compensation amount based on the signal synthesis node, generate an anti-disturbance control quantity and output it to the controlled object, thereby suppressing external interference and internal parameter drift of the controlled object.

[0129] Accordingly, a third aspect of the present invention provides an electronic device, including: at least one processor; and a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the above-described active disturbance rejection controller control method.

[0130] Accordingly, a fourth aspect of the present invention provides a computer-readable storage medium having computer instructions stored thereon, which, when executed by a processor, implement the above-described active disturbance rejection controller control method.

[0131] The embodiments of the present invention aim to protect a method and apparatus for an active disturbance rejection controller, which has the following effects:

[0132] 1. By using a third-order linear extended state observer, the total system disturbance is decomposed into three state variables: output tracking, differential change, and extended disturbance, for multi-dimensional modeling. An error correction mechanism corrects the output tracking deviation in real time to improve steady-state accuracy. The control quantity is coupled and injected to quantize the contribution of historical control signals to the disturbance. Differential gain compensation accelerates the response to sudden disturbances. Under the scenarios of power grid harmonic interference and motor parameter drift, the disturbance estimation accuracy is improved by more than 40%, fundamentally solving the overshoot problem caused by the feedback lag of traditional PID.

[0133] 2. An algorithm for alternating adjustment of observer and controller bandwidth: The stability limit is dynamically located using the system oscillation threshold and overshoot threshold as boundaries. Overshoot triggers a cyclical adjustment of the observer bandwidth backoff to avoid the risk of manual trial and error, ultimately achieving a stability of ω0 = 5ω. c The ratio locks in the optimal parameters; shortens the commissioning cycle by 60%, and achieves a load fluctuation response time of ≤50ms and an overshoot close to zero in power frequency regulation scenarios;

[0134] 3. By eliminating the influence of the order difference of the controlled object through normalized disturbance compensation, the negative synthesis mechanism feeds forward the dynamic compensation amount and the basic PD control signal to achieve physical decoupling of target tracking and disturbance cancellation. This architecture breaks through the coupling bottleneck of traditional control loop, effectively solves the gain mismatch problem when low-order ADRC controls high-order objects, and maintains a positioning accuracy of ±1μm under parameter drift in motor servo scenarios.

[0135] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0136] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0137] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0138] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for controlling an active disturbance rejection controller, characterized in that, The active disturbance rejection controller includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control method includes the following steps: Based on the linear extended state observer receiving the actual output feedback signal of the controlled object and the historical output control signal of the controller, the total disturbance of the system is calculated in real time through the third-order state space equation to obtain the disturbance estimate; Based on the disturbance estimate and control input gain coefficient received by the compensator, a dynamic compensation amount is calculated to offset the total disturbance of the system, and the dynamic compensation amount is output to the signal synthesis node; Based on the deviation signal between the setpoint signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, an anti-disturbance basic control signal is generated by combining the controller bandwidth. Based on the signal synthesis node, the anti-disturbance basic control signal is subtracted from the dynamic compensation amount to generate an active anti-disturbance control quantity, which is then output to the controlled object to suppress external interference and internal parameter drift of the controlled object.

2. The active disturbance rejection controller control method according to claim 1, characterized in that, Also includes: The observer bandwidth of the linearly extended state observer and the controller bandwidth of the controller are dynamically coordinated and adjusted, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, with overshoot / oscillation as the backoff trigger condition.

3. The active disturbance rejection controller control method according to claim 2, characterized in that, The dynamic coordinated adjustment of the observer bandwidth of the linearly extended state observer and the controller bandwidth of the controller, alternately increasing the observer bandwidth to the system critical point and the controller bandwidth to the performance critical point, includes: Initialize the controller bandwidth to a preset fixed value, and gradually increase the observer bandwidth with a preset first adjustment step size until the system output amplitude exceeds the preset oscillation threshold. The current observer bandwidth is fixed, and the controller bandwidth is gradually increased with a preset second adjustment step size; If the overshoot of the actual output feedback signal of the controlled object exceeds the preset overshoot threshold or the amplitude exceeds the oscillation threshold, the observer bandwidth will be reduced by a preset backoff ratio, and then the controller bandwidth will be increased. Repeat the above adjustment process so that the controlled object simultaneously satisfies that the step response overshoot is less than the preset overshoot threshold and the adjustment time is less than the preset time threshold. When the observer bandwidth reaches a preset multiple of the controller bandwidth, the parameter adjustment process is terminated and the corresponding controller bandwidth and observer bandwidth are locked. The preset oscillation threshold, preset overshoot threshold, and preset time threshold are all dynamically related to the controlled object.

4. The active disturbance rejection controller control method according to claim 3, characterized in that, The preset multiplier is 5 times.

5. The active disturbance rejection controller control method according to any one of claims 1-4, characterized in that, The linear extended state observer receives the actual output feedback signal of the controlled object and the historical output control signal of the controller, and calculates the total system disturbance in real time through the third-order state-space equation to obtain the disturbance estimate, including: The third-order state-space equation of the linear extended state observer is constructed, and its state variables include: output tracking state variable that tracks the actual output of the controlled object, differential state variable that characterizes the rate of change of the actual output of the controlled object, and extended state variable that characterizes the total disturbance of the system. Calculate the real-time deviation between the actual output feedback signal and the output tracking state variable, correct the real-time deviation using the observer error gain coefficient, and generate the first state update quantity; The historical output control signal is coupled to the extended state variable through the control input gain coefficient to generate a second state update quantity; Based on the real-time rate of change of the differential state variable, the estimation error is compensated by the differential gain coefficient of the observer to generate the third state update quantity. By integrating the first state update, the second state update, and the third state update, the state variables of the third-order state space equation are updated in real time, and the values ​​of the expanded state variables are output as disturbance estimates.

6. The active disturbance rejection controller control method according to claim 5, characterized in that, The process of receiving the disturbance estimate and compensation gain parameters from the compensator, and calculating the dynamic compensation amount to offset the total system disturbance, includes: Receive the disturbance estimate and the pre-configured control input gain coefficient, wherein the control input gain coefficient represents the nominal value of the control quantity gain of the controlled object; Divide the disturbance estimate by the control input gain coefficient to generate the normalized disturbance compensation amount; The normalized disturbance compensation amount is negatively charged to generate the dynamic compensation amount.

7. The active disturbance rejection controller control method according to claim 6, characterized in that, The method of generating an anti-disturbance basic control signal based on the deviation signal between the setpoint signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth, includes: Calculate the real-time deviation signal between the setpoint signal and the actual output feedback signal of the controlled object; The real-time deviation signal is proportionally calculated, and the calculation coefficient is the proportional gain coefficient of the controller to generate a proportional control component; The real-time deviation signal is differentiated, and the operation coefficients are the differential gain coefficients of the controller, to generate differential control components; The proportional control component and the derivative control component are superimposed to generate an anti-disturbance basic control signal.

8. The active disturbance rejection controller control method according to claim 7, characterized in that, The step of subtracting the basic anti-disturbance control signal from the dynamic compensation amount based on the signal synthesis node to generate an active anti-disturbance control quantity and outputting it to the controlled object includes: The signal synthesis node receives the anti-interference basic control signal and the dynamic compensation amount; The dynamic compensation amount is negativeized to generate a negative compensation amount; The disturbance rejection basic control signal is added to the negative compensation amount to generate the active disturbance rejection control quantity; The active disturbance rejection control quantity is output to the controlled object to drive the controlled object to track the target state.

9. A self-disturbance rejection controller control device, characterized in that, The active disturbance rejection controller is controlled based on the control method of any one of claims 1-8. The active disturbance rejection controller includes: a linear extended state observer, a compensator, a controller, and a signal synthesis node. The control device includes: The disturbance estimation calculation module is used to receive the actual output feedback signal of the controlled object and the historical output control signal of the controller based on the linear extended state observer, and calculate the total disturbance of the system in real time through the third-order state space equation to obtain the disturbance estimate. The dynamic compensation calculation module is used to calculate the dynamic compensation amount to offset the total system disturbance based on the disturbance estimate received by the compensator and the control input gain coefficient, and output the dynamic compensation amount to the signal synthesis node; The first signal compensation module is used to generate an anti-disturbance basic control signal based on the deviation signal between the set value signal received by the controller to define the target state of the controlled object and the actual output feedback signal of the controlled object, combined with the controller bandwidth. The second signal compensation module is used to subtract the anti-disturbance basic control signal from the dynamic compensation amount based on the signal synthesis node, generate an anti-disturbance control amount, and output it to the controlled object to suppress external interference and internal parameter drift of the controlled object.

10. An electronic device, characterized in that, include: At least one processor; And a memory connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to cause the at least one processor to perform the active disturbance rejection controller control method as described in any one of claims 1-8.

Images