A control method and device for a magnetic-reactance dynamic voltage restorer

Abstract

The application provides a magnetic resistance type dynamic voltage restorer control method and device, relates to the technical field of voltage restorer control, and synchronously collects power grid voltage and extracts vector information thereof, compares the vector information with a target value to obtain voltage vector deviation containing amplitude and phase deviation, then combines the deviation with real-time load current, and calculates equivalent inductance and capacitance target values of a series branch required for voltage restoration through a circuit model; according to a nonlinear mapping relationship of a magnetic control reactor pre-labeled, the equivalent inductance target values are reversely solved into corresponding direct-current excitation current reference values, finally, high-dynamic current tracking is used to drive an excitation winding, so that actual current tracks the reference value, thereby changing the magnetic permeability of the core to adjust the equivalent inductance of the alternating current side, and through closed-loop iteration, the power grid voltage is restored to a target range.

Description

Technical Field

[0001] This invention relates to the field of voltage restorer control technology, specifically to a magnetic reactive dynamic voltage restorer control method and device. Background Technology

[0002] In the field of power quality management, the magnetically controlled dynamic voltage restorer has significant advantages such as high reliability, strong overload capacity, and low operating loss due to its main power circuit being based on a passive component, the non-magnetically controlled reactor. It is particularly suitable for industrial applications with stringent requirements for power supply continuity, such as semiconductor manufacturing and precision machining, to cope with disturbances such as voltage dips and rises in the power grid and ensure the stable operation of sensitive loads. Its core principle is to continuously change the equivalent inductive reactance value on the AC side by adjusting the DC excitation current of the magnetically controlled reactor, thereby changing the overall load impedance characteristics of the line in a series connection manner and realizing the adjustment of the voltage amplitude at the load end. Traditional control methods directly calculate and output the corresponding excitation current command based on the detected voltage amplitude deviation through a fixed control algorithm.

[0003] However, these existing control methods have fundamental limitations when dealing with complex voltage quality problems. First, in the voltage disturbance characteristic identification stage, traditional methods mostly only focus on the amplitude deviation of the fundamental voltage, failing to extract and quantitatively describe the phase jump, negative sequence component, and specific subharmonic components that coexist in the voltage vector. This results in an incomplete error profile obtained by the controller, which cannot provide a comprehensive basis for subsequent accurate compensation. Second, when deciding how to adjust the magnetic impedance, existing technologies often use a relatively isolated calculation method. That is, different types of voltage problems, such as insufficient amplitude and harmonic distortion, generate impedance adjustment requirements independently. There is a lack of a coordination mechanism to assess whether these multi-band, multi-objective impedance requirements can be achieved simultaneously within the physical capabilities of a single magnetically controlled reactor. This can easily lead to control commands exceeding the linear operating range of the device, causing saturation, slow response, or compensation failure. Furthermore, in the process of generating the final execution command, traditional methods rely on the steady-state, idealized current-inductive reactance mapping relationship of the magnetically controlled reactor, ignoring the inherent nonlinear dynamics of the magnetic circuit, such as hysteresis and eddy currents, as well as the disturbances introduced by load current coupling. This results in a deviation between the theoretically calculated excitation current reference value and the actual current value required to achieve the expected impedance effect, which restricts the compensation accuracy and dynamic response speed. It is precisely because of these shortcomings in the entire chain from deviation perception to target setting to command generation, such as incomplete perception, uncoordinated planning, and inaccurate models, that existing magnetically controlled dynamic voltage restorers cannot meet the requirements of high-end industrial applications in terms of compensation effect, response speed, and stability when facing complex voltage disturbances that simultaneously include amplitude drops, phase changes, and harmonic pollution.

[0004] The information disclosed in the background section is only intended to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0005] The purpose of this invention is to provide a control method and device for a magnetic reactive dynamic voltage restorer to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A control method for a magnetically reactive dynamic voltage restorer includes the following steps: Step 1: Real-time acquisition of the actual instantaneous voltage value at the grid connection point; extraction of the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value; and calculation of the voltage vector deviation including amplitude difference information and phase difference information by combining the preset target voltage vector. Step 2: Establish an analysis model based on the circuit principle, combine the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve for the equivalent inductive reactance target value and equivalent capacitive reactance target value that make the grid voltage return to the target range and the series branch where the magnetically controlled reactor is located. Step 3: Based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; Step 4: Drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Continue to execute steps 1 to 4 and perform iterative adjustments until the grid voltage is restored to the target range.

[0007] Furthermore, the amplitude and phase information of the actual voltage vector are extracted from the actual instantaneous voltage value, and the voltage vector deviation is calculated, specifically including: A frequency-locked loop based on a dual second-order generalized integrator is used to perform positive and negative sequence separation and phase-locked processing on the real-time acquired instantaneous voltage values. The instantaneous amplitude and instantaneous phase angle of the fundamental positive sequence voltage component are separated and extracted, and used as the amplitude and phase information of the actual voltage vector, respectively. The instantaneous value of the actual voltage is synchronously analyzed by a set of resonant observers configured in parallel, so as to extract the instantaneous amplitude of the fundamental negative sequence voltage component and the instantaneous amplitude of each pre-selected harmonic voltage component. The center frequency of each resonant observer is set as the fundamental negative sequence frequency and the selected harmonic frequency, respectively. The fundamental positive sequence voltage component's instantaneous amplitude is subtracted from the preset rated voltage amplitude reference value to obtain the fundamental positive sequence voltage amplitude difference information. The fundamental positive sequence voltage component's instantaneous phase angle is subtracted from the preset rated phase reference value to obtain the fundamental positive sequence voltage phase difference information. The preset rated voltage amplitude reference value and the preset rated phase reference value constitute the preset target voltage vector. The instantaneous amplitude of the extracted fundamental negative-sequence voltage component is used as the negative-sequence voltage amplitude difference information to be suppressed, and the instantaneous amplitude of each harmonic voltage component is used as the corresponding harmonic voltage amplitude difference information to be suppressed. The fundamental positive-sequence voltage amplitude difference information, the fundamental positive-sequence voltage phase difference information, the negative-sequence voltage amplitude difference information, and the harmonic voltage amplitude difference information together constitute the voltage vector deviation.

[0008] Furthermore, a circuit equation model is established with grid voltage, load current and series branch impedance of the magnetically controlled reactor as variables. The information of each component in the voltage vector deviation, combined with the real-time monitored load current vector information, is substituted into the circuit equations established for different frequency components for calculation. The solution process employs a combination of frequency band calculation and multi-objective optimization, including: For the fundamental positive sequence voltage component, the fundamental positive sequence voltage amplitude difference information and fundamental positive sequence voltage phase difference information are used as inputs. The calculation is performed through a discretized proportional-integral-derivative controller, and a fundamental positive sequence equivalent reactance demand value for compensating for voltage deviation is output in real time. For the fundamental negative sequence voltage component and each selected harmonic voltage component, the corresponding negative sequence voltage amplitude difference information and harmonic voltage amplitude difference information are used as inputs, and the equivalent reactance requirements of each frequency band aimed at suppressing the corresponding frequency voltage component are calculated in real time by multiplying them by a preset high gain coefficient. Based on the pre-acquired physical characteristic data of the magnetically controlled reactor, the fundamental positive sequence equivalent reactance demand value and the equivalent reactance demand value of each frequency band calculated in real time are used as optimization inputs to construct and solve a multi-objective optimization problem. The optimization objective of the multi-objective optimization problem is to find a set of optimal equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​within the safe operating range of the excitation current of the magnetically controlled reactor, so that the overall weighted deviation between these target values ​​and the equivalent reactance demand values ​​of all frequency bands is minimized; the output is the equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​obtained by solving and presented in real time in the current control cycle.

[0009] Furthermore, based on the pre-calibrated nonlinear mapping relationship, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; wherein, the nonlinear mapping relationship is the correspondence curve between the DC excitation current and the fundamental equivalent inductive reactance value obtained by offline testing of the magnetically controlled reactor. The reverse calculation process also includes feedforward compensation for dynamic disturbances in the magnetic circuit, specifically as follows: The corresponding steady-state excitation current reference value is obtained by querying the relationship curve based on the equivalent inductive reactance target value; an extended state observer is constructed, which takes the actual driving voltage applied to the excitation winding and the actual measured excitation current value as input, is based on a dynamic model including the resistance and inductance parameters of the excitation winding, and introduces an extended state variable to characterize the total system disturbance. The total disturbance estimate is estimated in real time through the preset observer gain coefficient. Finally, the total disturbance value estimated by the extended state observer is used to dynamically feedforward compensate the steady-state excitation current reference value to generate the final DC excitation current reference value.

[0010] Furthermore, the construction of the extended state observer and the estimation of perturbations are carried out in the following manner: A discrete-time state-space model is established with the excitation current estimate and the total disturbance estimate as state variables. In each control cycle, the extended state observer receives the excitation drive voltage and measured excitation current data collected in real time during that control cycle. Based on the state-space model and using the preset observer gain coefficient, the estimated value of the excitation current and the estimated value of the total disturbance are updated by numerical integration. After the update, the extended state observer outputs the predicted value of the excitation current for the next control cycle and the estimated value of the total disturbance in the current control cycle.

[0011] Furthermore, the excitation winding of the magnetically controlled reactor is driven by current tracking with high dynamic response, specifically employing a composite control strategy combining feedforward control and sliding mode variable structure feedback control, including: Based on the resistance and inductance parameters of the excitation winding of the magnetically controlled reactor, as well as the reference value of the DC excitation current and its rate of change, the feedforward drive voltage component is calculated; based on the deviation between the DC excitation current reference value and the actual measured value of the excitation current and its integral, a sliding mode surface function is constructed; based on the sliding mode surface function, the feedback drive voltage component used to suppress tracking error is calculated through a control law with a boundary layer saturation function. The feedforward drive voltage component and the feedback drive voltage component are added together to generate the final excitation drive voltage command; the excitation magnetic field strength is changed by applying a DC excitation current reference value, thereby linearly adjusting the permeability of the magnetically controlled reactor core. Iterative fine-tuning also includes a self-optimization process for control parameters, specifically as follows: During the self-optimization of control parameters, historical data of voltage vector deviation, DC excitation current reference value, and current tracking error are continuously recorded; the historical data are analyzed periodically, and the control parameters are automatically optimized and adjusted based on the analysis results; the control parameters include: the gain coefficient of the proportional-integral-derivative controller, the weight coefficient of each frequency band in the multi-objective optimization problem, and the gain coefficient and boundary layer thickness parameter of the sliding mode variable structure feedback control strategy.

[0012] Furthermore, the grid voltage recovering to the target range means that the amplitude of the actual voltage vector is continuously within the preset voltage amplitude allowable deviation band, and the phase is within the preset phase allowable deviation band. After the grid voltage recovers to the target range, the exit procedure is initiated. The exit procedure controls the DC excitation current reference value to smoothly decrease to zero according to a linear ramp function with a preset slope. Once the actual excitation current of the magnetically controlled reactor drops to zero, the bypass switch is closed to bypass the magnetically controlled reactor and its series capacitor from the main circuit. Subsequently, the system switches to a low-power mode that only monitors voltage.

[0013] The present invention also provides a control device for a magnetic reactive dynamic voltage restorer, the device being used to execute the above-described control method for a magnetic reactive dynamic voltage restorer, comprising: The deviation calculation module is used to collect the actual instantaneous voltage value of the grid connection point in real time, extract the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value, and calculate the voltage vector deviation containing amplitude difference information and phase difference information by combining it with the preset target voltage vector. The target planning module is used to establish an analysis model based on circuit principles. It combines the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve the equivalent inductive reactance target value and equivalent capacitive reactance target value that enable the grid voltage to return to the target range and the series branch where the magnetically controlled reactor is located to achieve the real-time equivalent inductive reactance target value. The reverse calculation module is used to reverse calculate the equivalent inductive reactance target value into a DC excitation current reference value based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target. The iterative adjustment module is used to drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Steps 1 to 4 are continuously executed and iterative adjustments are made until the grid voltage is restored to the target range.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention effectively overcomes the inherent defects of traditional magnetically controlled dynamic voltage restorers in dealing with complex voltage disturbances by constructing a fully closed-loop control chain from precise perception and collaborative decision-making to robust execution, and achieves a significant improvement in compensation performance. Specifically, at the perception level, by simultaneously extracting the amplitude and phase information of the fundamental voltage and simultaneously analyzing the negative sequence and selected subharmonic components, this invention constructs a complete voltage deviation profile containing amplitude, phase, and spectrum information, overcoming the perception blind spot caused by traditional methods that only focus on the fundamental amplitude, and laying a precise data foundation for subsequent targeted compensation. This invention does not simply convert compensation requirements of different frequency bands into isolated impedance commands. Instead, it establishes a multi-frequency band independent calculation and collaborative optimization mechanism. First, it generates preliminary impedance requirements for voltage problems of different natures, such as fundamental, negative sequence, and harmonics. Then, it inputs these multi-objective requirements into an optimization model constrained by the physical characteristics of the magnetically controlled reactor for solution. This process ensures that the final generated equivalent inductive reactance and equivalent capacitive reactance target values ​​are a set of feasible solutions within the actual working capacity of the magnetically controlled reactor and that can optimally balance the various compensation requirements as a whole. This fundamentally avoids the control failure problem caused by commands exceeding the physical limits of the device. In the process of reversely solving the excitation current, this invention introduces a feedforward compensation stage based on an extended state observer. This stage observes and estimates the total disturbance caused by magnetic circuit nonlinearity, eddy currents, and load coupling in real time, and dynamically corrects the reference excitation current calculated based on the steady-state mapping relationship. Combined with a composite current tracking strategy with high dynamic response, this design significantly reduces the impact of the dynamic characteristics of the magnetically controlled reactor itself and external disturbances on the control accuracy, enabling the actual output equivalent inductive reactance to track the target value quickly and accurately. Finally, through the closed-loop iteration of the above stages, this invention can achieve comprehensive, rapid, and stable management of grid voltage amplitude, phase, and specific harmonics, improving the applicability and reliability of the device under complex grid conditions. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the overall method flow of the present invention; Figure 2 This is a bar chart showing the steady-state excitation current reference value, the total disturbance estimate, and the DC excitation current reference value of this invention. Figure 3 This is a scatter plot of the total disturbance estimate and the DC excitation current reference value of the present invention. Figure 4 This is a curve showing the fitting of the steady-state excitation current reference value and the DC excitation current reference value according to the present invention. Figure 5 This is a stacked diagram of the feedforward compensation coefficient and DC excitation current reference value of the present invention. Figure 6This is a flowchart of the overall device structure of the present invention. Detailed Implementation

[0016] 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.

[0017] It should be noted that, unless otherwise defined, the technical or scientific terms used in this invention should have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains. The terms "first," "second," and similar terms used in this invention do not indicate any order, quantity, or importance, but are merely used to distinguish different components. Terms such as "comprising" or "including" mean that the element or object preceding the word encompasses the elements or objects listed following the word and their equivalents, without excluding other elements or objects. Terms such as "connected" or "linked" are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. Terms such as "upper," "lower," "left," and "right" are used only to indicate relative positional relationships; when the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0018] Example: Please see Figures 1-5 The present invention provides a technical solution: A control method for a magnetically reactive dynamic voltage restorer includes the following steps: Step 1: Real-time acquisition of the actual instantaneous voltage value at the grid connection point; extraction of the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value; and calculation of the voltage vector deviation including amplitude difference information and phase difference information by combining the preset target voltage vector. In a specific implementation, the core purpose of step 1 is to quantify the multidimensional deviation between the actual operating voltage of the power grid and the ideal target state in real time and accurately, so as to provide input for subsequent accurate compensation. The entire process begins with the acquisition of the real-time three-phase voltage at the power grid connection point. The amplitude and phase information of the actual voltage vector are extracted from the actual instantaneous voltage value, and the voltage vector deviation is calculated. Specifically, this includes: A frequency-locked loop based on a dual second-order generalized integrator is used to perform positive and negative sequence separation and phase-locked processing on the real-time acquired instantaneous voltage values. The instantaneous amplitude and instantaneous phase angle of the fundamental positive sequence voltage component are separated and extracted, and used as the amplitude and phase information of the actual voltage vector, respectively. The reason for using a frequency-locked loop (FLL) based on a DSOGI (Dual Second-Order Generalized Integrator) to process the actual instantaneous voltage value is that the grid voltage contains negative sequence and harmonic components under fault or disturbance conditions. The DSOGI-FLL structure can achieve accurate phase locking and rapid extraction of the fundamental positive sequence component under such conditions, ensuring that subsequent control is based on the correct phase reference. Specifically, it processes the synchronously acquired three-phase real-time voltage instantaneous values... Obtained by Clarke transform Components in stationary coordinate system ,in These refer to the instantaneous voltage signals of the three phases at grid connection points A, B, and C, which are synchronously acquired through sensors. for Voltage components in the stationary coordinate system of the axis; DSOGI pair and The signal is processed to generate a set of orthogonal signals; through a built-in positive-sequence separation calculation network, the voltage of the rotating coordinate system (dq axis) containing only the fundamental positive-sequence component is decoupled. and The frequency-locked loop adjusts its internal angular frequency to make Controlled to zero; It is the direct-axis component of the fundamental positive-sequence voltage, which refers to the voltage component aligned with the reference phase axis (d-axis) in a rotating coordinate system. Its steady-state value represents the amplitude information of the fundamental positive-sequence voltage. This refers to the quadrature-axis component of the fundamental positive-sequence voltage, specifically the voltage component orthogonal to the d-axis (q-axis) in a rotating coordinate system. After phase-locking is completed, its value is controlled to zero for precise voltage phase tracking. A frequency-locked loop (LLL) is a control loop that locks the phase of the input signal by adjusting the frequency of its internal oscillator. It achieves this by adjusting the internal angular frequency... The value is zero, thus achieving precise phase locking; at this time, The amplitude is the instantaneous amplitude of the extracted fundamental positive sequence voltage component. The phase angle output of the integral inside the frequency-locked loop is the instantaneous phase angle of the fundamental positive-sequence voltage component. ,Will Defined as the magnitude and phase information of the actual voltage vector; The instantaneous value of the actual voltage is synchronously analyzed by a set of resonant observers configured in parallel, so as to extract the instantaneous amplitude of the fundamental negative sequence voltage component and the instantaneous amplitude of each pre-selected harmonic voltage component. The center frequency of each resonant observer is set as the fundamental negative sequence frequency and the selected harmonic frequency, respectively. It should be noted that, in order to comprehensively perceive voltage quality issues, this invention analyzes the imbalance and harmonic components in the voltage in parallel while extracting the fundamental positive sequence component. This is achieved through a set of parallel resonant observers. Each resonant observer is designed as a bandpass filter for a specific frequency, such as the fundamental negative sequence frequency -50Hz, the 5th harmonic frequency 250Hz, the 7th harmonic frequency 350Hz, etc., with a high gain at its center frequency, for example, above 40dB, thereby amplifying and separating the amplitude information of the specific frequency component from the total voltage signal. Specifically, the same The instantaneous value is input into the resonant observer array. After independent processing and amplitude calculation of each channel, the instantaneous amplitude of the fundamental negative sequence voltage component can be obtained at the output. And the instantaneous amplitude of selected harmonic voltage components (e.g., 5th, 7th). , wait; The fundamental positive sequence voltage component's instantaneous amplitude is subtracted from the preset rated voltage amplitude reference value to obtain the fundamental positive sequence voltage amplitude difference information. The fundamental positive sequence voltage component's instantaneous phase angle is subtracted from the preset rated phase reference value to obtain the fundamental positive sequence voltage phase difference information. The preset rated voltage amplitude reference value and the preset rated phase reference value constitute the preset target voltage vector. After obtaining the information of each component, it is necessary to calculate their deviation from the target value. The preset target voltage vector is obtained from the rated voltage amplitude reference value. (e.g., peak voltage of 220V or 380V) and rated phase reference value Composition, in which Provided by an ideal phase generator synchronized with the ideal grid frequency (50Hz), the deviation of the fundamental positive sequence voltage is calculated using the difference method: amplitude difference information. Phase difference information These two directly reflect the degree to which the voltage deviates from its rated value in terms of amplitude and phase; For the fundamental negative sequence and all harmonic components, the ideal target value should be zero. Therefore, the instantaneous amplitude directly extracted by the resonance observer is... , , These are respectively used as information on the magnitude difference of the negative sequence voltage to be suppressed. Harmonic voltage amplitude difference information , ,Right now , This means that the very existence of these components is considered a deviation that needs to be eliminated; The instantaneous amplitude of the extracted fundamental negative-sequence voltage component is used as the negative-sequence voltage amplitude difference information to be suppressed, and the instantaneous amplitude of each harmonic voltage component is used as the corresponding harmonic voltage amplitude difference information to be suppressed; the fundamental positive-sequence voltage amplitude difference information, the fundamental positive-sequence voltage phase difference information, the negative-sequence voltage amplitude difference information, and the harmonic voltage amplitude difference information together constitute the voltage vector deviation; Finally, all the calculated and acquired deviation information is collected to form a complete voltage vector deviation set. This set comprehensively and quantitatively describes the gap between the current grid voltage and the target high-quality voltage from multiple dimensions such as fundamental amplitude, three-phase imbalance, and harmonic distortion rate, laying a data foundation for generating accurate compensation instructions in the next step.

[0019] Step 2: Establish an analysis model based on the circuit principle, combine the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve for the equivalent inductive reactance target value and equivalent capacitive reactance target value that make the grid voltage return to the target range and the series branch where the magnetically controlled reactor is located. In a specific implementation, the core objective of this step is to transform the voltage vector deviation detected in step 1 into a specific and executable impedance adjustment command. This command clearly tells the series branch where the magnetically controlled reactor is located: at the current moment, what equivalent inductive reactance target value and equivalent capacitive reactance target value are required to restore the grid voltage to the target range. This is a process based on a combination of circuit model, real-time calculation and system optimization. A circuit equation model is established with grid voltage, load current and series branch impedance of magnetically controlled reactor as variables. The information of each component in voltage vector deviation, combined with the real-time monitored load current vector information, is substituted into the circuit equations established for different frequency components for calculation. First, an analytical model that can describe the electrical relationships of the system needs to be established. This model is based on Kirchhoff's voltage law and associates the grid voltage at the point of common coupling, the vector information of the load current flowing through it, and the impedance of the series branch of the magnetically controlled reactor (including the equivalent inductive reactance and fixed capacitive reactance to be determined). Since the impedance characteristics of the magnetically controlled reactor and the capacitor are different for currents of different frequencies, and the system needs to deal with the deviation of multiple frequency components, in actual calculations, it is necessary to establish the corresponding complex circuit equations for the fundamental positive sequence, the fundamental negative sequence, and each harmonic frequency. The solution process employs a combination of frequency band calculation and multi-objective optimization, including: For the fundamental positive sequence voltage component, the fundamental positive sequence voltage amplitude difference information and fundamental positive sequence voltage phase difference information are used as inputs. The calculation is performed through a discretized proportional-integral-derivative controller, and a fundamental positive sequence equivalent reactance demand value for compensating for voltage deviation is output in real time. The fundamental positive-sequence voltage amplitude difference and phase difference information are input into a discretized proportional-integral-derivative (PID) controller. The proportional unit of this controller reacts quickly to instantaneous deviations, the integral unit eliminates steady-state errors, and the derivative unit predicts trends. The three work together to output a dynamically changing fundamental positive-sequence equivalent reactance requirement. This value reflects the total reactance (positive for inductive, negative for capacitive) that the series branch needs to exhibit at the fundamental frequency in the current control cycle to correct the amplitude and phase deviations of the fundamental voltage. For the fundamental negative sequence voltage component and each selected harmonic voltage component, the corresponding negative sequence voltage amplitude difference information and harmonic voltage amplitude difference information are used as inputs, and the equivalent reactance requirements of each frequency band aimed at suppressing the corresponding frequency voltage component are calculated in real time by multiplying them by a preset high gain coefficient. For the fundamental negative sequence voltage component and each selected harmonic voltage component, the control objective is mainly to suppress (or block) the flow of these unbalanced or distorted currents. Therefore, the processing strategy is more direct: multiply the corresponding negative sequence voltage amplitude difference information or harmonic voltage amplitude difference information by a preset, large high gain coefficient, for example, set to more than 10 times the equivalent impedance of the system at this frequency. The calculation result is the equivalent reactance requirement value of each frequency band. Its physical meaning is to suppress the current at the corresponding frequency to an extremely low level. The series branch needs to present a very large reactance value at this specific frequency to form a high impedance path. Based on the pre-acquired physical characteristic data of the magnetically controlled reactor, the fundamental positive sequence equivalent reactance demand value and the equivalent reactance demand value of each frequency band calculated in real time are used as optimization inputs to construct and solve a multi-objective optimization problem. The optimization objective of the multi-objective optimization problem is to find a set of optimal equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​within the safe operating range of the excitation current of the magnetically controlled reactor, so as to minimize the overall weighted deviation between these target values ​​and the equivalent reactance demand values ​​of all frequency bands; and output the equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​obtained by solving the problem and presented in real time in the current control cycle. However, a real magnetically controlled reactor series branch can only present a certain inductive reactance value and a fixed capacitive reactance value at any given time, while the above calculation generates reactance requirements at multiple frequencies, which constitutes a contradiction: a limited number of actuators need to simultaneously meet multiple control objectives. To resolve this contradiction, a multi-objective optimization solution method is introduced. The input to the optimization problem is the equivalent reactance demand value for all frequency bands obtained in real time above. The decision variables of the optimization problem are the unique equivalent inductive reactance target value and equivalent capacitive reactance target value that need to be solved. The optimization objective function is designed to minimize the weighted sum of squares of the differences between the target value and the demand value for all frequency bands. The weight coefficients can be configured according to the severity of the power quality problem or the priority of governance. For example, the fundamental positive sequence compensation is given a higher weight. The key constraints stem from the physical reality of the magnetically controlled reactor: the target value of the equivalent inductive reactance must be achieved by adjusting the DC excitation current, and the excitation current must be within the safe operating range allowed by the device; therefore, it is necessary to embed the physical characteristic data reflecting the relationship between the excitation current and the fundamental equivalent inductive reactance, which are obtained through experiments in advance, as constraints into the optimization problem. By solving this constrained optimization problem, a set of equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​are finally output in real time within the current control cycle. This set of solutions is the globally optimal compromise for compensation requirements of all frequency bands under the premise of physical feasibility. It enables the series branch to use a fixed impedance configuration to deal with multiple problems such as voltage deviation, imbalance and harmonics at the same time and as well as possible. For the final equivalent inductive resistance target value The optimization objective formula is: ; in, This represents the set of indices of all frequency components that need to be considered, for example: ,in Represents the fundamental positive sequence. The fundamental frequency represents the negative sequence, and 5 and 7 represent the 5th and 7th harmonics, respectively. It is an index of the frequency components; It is the decision variable in the optimization problem, that is, the equivalent inductive reactance target value to be found, and attention inductive reactance. ,in The angular frequency is used here, and the inductance value is used. As a target, it is more in line with the control essence of the magnetically controlled reactor itself; It is aimed at the first The equivalent reactance requirement calculated for each frequency component is a complex number. For the fundamental positive sequence, it is output by the PID controller and includes an imaginary part designed to compensate for the phase difference; for the negative sequence and harmonics, it is a purely imaginary number (reactance) and is very large (calculated from a high gain coefficient), designed to block the current at that frequency. It is an impedance prediction function that, based on the physical characteristics of a magnetically controlled reactor, predicts the fundamental inductive reactance when the device is controlled to exhibit this value. At that time, it was in the first The actual impedance value that can be presented at each frequency component is a key modeling step, and its input is... The output is a prediction of the complex impedance for each frequency component; Is assigned to the first The weighting coefficients of each frequency component reflect the priority of addressing that frequency issue, for example... During voltage sag, fundamental frequency weight It can be set to the maximum; The constraints are: ; in, That is, based on the nonlinear mapping relationship, by The estimated DC excitation current calculated in reverse process; and These are the lower and upper limits of the safe operating range of the excitation current of the magnetically controlled reactor. This constraint ensures that the calculated target value is physically safe to achieve. This optimization problem formally resolves a fundamental contradiction: an actuator (MCR) can only present one inductive reactance value at a time, but it needs to simultaneously satisfy impedance requirements at multiple frequencies, even conflicting ones. This is achieved by introducing an impedance prediction function. The complex frequency-varying impedance characteristics of MCR are incorporated into the decision-making process; the globally optimal trade-off among all frequency band requirements is sought through a weighted least squares objective function; and the engineering feasibility of the scheme is ensured through current safety constraints. It is important to note that the target value of equivalent capacitive reactance is determined by the parameters of the fixed capacitor in the series branch. Therefore, it is a constant in a specific device. The optimization process is essentially about finding the optimal, real-time adjustable target value of equivalent inductive reactance. By changing it, the total impedance characteristics of the branch are adjusted to adapt to dynamically changing compensation requirements. The entire calculation process is completed and refreshed within each control cycle (e.g., 100 microseconds or 1 millisecond), thereby ensuring a rapid response to grid voltage disturbances and real-time compensation capabilities.

[0020] Step 3: Based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; In a specific implementation, the core objective of this step is to accurately and quickly convert the target value of the equivalent inductive reactance calculated in real time in step 2 into a control signal that can be directly applied to the excitation winding of the magnetically controlled reactor, namely the reference value of the DC excitation current. This conversion is the key link to achieve dynamic voltage recovery, because the equivalent inductive reactance presented by the magnetically controlled reactor is essentially a function of the permeability of its iron core, and the permeability can be continuously controlled by adjusting the DC excitation current. Based on a pre-calibrated nonlinear mapping relationship, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; wherein, the nonlinear mapping relationship is a curve relating the DC excitation current and the fundamental equivalent inductive reactance value obtained by offline testing of the magnetically controlled reactor. The aforementioned nonlinear mapping relationship specifically refers to the static characteristic of the one-to-one correspondence between the DC excitation current of a magnetically controlled reactor and the fundamental equivalent inductive reactance value presented by its AC winding at a specific power frequency (e.g., 50Hz). This relationship is nonlinear, and the fundamental reason lies in the saturation characteristics of the ferromagnetic material (core): when the DC excitation current is small, the core operates in the unsaturated region with high permeability, resulting in a large equivalent inductive reactance value; as the DC excitation current increases, the core tends to saturate, the permeability decreases significantly, and the equivalent inductive reactance value decreases accordingly. To obtain this relationship, a rigorous offline calibration test must be performed before the device is put into use. The specific operation is as follows: In a laboratory environment, a series of precisely known steady-state DC currents covering its allowable operating range are applied to the excitation winding of the magnetically controlled reactor, while the rated power frequency voltage is applied to the AC winding. The fundamental current is measured and calculated at this time, and then the corresponding fundamental equivalent inductive reactance value is calculated based on the ratio of voltage to fundamental current. All data points (DC excitation current, fundamental equivalent inductive reactance value) are recorded, and a continuous characteristic curve is formed by mathematical fitting (such as polynomial fitting or spline interpolation), or directly stored as a data lookup table. This curve or data table constitutes the nonlinear mapping relationship, which is the cornerstone for realizing the reverse mapping from the inductive reactance target to the current command. The reverse calculation process also includes feedforward compensation for dynamic disturbances in the magnetic circuit, specifically as follows: The corresponding steady-state excitation current reference value is obtained by querying the relationship curve based on the equivalent inductive reactance target value; an extended state observer is constructed, which takes the actual driving voltage applied to the excitation winding and the actual measured excitation current value as input, is based on a dynamic model including the resistance and inductance parameters of the excitation winding, and introduces an extended state variable to characterize the total system disturbance. The total disturbance estimate is estimated in real time through the preset observer gain coefficient. Finally, the total disturbance value estimated by the extended state observer is used to perform dynamic feedforward compensation on the steady-state excitation current reference value to generate the final DC excitation current reference value. In each control cycle, when the control system receives the latest equivalent inductive reactance target value from step 2, it immediately queries the aforementioned pre-stored relationship curve or data table, and through interpolation calculation, finds the DC excitation current value corresponding to the equivalent inductive reactance target value. This value is the steady-state excitation current reference value, which represents the magnitude of the DC current that the excitation winding should theoretically maintain in order for the magnetically controlled reactor to present a specified inductive reactance value under ideal steady-state conditions, ignoring all dynamic factors. However, in actual operation, the electromagnetic system of the magnetically controlled reactor is affected by various dynamic disturbances. These dynamic disturbances mainly include the modulation effect of the alternating magnetic flux generated by the AC load current in the iron core on the DC bias (i.e., load coupling), the additional losses caused by the iron core hysteresis and eddy current effect, the drift of silicon steel sheet parameters with temperature and frequency, and the small changes in the resistance and inductance parameters of the excitation circuit. These factors cause the instantaneous excitation current required to achieve the same inductive reactance value in the actual dynamic process to differ from the steady-state excitation current reference value given by the static characteristic curve. If only this steady-state value is used for control, the actual equivalent inductive reactance of the magnetically controlled reactor will not be able to accurately track the rapidly changing target value, thereby affecting the real-time performance and accuracy of voltage recovery. To overcome the above problems, an extended state observer is introduced to estimate and compensate for these disturbances in real time. The extended state observer is an advanced state and disturbance observation technique. Its core idea is to extend the unknown total disturbance (i.e., the total disturbance) in the system model into a new state variable, which is then estimated in real time along with the original system state (in this case, the excitation current). In this application, a dynamic electrical model of the excitation winding of the magnetically controlled reactor is established. This model includes the resistance and inductance parameters of the excitation winding, and an extended state variable is introduced specifically to characterize the cumulative effect (i.e., the total disturbance) of all the above dynamic disturbances. The specific operating mechanism of this extended state observer is as follows: In each control cycle, the observer receives two real-time signals as inputs: one is the actual driving voltage applied to the excitation winding by the power amplifier, and the other is the excitation current value actually measured by the current sensor. Based on the established extended state space model, and using a set of pre-designed observer gain coefficients, the observer updates the estimated values ​​of the excitation current state and the total disturbance state simultaneously through numerical integration algorithms (such as the Euler method or the Runge-Kutta method). Its output includes: the predicted value of the excitation current at the next moment, and the estimated value of the total disturbance at the current moment. This estimated value of the total disturbance quantitatively reflects the magnitude and direction of the influence of all unmodeled dynamics and external disturbances on the control loop. Finally, the estimated total disturbance output by the extended state observer in real time is multiplied by the feedforward compensation coefficient (which is determined by the system model parameters) to obtain a compensation current. This compensation current is then superimposed with the previously obtained steady-state excitation current reference value to generate the final DC excitation current reference value. This operation essentially feeds the estimated disturbance effect forward to the control command to cancel it in advance.

[0021] in, It is in the The final DC excitation current reference value generated in each control cycle is the instruction directly sent to the current tracking step 4. It is in the Each control cycle, based on the equivalent inductive impedance target value Query non-linear mapping relationships The steady-state excitation current reference value obtained from the inverse function is, i.e. It is based on static characteristics and does not consider dynamic disturbances; It is in the The total disturbance estimate, calculated in real time by the extended state observer over each control cycle, quantifies the impact of all unmodeled dynamics, such as magnetic circuit nonlinearity and load coupling, on the current-inductive reactance relationship. The feedforward compensation coefficient is a design parameter used to adjust the disturbance estimate. Converted to an equivalent current compensation, ideally, if the extended state observer is accurate, It can be set to 1. In practice, It can be fine-tuned to optimize the compensation effect; Table 1 shows the specific data for some cycle numbers and DC excitation current reference values.

[0022] Table 1 Data Statistics Table

[0023] Analysis of the provided control cycle data reveals a clear and logical linkage between the steady-state excitation current command, dynamic disturbance estimation, and final output command. The data shows that the steady-state excitation current reference value gradually increases as the cycle progresses, reflecting the control background of persistent grid voltage deviation and increasing compensation requirements. During this process, the total disturbance value estimated by the system in real time exhibits a characteristic of fluctuating around zero, sometimes positive and sometimes negative. For example, in cycle 4, the larger positive disturbance estimate has a significant increasing effect on the final command; while in cycles 3 and 7, the negative disturbance estimate significantly offsets part of the steady-state command, causing the final command value to decrease. The small adjustment of the feedforward compensation coefficient around the value of one zero further finely adjusts the intensity of disturbance compensation. This data relationship confirms the core mechanism of the control strategy: the final DC excitation current command does not simply depend on the steady-state calculation value, but is the result of its combined effect with the disturbance estimate after coefficient modulation. When the system detects a positive disturbance, such as a sudden load change that exacerbates core saturation, it will actively increase the output current to offset it; conversely, it will reduce the current. This feedforward compensation mechanism enables the system to pre-correct before the disturbance actually affects the performance of the magnetically controlled reactor, thereby significantly improving the accuracy and response speed of core permeability adjustment and ensuring that the equivalent inductive reactance value can quickly and accurately track the target value calculated in the previous step. This is a key technical guarantee for achieving high-quality voltage dynamic recovery.

[0024] This formula will transform traditional open-loop lookup table control. The system has been upgraded to a composite command generation strategy that combines static feedforward and dynamic disturbance feedforward. This strategy enables the control system to anticipate and counteract the dynamic disturbances that the system will be subjected to before the command is issued, thereby ensuring the tracking accuracy of the magnetically controlled reactor for rapidly changing target inductive reactance. This is the basis for achieving high-quality dynamic compensation. Through the above steps, the final DC excitation current reference value not only includes the basic instructions based on static characteristics, but also incorporates forward-looking compensation for real-time dynamic disturbances. Sending this reference value to the current tracking in step 4 can drive the excitation current of the magnetically controlled reactor, thereby adjusting the core permeability more accurately and quickly, ensuring that the equivalent inductive reactance value presented to the power grid by its AC winding closely follows the change of the equivalent inductive reactance target value.

[0025] Step 4: Drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Continue to execute steps 1 to 4 and perform iterative adjustments until the grid voltage is restored to the target range. In a specific implementation, this step is the final execution of the control command and the closed-loop formation link. Its core task is to drive the excitation winding current of the magnetically controlled reactor with high dynamic response speed and high precision, so that its actual value closely tracks the DC excitation current reference value generated by step 3, thereby realizing the real-time control of the core permeability and finally completing the closed-loop recovery of the grid voltage. The construction and perturbation estimation of the extended state observer are carried out in the following manner: A discrete-time state-space model is established with the excitation current estimate and the total disturbance estimate as state variables. In each control cycle, the extended state observer receives the excitation drive voltage and measured excitation current data collected in real time during that control cycle. Based on the state-space model and using the preset observer gain coefficient, the estimated value of the excitation current and the estimated value of the total disturbance are updated by numerical integration. After the update, the extended state observer outputs the predicted value of the excitation current for the next control cycle and the estimated value of the total disturbance in the current control cycle. In the entire control system, all calculations and control actions are executed cyclically in units of a fixed, extremely short time interval, which is called the control cycle. For example, a common value is 100 microseconds or 1 millisecond. In each control cycle, the system sequentially executes all calculations and output updates from step 1 to step 4. Therefore, the current control cycle refers to the calculation and output cycle that is currently being executed; the next control cycle refers to the next cycle that will begin immediately afterward. This discretized and periodic operation mode is a typical characteristic of digital control systems, ensuring the real-time performance and orderliness of control. The reason for constructing the extended state observer as a discrete-time state-space model and running iteratively in each control cycle is that the entire control algorithm ultimately runs in the digital controller with a fixed control cycle. It is necessary to transform the continuous-time extended state observer model into a discrete form suitable for digital computation. In specific implementation, within each control cycle, the extended state observer receives the actual applied excitation drive voltage and the sampled measured excitation current data in that cycle. Based on the discrete state-space model and the preset gain coefficient, it updates the estimated values ​​of excitation current and total disturbance through numerical iteration, thereby outputting the prediction of the current in the next control cycle and the estimate of the total disturbance in the current control cycle in real time. This design enables the observer to operate stably and efficiently in the digital system, providing real-time and accurate disturbance estimation for feedforward compensation. Furthermore, the excitation winding of the magnetically controlled reactor is driven by current tracking with high dynamic response, specifically employing a composite control strategy combining feedforward control and sliding mode variable structure feedback control, including: Based on the resistance and inductance parameters of the excitation winding of the magnetically controlled reactor, as well as the reference value of the DC excitation current and its rate of change, the feedforward drive voltage component is calculated; based on the deviation between the DC excitation current reference value and the actual measured value of the excitation current and its integral, a sliding mode surface function is constructed; based on the sliding mode surface function, the feedback drive voltage component used to suppress tracking error is calculated through a control law with a boundary layer saturation function. The feedforward drive voltage component and the feedback drive voltage component are added together to generate the final excitation drive voltage command; the excitation magnetic field strength is changed by applying a DC excitation current reference value, thereby linearly adjusting the permeability of the magnetically controlled reactor core. The calculation of the feedforward drive voltage component aims to provide the basic driving force to follow the trend of command changes. Its principle is to make open-loop prediction based on the mathematical model of the excitation winding of the magnetically controlled reactor. Given the resistance and inductance parameters of the winding, as well as the reference value of the DC excitation current given in the current cycle and its rate of change estimated by differentiation, a theoretical feedforward drive voltage component is directly calculated according to the circuit equation. This component attempts to counteract the electrical inertia of the winding itself, so that the current can ideally follow the trajectory of the reference value. However, due to model errors, parameter drift, and external disturbances, feedforward control alone cannot guarantee accurate tracking. Therefore, a robust sliding mode variable structure feedback control is introduced to suppress errors. The core of this feedback control is to construct a sliding surface function, defined as a linear combination of the current tracking error and its integral term. The goal of the sliding mode control law is to drive the system state (i.e., current deviation) to reach and maintain this sliding surface. Once reached, the system will slide along the sliding surface to the equilibrium point (where the error is zero). At this point, the system is not sensitive to parameter changes and disturbances, exhibiting strong robustness. The control law with boundary layer saturation function is an improvement on ideal sliding mode control. It sets a thin boundary layer near the sliding surface and uses a continuous saturation function (such as a continuous approximation of the sign function) to replace the discontinuous switching function within the boundary layer. This effectively reduces the chattering phenomenon caused by high-frequency switching and generates a smooth feedback drive voltage component. The feedforward and feedback components are combined to obtain the final excitation drive voltage command. This command is sent to a power amplifier, such as an H-bridge or Buck converter, which generates the actual power voltage applied to the excitation winding. In this way, the actual excitation current value can track the DC excitation current reference value at high speed and accurately. Iterative fine-tuning also includes a self-optimization process for control parameters, specifically as follows: During the self-optimization of control parameters, historical data of voltage vector deviation, DC excitation current reference value, and current tracking error are continuously recorded; the historical data are analyzed periodically, and the control parameters are automatically optimized and adjusted based on the analysis results; the control parameters include: the gain coefficient of the proportional-integral-derivative controller, the weight coefficient of each frequency band in the multi-objective optimization problem, and the gain coefficient and boundary layer thickness parameter of the sliding mode variable structure feedback control strategy; Changes in the excitation current directly regulate the permeability of the iron core. The permeability of the iron core is a physical quantity that measures its magnetic conductivity. When the reference value of the DC excitation current increases, the DC bias magnetic field strengthens, pushing the iron core's operating point towards the saturation region, resulting in a decrease in its permeability; conversely, the permeability increases. Since the inductance of the AC winding of the magnetically controlled reactor is proportional to the permeability of the iron core, the inductance value changes accordingly. Ultimately, the equivalent inductive reactance presented to the power grid is rapidly and continuously adjusted in a controlled manner, thereby dynamically changing the impedance characteristics of the series branch and performing voltage compensation. To achieve optimal long-term performance, the system integrates a background-running self-optimization process for control parameters. This process is a slow, periodic learning and adjustment loop that continuously records key historical operating data, including voltage vector deviation, DC excitation current reference value, and current tracking error for each control cycle. This data is analyzed periodically (e.g., after each hour of operation), using methods such as statistical analysis, heuristic rules, or simple optimization algorithms. Based on the analysis results, the system automatically fine-tunes key control parameters: for example, adjusting the gain coefficient of the proportional-integral-derivative controller according to the convergence speed of the voltage vector deviation; dynamically adjusting the weight coefficients of each frequency band in the multi-objective optimization problem based on the frequency and severity of power quality problems in each frequency band; and optimizing the gain coefficient and boundary layer thickness parameters of the sliding mode variable structure feedback control strategy based on the statistical characteristics of the current tracking error. This enables the control system to adapt to changes in grid operating conditions and maintain optimal control performance at all times. The iterative adjustment is a dynamic feedback loop, from sensing the voltage deviation in step 1 to calculating the required impedance target in step 2, then to solving the corresponding current command in step 3, and finally to executing the current command and adjusting the actual impedance in step 4. The change in impedance, in turn, affects the grid voltage (which is sensed again in step 1). This cycle repeats itself, with a closed-loop correction performed in each control cycle, driving the grid voltage to continuously approach the target range.

[0026] Specifically, the grid voltage recovers to the target range when the amplitude of the actual voltage vector is continuously within the preset voltage amplitude allowable deviation band and the phase is within the preset phase allowable deviation band. After the grid voltage recovers to the target range, the exit procedure is started. The exit procedure controls the DC excitation current reference value to smoothly decrease to zero according to a linear ramp function with a preset slope. After the actual excitation current value of the magnetically controlled reactor drops to zero, the bypass switch is closed to bypass the magnetically controlled reactor and its series capacitor from the main circuit. Then the system switches to a low-power mode that only monitors voltage. The target range is explicitly quantified as the actual voltage vector amplitude remaining (e.g., for 1 consecutive second) at the rated voltage amplitude. (e.g., within ±1%), and its phase remains within the rated phase. Within ±1°, once the system determines that this condition is met, it considers the compensation task to be completed and starts the exit procedure. This procedure first reduces the DC excitation current reference value to zero at a gentle slope, such as 10% of the rated value per second, in order to avoid the impact of sudden current changes on the magnetic circuit. In power grid voltage quality control, the target range is quantified as the rated voltage amplitude. Within and rated phase Within a certain range, this establishes a precise and quantifiable criterion for voltage recovery. (For example, ±1%) defines the allowable deviation range of the voltage amplitude, which means that the actual voltage amplitude must be stable between 99% and 101% of the rated value. This ensures the stability of the supply voltage and avoids the reduction in efficiency, overheating or damage to electrical equipment due to excessively high or low voltage. The degree (e.g. ±1°) defines the allowable deviation range of the voltage phase, meaning that the difference between the actual voltage phase and the ideal standard phase (synchronized with the system frequency) must be controlled within ±1 degree. This ensures the synchronization of the voltage waveform, which is crucial for the stable operation of motor loads, the system power factor, and the accurate power distribution among parallel devices. The reasons for adopting such settings (such as ±1% and ±1°) are twofold: firstly, they adhere to stringent international or industry standards for power quality, which are based on the safe operating tolerance range of a large number of electrical devices; secondly, in engineering practice, this setting achieves the best balance between measurement accuracy, control performance, and system cost: overly lenient limits (such as ±5%) cannot provide effective quality assurance; while overly stringent limits (such as ±0.1%) place unrealistic demands on the response speed of measurement sensors, control algorithms, and the accuracy of actuators, significantly increasing system complexity and cost; therefore, ±1% and ±1° are a widely accepted and reasonable tolerance range that balances high-standard power quality requirements with the feasibility of existing technologies. This allows the control system to clearly distinguish between transient disturbances and steady-state recovery, thereby making correct control decisions. Once the actual excitation current value of the magnetically controlled reactor is confirmed to have dropped to zero, a command is issued to close the bypass switch, such as a mechanical contactor or solid-state switch, completely bypassing the magnetically controlled reactor and its series-connected fixed capacitor from the main current path. At this point, the device no longer consumes power, and the system then switches to a low-power monitoring mode, retaining only basic voltage acquisition and logic judgment functions to prepare for the next voltage event. This completes the full operation cycle from dynamic compensation to static standby.

[0027] Please see Figure 6 The present invention also provides a control device for a magnetic reactive dynamic voltage restorer, the device being used to execute the above-described control method for a magnetic reactive dynamic voltage restorer, comprising: The deviation calculation module is used to collect the actual instantaneous voltage value of the grid connection point in real time, extract the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value, and calculate the voltage vector deviation containing amplitude difference information and phase difference information by combining it with the preset target voltage vector. The target planning module is used to establish an analysis model based on circuit principles. It combines the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve the equivalent inductive reactance target value and equivalent capacitive reactance target value that enable the grid voltage to return to the target range and the series branch where the magnetically controlled reactor is located to achieve the real-time equivalent inductive reactance target value. The reverse calculation module is used to reverse calculate the equivalent inductive reactance target value into a DC excitation current reference value based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target. The iterative adjustment module is used to drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Steps 1 to 4 are continuously executed and iterative adjustments are made until the grid voltage is restored to the target range.

[0028] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0029] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented in software, the above embodiments can be implemented, in whole or in part, as a computer program product. Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution.

[0030] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0031] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A control method for a magnetically reactive dynamic voltage restorer, characterized in that, The specific steps include: Step 1: Real-time acquisition of the actual instantaneous voltage value at the grid connection point; extraction of the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value; and calculation of the voltage vector deviation including amplitude difference information and phase difference information by combining the preset target voltage vector. Step 2: Establish an analysis model based on the circuit principle, combine the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve for the equivalent inductive reactance target value and equivalent capacitive reactance target value in the series branch where the magnetically controlled reactor is located so that the grid voltage can be restored to the target range. Step 3: Based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; Step 4: Drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Continue to execute steps 1 to 4 and perform iterative adjustments until the grid voltage is restored to the target range. The amplitude and phase information of the actual voltage vector are extracted from the actual instantaneous voltage value, and the voltage vector deviation is calculated. Specifically, this includes: A frequency-locked loop based on a dual second-order generalized integrator is used to perform positive and negative sequence separation and phase-locked processing on the real-time acquired instantaneous voltage values. The instantaneous amplitude and instantaneous phase angle of the fundamental positive sequence voltage component are separated and extracted, and used as the amplitude and phase information of the actual voltage vector, respectively. The instantaneous value of the actual voltage is synchronously analyzed by a set of resonant observers configured in parallel, so as to extract the instantaneous amplitude of the fundamental negative sequence voltage component and the instantaneous amplitude of each pre-selected harmonic voltage component. The center frequency of each resonant observer is set as the fundamental negative sequence frequency and the selected harmonic frequency, respectively. The fundamental positive sequence voltage component's instantaneous amplitude is subtracted from the preset rated voltage amplitude reference value to obtain the fundamental positive sequence voltage amplitude difference information. The fundamental positive sequence voltage component's instantaneous phase angle is subtracted from the preset rated phase reference value to obtain the fundamental positive sequence voltage phase difference information. The preset rated voltage amplitude reference value and the preset rated phase reference value constitute the preset target voltage vector. The instantaneous amplitude of the extracted fundamental negative-sequence voltage component is used as the negative-sequence voltage amplitude difference information to be suppressed, and the instantaneous amplitude of each harmonic voltage component is used as the corresponding harmonic voltage amplitude difference information to be suppressed; the fundamental positive-sequence voltage amplitude difference information, the fundamental positive-sequence voltage phase difference information, the negative-sequence voltage amplitude difference information, and the harmonic voltage amplitude difference information together constitute the voltage vector deviation; A circuit equation model is established with grid voltage, load current and series branch impedance of magnetically controlled reactor as variables. The information of each component in voltage vector deviation, combined with the real-time monitored load current vector information, is substituted into the circuit equations established for different frequency components for calculation. The solution process employs a combination of frequency band calculation and multi-objective optimization, including: For the fundamental positive sequence voltage component, the fundamental positive sequence voltage amplitude difference information and fundamental positive sequence voltage phase difference information are used as inputs. The calculation is performed through a discretized proportional-integral-derivative controller, and a fundamental positive sequence equivalent reactance demand value for compensating for voltage deviation is output in real time. For the fundamental negative sequence voltage component and each selected harmonic voltage component, the corresponding negative sequence voltage amplitude difference information and harmonic voltage amplitude difference information are used as inputs, and the equivalent reactance requirements of each frequency band aimed at suppressing the corresponding frequency voltage component are calculated in real time by multiplying them by a preset high gain coefficient. Based on the pre-acquired physical characteristic data of the magnetically controlled reactor, the fundamental positive sequence equivalent reactance demand value and the equivalent reactance demand value of each frequency band calculated in real time are used as optimization inputs to construct and solve a multi-objective optimization problem. The optimization objective of the multi-objective optimization problem is to find a set of optimal equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​within the safe operating range of the excitation current of the magnetically controlled reactor, so as to minimize the overall weighted deviation between these target values ​​and the equivalent reactance demand values ​​of all frequency bands; and output the equivalent inductive reactance target values ​​and equivalent capacitive reactance target values ​​obtained by solving the problem and presented in real time in the current control cycle. Based on a pre-calibrated nonlinear mapping relationship, the equivalent inductive reactance target value is inversely calculated into a DC excitation current reference value; wherein, the nonlinear mapping relationship is a curve relating the DC excitation current and the fundamental equivalent inductive reactance value obtained by offline testing of the magnetically controlled reactor. The reverse calculation process also includes feedforward compensation for dynamic disturbances in the magnetic circuit, specifically as follows: The corresponding steady-state excitation current reference value is obtained by querying the relationship curve based on the equivalent inductive reactance target value; an extended state observer is constructed, which takes the actual driving voltage applied to the excitation winding and the actual measured excitation current value as input, is based on a dynamic model including the resistance and inductance parameters of the excitation winding, and introduces an extended state variable to characterize the total system disturbance. The total disturbance estimate is estimated in real time through the preset observer gain coefficient. Finally, the total disturbance value estimated by the extended state observer is used to perform dynamic feedforward compensation on the steady-state excitation current reference value to generate the final DC excitation current reference value. The excitation winding of the magnetically controlled reactor is driven by current tracking with high dynamic response. Specifically, a composite control strategy combining feedforward control and sliding mode variable structure feedback control is adopted, including: Based on the resistance and inductance parameters of the excitation winding of the magnetically controlled reactor, as well as the reference value of the DC excitation current and its rate of change, the feedforward drive voltage component is calculated; based on the deviation between the DC excitation current reference value and the actual measured value of the excitation current and its integral, a sliding mode surface function is constructed; based on the sliding mode surface function, the feedback drive voltage component used to suppress tracking error is calculated through a control law with a boundary layer saturation function. The feedforward drive voltage component and the feedback drive voltage component are added together to generate the final excitation drive voltage command; the excitation magnetic field strength is changed by applying a DC excitation current reference value, thereby linearly adjusting the permeability of the magnetically controlled reactor core. Iterative fine-tuning also includes a self-optimization process for control parameters, specifically as follows: During the self-optimization of control parameters, historical data of voltage vector deviation, DC excitation current reference value, and current tracking error are continuously recorded; the historical data are analyzed periodically, and the control parameters are automatically optimized and adjusted based on the analysis results; the control parameters include: the gain coefficient of the proportional-integral-derivative controller, the weight coefficient of each frequency band in the multi-objective optimization problem, and the gain coefficient and boundary layer thickness parameter of the sliding mode variable structure feedback control strategy.

2. The control method for a magnetic reactive dynamic voltage restorer according to claim 1, characterized in that: The construction and perturbation estimation of the extended state observer are carried out in the following manner: A discrete-time state-space model is established with the excitation current estimate and the total disturbance estimate as state variables. In each control cycle, the extended state observer receives the excitation drive voltage and measured excitation current data collected in real time during that control cycle. Based on the state-space model and using the preset observer gain coefficient, the estimated value of the excitation current and the estimated value of the total disturbance are updated by numerical integration. After the update, the extended state observer outputs the predicted value of the excitation current for the next control cycle and the estimated value of the total disturbance in the current control cycle.

3. The control method for a magnetic reactive dynamic voltage restorer according to claim 1, characterized in that: The grid voltage recovers to the target range when the amplitude of the actual voltage vector is continuously within the preset voltage amplitude allowable deviation band and the phase is within the preset phase allowable deviation band. After the grid voltage recovers to the target range, the exit procedure is started. The exit procedure controls the DC excitation current reference value to smoothly decrease to zero according to a linear ramp function with a preset slope. Once the actual excitation current of the magnetically controlled reactor drops to zero, the bypass switch is closed to bypass the magnetically controlled reactor and its series capacitor from the main circuit. Subsequently, the system switches to a low-power mode that only monitors voltage.

4. A control device for a magnetic reactance-type dynamic voltage restorer, characterized in that: The device is used to execute a magnetic reactive dynamic voltage restorer control method according to any one of claims 1-3, comprising: The deviation calculation module is used to collect the actual instantaneous voltage value of the grid connection point in real time, extract the amplitude and phase information of the actual voltage vector from the actual instantaneous voltage value, and calculate the voltage vector deviation containing amplitude difference information and phase difference information by combining it with the preset target voltage vector. The target planning module is used to establish an analysis model based on circuit principles. It combines the amplitude difference information and phase difference information with the real-time monitored load current vector information to calculate and solve the equivalent inductive reactance target value and equivalent capacitive reactance target value that enable the grid voltage to return to the target range and the series branch where the magnetically controlled reactor is located to achieve the real-time equivalent inductive reactance target value. The reverse calculation module is used to reverse calculate the equivalent inductive reactance target value into a DC excitation current reference value based on the nonlinear mapping relationship between the DC excitation current of the magnetically controlled reactor and its externally presented fundamental equivalent inductive reactance target. The iterative adjustment module is used to drive the excitation winding of the magnetically controlled reactor through high dynamic response current tracking, so that its actual current value tracks the DC excitation current reference value, thereby changing the permeability of the core of the magnetically controlled reactor, thereby adjusting the equivalent inductive reactance value presented to the power grid by the AC winding of the magnetically controlled reactor. Steps 1 to 4 are continuously executed and iterative adjustments are made until the grid voltage is restored to the target range.

Images