Power distribution network impedance rapid identification method and system based on adaptive extremum search

Abstract

The invention discloses a power distribution network impedance rapid identification method based on adaptive extremum search. The method comprises the following steps: A, constructing an equivalent circuit model of a power distribution network node; b, establishing an extreme value relation between an injection current angle and a node voltage amplitude, and performing impedance calculation; c, a disturbance signal is injected into the equivalent circuit model, and then the signal is demodulated; d, adjusting the amplitude of the demodulation signal; and E, calculating the impedance of the power distribution network. According to the method, the defects in the prior art can be overcome, the method is suitable for various application scenes with rapid convergence requirements for impedance recognition, the problems of convergence speed, stability, dependence on a power grid model and the like in the prior art are effectively solved, and efficient and reliable recognition of the impedance of the power distribution network is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system analysis and control, and in particular to a method and system for rapid identification of distribution network impedance based on adaptive extremum seeking. Background Art

[0002] The configuration of the power system and the continuous change of the load cause the impedance of the distribution network to change continuously. This change has an impact on the stable operation of the power grid. In aspects such as power grid stability analysis, relay protection decision-making, voltage optimization regulation, island detection, and short-circuit calculation, it is crucial to accurately grasp the impedance change situation.

[0003] Passive methods do not require injecting signal disturbances into the power grid, but face many challenges. For example, in practical applications, the Extended Kalman Filter (EKF) needs to perform complex adjustments on unknown noise parameters to obtain accurate results; the Recursive Least Squares (RLS) relies on the power grid model and accurate phase and frequency estimation to achieve the dq reference coordinate system transformation, otherwise it is difficult to accurately estimate the impedance.

[0004] Although active methods can estimate the impedance more accurately by injecting disturbances and have relatively low complexity, they also have defects. Some active methods need to inject single-frequency signals, voltage or current pulses, or even full-spectrum signals, and may require complex post-processing, high pulse power injection, and long estimation time to obtain accurate results. Some methods based on Akagi's active and reactive power injection technology (p-q technology) can estimate the impedance, but there are still various limitations.

[0005] As a model-free control technology, Extremum Seeking Control (ESC) relies on feedback to find the extremum of a static mapping or optimize the parameters of a dynamic system. Since the mid-1990s, it has been applied in many fields, such as process optimization, anti-lock braking systems, and intelligent converter voltage regulation. In terms of impedance estimation applications, although ESC has certain advantages, previous studies have shown that there are problems in adjusting its convergence speed. Changing the convergence speed may lead to a decrease in output quality, or inject unnecessary large disturbances into the system, affecting system stability and the accuracy of impedance estimation.

[0006] In the actual operation of the distribution network, for some applications (such as stability margin calculation, voltage regulation, fault analysis, etc.), it is necessary to calculate the impedance value as soon as possible. However, the existing impedance estimation methods and technologies still have deficiencies in meeting the requirements of quickly and accurately identifying the impedance of the distribution network.

[0007] Therefore, there is an urgent need for a new method to improve the speed and accuracy of distribution network impedance identification to meet the requirements of stable operation and optimal control of the distribution network. Summary of the Invention

[0008] The technical problem to be solved by the present invention is to provide a method and system for quickly identifying the impedance of a distribution network based on adaptive extremum search, which can solve the deficiencies of the existing technology, is applicable to various application scenarios with a fast convergence requirement for impedance identification, effectively solves the problems of the existing technology in terms of convergence speed, stability, and dependence on the power grid model, and realizes the efficient and reliable identification of the impedance of the distribution network.

[0009] To solve the above technical problems, the technical inventions adopted by the present invention are as follows.

[0010] A method for quickly identifying the impedance of a distribution network based on adaptive extremum search includes the following steps:

[0011] A. Construct an equivalent circuit model of the nodes in the distribution network;

[0012] B. Establish an extremum relationship between the injection current angle and the node voltage amplitude and perform impedance calculation;

[0013] C. Inject a perturbation signal into the equivalent circuit model and then perform demodulation processing on the signal;

[0014] D. Adjust the amplitude of the demodulated signal;

[0015] E. Calculate the impedance of the distribution network.

[0016] Preferably, the construction process of the equivalent circuit model includes the following steps.

[0017] At node j of the distribution network, the grid-connected inverter is modeled as a controlled current source, and the controlled current source injects current into the power grid. The current is expressed as I j ∠θ, where θ is the angle of the injected current; node j is equivalent to a Thevenin voltage source Vj with an angle of reference 0 degrees 0 in series with a Thevenin impedance Z th ∠α; the voltage expression of node j is Convert the above formula into an expression of the voltage amplitude, and the expression of the voltage amplitude is obtained as Using the double-angle formula of trigonometric function cos 2x = 1 - 2sin 2 x to convert cos(θ + α) in the above formula, we get α is the angle of the Thevenin impedance Z th , and I j is the amplitude of the current injected by the grid-connected inverter at node j.

[0018] Preferably, establishing the extremum relationship between the injection current angle and the node voltage amplitude includes the following steps.

[0019] According to the expression of the voltage amplitude, when i.e., θ = -α, the voltage amplitude (Vj ) 2 reaches the maximum value, at which time Thus, by finding the injection current angle that maximizes the voltage amplitude, the Thevenin impedance angle α is then determined.

[0020] Preferably, the impedance calculation includes the following steps.

[0021] At the extreme point By transposing, we get The problem of identifying the impedance of the distribution network is transformed into the problem of finding the extreme points in the voltage balance diagram; in the process of finding the extreme points, the simulated annealing algorithm is used to optimize the search process, and a worse solution is accepted with a preset probability during the search to jump out of the local optimal solution, so as to find the injection current angle that maximizes the voltage amplitude; combined with the impedance change trend predicted by the recurrent neural network under different operating conditions, the calculated impedance amplitude is verified and corrected.

[0022] Preferably, injecting a perturbation signal into the equivalent circuit model and demodulating the signal include the following steps.

[0023] First, use a signal generator to inject a fixed frequency into the equivalent circuit model. The perturbation signal is selected in the form of a sine wave of asin(ωt), where a is the amplitude of the perturbation signal. This perturbation signal is modulated with the input of the equivalent circuit model, and the input of the equivalent circuit model is the current injected by the grid-connected inverter.

[0024] The node voltage output signal of the equivalent circuit model after injecting the perturbation is monitored in real time, and the node voltage is set as V; the node voltage output signal is first processed by a high-pass filter. The cut-off frequency of the high-pass filter is higher than the perturbation signal frequency ω, which is used to remove the low-frequency components in the output signal. The low-frequency components include the fundamental wave signal and low-frequency interference signal in the power grid, and then a high-frequency signal containing perturbation-related information is obtained.

[0025] Then, multiply the signal after high-pass filtering by the signal of the same frequency as the injected perturbation, sin(ωt), for signal demodulation, so as to extract the amplitude and phase information related to the perturbation, and then obtain the perturbation signal ξ.

[0026] In order to extract the gradient information, the signal after demodulation processing is passed through a high-pass filter again to remove the DC component, and then the filtered signal is multiplied by the signal of the same frequency as the injected perturbation to obtain a signal related to the gradient.

[0027] Preferably, adjusting the amplitude of the demodulated signal includes the following steps.

[0028] If the change of the signal related to the gradient exceeds the set value, it indicates that the equivalent circuit model is experiencing external interference. When the external interference ends, increase the amplitude of the demodulated signal; the function describing the relationship between the output and input of the equivalent circuit model is where f (2) > 0 represents a convex function, corresponding to a minimization problem, and f (2) < 0 represents a concave function, corresponding to a maximization problem, and the optimization objective is to minimize (θ - θ * ), so that the objective function f(θ) approaches the minimum value f * ;

[0029] Let be the estimated value of the optimal θ * The estimation error After adding the perturbation signal Substitute the expression of θ into f(θ) to obtain the expression of, θ * is the optimal current injection angle converged through the adaptive extremum search process in the distribution network impedance identification;

[0030] Ignoring the high-order terms, when the dynamics increase, the convergence speed of the equivalent circuit model increases and finally converges to the optimal value When Kab(t) > 0, the equivalent circuit model is stable and tends to the minimum value.

[0031] Preferably, calculating the distribution network impedance includes the following steps

[0032] When the equivalent circuit model converges to the optimal current injection angle θ * = -α, measure the final voltage value V at the point of common coupling j , given the injection current value I at this time j and the Thevenin voltage calculated through the initial state of the equivalent circuit model Substitute into After that, calculate the magnitude of the Thevenin impedance, so as to obtain the value of the Thevenin impedance at the point of common coupling and complete the identification process of the distribution network impedance.

[0033] A distribution network impedance fast identification system based on adaptive extremum search is used to implement the above-mentioned distribution network impedance fast identification method based on adaptive extremum search, including

[0034] A modeling module, which is used to model the grid-connected inverter as a controlled current source and describe the current output behavior of the grid-connected inverter by establishing a corresponding mathematical model; according to the Thevenin equivalent circuit principle of the grid node, establish an equivalent circuit model between the node voltage, current and impedance, determine the parameters of the Thevenin voltage source and Thevenin impedance for each node, and construct the expression of the node voltage;

[0035] A perturbation injection module, which is used to generate and inject a perturbation signal with a fixed frequency into the equivalent circuit model, and the parameters of the perturbation signal;

[0036] A signal processing module, which is used to remove low-frequency components through a high-pass filter, and then perform a demodulation operation, multiply the filtered signal by a signal with the same frequency as the injected disturbance, and extract disturbance information ξ and gradient information;

[0037] A demodulated signal amplitude adjustment module, which is used to monitor the gradient change of the equivalent circuit model, judge whether the equivalent circuit model is affected by external interference by comparing with a preset threshold; when it is detected that the external interference causes abnormal gradient change, wait until the interference ends, that is, after the gradient returns to the normal range, adaptively increase the demodulated signal amplitude according to a predetermined strategy; during the adjustment process, ensure the stability and convergence of the entire system by controlling the increase amplitude and rate of the demodulated signal amplitude, and achieve the best impedance identification effect; when a power grid fault or abnormal working condition is detected, adjust the parameters and strategies of the adaptive extreme value search control algorithm; for different types of faults and abnormal working conditions, establish a corresponding impedance identification model library in advance, and select a suitable model for impedance identification according to the fault type;

[0038] An impedance calculation module, which is used to calculate the impedance after the equivalent circuit model converges through the adaptive extreme value search process, using the measured voltage value V at the common coupling point j and the known injected current value I j and the related Thevenin voltage value to calculate the impedance.

[0039] The beneficial effects brought by adopting the above technical invention are as follows: By proposing a method for adaptively adjusting the demodulated signal amplitude, the present invention dynamically changes the demodulated signal amplitude according to the external interference situation of the system while keeping the injected disturbance amplitude constant, thereby accelerating the convergence speed of the system, enabling the system to more quickly identify the impedance of the distribution network, and providing accurate impedance data for power grid operation control in a timely manner. When adjusting the convergence speed, the present invention only changes the demodulated signal amplitude while keeping the disturbance amplitude unchanged, avoiding the adverse effects on the system stability caused by the change of disturbance power. At the same time, by adding a module to observe the response changes of the system to "injected disturbance (IP) + external disturbance (ED)" and only to IP, and increasing the demodulated signal amplitude in a timely manner after the external disturbance ends, it can not only accelerate the convergence but also ensure the system stability, enabling the system to operate stably throughout the impedance identification process. The present invention is based on adaptive extreme value search control and does not require an accurate power grid model, complex estimation techniques, phasor measurement units or high-power pulse injection to estimate the power grid impedance. By utilizing the extreme value relationship between the injected current angle and the voltage at the common coupling point (PCC), the impedance information is directly extracted from the system response, simplifying the calculation process, reducing the dependence on the power grid model and other complex devices and technologies, and improving the generality and practicality of the method. Description of the Drawings

[0040] Figure 1It is a schematic diagram of the equivalent circuit model of the present invention.

[0041] Figure 2 It is a voltage extreme value relationship diagram with one extreme value of the voltage at node j in the present invention.

[0042] Figure 3 It is a schematic diagram of the extreme value search controller with a gain adjustment module for demodulating signals in the present invention.

[0043] Figure 4 It is a schematic diagram of the basic extreme value search controller with a time-varying demodulation signal amplitude in the present invention. Detailed implementation manners

[0044] I. System modeling and problem transformation.

[0045] Model the grid-connected inverter as a controlled current source, and establish a relationship model between the node voltage, the injected current, the Thevenin voltage, and the Thevenin impedance based on the Thevenin equivalent circuit. Through the analysis of this model, determine the extreme value relationship between the injected current angle and the node voltage amplitude (when the injected current angle θ = -α, the node voltage amplitude reaches the maximum value), thereby transforming the problem of identifying the distribution network impedance into the problem of finding the extreme value point in the voltage balance diagram, laying a foundation for subsequent impedance identification based on extreme value search. The accuracy of this link directly affects whether the optimal current injection angle can be accurately found subsequently to calculate the impedance.

[0046] At node j of the distribution network, model the grid-connected inverter as a controlled current source. The controlled current source injects current into the grid, and its current is expressed as I j ∠θ, where θ is the angle of the injected current; equivalent node j to a Thevenin voltage source V with an angle of reference 0 degrees j 0 in series with a Thevenin impedance Z th ∠α; the voltage expression of node j is Convert the above formula into an expression of the voltage amplitude, and the expression of the voltage amplitude is obtained as

[0047] Use the double-angle formula of trigonometric function cos 2x = 1 - 2sin 2 x to convert cos(θ + α) in the above formula, and get ɑ is the angle of the Thevenin impedance Z th and is determined by finding the injected current angle when the voltage amplitude is the largest in the identification of the distribution network impedance, and there is an extreme value relationship with the injected current angle θ (when (θ = -ɑ), the voltage amplitude reaches the maximum value). I j is the amplitude of the current injected by the grid-connected inverter at node j and is a key electrical quantity when analyzing the relationship between the node voltage and the impedance. Z this the Thevenin impedance at the node, which reflects the impedance characteristics of the equivalent circuit of the distribution network node. It is the target quantity for the impedance identification of the distribution network and can be calculated through specific voltage and current relationships. Additionally, by integrating meteorological data such as environmental temperature and humidity and equipment operation status data (such as transformer oil temperature, winding temperature, etc.), using the Bayesian fusion algorithm, comprehensively considering the influence of various factors on the relationship between node voltage and impedance, a node equivalent circuit model that better fits the actual operating conditions is constructed.

[0048] According to the expression of the voltage amplitude, when i.e., θ = -α, the voltage amplitude (V j ) 2 reaches the maximum value. At this time, Thus, by finding the injection current angle that maximizes the voltage amplitude, the Thevenin impedance angle α is then determined.

[0049] At the extreme point By transposing terms, we get The problem of distribution network impedance identification is transformed into the problem of finding the extreme points in the voltage balance diagram; during the process of finding the extreme points, the simulated annealing algorithm is used to optimize the search process. When searching, a worse solution is accepted with a preset probability to jump out of the local optimal solution, so as to find the injection current angle that maximizes the voltage amplitude; combined with the impedance change trend predicted by the recurrent neural network under different operating conditions, the calculated impedance amplitude is verified and corrected.

[0050] II. Impedance identification based on adaptive extreme value search and adaptive adjustment of the demodulation signal amplitude.

[0051] The core of this method is to inject a fixed-frequency perturbation signal into the system and, combined with a series of signal processing and adaptive adjustment mechanisms, enable the system to quickly and accurately find the impedance extreme points in the dynamically changing distribution network environment, thereby achieving impedance identification. This adaptive adjustment mechanism is different from the traditional ESC method and can dynamically change the demodulation signal amplitude according to the external interference situation of the system. On the premise of keeping the perturbation amplitude constant, it improves the convergence speed of the system and ensures the system stability, effectively solving the problems of slow convergence speed and affected stability of the traditional method when facing impedance changes.

[0052] First, a signal generator is used to inject a fixed frequency into the equivalent circuit model. The perturbation signal selects the sine wave form of asin(ωt), where a is the amplitude of the perturbation signal. This perturbation signal is modulated with the input of the equivalent circuit model, and the input of the equivalent circuit model is the current injected by the grid-connected inverter;

[0053] The node voltage output signal of the equivalent circuit model after injecting a perturbation is monitored in real time, and the node voltage is set as V. The node voltage output signal is first processed by a high-pass filter. The cut-off frequency of the high-pass filter is higher than the perturbation signal frequency ω, which is used to remove the low-frequency components in the output signal. The low-frequency components include the fundamental wave signal and low-frequency interference signal in the power grid, and then a high-frequency signal containing perturbation-related information is obtained.

[0054] Then, the signal after high-pass filtering is multiplied by the signal sin(ωt) with the same frequency as the injected perturbation for signal demodulation, so as to extract the amplitude and phase information related to the perturbation, and then the perturbation signal ξ is obtained.

[0055] In order to extract the gradient information, the signal after demodulation processing is passed through a high-pass filter again to remove the DC component, and then the filtered signal is multiplied by the signal with the same frequency as the injected perturbation to obtain a signal related to the gradient.

[0056] If the change of the signal related to the gradient exceeds the set value, it indicates that the equivalent circuit model is experiencing external interference. When the external interference ends, the amplitude of the demodulated signal is increased. The function describing the relationship between the output and input of the equivalent circuit model is where f (2) > 0 represents a convex function, corresponding to a minimization problem, f (2) < 0 represents a concave function, corresponding to a maximization problem. The optimization objective is to minimize (θ - θ * ), so that the objective function f(θ) approaches the minimum value f * ;

[0057] Let be the estimated value of the optimal θ * , and the estimation error After adding the perturbation signal Substitute the expression of θ into f(θ) to obtain the expression of, θ * is the optimal current injection angle converged through an adaptive extreme value search process in the distribution network impedance identification;

[0058] Ignoring the high-order terms, when increasing dynamically, the convergence speed of the equivalent circuit model increases and finally converges to the optimal value When Kab(t) > 0, the equivalent circuit model is stable and tends to the minimum value.

[0059] Meanwhile, during the entire adaptive adjustment process, the amplitude of the injected disturbance signal a is kept unchanged all the time, thus avoiding the adverse effects on the system stability caused by the change of disturbance power. The variation of the demodulation signal amplitude with time in the basic extremum search controller helps to intuitively understand the dynamic variation process of the demodulation signal amplitude (b) with time in the adaptive ESC method of the present invention, and how this variation is associated with the improvement of the system convergence speed, and supplements the description of the action mechanism of dynamically increasing b on improving the convergence speed. Meanwhile, when a grid fault (such as a short-circuit fault or a disconnection fault) or an abnormal condition (such as a voltage sag or harmonic over-standard) is detected, the fault detection and diagnosis module automatically adjusts the parameters and strategies of the adaptive extremum search control algorithm. For example, when a short-circuit fault occurs, the amplitude of the disturbance signal is appropriately increased to enhance the sensitivity of the system to the impedance change after the fault, and the demodulation signal amplitude adjustment strategy is adjusted to accelerate the system convergence. For different types of faults and abnormal conditions, a corresponding impedance identification model library is established in advance, and a suitable model is selected according to the fault type for impedance identification.

[0060] The technical points are as follows:

[0061] Disturbance injection and signal modulation technology: precisely control parameters such as the frequency and amplitude of the disturbance signal to effectively modulate it with the system input, providing a basic signal source for subsequent signal processing and impedance identification. For example, select a suitable form of sine wave disturbance signal and determine its initial amplitude and other parameters to stimulate an observable response of the system without affecting the normal operation of the system.

[0062] Signal processing and demodulation technology: perform operations such as high-pass filtering and demodulation on the system output signal to extract information related to the disturbance (such as the disturbance signal ξ) and gradient information. This involves the design of a high-pass filter, whose cut-off frequency needs to be reasonably set according to the disturbance signal frequency to accurately remove the low-frequency components and retain the high-frequency signals related to impedance identification; at the same time, operations such as multiplication in the demodulation process need to ensure accuracy to ensure that useful information can be extracted from complex signals.

[0063] Adaptive adjustment of demodulation signal amplitude technology: The new module observes the response changes of the system to "injected disturbance (IP) + external disturbance (ED)" and only to IP. After the external disturbance ends, according to the gradient change of the system, it adaptively increases the demodulation signal amplitude b. The key to this technology lies in accurately judging the occurrence and end of the external disturbance, and reasonably designing the increasing strategy of b so that it can not only accelerate the convergence speed but also not affect the system stability. For example, detect the external disturbance by setting a reasonable gradient change threshold, and determine the increasing method (such as linear increase, exponential increase, etc.) and amplitude range of b according to theoretical derivation and simulation experiments.

[0064] III. Impedance calculation and result output.

[0065] This process involves multiple signal processing steps and parameter calculations, which is a key link to achieve accurate impedance identification.

[0066] When the equivalent circuit model converges to the optimal current injection angle θ * = -α, measure the final voltage value V at the point of common coupling j , knowing the injected current value I at this time j and the Thevenin voltage calculated from the initial state of the equivalent circuit model Substitute into to calculate the magnitude of the Thevenin impedance, thereby obtaining the value of the Thevenin impedance at the point of common coupling and completing the identification process of the distribution network impedance.

[0067] The technical points are as follows:

[0068] Extreme value relationship modeling and analysis: Accurately construct a node voltage model based on the Thevenin equivalent circuit, and determine the extreme value relationship between the injection current angle and the voltage magnitude through mathematical derivation and analysis. This requires a deep understanding of circuit principles and mathematical analysis methods to ensure that the model can accurately reflect the characteristics of the actual system and the correctness and uniqueness of the extreme value relationship. For example, correctly apply Kirchhoff's voltage law, complex number operations, etc. in the derivation process, and reasonably simplify and analyze the node voltage expression to obtain clear extreme value conditions.

[0069] Impedance calculation algorithm: Design an accurate impedance calculation algorithm based on the determined extreme value relationship and the measured voltage value at the PCC and the known injected current value. This algorithm needs to consider the influence of actual factors such as measurement errors and system noise to ensure the accuracy and reliability of the calculation results. For example, data filtering, error correction and other technologies may be required in the algorithm to improve the accuracy of the calculation results.

[0070] IV. Composition of the distribution network impedance identification system.

[0071] Modeling module, used to model the grid-connected inverter as a controlled current source, describe the current output behavior of the grid-connected inverter by establishing a corresponding mathematical model; according to the Thevenin equivalent circuit principle of the power grid node, establish an equivalent circuit model between the node voltage, current and impedance, determine the parameters of the Thevenin voltage source and Thevenin impedance for each node, and construct an expression of the node voltage;

[0072] Perturbation injection module, used to generate and inject a perturbation signal with a fixed frequency into the equivalent circuit model, the parameters of the perturbation signal;

[0073] Signal processing module, used to remove the low-frequency components through a high-pass filter, then perform a demodulation operation, multiply the filtered signal by the signal with the same frequency as the injected perturbation, and extract the perturbation information ξ and gradient information;

[0074] The demodulation signal amplitude adjustment module is used to monitor the gradient change of the equivalent circuit model. By comparing with a preset threshold, it determines whether the equivalent circuit model is affected by external interference. When it detects that the external interference causes abnormal gradient change, it waits for the interference to end, that is, when the gradient returns to the normal range, and then adaptively increases the demodulation signal amplitude according to a predetermined strategy. During the adjustment process, by controlling the increase amplitude and rate of the demodulation signal amplitude, it ensures the stability and convergence of the entire system and achieves the best impedance identification effect. When it detects a power grid fault or abnormal working condition, it adjusts the parameters and strategies of the adaptive extreme value search control algorithm. For different types of faults and abnormal working conditions, it establishes a corresponding impedance identification model library in advance and selects a suitable model for impedance identification according to the fault type.

[0075] The impedance calculation module is used to calculate the impedance by using the measured voltage value V at the common coupling point j and the known injected current value I j and the related Thevenin voltage value after the equivalent circuit model converges through the adaptive extreme value search process.

[0076] The technical key points are as follows:

[0077] Module function implementation: Ensure that each module can accurately implement its predetermined function. For example, the modeling module needs to accurately model the grid-connected inverter as a controlled current source and establish a correct voltage-current-impedance relationship model; each sub-unit (perturbation injection unit, signal processing unit, demodulation signal amplitude adaptive adjustment unit) in the adaptive extreme value search module needs to work together to complete operations such as perturbation injection, signal processing, and adaptive adjustment; the impedance calculation module needs to obtain accurate voltage and current values at the right time and calculate the impedance using the correct algorithm.

[0078] Module interface and data transfer: Design clear and efficient module interfaces to ensure accurate and timely data transfer between modules. For example, the adaptive extreme value search module needs to transfer information such as grid model parameters and injected current to the modeling module, and transfer data such as the converged current injection angle and voltage value to the impedance calculation module to ensure the smooth flow of information in the entire system and avoid system failures or calculation errors caused by data loss or incorrect transfer.

[0079] Accurately calculate the magnitude of the Thevenin impedance, thereby obtaining the specific value of the Thevenin impedance at the PCC, and output the impedance identification result to provide key impedance data for the operation analysis and control decision-making of the distribution network. At the same time, increase the redundant configuration of sensors. When the sensors responsible for measuring the voltage and injected current at the PCC fail or have abnormal measurements, other redundant sensors can replace them in time to ensure the continuous and stable operation of the system.

[0080] The present invention has the following many advantages.

[0081] 1. Fast and accurate impedance identification

[0082] Improve the convergence speed: By adaptively adjusting the amplitude of the demodulation signal, the present invention can accelerate the system's convergence to the optimal impedance value while keeping the amplitude of the injection disturbance constant. In the distribution network, whether it is impedance monitoring under normal operating conditions or in the face of impedance changes caused by power grid faults, load changes, etc., accurate impedance information can be obtained more quickly. For example, in application scenarios with extremely high real-time requirements such as stability margin calculation, voltage regulation, and fault analysis, compared with traditional methods (such as the traditional extremum search control method with slow convergence speed and inability to adapt to impedance changes in a timely manner), the present invention can converge faster and provide accurate impedance data for power grid operation control in a timely manner, thus ensuring the stable operation of the power grid.

[0083] Accurately identify the impedance: Based on the effective utilization of the relationship between the angle of the injected current and the voltage extremum at the point of common coupling (PCC), as well as precise signal processing and adaptive adjustment mechanisms, the present invention can accurately calculate the impedance value of the distribution network. In the actual power grid, accurate impedance identification is crucial for analyzing power grid characteristics, optimizing control strategies, etc., avoiding problems such as low power grid operation efficiency and incorrect control decisions caused by inaccurate impedance estimation.

[0084] 2. Enhance system stability

[0085] Avoid the influence of disturbances on stability: During the process of adjusting the convergence speed, the present invention keeps the amplitude of the injection disturbance unchanged and only changes the amplitude of the demodulation signal. This feature effectively avoids the system instability problems that may be brought about by traditional methods when changing the convergence speed by increasing the disturbance power injected into the system, such as increased output signal fluctuations and increased harmonic content. In a complex distribution network environment, stable system operation is of great significance for ensuring power quality and protecting power equipment.

[0086] Adaptively cope with external disturbances: The newly added module can keenly observe the system's response to external disturbances. After the dynamic changes caused by external disturbances end, it timely increases the amplitude of the demodulation signal, enabling the system to quickly recover and stably converge to the optimal value after the disturbance. This makes the system have stronger adaptability and stability in the face of common external disturbances in the power grid (such as intermittent fluctuations of distributed power sources, sudden changes in loads, etc.), ensuring the continuity and accuracy of the impedance identification process.

[0087] 3. Reduce complexity and model dependence

[0088] Simplify the calculation process: The present invention does not require complex adjustment of unknown noise parameters like the Extended Kalman Filter (EKF), nor does it need to rely on an accurate power grid model and phase and frequency estimation as in the Recursive Least Squares (RLS). Based on adaptive extremum seeking control, impedance information is directly extracted from the system response, simplifying the calculation process, reducing the consumption of computing resources, and improving the real-time performance and practicality of the method.

[0089] Improve generality: Since it does not depend on a complex power grid model and specific devices (such as phasor measurement units) or high-power pulse injection, the present invention can be more conveniently applied to distribution networks with different structures and operating states. Whether it is a small-scale distributed distribution network or a large-scale complex distribution network, the impedance identification can be achieved using the present invention, which has broad applicability and generality, providing strong support for the intelligent development of distribution networks.

[0090] 4. Effectively support various application scenarios

[0091] The adaptive technology proposed by the present invention has strong generality and can be easily applied to any ESC controller system that requires fast convergence, not limited to the field of distribution network impedance identification. It can also provide useful references and effective solutions for the optimal control of other similar systems, having broad application prospects and being able to play its advantages in different system environments, promoting the development and progress of related technical fields.

[0092] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "longitudinal", "transverse", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as a limitation to the present invention.

[0093] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. What is described in the above embodiments and the specification only illustrates the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A fast impedance identification method for distribution networks based on adaptive extreme value search, characterized in that The method includes the following steps: A. Construct an equivalent circuit model of the distribution network nodes; B. Establish the extreme value relationship between the injection current angle and the node voltage amplitude and perform impedance calculation; C. Inject a disturbance signal into the equivalent circuit model and then perform demodulation processing on the signal; D. Adjust the amplitude of the demodulated signal; E. Calculate the distribution network impedance.

2. The fast impedance identification method for a distribution network based on adaptive extremum search according to claim 1, characterized in that: The construction process of the equivalent circuit model includes the following steps: At the node j of the distribution network, the grid-connected inverter is modeled as a controlled current source that injects current into the grid, and the current is represented as I j ∠θ, where θ is the angle of the injected current; Equivalent node j to a Thevenin voltage source V with an angle of reference 0 degrees j 0 in series with a Thevenin impedance Z th ∠α; the voltage expression of node j is Convert the above formula to an expression of voltage amplitude, and the expression of voltage amplitude is obtained as Use the double-angle formula of trigonometric function cos2x = 1 - 2sin 2 x to convert cos(θ + α) in the above formula, and we get α is the angle of the Thevenin impedance Z th and I j is the amplitude of the current injected by the grid-connected inverter at node j.

3. The fast impedance identification method for a distribution network based on adaptive extremum search according to claim 2, characterized in that: The establishment of the extreme value relationship between the injection current angle and the node voltage amplitude includes the following steps: According to the expression of the voltage amplitude, when i.e., θ = -α, the voltage amplitude (V j ) 2 reaches the maximum value. At this time Therefore, by finding the injection current angle that maximizes the voltage amplitude, the Thevenin impedance angle α can be determined.

4. The method for quickly identifying the impedance of a distribution network based on adaptive extremum search according to claim 3, characterized in that: The impedance calculation includes the following steps: At the extreme point By transposing terms, we get The problem of identifying the impedance of the distribution network is transformed into the problem of finding the extreme points in the voltage balance diagram. During the process of finding the extreme points, the simulated annealing algorithm is used to optimize the search process. When searching, a worse solution is accepted with a preset probability to jump out of the local optimal solution, so as to find the injection current angle that maximizes the voltage amplitude. Combined with the impedance change trend under different operating conditions predicted by the recurrent neural network, verify and correct the calculated impedance amplitude.

5. The fast impedance identification method for a distribution network based on adaptive extremum search according to claim 4, wherein: Injecting a disturbance signal into the equivalent circuit model and performing demodulation processing on the signal includes the following steps: First, use a signal generator to inject a fixed frequency into the equivalent circuit model. The disturbance signal is in the form of a sine wave of asin(ωt), where a is the amplitude of the disturbance signal. This disturbance signal is modulated with the input of the equivalent circuit model, and the input of the equivalent circuit model is the current injected by the grid-connected inverter; The node voltage output signal of the equivalent circuit model after injecting the disturbance is monitored in real time. The node voltage is set as V. The node voltage output signal is first processed by a high-pass filter. The cut-off frequency of the high-pass filter is higher than the disturbance signal frequency ω, which is used to remove the low-frequency components in the output signal. The low-frequency components include the fundamental wave signal and low-frequency interference signal in the power grid, so as to obtain a high-frequency signal containing disturbance-related information; Then, multiply the high-pass filtered signal by the signal of the same frequency as the injected disturbance, sin(ωt), for signal demodulation, so as to extract the amplitude and phase information related to the disturbance, and then obtain the disturbance signal ξ; In order to extract the gradient information, the signal after demodulation processing is passed through a high-pass filter again to remove the DC component, and then the filtered signal is multiplied by the signal of the same frequency as the injected disturbance to obtain a signal related to the gradient.

6. The method for quickly identifying the impedance of a distribution network based on adaptive extremum search according to claim 5, wherein: Adjusting the amplitude of the demodulated signal includes the following steps: If the change of the signal related to the gradient exceeds the set value, it indicates that the equivalent circuit model is experiencing external interference. When the external interference ends, increase the amplitude of the demodulated signal; The function that describes the relationship between the output and input of the equivalent circuit model is where f (2) > 0 represents a convex function, corresponding to a minimization problem, and f (2) > 0 represents a concave function, corresponding to a maximization problem. The optimization goal is to minimize (θ - θ * ), so that the objective function f(θ) approaches the minimum value f * ; Let be the estimated value of the optimal θ * and the estimation error After adding the disturbance signal Substitute the expression of θ into f(θ), and we get the expression of * θ is the optimal current injection angle converged through the adaptive extreme value search process in the distribution network impedance identification; Ignoring the higher-order terms, when increasing dynamically, the convergence rate of the equivalent circuit model increases and finally converges to the optimal value. When Kab(t) > 0, the equivalent circuit model is stable and tends to the minimum value.

7. The fast impedance identification method for a distribution network based on adaptive extremum search according to claim 6, characterized in that: Calculating the distribution network impedance includes the following steps: When the equivalent circuit model converges to the optimal current injection angle θ * = -α, measure the final voltage value V j at the common coupling point. Given the injection current value I j at this time and the Thevenin voltage calculated from the initial state of the equivalent circuit model Substitute into to calculate the magnitude of the Thevenin impedance, and thus obtain the value of the Thevenin impedance at the common coupling point, completing the identification process of the distribution network impedance.

8. A fast impedance identification system for a distribution network based on adaptive extremum search, which is used to implement the fast impedance identification method for a distribution network based on adaptive extremum search according to any one of claims 1-7, characterized in that: Including: A modeling module for modeling the grid-connected inverter as a controlled current source and describing the current output behavior of the grid-connected inverter by establishing a corresponding mathematical model; According to the Thevenin equivalent circuit principle of the power grid node, establish an equivalent circuit model between the node voltage, current and impedance. For each node, determine the parameters of the Thevenin voltage source and Thevenin impedance, and construct the expression of the node voltage; A disturbance injection module for generating and injecting a disturbance signal with a fixed frequency into the equivalent circuit model, and the parameters of the disturbance signal; A signal processing module for removing low-frequency components through a high-pass filter, then performing demodulation operations, multiplying the filtered signal by the signal of the same frequency as the injected disturbance, and extracting the disturbance information ξ and gradient information; The demodulation signal amplitude adjustment module is used to monitor the gradient change of the equivalent circuit model. By comparing with a preset threshold, it determines whether the equivalent circuit model is affected by external interference. When it detects that the gradient change is abnormal due to external interference, it waits for the interference to end, that is, after the gradient returns to the normal range, it adaptively increases the demodulation signal amplitude according to a predetermined strategy. During the adjustment process, by controlling the increase amplitude and rate of the demodulation signal amplitude, it ensures the stability and convergence of the entire system and achieves the best impedance identification effect. When it detects a power grid fault or abnormal working condition, it adjusts the parameters and strategies of the adaptive extreme value search control algorithm. For different types of faults and abnormal working conditions, a corresponding impedance identification model library is established in advance, and a suitable model is selected according to the fault type for impedance identification. An impedance calculation module, which is used to calculate the impedance by using the measured voltage value V at the common coupling point, the known injected current value I, and the relevant Thevenin voltage value after the equivalent circuit model converges through the adaptive extreme value search process. j and the known injected current value I j and the relevant Thevenin voltage value to calculate the impedance.

Images