Power quality disturbance tracing and impedance matching method, system and device and medium

Abstract

The invention discloses a power quality disturbance traceability and impedance matching method, system and device and a medium, and the method comprises the steps: synchronously collecting voltage and current waveform data through deploying a harmonic monitoring device, constructing a dynamic impedance matrix through employing a node admittance matrix inversion method, and compensating line parasitic capacitance through employing a spatial-temporal characteristic compensation algorithm; based on the feature vector of the dynamic impedance matrix and the improved Pearson's correlation coefficient, calculating the harmonic wave responsibility proportion of the user side and the system side, and performing multi-objective optimization on the dynamic impedance matrix through a chaotic optimization particle swarm algorithm; an inverted L-shaped adaptive topology is adopted to adjust a parallel capacitor and a series inductor, and a deep Q network dynamically adjusts a parameter combination by taking a fixed reflection coefficient as an optimization target; a disturbance interval is segmented by using an atomic decomposition algorithm, a voltage sag and a harmonic superposition event are distinguished according to an S transformation modular matrix slope difference, a responsibility subject is judged in combination with an equivalent impedance real part symbol, and a non-tampering traceability report based on the block chain technology is generated.

Description

Technical Field

[0001] This invention relates to the field of power quality management technology, and in particular to a method, system, device and medium for tracing the source of power quality disturbances and impedance matching. Background Technology

[0002] Current power quality disturbance mitigation primarily relies on harmonic source localization techniques such as the impedance matrix method and the equivalent impedance polarity method. During implementation, the direction of the disturbance is determined by constructing a node admittance matrix or measuring impedance changes before and after the disturbance. For impedance matching, L-type / inverse-Γ-type passive networks and transformer-inverter composite topologies with mechanical switching are widely used, combined with traditional PID control to achieve impedance adjustment within a limited frequency band. Regarding disturbance detection and identification, voltage sag detection is mostly based on S-transform time-frequency analysis or the disturbance power energy method, relying on the linear system assumption to extract sag features. Disturbance identification technology classifies time-frequency features using wavelet neural networks or deep learning models. Furthermore, dynamic impedance regulation mainly relies on SiCMOSFET fast switching to achieve four-quadrant impedance switching, but this is limited by fixed control algorithms.

[0003] Current technologies suffer from several shortcomings. First, harmonic source localization relies heavily on high-precision line impedance models. In practical applications, factors such as line distributed parameters and parasitic capacitance make accurate modeling difficult, leading to large localization errors. Furthermore, calculations are complex and real-time performance is poor in multi-source coupling scenarios. Second, traditional passive impedance matching networks are limited by the inherent characteristics of LC components, easily inducing resonance at high frequencies. Moreover, mechanical switches have slow responses and cannot handle millisecond-level disturbances. Third, voltage sag detection methods fail in nonlinear systems, have weak noise immunity, and struggle to distinguish between concurrent multi-event scenarios. Disturbance identification technologies suffer from feature redundancy, rely on large amounts of labeled data, and have high hardware deployment costs. In addition, while dynamic impedance adjustment devices offer faster responses, they suffer from significant energy efficiency losses, are prone to parameter drift due to temperature and humidity changes, and have insufficient adaptability in their control algorithms.

[0004] Therefore, a method for tracing the source of power quality disturbances and impedance matching is needed to meet the urgent need for intelligent and precise power quality management in new power systems. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, this invention provides a method, system, device, and medium for tracing the source of power quality disturbances and impedance matching, which solves the problems of low harmonic source location accuracy, slow dynamic impedance matching response speed, insufficient ability to handle multiple disturbances concurrently, reliance on manual experience for responsibility division, and limited wide-frequency impedance adjustment range in existing power quality management technologies.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution: In a first aspect, the present invention provides a method for tracing the source of power quality disturbances and impedance matching, comprising: By deploying harmonic monitoring devices to synchronously collect voltage and current waveform data, a dynamic impedance matrix is ​​constructed using the inverse method of the nodal admittance matrix, and a spatiotemporal characteristic compensation algorithm is used to compensate for the parasitic capacitance of the line. Based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, the harmonic responsibility ratio between the user side and the system side is calculated, and the dynamic impedance matrix is ​​optimized for multiple objectives using a chaotic optimization particle swarm algorithm. Based on the harmonic source location results, the parallel capacitor and series inductor are adjusted using an inverse Γ-type adaptive topology, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target. The perturbation interval is segmented using an atomic decomposition algorithm. Voltage sag and harmonic superposition events are distinguished based on the slope difference of the S-transform modulus matrix. The responsible party is determined by combining the sign of the real part of the equivalent impedance. An immutable traceability report based on blockchain technology is generated.

[0008] As a preferred embodiment of the power quality disturbance tracing and impedance matching method of the present invention, the step of distinguishing voltage sag and harmonic superposition events based on the slope difference of the S-transform matrix includes: For the waveform data within each perturbation event interval segmented by the atomic decomposition algorithm, perform an S-transform to obtain the time-frequency domain modulus matrix; Calculate the slope of the time-frequency modulus matrix, and distinguish between voltage sag events and harmonic superposition events based on the difference in the slope.

[0009] As a preferred embodiment of the power quality disturbance tracing and impedance matching method described in this invention, the generation of an immutable tracing report based on blockchain technology includes: Based on the determination of the sign of the real part of the equivalent impedance and the type of disturbance analyzed, the responsible party is determined; The edge node packages the disturbance type, occurrence time, the result of the determination of the real part sign of the equivalent impedance of the responsible party, and the comparison data before and after the governance, generates a hash digest, and uploads it to the cloud blockchain network for distributed storage and consensus verification. Based on the verified data stored in the cloud blockchain network, an electronic traceability report with timestamps, on-chain addresses, and data fingerprints is automatically generated.

[0010] As a preferred embodiment of the power quality disturbance tracing and impedance matching method described in this invention, the method of compensating for line parasitic capacitance using a spatiotemporal characteristic compensation algorithm includes: By analyzing the transmission delay and attenuation characteristics of the voltage and current waveform data between different nodes in the distribution network, the phase delay characteristics in the time dimension are extracted; by comparing the voltage and current amplitude differences of nodes at different spatial locations at the same time, the amplitude distribution characteristics in the spatial dimension are extracted. Based on the phase delay characteristics and amplitude distribution characteristics, a line model including a distributed capacitance correction term is established. Based on the aforementioned line model, a frequency-dependent complex compensation factor is introduced to dynamically correct the corresponding elements in the node admittance matrix.

[0011] As a preferred embodiment of the power quality disturbance tracing and impedance matching method of the present invention, the step of performing multi-objective optimization of the dynamic impedance matrix using a chaotic optimization particle swarm optimization algorithm includes: The encoding dimension of each particle is set to three times the number of harmonic sources, corresponding to the amplitude, phase, and position information of each harmonic source respectively; The mean square error between the theoretical and measured values ​​of node voltage distortion rate is used as the fitness function. When the change in the fitness of the optimal population over several consecutive generations is less than a preset threshold, a sequence perturbation based on the Tent chaotic mapping is applied to the current local optimum. The harmonic limit constraints are processed by combining the Lagrange function. The harmonic limit constraints are integrated into the fitness function in the form of a penalty function. Iterative optimization is performed based on the integrated fitness function until the preset convergence condition is met, and the optimized harmonic source parameters are output.

[0012] The beneficial effects of this preferred technical solution are: by utilizing the spatiotemporal feature compensation algorithm and the chaotic optimized particle swarm algorithm, the accuracy of harmonic source localization and the processing capability of multi-source concurrent scenarios are significantly improved.

[0013] As a preferred embodiment of the power quality disturbance tracing and impedance matching method described in this invention, the method of adjusting the parallel capacitor and series inductor using an inverse Γ-type adaptive topology includes: A capacitor array adjustable by a digital potentiometer is set in the parallel branch, wherein the capacitance adjustment step of the capacitor array is not greater than a preset minimum capacitance step threshold. An inductor array composed of magnetically coupled resonant coils is set in a series branch, and the connection and disconnection of the inductor array are controlled by a wide-bandgap semiconductor fast switch. Extract the disturbance components in the two-phase stationary coordinate system from the voltage and current waveform data; Based on the disturbance components, the real and imaginary parts of the equivalent impedance of the current system are identified online using the least squares method, and the real and imaginary parts of the equivalent impedance, the current state of the capacitor array, and the current switching state of the inductor array are input into the deep Q network.

[0014] As a preferred embodiment of the power quality disturbance tracing and impedance matching method described in this invention, the step of dynamically adjusting the parameter combination by a deep Q-network with a fixed reflection coefficient as the optimization target includes: The state space of the deep Q network is defined as a set including the current grid voltage and current waveforms, harmonic amplitudes, phases, and current parameters of the capacitor array and inductor array. The action space is defined as the digital adjustment commands for the corresponding capacitor array and the on / off commands for the wide bandgap semiconductor fast switches of the inductor array. The reward function of the deep Q-network is set so that the action output by the network can drive the reflection coefficient of the system to approach a preset fixed target value; The deep Q-network infers based on the current state and the identified impedance, and outputs the optimal action combination that maximizes the reward function.

[0015] The beneficial effects of this preferred technical solution are: to achieve ultra-high-speed dynamic impedance matching and wide-frequency domain adjustment.

[0016] Secondly, the present invention provides a system for tracing the source of power quality disturbances and impedance matching, comprising: The impedance matrix construction module is used to synchronously collect voltage and current waveform data by deploying harmonic monitoring devices, construct a dynamic impedance matrix using the inversion method of nodal admittance matrix, and compensate for line parasitic capacitance using a spatiotemporal characteristic compensation algorithm. The harmonic source localization module is used to calculate the harmonic responsibility ratio between the user side and the system side based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, and to perform multi-objective optimization of the dynamic impedance matrix through a chaotic optimization particle swarm algorithm. The impedance matching optimization module is used to adjust the parallel capacitor and series inductor using an inverse Γ-type adaptive topology based on the harmonic source location results, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target. The traceability report generation module is used to segment the disturbance interval using an atomic decomposition algorithm, distinguish voltage sag and harmonic superposition events based on the slope difference of the S-transform modulus matrix, determine the responsible party by combining the sign of the real part of the equivalent impedance, and generate an immutable traceability report based on blockchain technology.

[0017] Thirdly, the present invention provides an electronic device, including a memory and a processor; the memory is used to store computer-executable instructions, and the processor, when executing the computer-executable instructions, implements the steps of a method for tracing the source of power quality disturbances and impedance matching.

[0018] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of a method for tracing the source of power quality disturbances and impedance matching.

[0019] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention dynamically corrects distributed capacitance through a spatiotemporal feature compensation algorithm and introduces a chaotic optimization particle swarm optimization algorithm. It employs a Tent chaotic perturbation mechanism to break local convergence limitations and combines multi-objective constraint optimization to achieve rapid and accurate positioning within 50ms in multi-harmonic source scenarios. This invention achieves high matching efficiency by dynamically optimizing parameter combinations with a reflection coefficient Γ=0.05 through inverse Γ-type adaptive topology (parallel capacitor digital adjustment + series inductor array switching) and deep Q-network reinforcement learning control, and further improves this through nonlinear correction of magnetically coupled resonant coils. This invention constructs a wide-area panoramic synchronous perception system, integrating magnetically coupled resonant topology and deep reinforcement learning algorithms to solve the problem of quantifying the responsibility for concurrent multi-disturbance events. Simultaneously, through negative resistance mode energy efficiency optimization and adaptive compensation of environmental parameters, it overcomes the high-frequency resonance bottleneck of traditional LC networks, forming an integrated solution covering harmonic suppression, voltage transient compensation, and multi-responsibility collaborative governance. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the overall process logic of a method for tracing the source of power quality disturbances and impedance matching provided in an embodiment of the present invention.

[0022] Figure 2 The flowchart illustrates the execution of a chaotic optimization particle swarm optimization algorithm for a power quality disturbance tracing and impedance matching method provided in one embodiment of the present invention.

[0023] Figure 3 The present invention provides a DQN-controlled inverse Γ-type adaptive topology diagram for a power quality disturbance tracing and impedance matching method according to an embodiment of the present invention.

[0024] Figure 4This is a system architecture diagram of a method for tracing the source of power quality disturbances and impedance matching according to an embodiment of the present invention. Detailed Implementation

[0025] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0026] Example 1, referring to Figure 1 As one embodiment of the present invention, a method for tracing the source of power quality disturbances and impedance matching is provided, such as... Figure 1 The specific steps shown are as follows: S100: Voltage and current waveform data are collected synchronously by deploying harmonic monitoring devices, a dynamic impedance matrix is ​​constructed using the inversion method of the node admittance matrix, and the line parasitic capacitance is compensated by the spatiotemporal characteristic compensation algorithm. S200: Based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, the harmonic responsibility ratio between the user side and the system side is calculated, and the dynamic impedance matrix is ​​optimized for multiple objectives through a chaotic optimization particle swarm algorithm. S300: Based on the harmonic source location results, the parallel capacitor and series inductor are adjusted using an inverse Γ type adaptive topology, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target; S400: It uses an atomic decomposition algorithm to segment the disturbance interval, distinguishes voltage sag and harmonic superposition events based on the slope difference of the S-transform modulus matrix, determines the responsible party by combining the real part sign of the equivalent impedance, and generates an immutable traceability report based on blockchain technology.

[0027] It should be noted that, to address the problems of low harmonic source location accuracy, slow dynamic impedance matching response, insufficient concurrent processing capability for multi-source disturbances, reliance on manual experience for responsibility allocation, and limited wide-frequency impedance adjustment range in existing power quality management technologies, steps S100~S400 dynamically correct distributed capacitance using a spatiotemporal feature compensation algorithm and introduce a chaotic optimization particle swarm optimization algorithm. The Tent chaotic perturbation mechanism is used to break local convergence limitations, combined with multi-objective constraint optimization, to achieve rapid and accurate location within 50ms in multi-harmonic source scenarios. This invention achieves high matching efficiency by dynamically optimizing parameter combinations with a reflection coefficient Γ=0.05 using an inverse Γ-type adaptive topology (parallel capacitor digital adjustment + series inductor array switching) and deep Q-network reinforcement learning control, and further improves this through nonlinear correction of the magnetically coupled resonant coil. This invention constructs a wide-area panoramic synchronous perception system, integrating magnetically coupled resonant topology and deep reinforcement learning algorithm to solve the problem of quantifying the responsibility of multiple disturbance events concurrently. At the same time, through negative resistance mode energy efficiency optimization and adaptive compensation of environmental parameters, it breaks through the high-frequency resonance bottleneck of traditional LC networks, forming an integrated solution covering harmonic suppression, voltage transient compensation and multi-responsible party collaborative governance.

[0028] Example 2, refer to Figures 2-4 This embodiment provides a specific implementation of a method for tracing the source of power quality disturbances and impedance matching, and explains the technical means used in this method.

[0029] In this embodiment of the invention, step S100, which involves deploying a harmonic monitoring device to synchronously collect voltage and current waveform data, constructing a dynamic impedance matrix using the nodal admittance matrix inversion method, and compensating for line parasitic capacitance using a spatiotemporal characteristic compensation algorithm, includes the following sub-steps A1 to A3: In A1: Voltage and current waveform data are collected synchronously by deploying a harmonic monitoring device; Specifically, in order to capture the ever-changing state of the power grid, high-performance harmonic monitoring devices are deployed at key nodes of the distribution network, including substation outlets, distributed power generation grid connection points, and important sensitive load terminals.

[0030] It is important to note that high-performance harmonic monitoring devices must possess high-precision clock synchronization capabilities. Considering the phase sensitivity of high-frequency harmonics, even microsecond-level time errors can lead to significant phase drift. Therefore, the system must employ a timing mechanism based on GPS / BeiDou or the IEEE 1588 PTP protocol to ensure nanosecond-level synchronization of data across the entire network. The acquired physical quantities are not limited to the effective values ​​of voltage and current; they must also include complete waveform data. Only raw waveform data can support subsequent high-order spectral analysis and transient feature extraction.

[0031] In A2: The dynamic impedance matrix is ​​constructed using the inversion method of the nodal admittance matrix; Specifically, a node admittance matrix is ​​constructed based on the collected synchronous voltage and current data. According to circuit theory, the node impedance matrix It is the inverse of the admittance matrix, i.e. .

[0032] It should be noted that the innovation of this embodiment lies in not stopping at static matrix calculation, but introducing a "dynamic" perspective. Since the topology of the distribution network may change with switching operations, and load characteristics fluctuate over time, the admittance matrix... It is time-varying. The system needs to be updated in real time. The matrix is ​​calculated, and the inversion process is accelerated using sparse matrix techniques, thereby obtaining the real-time dynamic impedance matrix. This not only reflects the topological connections of the power grid but also implicitly contains the electrical coupling strength of each node.

[0033] In A3: a spatiotemporal feature compensation algorithm is used to compensate for line parasitic capacitance; the detailed steps include: By analyzing the transmission delay and attenuation characteristics of voltage and current waveform data between different nodes in the distribution network, the phase delay characteristics in the time dimension are extracted; by comparing the voltage and current amplitude differences of nodes at different spatial locations at the same time, the amplitude distribution characteristics in the spatial dimension are extracted. Based on phase delay characteristics and amplitude distribution characteristics, a line model including a distributed capacitance correction term is established. Based on the line model, a frequency-dependent complex compensation factor is introduced. The corresponding elements in the nodal admittance matrix are dynamically corrected, as expressed by the formula: in, This represents the admittance increment caused by parasitic capacitance. Represents the initial nodal admittance matrix. This represents the compensated nodal admittance matrix.

[0034] It should be noted that the compensated impedance matrix can more accurately describe the propagation behavior of high-frequency harmonics in the power grid, reducing the model error from more than 10% to less than 1%, laying a solid physical foundation for subsequent high-precision positioning.

[0035] In an optional embodiment, the step of compensating for parasitic capacitance of a line can also be to establish a frequency-varying transmission line model of a distributed parameter line, directly calculate its resistance per unit length, inductance, and capacitance to ground parameters based on the line's geometry, material properties, and operating frequency, and integrate this frequency-varying parameter model into the construction process of the node admittance matrix to achieve physical modeling compensation for parasitic capacitance.

[0036] In an optional embodiment, the step of compensating for line parasitic capacitance can also employ a data-driven parameter identification method. This involves injecting multiple test signals of known frequencies into the system, measuring the voltage and current responses between different nodes, and using an optimization algorithm to backfit the equivalent distributed capacitance parameters of the line, thereby correcting the admittance matrix.

[0037] In embodiments of the present invention, such as Figure 2 As shown, the step S200 above, which involves multi-objective optimization of the dynamic impedance matrix using a chaotic particle swarm optimization algorithm, includes: The encoding dimension of each particle is set to three times the number of harmonic sources, corresponding to the amplitude, phase, and position information of each harmonic source respectively; The mean square error between the theoretical and measured values ​​of node voltage distortion rate is used as the fitness function. When the change in the fitness of the optimal population over several consecutive generations is less than a preset threshold, a sequence perturbation based on the Tent chaotic mapping is applied to the current local optimum. By combining the Lagrange function to handle the harmonic limit constraints, the harmonic limit constraints are integrated into the fitness function in the form of a penalty function. Iterative optimization is performed based on the integrated fitness function until the preset convergence condition is met, and the optimized harmonic source parameters are output.

[0038] Specifically, in conjunction with the Lagrange function To address harmonic limit constraints, the fitness function is improved, and the formula is expressed as follows: in, and These are the theoretical and actual measured values ​​of the node voltage distortion rate. For constraint term coefficients, This represents the actual short-circuit capacity. As a reference short-circuit capacity, The reference current needs to be obtained from a table, where N represents the total number of nodes. Represents the Lagrange multipliers used in balance constraints. This represents the harmonic limit constraint function.

[0039] It should be noted that the preset threshold is usually set based on a comprehensive consideration of algorithm convergence and computational efficiency. For example, it can be set so that the relative improvement of the global optimal fitness value in 5 to 10 consecutive iterations is less than one-thousandth to one-ten-thousandth. The threshold must ensure that the algorithm can identify and escape convergence stagnation in a timely manner while fully exploring the solution space, avoiding premature perturbation due to a threshold that is too small, which would affect the optimization stability, or excessively large, which would prolong the invalid computation time.

[0040] In an optional embodiment, multi-objective optimization can also be performed using a decomposition-based multi-objective evolutionary algorithm, which decomposes the original problem into several scalar quantum problems and optimizes them collaboratively, guiding the search through a preset weight vector to approximate the Pareto front.

[0041] In an optional embodiment, multi-objective optimization can also be performed using a multi-objective genetic algorithm, which utilizes non-dominated sorting and crowding calculation to maintain population diversity, and iteratively searches for a multi-objective non-dominated solution set through genetic operations such as selection, crossover, and mutation.

[0042] In this embodiment of the invention, step S300 includes the following sub-steps C1 and C2: In C1: Based on the harmonic source location results, an inverse Γ-type adaptive topology is used to adjust the parallel capacitor and series inductor; detailed steps include: A capacitor array adjustable by a digital potentiometer is set in the parallel branch, and the capacitance adjustment step of the capacitor array is not greater than the preset minimum capacitance step threshold. An inductor array composed of magnetically coupled resonant coils is set in the series branch, and the connection and disconnection of the inductor array are controlled by a wide-bandgap semiconductor fast switch. Extract disturbance components in a two-phase stationary coordinate system from voltage and current waveform data; Based on the disturbance components, the real and imaginary parts of the equivalent impedance of the current system are identified online using the least squares method, and the real and imaginary parts of the equivalent impedance, the current state of the capacitor array, and the current switching state of the inductor array are input into the deep Q network.

[0043] It should be noted that, as Figure 3 As shown, the inverse Γ-type adaptive topology consists of parallel capacitor branches ( ) and series inductor branch ( This is composed of several components, forming a second-order filter structure. Its core advantages lie in its extremely wide bandwidth coverage and extremely high tuning accuracy.

[0044] Specifically, a capacitor array with digital potentiometer adjustment is set in the parallel branch. Its capacitance adjustment step size is ≤1nF, covering the 0.1-5kHz frequency band; It should be noted that the parallel branch is responsible for providing a low-impedance path to filter out high-frequency harmonics. A binary-coded capacitor array is used, and continuous capacitance adjustment is achieved through the parallel combination of multiple small-value capacitors. This embodiment requires a capacitance adjustment step size... This allows the system to be precisely tuned to a specific harmonic frequency, avoiding the problem of "undercompensation or overcompensation due to excessive step size" in traditional capacitor banks.

[0045] Specifically, the series branch includes an FPGA-controlled inductor. The array employs magnetically coupled resonant coils and SiC MOSFETs for fast switching to achieve dynamic nanosecond-level switching, where the equivalent inductance of the magnetically coupled resonant coil is... satisfy: in, The permeability of free space, The relative permeability, N The number of coil turns. The cross-sectional area of ​​the magnetic core is... The length of the magnetic circuit. k This is a nonlinear correction coefficient used to compensate for magnetic saturation effects under high-frequency current. This represents the change in current. The reference current value representing the normalized change in current.

[0046] It's important to note that to achieve a dynamic response time of 0.1ms, traditional mechanical switches have been completely abandoned, replaced by third-generation wide-bandgap semiconductors—silicon carbide (SiC) MOSFETs. SiC material has a bandgap three times that of silicon and a breakdown field strength ten times greater. This allows SiC MOSFETs to withstand higher voltages while possessing extremely low on-resistance and nanosecond-level switching speeds. Utilizing the high-speed switching characteristics of SiC, the system can switch or reconfigure the inductor branch the instantaneously (in microseconds) a disturbance is detected, truly achieving real-time control.

[0047] In C2: The deep Q-network dynamically adjusts the parameter combination with a fixed reflection coefficient as the optimization target; the detailed steps include: The state space of a deep Q-network is defined as a set containing the current grid voltage and current waveforms, harmonic amplitudes, phases, and current parameters of the capacitor array and inductor array. The action space is defined as the digital adjustment commands for the capacitor array and the on / off commands for the wide-bandgap semiconductor fast switches of the inductor array. Set the reward function of the deep Q network so that the action output by the network can drive the system's reflection coefficient to approach a preset fixed target value; Deep Q-networks reason based on the current state and identified impedances, and output the optimal action combination that maximizes the reward function.

[0048] It should be noted that the deep Q-network dynamically adjusts the parameter combination with the reflection coefficient Γ=0.05 as the optimization objective. When the system takes a certain action... When the value decreases, a positive reward is given; conversely, a negative penalty is given. After the system performs an action, it observes the new state and reward again, continuously updating the Q-value network, thus enabling the system to have the ability to learn and evolve adaptively.

[0049] In an alternative embodiment, the optimal action combination can also be obtained through a deep deterministic policy gradient algorithm based on deterministic policy gradient. Through an actor-critic architecture, the actor network directly outputs deterministic actions, and the critic network evaluates the value of the actions and guides the actor network parameter updates.

[0050] In an optional embodiment, the optimal action combination can also be obtained through a proximal policy optimization algorithm based on policy optimization. By constraining the magnitude of policy updates, the random policy network is optimized in a way with high sampling efficiency while ensuring training stability, so that the probability distribution of its output actions is concentrated in the high reward region.

[0051] In this embodiment of the invention, step S400 includes the following sub-steps D1 to D3: In D1: The perturbation interval is divided using the atomic decomposition algorithm; Specifically, the atom decomposition algorithm is used to segment continuous waveform data in the time domain. This algorithm can accurately locate the start and end points of disturbances based on the energy changes of the signal, decomposing complex long processes into several independent disturbance event atoms.

[0052] In an optional embodiment, the segmentation of the disturbance interval can also employ a time-frequency segmentation method based on variational mode decomposition, which adaptively decomposes the signal into multiple intrinsic mode functions and identifies the disturbance boundary based on the instantaneous frequency and energy mutation point of each component.

[0053] In an optional embodiment, the segmentation of the disturbance interval can also employ a statistical segmentation method based on abrupt change detection. This method calculates the local statistical characteristics of the signal through a sliding window and uses hypothesis testing or information entropy changes to detect structural abrupt changes in the sequence, thereby achieving automatic segmentation of the disturbance event.

[0054] In D2: Voltage sag and harmonic superposition events are distinguished based on the difference in the slope of the S-transform modulus matrix; Specifically, for the waveform data within each disturbance event interval segmented by the atomic decomposition algorithm, an S-transform is performed to obtain the time-frequency domain modulus matrix; the slope of the time-frequency modulus matrix is ​​calculated, and the voltage sag event and harmonic superposition event are distinguished based on the difference in slope.

[0055] It should be noted that for each segmented event atom, time-frequency analysis is performed using the S-transform. The S-transform combines the advantages of the short-time Fourier transform and wavelet transform, exhibiting excellent time-frequency localization characteristics. Different types of perturbations show different energy distribution characteristics in the modulus matrix of the S-transform. The slope difference of the modulus matrix is ​​used to distinguish voltage sags and harmonic superposition events. Voltage sags typically manifest as a sharp drop in energy across the entire frequency band or the fundamental frequency band, with a drastic change in slope; while harmonic superposition manifests as the continuous presence of energy in a specific high-frequency band, with a relatively gentle slope.

[0056] In an optional embodiment, event differentiation can also employ an end-to-end classification method based on a deep convolutional neural network, where the original perturbation waveform or time spectrum is input into the network, and discriminative features are automatically extracted and classified through convolutional layers.

[0057] In an optional embodiment, event differentiation can also employ a feature classification method based on multi-scale entropy and support vector machines. First, the sample entropy of the signal at different time scales is calculated to form a feature vector. Then, a classification hyperplane is constructed in the feature space using a support vector machine to differentiate event types.

[0058] In D3: The responsible party is determined by combining the sign of the real part of the equivalent impedance, and an immutable traceability report based on blockchain technology is generated; detailed steps include: Based on the determination of the sign of the real part of the equivalent impedance and the type of disturbance analyzed, the responsible party is identified; Edge nodes package the disturbance type, occurrence time, the result of determining the sign of the real part of the equivalent impedance of the responsible party, and the comparison data before and after governance to generate a hash digest, and upload it to the cloud blockchain network for distributed storage and consensus verification; Based on verified data stored in the cloud blockchain network, an electronic traceability report with timestamps, on-chain addresses, and data fingerprints is automatically generated.

[0059] As described above, this invention dynamically corrects distributed capacitance through a spatiotemporal feature compensation algorithm and introduces a chaotic optimization particle swarm optimization algorithm. It employs a Tent chaotic perturbation mechanism to break local convergence limitations and combines multi-objective constraint optimization to achieve rapid and accurate positioning within 50ms in multi-harmonic source scenarios. This invention utilizes an inverse Γ-type adaptive topology (parallel capacitor digital adjustment + series inductor array switching) and deep Q-network reinforcement learning control to dynamically optimize parameter combinations with a reflection coefficient Γ=0.05 as the target, achieving high matching efficiency. Nonlinear correction is achieved through magnetically coupled resonant coils. This invention constructs a wide-area panoramic synchronous sensing system, integrating magnetically coupled resonant topology and deep reinforcement learning algorithms to solve the problem of concurrent responsibility quantification for multiple perturbation events. Simultaneously, through negative resistance mode energy efficiency optimization and adaptive compensation of environmental parameters, it overcomes the high-frequency resonance bottleneck of traditional LC networks, forming an integrated solution covering harmonic suppression, voltage transient compensation, and multi-responsibility collaborative governance.

[0060] Example 3: This example provides a system for tracing the source of power quality disturbances and impedance matching, including: The impedance matrix construction module is used to synchronously collect voltage and current waveform data by deploying harmonic monitoring devices, construct a dynamic impedance matrix using the inversion method of nodal admittance matrix, and compensate for line parasitic capacitance using a spatiotemporal characteristic compensation algorithm. The harmonic source localization module is used to calculate the harmonic responsibility ratio between the user side and the system side based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, and to perform multi-objective optimization of the dynamic impedance matrix through a chaotic optimization particle swarm algorithm. The impedance matching optimization module is used to adjust the parallel capacitor and series inductor using an inverse Γ-type adaptive topology based on the harmonic source location results, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target. The traceability report generation module is used to segment the disturbance interval using an atomic decomposition algorithm, distinguish voltage sag and harmonic superposition events based on the slope difference of the S-transform modulus matrix, determine the responsible party by combining the sign of the real part of the equivalent impedance, and generate an immutable traceability report based on blockchain technology.

[0061] It should be noted that the technical solution of the power quality disturbance tracing and impedance matching system is based on the same concept as the technical solution of the power quality disturbance tracing and impedance matching method described above. For details not described in detail in the technical solution of the power quality disturbance tracing and impedance matching system in this embodiment, please refer to the description of the technical solution of the power quality disturbance tracing and impedance matching method described above.

[0062] The above-mentioned unit modules can be embedded in the processor of the electronic device in hardware form or independent of it, or they can be stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0063] This embodiment also provides an electronic device, which includes a processor, a memory, a communication interface, a display screen, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for tracing power quality disturbances and impedance matching. The display screen can be a liquid crystal display (LCD) or an e-ink display. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0064] This embodiment also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method proposed in the above embodiments.

[0065] The storage medium proposed in this embodiment belongs to the same inventive concept as the method proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0066] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory, random access memory, flash memory, hard disk, or optical disk, and includes several instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute the method of the embodiments of the present invention.

[0067] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention 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 the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for tracing the source of power quality disturbances and impedance matching, characterized in that, include: By deploying harmonic monitoring devices to synchronously collect voltage and current waveform data, a dynamic impedance matrix is ​​constructed using the inverse method of the nodal admittance matrix, and a spatiotemporal characteristic compensation algorithm is used to compensate for the parasitic capacitance of the line. Based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, the harmonic responsibility ratio between the user side and the system side is calculated, and the dynamic impedance matrix is ​​optimized for multiple objectives using a chaotic optimization particle swarm algorithm. Based on the harmonic source location results, the parallel capacitor and series inductor are adjusted using an inverse Γ-type adaptive topology, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target. The perturbation interval is segmented using an atomic decomposition algorithm. Voltage sag and harmonic superposition events are distinguished based on the slope difference of the S-transform modulus matrix. The responsible party is determined by combining the sign of the real part of the equivalent impedance. An immutable traceability report based on blockchain technology is generated.

2. The method for tracing the source of power quality disturbances and impedance matching as described in claim 1, characterized in that, The method of distinguishing voltage sag and harmonic superposition events based on the slope difference of the S-transform matrix includes: For the waveform data within each perturbation event interval segmented by the atomic decomposition algorithm, perform an S-transform to obtain the time-frequency domain modulus matrix; Calculate the slope of the time-frequency modulus matrix, and distinguish between voltage sag events and harmonic superposition events based on the difference in the slope.

3. The method for tracing the source of power quality disturbances and impedance matching as described in claim 2, characterized in that, The generation of an immutable traceability report based on blockchain technology includes: Based on the determination of the sign of the real part of the equivalent impedance and the type of disturbance analyzed, the responsible party is determined; The edge node packages the disturbance type, occurrence time, the result of the determination of the real part sign of the equivalent impedance of the responsible party, and the comparison data before and after the governance, generates a hash digest, and uploads it to the cloud blockchain network for distributed storage and consensus verification. Based on the verified data stored in the cloud blockchain network, an electronic traceability report with timestamps, on-chain addresses, and data fingerprints is automatically generated.

4. The method for tracing the source of power quality disturbances and impedance matching as described in claim 1, characterized in that, The method of using a spatiotemporal feature compensation algorithm to compensate for line parasitic capacitance includes: By analyzing the transmission delay and attenuation characteristics of the voltage and current waveform data between different nodes in the distribution network, the phase delay characteristics in the time dimension are extracted; by comparing the voltage and current amplitude differences of nodes at different spatial locations at the same time, the amplitude distribution characteristics in the spatial dimension are extracted. Based on the phase delay characteristics and amplitude distribution characteristics, a line model including a distributed capacitance correction term is established. Based on the aforementioned line model, a frequency-dependent complex compensation factor is introduced to dynamically correct the corresponding elements in the node admittance matrix.

5. The method for tracing the source of power quality disturbances and impedance matching as described in claim 4, characterized in that, The multi-objective optimization of the dynamic impedance matrix using the chaotic particle swarm optimization algorithm includes: The encoding dimension of each particle is set to three times the number of harmonic sources, corresponding to the amplitude, phase, and position information of each harmonic source respectively; The mean square error between the theoretical and measured values ​​of node voltage distortion rate is used as the fitness function. When the change in the fitness of the optimal population over several consecutive generations is less than a preset threshold, a sequence perturbation based on the Tent chaotic mapping is applied to the current local optimum. The harmonic limit constraints are processed by combining the Lagrange function. The harmonic limit constraints are integrated into the fitness function in the form of a penalty function. Iterative optimization is performed based on the integrated fitness function until the preset convergence condition is met, and the optimized harmonic source parameters are output.

6. The method for tracing the source of power quality disturbances and impedance matching as described in claim 5, characterized in that, The method of using anti-Γ type adaptive topology to adjust the parallel capacitor and series inductor includes: A capacitor array adjustable by a digital potentiometer is set in the parallel branch, wherein the capacitance adjustment step of the capacitor array is not greater than a preset minimum capacitance step threshold. An inductor array composed of magnetically coupled resonant coils is set in a series branch, and the connection and disconnection of the inductor array are controlled by a wide-bandgap semiconductor fast switch. Extract the disturbance components in the two-phase stationary coordinate system from the voltage and current waveform data; Based on the disturbance components, the real and imaginary parts of the equivalent impedance of the current system are identified online using the least squares method, and the real and imaginary parts of the equivalent impedance, the current state of the capacitor array, and the current switching state of the inductor array are input into the deep Q network.

7. The method for tracing the source of power quality disturbances and impedance matching as described in claim 6, characterized in that, The dynamic adjustment of parameter combinations by the deep Q-network with a fixed reflection coefficient as the optimization objective includes: The state space of the deep Q network is defined as a set including the current grid voltage and current waveforms, harmonic amplitudes, phases, and current parameters of the capacitor array and inductor array. The action space is defined as the digital adjustment commands for the corresponding capacitor array and the on / off commands for the wide bandgap semiconductor fast switches of the inductor array. The reward function of the deep Q-network is set so that the action output by the network can drive the reflection coefficient of the system to approach a preset fixed target value; The deep Q-network infers based on the current state and the identified impedance, and outputs the optimal action combination that maximizes the reward function.

8. A system for tracing the source of power quality disturbances and impedance matching, comprising applying the method for tracing the source of power quality disturbances and impedance matching as described in any one of claims 1 to 7, characterized in that, include: The impedance matrix construction module is used to synchronously collect voltage and current waveform data by deploying harmonic monitoring devices, construct a dynamic impedance matrix using the inversion method of nodal admittance matrix, and compensate for line parasitic capacitance using a spatiotemporal characteristic compensation algorithm. The harmonic source localization module is used to calculate the harmonic responsibility ratio between the user side and the system side based on the eigenvectors of the dynamic impedance matrix and the improved Pearson correlation coefficient, and to perform multi-objective optimization of the dynamic impedance matrix through a chaotic optimization particle swarm algorithm. The impedance matching optimization module is used to adjust the parallel capacitor and series inductor using an inverse Γ-type adaptive topology based on the harmonic source location results, and the parameter combination is dynamically adjusted by a deep Q network with a fixed reflection coefficient as the optimization target. The traceability report generation module is used to segment the disturbance interval using an atomic decomposition algorithm, distinguish voltage sag and harmonic superposition events based on the slope difference of the S-transform modulus matrix, determine the responsible party by combining the sign of the real part of the equivalent impedance, and generate an immutable traceability report based on blockchain technology.

9. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store computer-executable instructions, and when the processor executes the computer-executable instructions, it implements the steps of the method for tracing the source of power quality disturbances and impedance matching as described in any one of claims 1 to 7.

10. A computer-readable storage medium having computer-executable instructions stored thereon, characterized in that: When the computer-executable instructions are executed by the processor, they implement the steps of the method for tracing the source of power quality disturbances and impedance matching as described in any one of claims 1 to 7.

Images