A method and system for suppressing voltage fluctuations in industrial power grids

Abstract

This application relates to the technical field of power system power quality management, and discloses a method and system for suppressing voltage fluctuations in industrial power grids. The method includes: firstly, establishing a system-level voltage and current forward propagation model composed of cascaded power units, filter circuits, and grid impedance; secondly, constructing a feedforward controller based on the voltage and current forward propagation model, generating a main compensation voltage command based on real-time detected grid disturbances; thirdly, acquiring the state parameters of each power unit and constructing differential equations, obtaining a comprehensive real-time health index through fuzzy inference; fourthly, dynamically correcting the equivalent impedance parameters in the forward propagation model using this index, and calculating the voltage allocation weights of each unit; and finally, decomposing the total compensation voltage command into modulation signals for each unit based on the weights. This invention achieves high-speed, accurate, and highly reliable suppression of voltage fluctuations in industrial power grids through systematic modeling, intelligent state assessment, and dynamic fault-tolerant allocation.

Description

Technical Field

[0001] This application relates to the technical field of power quality management in power systems, and discloses a method and system for suppressing voltage fluctuations in industrial power grids. Background Technology

[0002] In industrial power grids, the widespread use of large-capacity impulsive loads and nonlinear loads leads to frequent voltage dips, swells, and waveform distortion, which seriously affect the normal operation of precision equipment.

[0003] To address this problem, compensation devices such as static var generators (SVA) and dynamic voltage restorers based on cascaded H-bridge topologies have been widely used. However, existing control methods have significant limitations. For example, most methods design controllers based on local or simplified models, failing to fully consider the cascaded coupling relationships between the internal power units, filter circuits, and grid impedance of the compensation device, resulting in slow dynamic response and limited compensation accuracy. Secondly, traditional control strategies treat the compensation system as a fixed black box, unable to perceive changes in the system's internal state. When some power units degrade due to aging or faults, it not only affects the compensation effect but may also cause system instability. Furthermore, existing fault-tolerant controls often employ simple bypass strategies, leading to a sharp drop in system capacity and failing to achieve seamless derating operation under fault conditions. Summary of the Invention

[0004] The purpose of this section is to outline some aspects of the embodiments of this application and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents, and such simplifications or omissions should not be construed as limiting the scope of this application.

[0005] To address the aforementioned technical problems, this application provides a method and system for suppressing voltage fluctuations in industrial power grids.

[0006] On the one hand, this application provides a method for suppressing voltage fluctuations in an industrial power grid, including S1 using a voltage and current forward propagation model composed of power units, filter circuits, and transmission parameters in the grid impedance of a cascaded compensation system;

[0007] S2 detects the voltage and current disturbances in the power grid in real time and inputs them into the feedforward controller, which is built in reverse based on the voltage and current forward propagation model, to generate the main compensation voltage command.

[0008] S3 acquires the DC-side voltage deviation, output current distortion rate, and device temperature parameters of each power unit. It constructs a differential equation using the DC-side voltage deviation, output current distortion rate, and device temperature parameters, and obtains real-time health indicators through fuzzy reasoning via membership functions and reasoning criteria.

[0009] S4 corrects the equivalent impedance parameters in the voltage-current forward transfer model using the real-time health index and calculates the voltage allocation weights for each power unit.

[0010] S5 decomposes the total compensation voltage command into modulation signals for each unit through the voltage allocation weight, thereby modulating the industrial power grid.

[0011] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0012] The power unit, filter circuit, and grid impedance are respectively characterized as a two-port network of the power unit, a two-port network of the filter circuit, and a two-port network of the grid impedance. The voltage and current forward propagation model is obtained by calculating the product of the ABCD transmission matrices of each two-port network.

[0013] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0014] Obtain the total ABCD transfer matrix corresponding to the voltage-current forward transfer model;

[0015] The inverse of the total ABCD transmission matrix is ​​performed to obtain the inverse matrix model of ABCD.

[0016] The real-time detected grid voltage disturbance and current disturbance are combined to form an input vector, which is then used in matrix operation with the inverse matrix model. The voltage component of the output vector is the main compensation voltage command.

[0017] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0018] The methods for obtaining the DC-side voltage deviation, output current distortion rate, and device temperature parameters are as follows:

[0019] At a specific sampling rate higher than the base frequency, the instantaneous DC-side voltage value, instantaneous AC output current value, and device temperature sensor readings of each power unit are collected synchronously.

[0020] The DC-side voltage deviation is obtained by subtracting the average value of the DC-side voltage from the reference voltage value using a sliding window averaging algorithm.

[0021] The total harmonic distortion rate is calculated by performing a fast Fourier transform on the instantaneous value of the AC output current within the window, and is taken as the output current distortion rate.

[0022] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0023] When constructing the real-time health index, a differential equation characterizing the dynamic process of performance degradation is first constructed using the parameters. The method for constructing the differential equation is as follows:

[0024] S301 establishes a state equation with the DC side voltage deviation, output current distortion rate, and device junction temperature as state variables;

[0025] The state equation described in S302 comprises the following components: a linear term describing the decay trend of each state variable itself, a coupling term describing the nonlinear influence between each state variable, and an external input term characterizing the excitation effect of the current output power on the state change.

[0026] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, the solution result of the differential equation, namely the instantaneous comprehensive amplitude of the state variable and the comprehensive amplitude of its instantaneous rate of change, is used as the input variable for fuzzy inference.

[0027] Define membership functions covering a range from low to high for the two combined amplitudes respectively;

[0028] A fuzzy inference rule is established, which includes: when the instantaneous amplitude and rate of change of the state variable are both at a high level, the health is determined to be rapidly deteriorating; when both are at a low level, the health is determined to be good.

[0029] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0030] The output of the fuzzy inference system is defuzzified to obtain a preliminary health score.

[0031] This preliminary score is then weighted and fused with a remaining life assessment factor predicted based on the long-term integral of the aforementioned differential equation.

[0032] The weighting coefficients in the weighted fusion are dynamically adjusted based on the total output power currently borne by the compensation system, ultimately generating the real-time health index.

[0033] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0034] The real-time health index is mapped to the per-unit correction factor of the equivalent series impedance of the power unit in the forward pass model.

[0035] Based on the corrected equivalent series impedance of each power unit, calculate the reciprocal of its proportion in the total equivalent impedance of the system, and normalize the reciprocal of the proportion to obtain the initial voltage allocation weight of each power unit.

[0036] The initial voltage allocation weights are weighted geometrically and the real-time health indicators of the corresponding power units are used to generate voltage allocation weights.

[0037] Using the real-time health index, for power units whose health is below a preset threshold, a virtual admittance is connected in parallel with the corresponding equivalent impedance parameter to simulate the state of being partially bypassed.

[0038] Based on the corrected system equivalent impedance network, with the goal of maintaining the balance of the total output voltage of the system, the optimal voltage distribution weight of each power unit is solved so that the current stress flowing through each healthy unit tends to be consistent.

[0039] Establish a function with the optimization objectives of minimizing the system output impedance magnitude and achieving the most balanced voltage distribution among units;

[0040] The real-time health index is introduced as a constraint into the function of the optimization objective;

[0041] By solving the constrained optimization problem, the corrected values ​​of the system's equivalent impedance parameters and the optimal voltage allocation weights of each power unit in the forward transfer model are obtained simultaneously.

[0042] As a preferred embodiment of the industrial power grid voltage fluctuation suppression method of this application, wherein:

[0043] The total compensation voltage command is multiplied sequentially by the voltage allocation weight of each power unit obtained in step S4 to determine the specific output voltage command that each power unit needs to undertake.

[0044] Divide the specific output voltage command of each power unit by its measured DC-side voltage value to obtain the corresponding modulation wave signal;

[0045] Each modulated wave signal is compared with a carrier signal to generate a pulse width modulation signal for driving the switching devices in the corresponding power unit; wherein the carrier signal adopts an alternating anti-symmetrical triangular carrier, and the phase of the carrier of each power unit is evenly distributed between 0 and 360 degrees according to the unit number.

[0046] This application provides a novel system for suppressing voltage fluctuations in industrial power grids, comprising:

[0047] A novel industrial power grid voltage fluctuation suppression system comprises a compensation main circuit consisting of multiple cascaded power units, characterized in that:

[0048] The parameterization module includes a cascade unit and a calculation unit. The cascade unit models the compensation main circuit, output filter and known grid impedance parameters as multiple cascaded sub-units. The calculation module includes obtaining the voltage and current forward propagation model of the system by calculating the combination of transmission parameters of each functional module.

[0049] The feedforward module constructs a feedforward controller through the inverse matrix model of the voltage and current forward propagation model. The feedforward controller takes the detected instantaneous disturbances in grid voltage and current as inputs and directly generates a preliminary compensation voltage command.

[0050] The fault-tolerant module will optimize and allocate the synthesized total compensation voltage command in real time based on the number of power units currently operating normally and their performance parameters.

[0051] The beneficial effects of this application are as follows:

[0052] This application achieves near-instantaneous compensation for grid disturbances by constructing a feedforward controller based on the system-level voltage and current forward transfer model. By dynamically correcting the system model parameters through real-time health indicators, the control system always makes decisions based on the most realistic current system state, effectively avoiding compensation errors and potential oscillations caused by unit performance degradation, and ensuring stable operation under all operating conditions.

[0053] This application constructs a dynamic model of voltage deviation, distortion rate, and temperature changes over time using the aforementioned differential equation, which can quantitatively describe the attenuation trend and coupling relationship. Then, it uses fuzzy reasoning to address the problems of unclear boundaries and uncertainties in the changes of voltage deviation, distortion rate, and temperature over time, thereby achieving a unified quantitative prediction and qualitative assessment of the health status of power units and overcoming the shortcomings in the accuracy and adaptability of processing changes of voltage deviation, distortion rate, and temperature over time.

[0054] This application also extrapolates the future performance degradation trajectory of multiple power units based on the current state using differential equations, and combines the prediction results with the real-time state through fuzzy inference, so that this application can not only perceive the current voltage problem, but also predict the development of the voltage problem.

[0055] This application dynamically links the real-time health status of power units with voltage allocation weights. When the performance of some units deteriorates, the system can automatically reassign compensation tasks to healthy units, achieving seamless derating operation and improving the availability and reliability of the system. Attached Figure Description

[0056] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained through these drawings without creative effort. Wherein:

[0057] Figure 1 A schematic diagram of a method for suppressing voltage fluctuations in an industrial power grid provided in this application;

[0058] Figure 2 A flowchart of a new industrial power grid voltage fluctuation suppression system provided in this application;

[0059] Figure 3 A schematic diagram illustrating the voltage and current forward propagation model for an industrial power grid voltage fluctuation suppression method provided in this application;

[0060] Figure 4 The flowchart of the real-time health assessment algorithm for a novel method for suppressing voltage fluctuations in industrial power grids provided in this application is shown. Detailed Implementation

[0061] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the specific embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0062] Many specific details are set forth in the following description in order to provide a full understanding of this application. However, this application may also be implemented in other ways different from those described herein. Those skilled in the art can make similar extensions without departing from the spirit of this application. Therefore, this application is not limited to the specific embodiments disclosed below.

[0063] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of this application. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single embodiment or an embodiment selectively excluded from other embodiments.

[0064] Example 1

[0065] like Figure 1 As shown, an industrial power grid voltage fluctuation suppression method aims to maximize system resource utilization and minimize task latency, including:

[0066] S1 is a voltage and current forward propagation model composed of the transmission parameters in the power unit, filter circuit and grid impedance of the cascaded compensation system.

[0067] The power unit, filter circuit, and grid impedance are respectively characterized as a two-port network of the power unit, a two-port network of the filter circuit, and a two-port network of the grid impedance. The voltage-current forward propagation model is obtained by calculating the product of the ABCD transfer matrices of each two-port network, as shown below. Figure 3 As shown:

[0068] The specific implementation of establishing the system-level voltage and current forward propagation model is as follows: the power unit in the compensation system is characterized as a series two-port network of a controlled voltage source and an output impedance; the filter circuit is decomposed into a T-type two-port network composed of inductor and capacitor elements; and the grid impedance is characterized as a series RL-type two-port network.

[0069] By establishing ABCD transmission matrix models for each component, where the power unit matrix includes PWM modulation gain and equivalent impedance parameters for switching losses, the filter circuit matrix is ​​determined by inductance and capacitance values, and the grid impedance matrix includes system short-circuit impedance parameters, the ABCD transmission matrices of each level of the two-port network are multiplied sequentially according to the signal transmission order using the matrix multiplication method to obtain the voltage and current forward transmission model. The voltage and current forward transmission model is used to characterize the voltage and current transmission characteristics from the output of the compensation device to the grid connection point.

[0070] S2 detects the voltage and current disturbances in the power grid in real time and inputs them into the feedforward controller, which is built in reverse based on the voltage and current forward propagation model, to generate the main compensation voltage command.

[0071] Obtain the total ABCD transfer matrix corresponding to the voltage-current forward transfer model;

[0072] The inverse of the total ABCD transmission matrix is ​​performed to obtain the inverse matrix model of ABCD.

[0073] The real-time detected grid voltage disturbance and current disturbance are combined to form an input vector, which is then used in matrix operation with the inverse matrix model. The voltage component of the output vector is the main compensation voltage command.

[0074] A preferred implementation method for generating real-time compensation voltage commands for grid disturbances based on a feedforward inverse model includes:

[0075] S2.1 Obtain the total ABCD transfer matrix of the system-level voltage and current forward transfer model.

[0076] Based on the voltage and current forward propagation model, the power unit network, filter circuit network, and grid impedance network are constructed as three cascaded two-port networks. Each network is described using an ABCD parameter matrix, which is in the form of [A,B;C,D] and relates the voltage and current at the network input and output.

[0077] According to the two-port network cascading theory, the total ABCD transmission matrix [Atotal, Btotal; Ctotal, Dtotal] of the system, from the output of the compensation device (considered as the input of the cascading network) to the grid access point (considered as the output of the cascading network), is equal to the product of the ABCD matrices of each sub-network in the direction of signal transmission.

[0078] The specific calculation is as follows: Total matrix = Power unit matrix × Filter circuit matrix × Grid impedance matrix. Matrix multiplication is performed according to the standard matrix multiplication rules. The resulting total matrix is ​​used to characterize the voltage and current forward propagation characteristics of the system under given parameters.

[0079] S2.2: Construct the inverse matrix model. The inverse matrix model is used to calculate the output voltage to compensate for known disturbances on the power grid side, such as voltage and current.

[0080] Specifically, the 2x2 total ABCD transfer matrix [Atotal, Btotal; Ctotal, Dtotal] is inverted to obtain its inverse matrix [Ainv, Binv; Cinv, Dinv], which represents the reverse mapping of the original transfer relationship and constitutes the mathematical model of the feedforward controller. In practical implementation, this inversion operation can be performed using linear algebra library functions built into the embedded processor or digital signal processor.

[0081] S2.3: Real-time disturbance detection and input vector construction uses voltage and current sensors to collect instantaneous grid voltage U and instantaneous current I in real time at a high sampling rate at the common connection point or nearby point between the power grid and the compensation device.

[0082] Specifically, the detected instantaneous value is compared with the fundamental sinusoidal reference value to calculate the voltage disturbance Δu(t) and the current disturbance Δi(t).

[0083] Furthermore, the voltage disturbance Δu(t) and the current disturbance Δi(t) at the same moment are combined into a two-dimensional input column vector [Δu(t);Δi(t)].

[0084] S2.4: The feedforward calculation generates the main compensation voltage command, and performs matrix multiplication operation on the real-time disturbance input vector [Δu(t);Δi(t)] and the inverse matrix model [Ainv,Binv;Cinv,Dinv].

[0085] The operation format is: [Ucomp_cmd(t);Iaux(t)]=[Ainv,Binv;Cinv,Dinv]×[Δu(t);Δi(t)]

[0086] Where × represents matrix multiplication.

[0087] The initial compensation voltage command is the initial output of the feedforward module, while the main compensation voltage command is the final command after correction.

[0088] The output of the matrix operation is a two-dimensional column vector. The first element of the two-dimensional column vector, U_comp_cmd(t), is the main compensation voltage command to be calculated. The main compensation voltage command is a time-varying signal that can offset the influence of the currently detected grid disturbance on the voltage of the sensitive load side. The second element of the output vector, I_aux(t), is the auxiliary current used to help offset the influence of the current on the sensitive load side.

[0089] S3 acquires the DC-side voltage deviation, output current distortion rate, and device temperature parameters of each power unit. It constructs a differential equation using the DC-side voltage deviation, output current distortion rate, and device temperature parameters, and obtains a comprehensive real-time health index through fuzzy reasoning via membership function and reasoning criteria.

[0090] The methods for obtaining the DC-side voltage deviation, output current distortion rate, and device temperature parameters are as follows:

[0091] At a specific sampling rate higher than the base frequency, the instantaneous DC-side voltage value, instantaneous AC output current value, and device temperature sensor readings of each power unit are collected synchronously.

[0092] The DC-side voltage deviation is obtained by subtracting the average value of the DC-side voltage from the reference voltage value using a sliding window averaging algorithm.

[0093] The total harmonic distortion rate is calculated by performing a fast Fourier transform on the instantaneous value of the AC output current within the window, and is taken as the output current distortion rate.

[0094] By synchronously collecting the operating data of each power unit and using digital signal processing methods known in the art, the DC side voltage deviation, output current distortion rate, and device temperature are extracted.

[0095] The DC-side voltage deviation, output current distortion rate, and device temperature parameters are obtained in the following manner:

[0096] First, at a sampling rate higher than the base frequency of the power grid, the instantaneous DC-side voltage and AC output current of each power unit are collected synchronously, and the measured values ​​of the temperature sensor inside the power semiconductor device are read.

[0097] The methods for obtaining the DC-side voltage deviation include:

[0098] Instantaneous values ​​of DC-side voltage are continuously collected to form time series data. The time series data is processed using a sliding window averaging algorithm. The processing steps of the sliding window averaging algorithm are as follows: a data window covering a specific time length is set, and the data window slides forward with the arrival of new sampling points; at each processing moment, the arithmetic mean of all instantaneous DC-side voltage sampling values ​​within the current window is calculated as the estimated average value of the DC-side voltage at the current moment.

[0099] Furthermore, the calculated estimated average value is subtracted from the reference voltage value set by the power unit, and the difference is defined as the DC side voltage deviation, where the reference voltage value is the rated DC voltage of the power unit.

[0100] The method for obtaining the output current distortion rate includes:

[0101] The instantaneous values ​​of AC output current are continuously acquired to form time series data. Data containing an integer number of fundamental frequency cycles of the power grid is extracted from the time series data as an analysis window, and a fast Fourier transform is performed on the instantaneous value sequence of AC output current within the analysis window.

[0102] Specifically, the Fast Fourier Transform (FFT) is a method for converting a time-domain signal sequence into a frequency-domain spectrum. The FFT can determine the amplitude and phase information of the fundamental component and each harmonic component in the signal. Based on the spectrum results obtained from the FFT, the total harmonic distortion rate is calculated.

[0103] Furthermore, the method for calculating the total harmonic distortion rate is as follows: identify and calculate the effective value of the fundamental current component from the spectrum; calculate the effective value of all harmonic current components within a specified order range; take the square root of the sum of the squares of the effective values ​​of all harmonic current components, divide the result by the effective value of the fundamental current component, and express it as a percentage. The result obtained through this calculation process is the output current distortion rate.

[0104] The device temperature parameters are derived from the real-time readings of the temperature sensor and are synchronized with the acquisition time of the electrical parameters.

[0105] This application can acquire the three-dimensional state parameters of each power unit in real time and synchronously: voltage deviation reflecting the stability of DC power supply, current distortion rate reflecting the quality of AC output waveform, and temperature parameters that directly characterize the thermal load of device operation.

[0106] Those skilled in the art can obtain the parameters described in this step without any creative effort, based on the sliding window mean algorithm, fast Fourier transform, and total harmonic distortion rate calculation method disclosed in this section, which includes specific processing steps.

[0107] When constructing the real-time health index, a differential equation characterizing the dynamic process of performance degradation is first constructed using the parameters. The method for constructing the differential equation is as follows:

[0108] S301 establishes a state equation with the DC side voltage deviation, output current distortion rate, and device junction temperature as state variables;

[0109] The DC-side voltage deviation of the same power unit at the same time is denoted as ΔV, the output current distortion rate is denoted as THD, and the device junction temperature is denoted as T, which are used as a three-dimensional state vector.

[0110] Furthermore, a state equation is established to represent the evolution of the three-dimensional state vector over time. Mathematically, the state equation is expressed as a system of differential equations, which is used to express the functional relationship between the time derivative of the state vector, the state vector itself, and external stimuli.

[0111] The state equation described in S302 comprises the following components: a linear term describing the decay trend of each state variable itself, a coupling term describing the nonlinear influence between each state variable, and an external input term characterizing the excitation effect of the current output power on the state change.

[0112] The state equation includes linear terms, coupling terms, and external input terms.

[0113] The linear term describes the self-decay trend, which is the natural evolution trend of each state variable without external intervention or coupling effects. For example, the junction temperature of a device will naturally cool down after it stops operating, and the aging of the DC-side capacitor may cause its voltage sustaining capability to exhibit a slow, linear decline.

[0114] An example of a preferred linear term is the product of a state vector and a coefficient matrix, where the elements of the coefficient matrix describe the decay or drift rate of each state variable.

[0115] The coupling term describes the mutual nonlinear effects and is used to characterize the nonlinear mutual influence between the DC-side voltage deviation, output current distortion rate, and device junction temperature as state variables.

[0116] For example: the effect of temperature on electrical parameters: an increase in the junction temperature of a device may lead to an increase in switching losses, causing a more severe distortion of the output current waveform; the effect of electrical parameters on temperature: an increase in the output current distortion rate and the increase in harmonic current may lead to an increase in device conduction losses or core losses, and excite changes in junction temperature.

[0117] External input terms are used to characterize the excitation effect of the current output power. The current output power of the power unit is introduced into the equation as an external input quantity to identify the direct excitation or acceleration effect of the operating conditions on the performance degradation process.

[0118] For example, higher output power leads to more frequent current stress and switching operations, causing the excitation junction temperature to rise faster and exposing voltage deviations or current distortions caused by device aging more quickly. A preferred example of an external input term is the product of the output power P and the coefficient vector.

[0119] By adding the above linear terms, nonlinear coupling terms, and external input terms, the complete state equation is formed as a differential equation. This differential equation describes how the three core states of the power unit—voltage deviation, current distortion, and junction temperature—dynamically change due to their own characteristics and the complex interactions between them under a given output power (external excitation).

[0120] The solution results of the differential equation, namely the instantaneous comprehensive magnitude of the state variables and the comprehensive magnitude of their instantaneous rate of change, are used as the input variables for fuzzy inference.

[0121] Define membership functions covering a range from low to high for the two combined amplitudes respectively;

[0122] A fuzzy inference rule is established, which includes: when the instantaneous amplitude and rate of change of the state variable are both at a high level, the health is determined to be rapidly deteriorating; when both are at a low level, the health is determined to be good.

[0123] The state vector and its rate of change obtained by solving the differential equation are subjected to norm calculation to obtain the instantaneous comprehensive amplitude |X| characterizing the current operating state of the system and the comprehensive amplitude of the instantaneous rate of change reflecting the drasticness of the state change, respectively. .

[0124] Instantaneous composite amplitude |X| and instantaneous rate of change composite amplitude Triangular membership functions with low, medium, and high fuzziness levels are established respectively, where the universe of discourse of the instantaneous comprehensive amplitude |X| is set to [0,1] per unit value, and the instantaneous rate of change comprehensive amplitude is... The value can be set to [0, 0.1] per second based on the actual dynamic characteristics of the system.

[0125] It should be noted that a preferred method for selecting the triangular membership function can refer to existing mature solutions for health evaluation of substation monitoring and control devices.

[0126] A preferred example: The fuzzy rule base for constructing inference rules may include: if |X| is high and If |X| is high, the health output will show a sharp deterioration; if |X| is low and If the value is low, the health status output is "maintained good"; the remaining rules cover all intermediate state combinations. The inference rules are constructed based on the two-dimensional dimensions of state amplitude and change rate. Referring to the three-level health classification system of power equipment, good, deteriorated and faulty are used to form a 3×3 rule matrix to realize the full discrimination of the health status of power units in industrial power grids.

[0127] The output of the fuzzy inference system is defuzzified to obtain a preliminary health score.

[0128] This preliminary score is then weighted and fused with a remaining life assessment factor predicted based on the long-term integral of the differential equation.

[0129] The weighting coefficients in the weighted fusion are dynamically adjusted based on the total output power currently borne by the compensation system, ultimately generating the real-time health index.

[0130] In this application, a preferred method for generating the real-time health indicator includes:

[0131] Step 1: Defuzzify the output of the fuzzy inference system to obtain a preliminary health score.

[0132] Specifically, the health inference results, represented in the form of fuzzy sets, output by the fuzzy inference system are converted into scalar values ​​through defuzzification.

[0133] A preferred implementation of the defuzzification process is to use the centroid method to calculate the abscissa value corresponding to the centroid of the area enclosed by the membership function curve of the output fuzzy set and the abscissa axis. The abscissa value is the preliminary health score. The range of the preliminary health score is defined in the interval [0,1], where 1 represents an ideal health state and 0 represents complete failure. The defuzzification process is used to transform the qualitative evaluation of fuzzy logic into a quantitative score that can be calculated.

[0134] Step 2: Based on the long-term integral of the differential equation, predict the remaining life assessment factor.

[0135] The differential equation established in S3, with DC-side voltage deviation, output current distortion rate, and device junction temperature as state variables, is also a state equation.

[0136] Use the current state variable measurement as the initial condition; set a future time interval, such as the remaining part of the equipment's design life or a maintenance cycle, and within this interval, assume the future output power load curve of the system. A preferred simplification is to assume that the current output power level remains unchanged.

[0137] Furthermore, the differential equation is solved by numerical integration over a future time interval. Through integration, the evolution trajectory of the state variables over time under the aforementioned load conditions is simulated.

[0138] Furthermore, multiple failure thresholds are defined, such as the absolute upper limit of device junction temperature, the maximum allowable deviation of DC voltage deviation, and the acceptable limit of current distortion rate. By analyzing the state trajectory obtained by integration, the predicted time point when the state variable first touches or exceeds any failure threshold is determined. The duration from the current time to the predicted time point is compared with the rated design life of the power unit. The remaining life assessment factor, denoted as L, is calculated through a preset mapping function, such as a linear or exponential decay function. L is normalized to the interval [0, 1], where 1 indicates that the remaining life is close to new and 0 indicates that the life is about to end.

[0139] Step 3: The preliminary health score is weighted and integrated with the remaining life assessment factors, and the weighting coefficients are dynamically adjusted according to the total output power of the system.

[0140] A weighted fusion formula is pre-defined, for example: H = α * preliminary health score + (1-α) * remaining lifespan assessment factor, where H is the final real-time health index to be generated, and α is the dynamic weight coefficient.

[0141] The value of the dynamic weighting coefficient α is functionally related to the total output power currently borne by the compensation system.

[0142] A preferred implementation method for the suggested functional relationship is as follows: When the system is running under high load or high output power, the power unit is subjected to large electrical and thermal stresses, and its health status may change rapidly and drastically, making real-time monitoring of the status critical; under low load, the long-term aging and degradation process is the main process.

[0143] A preferred implementation is to define a normalized per-unit value for the current output power. ,in This represents the current total output power. The rated output power of the system is given. Then, α is determined using a monotonically increasing function, for example: α = αmin + (αmax - αmin). f(p). Where αmin and αmax are preset minimum and maximum weight coefficients, f(p) can be a linear function, a sigmoid function, or a piecewise function with a dead zone, etc., to ensure that under light load, α is biased towards αmin, which focuses more on the long-term lifetime factor; under heavy load, α is biased towards αmax, which focuses more on the real-time status score and the preliminary health score.

[0144] Step 4: Generate the final real-time health index. Substitute the calculated dynamic weight coefficient α, the preliminary health score, and the remaining life expectancy assessment factor into the weighted fusion formula to calculate the final real-time health index H.

[0145] When the initial health score shows a decline, the real-time health indicator H will drop rapidly, sending a priority maintenance signal.

[0146] When the initial health score is high, but the long-term model predicts a short remaining lifespan, the real-time health index H will also be appropriately reduced to indicate preventative maintenance.

[0147] By dynamically adjusting α, the evaluation focus is automatically adjusted under different operating conditions, avoiding evaluation failure under specific operating conditions with fixed weights.

[0148] S4 corrects the equivalent impedance parameters in the voltage-current forward transfer model using the real-time health index and calculates the voltage allocation weights for each power unit, such as... Figure 4 As shown:

[0149] The real-time health index is mapped to the per-unit correction factor of the equivalent series impedance of the power unit in the forward pass model.

[0150] By establishing a mapping relationship between health status and model parameters, and combining it with weighted average calculation, fast and efficient adaptive adjustment of the state can be achieved.

[0151] Specifically, the real-time health index of each power unit is converted into a per-unit correction coefficient of the equivalent series impedance of the power unit in the voltage-current forward transmission model through a preset mapping function that reflects the relationship between health and impedance change.

[0152] A preferred example of a mapping function is as follows: when the health index is high, the correction coefficient approaches one, indicating that the power unit impedance is normal and the voltage-current forward transfer model does not need to be significantly modified; when the health index decreases, the correction coefficient increases accordingly, indicating that the equivalent series impedance of the unit needs to be increased in the voltage-current forward transfer model to simulate the decrease in its output capability due to aging or minor faults.

[0153] Furthermore, by using the corrected equivalent series impedance of each power unit, the proportion of the impedance of each unit in the total equivalent impedance of all units in the system is calculated.

[0154] For each unit, the reciprocal of the ratio is taken. Units with increased equivalent impedance are those with lower health, and their reciprocals are smaller, so they should bear a lower voltage allocation share. The reciprocals of all units are normalized so that the sum of the values ​​of all units is one, thus obtaining the initial voltage allocation weight of each power unit.

[0155] Furthermore, the calculated initial voltage allocation weights are combined with the real-time health indicators of the corresponding power units to perform a weighted geometric average operation, generating the final voltage allocation weights.

[0156] If either the initial weight or the health index is low, the final weight will be significantly reduced, thus more proactively transferring the voltage compensation task from the poorly healthy unit to the healthy unit.

[0157] Based on the corrected equivalent series impedance of each power unit, calculate the reciprocal of its proportion in the total equivalent impedance of the system, and normalize the reciprocal of the proportion to obtain the initial voltage allocation weight of each power unit.

[0158] The initial voltage allocation weights are weighted geometrically and the real-time health indicators of the corresponding power units are used to generate voltage allocation weights.

[0159] Using the real-time health index, for power units whose health is below a preset threshold, a virtual admittance is connected in parallel with the corresponding equivalent impedance parameter to simulate the state of being partially bypassed. The threshold value can be set according to actual production needs.

[0160] When the real-time health index of a power unit is detected to be lower than the preset safety threshold, it is determined that the current power unit is no longer suitable for undertaking the task of full voltage output. In order to simulate the effect of the current power unit being partially bypassed in the system-level equivalent circuit, a virtual admittance is connected in parallel across the equivalent impedance parameter of the current power unit. The value of the virtual admittance is set to be negatively correlated with the real-time health index, that is, the lower the health, the larger the virtual admittance value. In other words, a low impedance path is connected in parallel across the degraded unit, which diverts the output current of the current power unit, thereby simulating the state of the output capability being partially bypassed.

[0161] After completing the equivalent impedance correction for all units, including healthy units and degraded units with parallel virtual admittance, an updated equivalent impedance network is formed. With the control objective of maintaining a balanced distribution of the total output voltage of the system among the units, circuit network theory, such as the nodal voltage method or the loop current method, can be selected to solve for the optimal voltage distribution weight that each power unit should bear under network constraints so that the current stress flowing through each normal operating unit is consistent.

[0162] By solving the circuit equations of the modified network, the electrical behavior of the system after the performance of some power units has been severely degraded is described more accurately, and a more reasonable power output task can be assigned.

[0163] Based on the corrected system equivalent impedance network, with the goal of maintaining the balance of the total output voltage of the system, the optimal voltage distribution weight of each power unit is solved so that the current stress flowing through each healthy unit tends to be consistent.

[0164] Establish a function with the optimization objectives of minimizing the system output impedance magnitude and achieving the most balanced voltage distribution among units;

[0165] The real-time health index is introduced as a constraint into the function of the optimization objective;

[0166] By solving the constrained optimization problem, the corrected values ​​of the system's equivalent impedance parameters and the optimal voltage allocation weights of each power unit in the forward transfer model are obtained simultaneously.

[0167] In this application, a preferred implementation method for a multi-objective optimization function includes:

[0168] A multi-objective optimization function with two objectives is established. The first objective is to minimize the overall output impedance amplitude to improve the dynamic response and stability of the compensation device. The second objective is to achieve the most balanced voltage distribution among the power units, i.e., to minimize the difference in voltage command values ​​allocated to each unit, thereby optimizing system lifetime and reliability. By assigning appropriate weighting coefficients to the two objectives, they are combined into a single comprehensive optimization objective.

[0169] Furthermore, the real-time health index of each power unit is introduced as a constraint into the optimization problem. The constraint is used to ensure that the voltage weight allocated to units whose health index is lower than a certain critical value does not exceed the upper limit that their current health status can safely support; or, it directly requires that the health of the units participating in the allocation must not be lower than a certain minimum safety limit.

[0170] Furthermore, by solving the above-mentioned constrained multi-objective optimization problem, the correction value of the system equivalent impedance parameter that most accurately reflects the current overall health state of the system in the voltage and current forward propagation model is obtained; while satisfying the health constraints of each unit, the optimal voltage allocation weight of each power unit is obtained so that the overall optimization objective of the system, namely, low output impedance and balanced distribution, is optimal.

[0171] The optimization problem can be solved by selecting an optimization algorithm, such as a sequential quadratic programming algorithm or a swarm intelligence optimization algorithm.

[0172] This application uses real-time health indicators as direct input to dynamically correct the established system-level voltage and current forward propagation model, enabling the mathematical model upon which the control algorithm relies to reflect changes in the internal state of the system in real time. This solves the problem of treating the system as a fixed black box in traditional methods. By dynamically linking the health status with voltage allocation weights, when the performance of some units deteriorates, the compensation task can be automatically and smoothly redistributed to healthier units, achieving truly seamless derating operation and improving the availability and long-term reliability of the system under non-ideal conditions.

[0173] S5 decomposes the total compensation voltage command into modulation signals for each unit through the voltage allocation weight, thereby modulating the industrial power grid.

[0174] The total compensation voltage command is multiplied sequentially by the voltage allocation weight of each power unit obtained in step S4 to determine the specific output voltage command that each power unit needs to undertake.

[0175] Divide the specific output voltage command of each power unit by its measured DC-side voltage value to obtain the corresponding modulation wave signal;

[0176] Each modulated wave signal is compared with a carrier signal to generate a pulse width modulation signal for driving the switching devices in the corresponding power unit. The carrier signal adopts an alternating anti-symmetrical triangular carrier, and the phase of the carrier of each power unit is evenly distributed between 0 and 360 degrees according to the unit number.

[0177] It should be noted that the uniform carrier phase distribution strategy refers to the cascaded H-bridge carrier phase shift modulation specification, and the phase difference between adjacent units is set to be between 0 and 360 degrees.

[0178] Determine the specific output voltage command that each power unit needs to handle. Then, multiply the total compensation voltage command by the voltage allocation weight of each power unit in sequence.

[0179] The voltage allocation weight of each power unit is used to identify the proportion of compensation tasks that the power unit should undertake under the current system state. Through multiplication, a specific output voltage command is generated for each power unit, realizing the dynamic and reasonable allocation of the system-level compensation target to each execution unit according to the health status and capability of each unit.

[0180] Furthermore, the modulated wave signal corresponding to each power unit is generated, and the command is converted into a modulated signal.

[0181] A preferred example is: dividing the specific output voltage command of each power unit by the measured DC side voltage value obtained in real time for the current power unit. The division operation is used to normalize the output voltage command. The calculation result is a modulation ratio signal, or modulation wave signal. The value range of the modulation wave signal is between -1 and 1, which is used to determine the ratio between the desired output AC voltage amplitude and the DC bus voltage.

[0182] Furthermore, the modulated wave signal of each power unit is compared with a carrier signal to generate a pulse width modulation signal.

[0183] The pulse width modulation signal is used as a control command to drive the fully controllable switching devices in each power unit to turn on or off. The carrier signal uses alternating anti-symmetrical triangular carriers, with adjacent power units using triangular carrier signals of opposite phase; that is, when the carrier of one unit is at its rising edge, the carrier of its adjacent unit is at its falling edge.

[0184] Furthermore, the phase of the carrier signal of all power units is uniformly distributed within a range of 0 to 360 degrees, according to the unit's serial number. For example, in a system cascaded with N units, the phase of the carrier signal of the k-th unit will be shifted... Spend.

[0185] This application generates a modulated wave by dividing the voltage command by the measured DC-side voltage, adaptively eliminating the influence of DC bus voltage fluctuations on the final output voltage. It uses alternating, anti-symmetrical, and uniformly phased triangular carrier waves for modulation, which can effectively increase the overall equivalent switching frequency of the system without increasing the switching frequency of individual units. This reduces the harmonic content in the compensation voltage, improves the output waveform quality, and the switching action times of each power unit are staggered, which helps to smooth out harmonic currents on the DC side and the grid side, disperse switching losses, and improve the overall reliability and efficiency of the system.

[0186] Example 2

[0187] A novel industrial power grid voltage fluctuation suppression system comprises a compensation main circuit consisting of multiple cascaded power units, such as... Figure 2 As shown, the input includes parameters related to the overall voltage and current transmission model of the system corresponding to the total matrix and state parameters of each power unit, such as DC side voltage, current distortion rate, and device temperature. At the same time, parameter extraction is also connected to filter and extract features from the original input parameters.

[0188] The parameter extraction module outputs multiple parallel signals. Each signal first passes through the signal allocation module, which dynamically allocates the proportion of compensation tasks based on the health status of the power unit and the system requirements.

[0189] The distributed signals are converted into modulation signals for each power unit, and finally aggregated into the collaborative output stage of each power unit, driving each power unit to synchronously generate pulse width modulation signals, thereby achieving collaborative suppression of power grid fluctuations.

[0190] The parameterization module includes cascade units and computation units. The cascade units are used to mathematically abstract the physical structure of the compensation system. Specifically, each power unit in the main compensation circuit, the output filter connected to the main circuit, and the pre-measured or estimated grid impedance parameters are independently modeled as a two-port network sub-unit. Each sub-unit is described by parameters ABCD. The cascade units are defined as cascaded according to the actual physical connection sequence of the signal flowing from the output of the compensation device to the grid side.

[0191] The computing unit calculates the combination of transmission parameters of each functional module. Specifically, it performs a series of multiplication operations on the ABCD parameter matrices of each two-port network subunit in the cascaded order to obtain the voltage and current forward transmission model of the compensation system from the output end to the grid access point.

[0192] The feedforward module is the core of high-speed compensation, and it is built directly based on the accurate mathematical model provided by the parameterization module.

[0193] The feedforward module first obtains the voltage and current forward propagation model calculated by the parameterization module. The feedforward module performs mathematical inverse operations to construct an inverse matrix model, which constitutes the feedforward controller.

[0194] During operation, the feedforward controller continuously receives instantaneous voltage and current disturbance signals from the grid voltage sensor and current sensor, and uses these two disturbance signals as input vectors to directly input them into the inverse matrix model for real-time calculation.

[0195] In the output of the inverse matrix model, the voltage component is directly extracted as the initial compensation voltage command required by the system. The command is designed to counteract the detected grid disturbances, thus achieving fast open-loop compensation.

[0196] The fault-tolerant module is used to achieve high system reliability and seamless derating operation, and to optimize instruction allocation under non-ideal conditions.

[0197] The fault-tolerant module receives the initial compensation voltage command from the feedforward module or the total compensation voltage command after fine-tuning by other closed-loop regulators.

[0198] The fault-tolerant module acquires information reflecting the internal state of the system in real time, including: the number of power units currently operating normally, and the performance parameters of each power unit obtained through online monitoring, such as real-time health indicators calculated based on DC-side voltage deviation, output current distortion rate, and device temperature.

[0199] The fault-tolerant module performs real-time optimization and allocation of the received total compensation voltage command.

[0200] Specifically, it employs an equivalent model of a dynamic health correction system and calculates optimal weights to dynamically and rationally decompose the total voltage compensation task into specific output voltage sub-instructions for each normally functioning power unit. When the performance of some units deteriorates, its allocation ratio is automatically reduced, and the task is transferred to healthy units, achieving fault tolerance and derating while maintaining continuous and stable system operation.

[0201] During system operation, the parameterization module first establishes an accurate system baseline model. The feedforward module then uses this baseline model to construct a controller, enabling high-speed feedforward compensation for grid disturbances and generating a general command. The fault-tolerant module optimally distributes the general command to each power unit based on its real-time performance parameters. This results in not only fast response and high compensation accuracy, but also automatic reconfiguration of the control strategy to maintain optimal compensation performance in the event of aging or minor faults in internal units. This improves the reliability, adaptability, and overall lifespan of the industrial grid voltage fluctuation suppression system.

[0202] It is important to note that the constructions and arrangements of this application shown in several different exemplary embodiments are merely illustrative. Although only two embodiments are described in detail in this disclosure, those who consult this disclosure will readily understand that many modifications are possible without substantially departing from the novel teachings and advantages of the subject matter described in this application. These modifications may include, for example, changes in the size, dimensions, structure, shape, and proportions of various elements, as well as parameter values ​​(e.g., temperature, pressure, etc.), installation arrangements, the use of materials, colors, orientations, etc. For example, an element shown as integrally formed may be composed of multiple parts or elements, the position of elements may be inverted or otherwise altered, and the nature or number or position of discrete elements may be changed or altered. Therefore, all such modifications are intended to be included within the scope of this application. The order or sequence of any process or method steps may be changed or rearranged by alternative embodiments. Any "apparatus plus function" clause is intended to cover, and not only structurally equivalent but also equivalent structures, the structures performing the functions described herein. Other substitutions, modifications, alterations, and omissions may be made in the design, operation, and arrangement of the exemplary embodiments without departing from the scope of this application. Therefore, this application is not limited to a particular embodiment, but extends to various modifications that still fall within the scope of the appended claims.

[0203] Furthermore, in order to provide a concise description of exemplary embodiments, not all features of actual embodiments (i.e., those features that are not relevant to the best mode of performing this application as currently considered, or those features that are not relevant to implementing this application) may be omitted.

[0204] It should be understood that numerous specific implementation decisions can be made during the development of any practical implementation, such as in any engineering or design project. Such development efforts may be complex and time-consuming, but for those of ordinary skill in the art who benefit from this disclosure, the development effort will be a routine task in design, manufacturing, and production without requiring extensive experimentation.

[0205] It should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application, and all such modifications and substitutions should be covered within the scope of the claims of this application.

Claims

1. A method for suppressing voltage fluctuations in industrial power grids, characterized in that, include: S1 is a voltage and current forward propagation model composed of the transmission parameters in the power unit, filter circuit and grid impedance of the cascaded compensation system. The power unit, filter circuit, and grid impedance are respectively characterized as a two-port network of the power unit, a two-port network of the filter circuit, and a two-port network of the grid impedance. The voltage and current forward propagation model is obtained by calculating the product of the ABCD transmission matrices of each two-port network. S2 detects the voltage and current disturbances in the power grid in real time and inputs them into the feedforward controller, which is built in reverse based on the voltage and current forward propagation model, to generate the main compensation voltage command. S3 acquires the DC-side voltage deviation, output current distortion rate, and device temperature parameters of each power unit. It constructs a differential equation using the DC-side voltage deviation, output current distortion rate, and device temperature parameters, and obtains real-time health indicators through fuzzy reasoning via membership functions and reasoning criteria. When constructing the real-time health index, a differential equation characterizing the dynamic process of performance degradation is first constructed using the parameters. The method for constructing the differential equation is as follows: S301 establishes a state equation with the DC side voltage deviation, output current distortion rate, and device junction temperature as state variables; The state equation described in S302 includes a linear term describing the decay trend of each state variable, a coupling term describing the nonlinear influence between each state variable, and an external input term characterizing the excitation effect of the current output power on the state change. S4 corrects the equivalent impedance parameters in the voltage-current forward transfer model using the real-time health index and calculates the voltage allocation weights for each power unit. S5 decomposes the total compensation voltage command into modulation signals for each unit through the voltage allocation weight, thereby modulating the industrial power grid.

2. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 1, characterized in that: Obtain the total ABCD transfer matrix corresponding to the voltage-current forward transfer model; The inverse of the total ABCD transmission matrix is ​​performed to obtain the inverse matrix model of ABCD. The real-time detected grid voltage disturbance and current disturbance are combined to form an input vector, which is then used in matrix operation with the inverse matrix model. The voltage component of the output vector is the main compensation voltage command.

3. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 1, characterized in that: The methods for obtaining the DC-side voltage deviation, output current distortion rate, and device temperature parameters are as follows: At a sampling rate higher than the base frequency, the instantaneous DC-side voltage, instantaneous AC output current, and device temperature sensor readings of each power unit are collected synchronously. The DC-side voltage deviation is obtained by subtracting the average value of the DC-side voltage from the reference voltage value using a sliding window averaging algorithm. The total harmonic distortion rate is calculated as the output current distortion rate by performing a fast Fourier transform on the instantaneous value of the AC output current within the window.

4. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 1, characterized in that: The solution results of the differential equation, namely the instantaneous comprehensive magnitude of the state variables and the comprehensive magnitude of their instantaneous rate of change, are used as the input variables for fuzzy inference. Define membership functions ranging from low to high for both instantaneous composite amplitude and composite amplitude; Establish fuzzy inference rules, which include: when the instantaneous amplitude and rate of change of the state variables are all at high levels, it is determined that the health is rapidly deteriorating; When the instantaneous amplitude and rate of change of the state variables are both at low levels, the health status is considered to be good.

5. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 4, characterized in that: The output of the fuzzy inference system is defuzzified to obtain a preliminary health score. This preliminary score is then weighted and fused with a remaining life assessment factor predicted based on the long-term integral of the aforementioned differential equation. The weighting coefficients in the weighted fusion are dynamically adjusted based on the total output power currently borne by the compensation system, ultimately generating the real-time health index.

6. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 5, characterized in that: The real-time health index is mapped to the per-unit correction factor of the equivalent series impedance of the power unit in the forward pass model. Based on the corrected equivalent series impedance of each power unit, calculate the reciprocal of its proportion in the total equivalent impedance of the system, and normalize the reciprocal of the proportion to obtain the initial voltage allocation weight of each power unit. The initial voltage allocation weights are weighted geometrically and the real-time health indicators of the corresponding power units are used to generate voltage allocation weights. Using the real-time health index, for power units whose health is below a preset threshold, a virtual admittance is connected in parallel with the corresponding equivalent impedance parameter to simulate the state of being partially bypassed. Based on the corrected system equivalent impedance network, with the goal of maintaining the balance of the total output voltage of the system, the optimal voltage distribution weight of each power unit is solved so that the current stress flowing through each healthy unit tends to be consistent. Establish a function with the optimization objectives of minimizing the system output impedance magnitude and achieving the most balanced voltage distribution among units; The real-time health index is introduced as a constraint into the function of the optimization objective; By solving the constrained optimization problem, the corrected values ​​of the system's equivalent impedance parameters and the optimal voltage allocation weights of each power unit in the forward transfer model are obtained simultaneously.

7. The method for suppressing voltage fluctuations in an industrial power grid as described in claim 6, characterized in that: The total compensation voltage command is multiplied sequentially by the voltage allocation weight of each power unit obtained in step S4 to determine the specific output voltage command that each power unit needs to undertake. Divide the specific output voltage command of each power unit by its measured DC-side voltage value to obtain the corresponding modulation wave signal; Each modulated wave signal is compared with a carrier signal to generate a pulse width modulation signal for driving the switching devices in the corresponding power unit; the carrier signal adopts an alternating anti-symmetrical triangular carrier, and the phase of the carrier of each power unit is evenly distributed between 0 and 360 degrees according to the unit number.

8. A novel system for suppressing voltage fluctuations in industrial power grids, characterized in that... This includes a method for suppressing voltage fluctuations in an industrial power grid as described in any one of claims 1-7; including a compensation main circuit composed of multiple cascaded power units; The parameterization module includes a cascade unit and a calculation unit. The cascade unit models the compensation main circuit, output filter and known grid impedance parameters as multiple cascade sub-units. The calculation module includes obtaining the voltage and current forward propagation model of the system by calculating the combination of transmission parameters of each functional module. The feedforward module constructs a feedforward controller through the inverse matrix model of the voltage and current forward propagation model. The feedforward controller takes the detected instantaneous disturbances in grid voltage and current as inputs and directly generates a preliminary compensation voltage command. The fault-tolerant module will optimize and allocate the synthesized total compensation voltage command in real time based on the number of power units currently operating normally and their performance parameters.

Images