A distributed control and linkage regulation system for dynamic monitoring of harmonics in power distribution networks

By networking and coordinating the control of distributed harmonic sensing nodes and decentralized governance units, the communication delay and parallel resonance problems in the distribution network harmonic control system are solved, achieving efficient harmonic governance and system linkage regulation.

CN123121618APending Publication Date: 2026-11-17WUXI VOCATIONAL INSTITUTE OF COMMERCE +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611510751.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-28
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing power distribution network harmonic control systems suffer from problems such as communication delays, overload shutdowns of individual decentralized governance units, parallel resonance risks, and interruptions in system linkage control processes.

Method used

Distributed harmonic sensing nodes and decentralized governance units are adopted, and a peer-to-peer communication network is built using edge computing chips and industrial Ethernet communication modules. Multicast data packets are generated by the edge computing chips, and the microprocessor performs capacity negotiation and current compensation to achieve targeted push and local control, thus avoiding parallel resonance.

Benefits of technology

It reduces communication latency, avoids overload shutdown of a single decentralized governance unit, maintains the system's coordinated control capability, and reduces the risk of parallel resonance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123121618A_ABST
    Figure CN123121618A_ABST
Patent Text Reader

Abstract

This invention relates to the field of power quality control in power systems, and discloses a distributed harmonic dynamic monitoring and decentralized control linkage system for distribution networks. The system includes distributed harmonic sensing nodes and decentralized governance units. The distributed harmonic sensing nodes integrate edge computing chips, which generate harmonic component parameters and generate multicast data packets when parameters exceed limits. The decentralized governance units receive these data packets and control the main circuit inverter topology to output AC current. The decentralized governance units establish a capacity allocation status table and generate command compensation current using a consensus iterative equation. The decentralized governance units calculate voltage prediction deviations based on line impedance parameters, and execute blocking actions to generate a reduced-order connectivity topology when deviations exceed limits. This invention reduces communication latency by extracting directional quantities through edge computing chips and using consensus iteration to complete capacity negotiation, avoiding local overload shutdowns. It also reconstructs the communication network to block resonance conditions and maintain linkage control operation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of power quality control in power systems, specifically to a distributed control and linkage regulation system for dynamic monitoring of harmonics in distribution networks. Background Technology

[0002] With the increase of nonlinear loads within power distribution networks, the problem of harmonic distortion generated in distribution lines is becoming increasingly prominent. Existing power distribution network harmonic control systems typically rely on a central server to collect data from various monitoring nodes and uniformly distribute mitigation commands. This communication mode, where the central server centrally receives data and broadcasts commands, consumes a large amount of network bandwidth, causing data transmission channel congestion and resulting in long-distance communication delays between monitoring nodes and mitigation nodes. These communication delays cause the control signals issued by the mitigation nodes to fail to match the transient changes in harmonic waveforms within the power distribution lines.

[0003] In distributed harmonic mitigation scenarios, independently deployed distributed mitigation units typically output compensation current based on local measurement data, lacking a mechanism for capacity negotiation with adjacent distributed mitigation units. When the harmonic distortion parameters of a local power distribution line exceed the rated compensation limit of a single distributed mitigation unit, the unit will trigger shutdown protection due to output overload, failing to complete the harmonic mitigation task for the corresponding area.

[0004] Existing multi-node collaborative governance systems do not perform local voltage prediction and assessment based on the equivalent harmonic impedance of the distribution lines before injecting compensation current. The high-frequency AC current synchronously output by multiple governance nodes can easily interact with the inherent capacitive reactance parameters of the distribution lines, leading to parallel resonance amplification and increasing the voltage waveform distortion rate within the distribution lines. Furthermore, when a local node is forcibly disconnected due to the risk of parallel resonance, the existing multi-node collaborative governance system lacks communication link reconfiguration logic. Nodes that have not disconnected cannot reallocate the remaining compensation tasks, causing interruptions in the localized coordinated control process of the distribution network. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a distributed control and linkage regulation system for dynamic monitoring of harmonics in power distribution networks. This system solves the problems of difficulty in real-time perception of harmonic fluctuations in power distribution networks, high communication delays in centralized control mechanisms, and the risk of high-frequency parallel resonance caused by multi-node collaborative compensation.

[0006] To achieve the above objectives, this invention provides a distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network, comprising: The distributed harmonic sensing node integrates an edge computing chip and a first industrial Ethernet communication module. The edge computing chip generates parameters for each harmonic component. When the parameters of each harmonic component are greater than a preset distortion threshold, the edge computing chip generates a trigger signal. The first industrial Ethernet communication module encapsulates the trigger signal to generate a multicast data packet. The distributed governance unit includes a main circuit inverter topology with multiple sets of insulated gate bipolar transistors inside, a DC-side capacitor, a second industrial Ethernet communication module, a control board and a microprocessor. The second industrial Ethernet communication module and the microprocessor are electrically connected to the control board. The distributed harmonic sensing nodes and the distributed governance unit form a peer-to-peer communication network inside the optical fiber ring network. The microprocessor generates pulse width modulation signals based on multicast data packets to control the on and off of multiple sets of insulated gate bipolar transistors. The main circuit inverter topology, combined with the DC voltage of the DC-side capacitor, generates AC current that is injected into the power distribution line.

[0007] Preferably, the distributed harmonic sensing node integrates a high-frequency sampling module, which continuously collects the instantaneous voltage waveform and instantaneous current waveform of the power distribution line and converts them into digital discrete signals; the edge computing chip calls the internally embedded discrete Fourier transform algorithm to calculate the digital discrete signals and generate parameters of each harmonic component.

[0008] Preferably, during the time period when the control board does not receive multicast data packets, the microprocessor continuously collects background harmonic voltage and background harmonic current data sequences from the power distribution line. The microprocessor extracts numerical fluctuation parameters from the data sequences and compares these parameters with an internally preset excitation dead-zone threshold. When the numerical fluctuation parameters exceed the excitation dead-zone threshold, the background harmonic voltage, background harmonic current, and equivalent harmonic impedance of the power distribution line are converted to per-unit values ​​for normalization. The normalized per-unit values ​​of the background harmonic voltage and background harmonic current are used as the calculation data source and input into an internally fixed recursive least squares algorithm model for continuous iterative calculation, outputting the equivalent harmonic impedance of the power distribution line. The recursive least squares algorithm model does not rely on externally injected test pulses and utilizes a high-quality calculation data source with a data filtering mechanism to perform iterative mathematical operations.

[0009] Preferably, the edge computing chip uses an internally set fixed sliding time window to perform data timing segmentation on the generated harmonic component parameters and extracts the harmonic current amplitude corresponding to the sliding time window; the edge computing chip performs differential operation on the harmonic current amplitude corresponding to two adjacent sliding time windows and calculates and outputs the harmonic amplitude change rate, which characterizes the severity of high-frequency waveform distortion in the power distribution line.

[0010] Preferably, the edge computing chip selects harmonic voltage phasors and harmonic current phasors from the parameters of each harmonic component and extracts the phase angle difference, performs a product operation to output the harmonic active power value; the edge computing chip determines the positive or negative sign of the harmonic active power value and extracts the harmonic active power direction quantity, which represents the propagation direction of the nonlinear power flow inside the distribution line.

[0011] Preferably, the edge computing chip performs dynamic division of the communication addressing range based on the harmonic active power direction quantity, and retrieves the downstream node address set or the upstream node address set from the self-distribution network topology mapping table; the edge computing chip selects the subscription end media access control address in the set as the target receiver identifier, and writes it into the addressing information segment of the multicast data packet to realize targeted push.

[0012] Preferably, the microprocessor reads the difference between the rated compensation capacity and the current output capacity of the main circuit inverter topology to generate the percentage of remaining available compensation capacity; the control board combines the percentage of remaining available compensation capacity with the node hardware identifier code set inside the distributed governance unit to generate a state initialization data frame; all distributed governance units in the local communication neighbor set send and receive state initialization data frames to each other; the microprocessor parses the state initialization data frame, establishes a local capacity allocation state table in memory, and writes the node hardware identifier code and the percentage of remaining available compensation capacity into the local capacity allocation state table.

[0013] Preferably, the microprocessor generates a restricted adaptive step size based on the local node connectivity and the preset harmonic amplitude change rate, extracts the remaining available compensation capacity ratio of adjacent decentralized governance units within the local communication neighbor set, and substitutes the restricted adaptive step size into the compensation weight consistency iterative state update equation to perform loop calculations. When the difference amplitude of the loop calculation is less than or equal to the preset convergence tolerance threshold, the microprocessor outputs the converged target compensation capacity allocation weight, combines it with the effective value output current parameter of the main circuit inverter topology under rated operating conditions, performs multiplication calculations, and generates the command compensation current.

[0014] Preferably, the microprocessor performs a phasor multiplication operation on the equivalent harmonic impedance of the distribution line and the command compensation current, outputting the local harmonic voltage prediction deviation. When it is determined that the local harmonic voltage prediction deviation is greater than the safe operating voltage distortion threshold and the duration calculated from the first continuous sampling exceeding the limit recorded by the internal clock is longer than the tolerance delay calculation period, the microprocessor generates an active blocking level signal, forcing multiple sets of insulated gate bipolar transistors to switch to the off state. The forced off state blocks the high-frequency parallel resonance amplification effect generated by the AC current output by the main circuit inverter topology and the inherent capacitive reactance parameters of the distribution line. After the control board executes the active blocking action, the microprocessor reads the node hardware identification code of the target distributed governance unit, concatenates the node hardware identification code with the blocking state feature code in a data format, generates a state update data frame, and the control board delivers the state update data frame to the fiber optic ring network via the second industrial Ethernet communication module.

[0015] Preferably, the control board in the unblocked distributed governance unit extracts the node hardware identifier code attached to the status update data frame. The microprocessor accesses the local capacity allocation status table to execute the matching and retrieval instruction of the node hardware identifier code, locks the distributed governance unit that sent the blocking status feature code, and deletes the node hardware identifier code and the remaining available compensation capacity ratio of the distributed governance unit that sent the blocking status feature code from the local capacity allocation status table. The microprocessor in the unblocked distributed governance unit disconnects the data transmission interaction link connecting the blocking node hardware identifier code in the peer-to-peer communication network, generates a reduced-order connectivity topology graph, and restarts the consistency iteration calculation.

[0016] This invention provides a distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network. It has the following beneficial effects: 1. This invention extracts the harmonic active power directional quantity by the edge computing chip in the distributed harmonic sensing node, retrieves the target receiver identifier from the distribution network topology mapping table by combining the harmonic active power directional quantity, and writes the target receiver identifier into the addressing information segment of the multicast data packet to achieve targeted push, thereby reducing the data transmission traffic inside the optical fiber ring network and reducing the communication delay between the distributed harmonic sensing node and the decentralized governance unit.

[0017] 2. This invention extracts the remaining available compensation capacity ratio of adjacent distributed governance units through the microprocessor within the distributed governance unit, substitutes the remaining available compensation capacity ratio into the compensation weight consistency iterative state update equation to perform loop calculation, and generates instruction compensation current by combining the converged target compensation capacity allocation weight. Capacity negotiation is completed using a peer-to-peer communication network, thus avoiding the overload shutdown problem caused by a single distributed governance unit reaching its upper limit.

[0018] 3. This invention uses a microprocessor to perform phasor multiplication of the equivalent harmonic impedance of the distribution line and the command compensation current to obtain the local harmonic voltage prediction deviation. When the local harmonic voltage prediction deviation is greater than the safe operating voltage distortion threshold, multiple sets of insulated gate bipolar transistors are forced to switch to the off state. A reduced-order connected topology is generated in the unblocked distributed governance unit to restart the consistency iterative calculation, thereby cutting off the parallel resonance amplification condition caused by the AC current output of the main circuit inverter topology and maintaining the system linkage control capability when a local node of the distribution network exits. Attached Figure Description

[0019] Figure 1 This is a diagram of the overall system architecture of the present invention; Figure 2 This is a block diagram of the internal structure of the distributed harmonic sensing node of the present invention; Figure 3 This is a block diagram of the hardware structure and main circuit topology of the distributed governance unit of the present invention; Figure 4 This is a data processing flowchart of the distributed harmonic sensing node of the present invention; Figure 5 This is a flowchart illustrating the control and iterative calculation process of the decentralized governance unit of the present invention. Figure 6 This is the prediction deviation calculation task and active locking logic diagram of the present invention; Figure 7 A comparison chart showing the convergence of the local consistency iterative calculation process for the compensation capacity allocation weights in this invention; Figure 8 This diagram illustrates the voltage distortion blocking control during the execution of the prediction deviation calculation task of this invention. Detailed Implementation

[0020] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] See attached document Figure 1 and attached Figure 2 The present invention provides a distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network, comprising distributed harmonic sensing nodes and distributed governance units; Distributed harmonic sensing nodes are deployed at the outgoing lines of substations, line branch nodes, and the low-voltage side of distribution transformers in the power distribution network. They integrate a high-frequency sampling module and an edge computing chip connected to the high-frequency sampling module. The high-frequency sampling module connects to the power distribution lines of the power distribution network via voltage and current transformers, continuously acquiring the instantaneous voltage and current waveforms of the power distribution lines. It converts these instantaneous voltage and current waveforms into discrete digital signals and transmits them to the edge computing chip. The edge computing chip uses a digital signal processor with a built-in Discrete Fourier Transform (DFT) algorithm. It receives the discrete digital signals, calls the DFT algorithm to calculate the parameters of each harmonic component, and generates these parameters.

[0022] The distributed harmonic sensing node also integrates a first industrial Ethernet communication module connected to an edge computing chip. The edge computing chip transmits the parameters of each harmonic component to the first industrial Ethernet communication module. The first industrial Ethernet communication module is connected to the existing fiber optic ring network of the power distribution network.

[0023] See attached document Figure 3 The distributed control unit is deployed at the nonlinear load access point within the distribution network, and includes a main circuit inverter topology, DC-side capacitors, filter inductors, and control boards. The main circuit inverter topology adopts a three-phase voltage-source bridge inverter circuit, which internally contains multiple sets of insulated-gate bipolar transistors. The AC side terminals are connected to one end of the filter inductor, and the other end of the filter inductor is connected in parallel to the distribution line. The DC side terminals are connected to both ends of the DC-side capacitor.

[0024] The control board is connected to the gates of multiple sets of insulated-gate bipolar transistors within the main circuit inverter topology. The distributed management unit contains a second industrial Ethernet communication module and a microprocessor, both electrically connected to the control board. The second industrial Ethernet communication module is also connected to a fiber optic ring network, and the first industrial Ethernet communication module establishes a communication connection with the second industrial Ethernet communication module via the fiber optic ring network.

[0025] The microprocessor receives multicast data packets via a second industrial Ethernet communication module and a control board. The microprocessor generates pulse-width modulation (PWM) signals based on the PWM data packets. The control board sends these PWM signals to the gates of multiple sets of insulated-gate bipolar transistors (IGBTs), controlling the switching on and off of the IGBTs. The main circuit inverter topology, combined with the DC voltage of the DC-side capacitor, generates AC current, which is then injected into the power distribution line via a filter inductor.

[0026] The first industrial Ethernet communication module and multiple second industrial Ethernet communication modules form a peer-to-peer communication network within the fiber optic ring network. The first industrial Ethernet communication module acts as the publisher and stores multiple sets of media access control addresses for the subscribers, while the second industrial Ethernet communication modules act as subscribers.

[0027] The edge computing chip internally stores a preset distortion threshold. It compares the parameters of each harmonic component with this threshold. When a harmonic component parameter exceeds the threshold, the chip generates a trigger signal and transmits it to the first industrial Ethernet communication module. The first industrial Ethernet communication module receives the trigger signal, encapsulates it using the subscriber's media access control address, generates a multicast data packet, and sends it to the fiber optic ring network.

[0028] The fiber optic ring network transmits multicast data packets. The second industrial Ethernet communication module listens to the data transmission traffic within the fiber optic ring network and verifies the data transmission traffic based on the internally set receiving address. When the receiving address matches the addressing information of the multicast data packet, the multicast data packet is extracted and transmitted to the control board. The microprocessor within the control board decodes and parses the multicast data packet, extracting the node hardware identifier code, time synchronization stamp, and harmonic amplitude change rate carried within the multicast data packet, which serve as the data source for subsequent closed-loop control.

[0029] See attached document Figure 5 During the time period when no multicast data packets are received, the control board in the distributed governance unit performs impedance identification calculations. The microprocessor in the control board continuously collects background harmonic voltage and background harmonic current in the distribution line through the AC side terminals of the main circuit inverter topology, and performs discretization numerical conversion on the background harmonic voltage and background harmonic current according to the internally set fixed sampling calculation period to generate a data sequence with corresponding discrete time index.

[0030] The microprocessor extracts numerical fluctuation parameters within the data sequence by calculating the amplitude of background harmonic current changes corresponding to two adjacent discrete-time indices. The microprocessor has a preset excitation dead-zone threshold that limits the validity of the data; the numerical fluctuation parameters are then compared to this threshold.

[0031] When the numerical fluctuation parameter is less than or equal to the excitation dead zone threshold, the microprocessor determines that the distribution network is in a steady state and lacks continuous excitation. In order to avoid the covariance matrix diverging, the microprocessor suspends the update of the internally fixed recursive least squares algorithm model and directly maintains and outputs the equivalent harmonic impedance of the distribution line at the previous moment.

[0032] When the numerical fluctuation parameter exceeds the excitation dead zone threshold, the microprocessor determines that a valid fluctuation excitation exists. Before performing calculations, the microprocessor converts the current background harmonic voltage, background harmonic current, and equivalent harmonic impedance of the distribution line into per-unit values ​​for normalization to eliminate dimensional conflicts between different physical quantities. The microprocessor inputs the normalized per-unit values ​​of the background harmonic voltage and background harmonic current as the calculation data source into its internally fixed recursive least squares algorithm model. The recursive least squares algorithm model does not rely on externally injected test pulses, but uses a high-quality calculation data source with a data filtering mechanism to perform iterative mathematical operations and continuously outputs the equivalent harmonic impedance of the distribution line.

[0033] The microprocessor maintains a rolling update state, storing the latest output of the equivalent harmonic impedance of the power distribution line in the microprocessor's internal memory.

[0034] The recursive least squares algorithm model internally includes a state update equation for online identification of the equivalent harmonic impedance of power distribution lines, specifically: ; In the formula, Indicates discrete-time index as The first time Equivalent harmonic impedance of secondary distribution lines; Indicates discrete-time index as The first time Equivalent harmonic impedance of secondary distribution lines; Indicates discrete-time index as The error correction gain is limited to a range of 0 to 1 by an internally preset forgetting factor. Indicates discrete-time index as The first time Per-unit value of sub-background harmonic voltage; Indicates discrete-time index as The first time Per-unit value of sub-background harmonic current.

[0035] The specific value is obtained through real-time iterative calculation of the internally preset forgetting factor and the covariance matrix of the input data. The specific calculation formula is as follows: ; In the formula, The forgetting factor is defined as being between 0 and 1. Indicates discrete-time index as The value of the covariance matrix at that time.

[0036] See attached document Figure 4The edge computing chip within the distributed harmonic sensing node continuously performs discretization extraction of data features during the time period when no trigger signal is generated. It uses an internally set fixed sliding time window to perform time-series segmentation of the generated harmonic component parameters, divides the data sets corresponding to different sliding time windows, and extracts the harmonic current amplitude corresponding to the sliding time window from the data set.

[0037] Before invoking the discrete differential algorithm, the edge computing chip performs time base alignment confirmation on two adjacent sliding time windows. After confirmation, the edge computing chip performs differential calculation on the harmonic current amplitudes corresponding to the two adjacent sliding time windows, and calculates and outputs the harmonic amplitude change rate, which characterizes the severity of high-frequency waveform distortion in the power distribution line. The harmonic amplitude change rate extraction equation set internally by the edge computing chip is as follows: ; In the formula, Indicates the first The first sliding time window within the Rate of change of subharmonic amplitude; Indicates the first The first sliding time window within the The amplitude of the second harmonic current; Indicates the first The first sliding time window within the The amplitude of the second harmonic current; This represents the calculation step size between adjacent sliding time windows. The lower limit of the calculation step size is constrained by the Nyquist sampling theorem, specifically set to a positive integer multiple of the sampling period of the high-frequency sampling module, and is forcibly set to a constant greater than 0.

[0038] The edge computing chip within the distributed harmonic sensing node calculates the harmonic active power directional quantity while extracting the rate of change of harmonic amplitude. It selects the harmonic voltage phasor and harmonic current phasor corresponding to the target frequency band from the parameters of each harmonic component, extracts the corresponding harmonic voltage phase angle and harmonic current phase angle to obtain the phase angle difference, and calls the internally integrated multiplier to perform a product operation on the cosine value of the harmonic voltage phasor amplitude, the harmonic current phasor amplitude and the phase angle difference, and outputs the harmonic active power value.

[0039] The edge computing chip extracts the sign of the harmonic active power value and generates a harmonic active power directional quantity based on the sign. This directional quantity characterizes the propagation direction of nonlinear power flow within the distribution line. A positive value indicates that the nonlinear power flow is propagating towards downstream nodes in the distribution network topology. A negative value indicates that the nonlinear power flow is propagating towards upstream nodes in the distribution network topology. The edge computing chip within the distributed harmonic sensing node stores the harmonic amplitude change rate and the harmonic active power directional quantity in a local cache sequence for later use.

[0040] The distributed harmonic sensing node is equipped with a continuous-cycle counter. The edge computing chip extracts the harmonic amplitude change rate from the local cache sequence, calculates the magnitude of the harmonic amplitude change rate, and compares the magnitude with a preset distortion threshold. When the magnitude is greater than the preset distortion threshold, the edge computing chip sends a counting command to the continuous-cycle counter, which then increments the count accordingly. When the magnitude is less than or equal to the preset distortion threshold, the edge computing chip sends a reset command to the continuous-cycle counter, which then resets its internal count value to 0.

[0041] The edge computing chip internally stores a preset period constant. When the count value inside the continuous period counter reaches the preset period constant, the edge computing chip determines that the power distribution line is in a harmonic abrupt change state and generates a trigger signal. In response to the trigger signal, the edge computing chip calls its internal memory to perform snapshot latching on the rate of change of harmonic amplitude and the direction of harmonic active power within the current sliding time window. While latching the data, the high-frequency sampling module and the discrete Fourier transform algorithm maintain a continuous loop in the background, continuously generating new parameters for each harmonic component to ensure the continuity of dynamic monitoring. The edge computing chip uses its internal interrupt mechanism to transfer the latched data to the data packet encapsulation task for processing.

[0042] When the edge computing chip generates a trigger signal, it reads the harmonic amplitude change rate and harmonic active power direction quantity from its local cache sequence. The distributed harmonic sensing node integrates a high-precision clock module and pre-stores node hardware identification codes. The edge computing chip obtains the current time synchronization stamp via the high-precision clock module, reads the pre-stored node hardware identification code from the distributed harmonic sensing node, and combines the node hardware identification code, time synchronization stamp, harmonic amplitude change rate, and harmonic active power direction quantity in sequence to generate a valid data segment. The first industrial Ethernet communication module receives the valid data segment, generates a communication frame header and a cyclic redundancy check (CRC) code, and encapsulates the communication frame header, valid data segment, and CRC code into a multicast data packet.

[0043] The first industrial Ethernet communication module sends multicast data packets to the fiber optic ring network. The packets only encapsulate the node hardware identifier, time synchronization stamp, harmonic amplitude change rate, and harmonic active power direction; they do not encapsulate the instantaneous voltage and current waveforms collected by the high-frequency sampling module. The multicast packet has a fixed byte length, occupying less channel transmission bandwidth in the fiber optic ring network than the instantaneous waveform data. This fixed byte length reduces encapsulation time within the distributed harmonic sensing nodes and queuing buffer time at the switching nodes within the fiber optic ring network, thus lowering the transmission delay from the distributed harmonic sensing nodes to the distributed governance unit.

[0044] Before the first industrial Ethernet communication module encapsulates multicast data packets, the edge computing chip within the distributed harmonic sensing node dynamically divides the communication addressing range based on the harmonic active power direction. The memory inside the distributed harmonic sensing node pre-stores a distribution network topology mapping table, which is divided into upstream node address sets and downstream node address sets. The upstream node address set records the subscription media access control addresses within multiple distributed governance units located on the power supply side of the distributed harmonic sensing node, while the downstream node address set records the subscription media access control addresses within multiple distributed governance units located on the load side of the distributed harmonic sensing node.

[0045] The edge computing chip reads the harmonic active power directional quantity from the local cache sequence and extracts and compares the numerical signs of the harmonic active power directional quantity. When the harmonic active power directional quantity is positive, it is determined that the nonlinear power flow is propagating to downstream nodes of the distribution network topology. The downstream node address set is retrieved from the distribution network topology mapping table, and the subscription end media access control address in the set is selected as the target receiver identifier. When the harmonic active power directional quantity is negative, it is determined that the nonlinear power flow is propagating to upstream nodes of the distribution network topology. The upstream node address set is retrieved from the distribution network topology mapping table, and the subscription end media access control address in the set is selected as the target receiver identifier.

[0046] The edge computing chip transmits the selected target receiver identifier to the first industrial Ethernet communication module. The first industrial Ethernet communication module receives the target receiver identifier and writes it into the addressing information segment of the multicast data packet. After completing the writing of the addressing information segment, it sends a multicast data packet carrying the target receiver identifier to the fiber optic ring network.

[0047] The fiber optic ring network transmits multicast data packets to matching distributed governance units based on the addressing information segment within the multicast data packet. When the receiving address of the second industrial Ethernet communication module does not match the addressing information segment, the second industrial Ethernet communication module intercepts and cleans up the multicast data packets transmitted through the fiber optic ring network. Distributed harmonic sensing nodes select different subscription end media access control addresses based on the positive and negative signs of the harmonic active power direction quantity, establishing a directional multicast communication topology within the fiber optic ring network to achieve directional push of multicast data packets to the distributed governance unit corresponding to the target receiving end.

[0048] The second industrial Ethernet communication module extracts multicast data packets from the fiber optic ring network. The control board within the distributed governance unit performs a communication handshake and state initialization. The microprocessor within the control board decodes and parses the multicast data packets, extracting the time synchronization stamps carried within them. These time synchronization stamps are used to align the time bases of different distributed governance units within the fiber optic ring network. Multiple distributed governance units receiving the same clock base form a local communication neighbor set within the fiber optic ring network.

[0049] The main circuit inverter topology has a pre-stored rated compensation capacity and monitors and generates the current output capacity in real time. The microprocessor in the distributed management unit reads the rated compensation capacity and the current output capacity of the main circuit inverter topology, performs a subtraction operation to obtain the difference between the rated compensation capacity and the current output capacity, and generates the remaining available compensation capacity. The microprocessor has a pre-set minimum positive tolerance. The microprocessor adds the rated compensation capacity and the minimum positive tolerance to form a corrected denominator, and performs a division operation with the remaining available compensation capacity and the corrected denominator to generate the percentage of remaining available compensation capacity. The control board combines the percentage of remaining available compensation capacity with the node hardware identifier code set inside the distributed management unit and encapsulates it to generate a status initialization data frame.

[0050] The control board transmits the state initialization data frame to the second industrial Ethernet communication module. The second industrial Ethernet communication module sends the state initialization data frame to the local communication neighbor set. All distributed governance units within the local communication neighbor set send and receive state initialization data frames to each other, completing the communication handshake interaction.

[0051] The second industrial Ethernet communication module receives status initialization data frames sent by other distributed governance units within the local communication neighbor set and transmits these frames to the control board. The microprocessor parses the status initialization data frames, extracts the remaining available compensation capacity percentage and node hardware identifier code of the other distributed governance units, establishes a local capacity allocation status table in the control board's memory, and writes the corresponding node hardware identifier codes and remaining available compensation capacity percentages of the other distributed governance units into the local capacity allocation status table, thus completing the status initialization parameter update.

[0052] After the distributed governance unit completes the state initialization parameter update, the microprocessor extracts the harmonic amplitude change rate from the multicast data packet, reads the node hardware identification code of other distributed governance units recorded in the local capacity allocation status table, constructs a connected topology graph of the local communication neighbor set based on the node hardware identification code, counts the number of directly adjacent nodes of the target distributed governance unit in the connected topology graph, and generates the local node connection degree.

[0053] The microprocessor internally stores an initial fixed step size constant and a step size mapping adjustment coefficient. The microprocessor calls the maximum value function to perform positive truncation processing on the harmonic amplitude change rate, that is, extracting the maximum value between the constant 0 and the harmonic amplitude change rate. The extracted truncation value is multiplied with the step size mapping adjustment coefficient to generate the step size adjustment increment. The initial fixed step size constant is added to the step size adjustment increment to generate an unconstrained adaptive step size. This positive truncation processing only extracts the positive change rate of the harmonic that increases sharply to participate in the step size adjustment, and forces the negative change rate to zero to prevent impedance estimation overshoot when the harmonic drops sharply.

[0054] The microprocessor divides the constant 1 by the sum of the local node connection degree plus 1 to generate the convergence boundary threshold. The denominator of the division operation consists of the local node connection degree plus the constant 1.

[0055] The microprocessor compares the unconstrained adaptive step size with the convergence boundary threshold. When the unconstrained adaptive step size is less than or equal to the convergence boundary threshold, it is used as the constrained adaptive step size. When the unconstrained adaptive step size is greater than the convergence boundary threshold, the convergence boundary threshold is used as the constrained adaptive step size, ensuring that the constrained adaptive step size remains within the numerical range defined by the convergence boundary threshold.

[0056] The microprocessor internally has a constrained adaptive step-size mapping equation as follows: ; In the formula, Indicates the first The first sliding time window within the Secondary constrained adaptive step size; This represents a function that takes the minimum value. This represents the initial fixed step size constant, with a value set between 0 and 0.1. This represents the step size mapping adjustment coefficient, whose value is set as the reciprocal of the preset maximum harmonic change rate, used to normalize the change rate to a reasonable step size increment range; This indicates that the value is 0 and the first... The maximum of the two rates of change of the amplitude of the second harmonic; Indicates the first The first sliding time window within the Rate of change of subharmonic amplitude; This represents the local node connectivity degree of the target decentralized governance unit. The formula uses the reciprocal of the denominator plus 1 as the convergence boundary, which absolutely guarantees the random matrix characteristics of the subsequent capacity allocation weight iteration matrix. This avoids the problem of negative weight coefficients and iteration divergence oscillation caused by large node degrees, while also avoiding the defects of difficulty in obtaining global eigenvalues ​​and suppression of step size limits.

[0057] After the microprocessor extracts the restricted adaptive step size, it initiates the local consistency iterative calculation of the compensation capacity allocation weight, reads the remaining available compensation capacity percentage recorded in the local capacity allocation status table, sets the remaining available compensation capacity percentage of the target decentralized governance unit as the initial iterative value of the compensation capacity allocation weight, extracts the remaining available compensation capacity percentage of adjacent decentralized governance units in the local communication neighbor set, and writes the remaining available compensation capacity percentage of adjacent decentralized governance units into the initial iterative value item of the adjacent decentralized governance units.

[0058] The microprocessor substitutes the constrained adaptive step size, the compensation capacity allocation weights of the target decentralized governance unit, and the compensation capacity allocation weights of adjacent decentralized governance units into the internally fixed compensation weight consistency iterative state update equation and performs loop calculations. The compensation weight consistency iterative state update equation is as follows: ; In the formula, Indicates the number of iterations. The weighting of compensation capacity allocation for decentralized governance units at different times; Indicates the number of iterations. The weighting of compensation capacity allocation for decentralized governance units at different times; Indicates a constrained adaptive step size; Represents a local communication neighbor set; Indicates the index number of adjacent decentralized governance units within the local communication neighbor set; This indicates that the target decentralized governance unit and the index number are... The communication topology adjacency coefficient between adjacent decentralized governance units is defined as follows: if the target decentralized governance unit has a direct communication link with the adjacent decentralized governance unit, the value is set to 1; if there is no direct communication link, the value is set to 0. Indicates the number of iterations. The time index number is The compensation capacity allocation weight of adjacent decentralized governance units.

[0059] After a single loop operation, the microprocessor calculates the number of iterations. The compensation capacity allocation weight and iteration number are as follows: The difference magnitude between the compensation capacity allocation weights is calculated. The microprocessor internally stores a preset convergence tolerance threshold, and compares the difference magnitude with this threshold. When the difference magnitude is greater than the preset threshold, the microprocessor continues with the next iteration. When the difference magnitude is less than or equal to the preset threshold, all distributed governance units within the local communication neighbor set reach data consensus, the control board interrupts the loop operation, and the microprocessor outputs the converged target compensation capacity allocation weights.

[0060] The microprocessor reads the pre-stored effective value output current parameters of the main circuit inverter topology under rated operating conditions from memory. It then multiplies the convergence target compensation capacity allocation weight with these parameters, setting the multiplication result as the command compensation current and generating a pulse width modulation (PWM) signal. The control board sends this PWM signal to the gates of multiple insulated-gate bipolar transistors (IGBTs) within the main circuit inverter topology. The main circuit inverter topology controls the conduction state of these IGBTs according to the PWM signal, outputting the corresponding AC current to the power distribution line.

[0061] See attached document Figure 6 Before generating the pulse width modulation signal, the control board within the distributed governance unit initiates a prediction deviation calculation task. The microprocessor within the control board accesses memory space, extracts the equivalent harmonic impedance of the distribution line reflecting the local grid characteristics of the distribution line, and synchronously retrieves the instruction compensation current generated by the local consistency iterative calculation, which characterizes the amount of adjustment current.

[0062] The microprocessor has a built-in equation for calculating the local harmonic voltage prediction deviation. It performs a phasor multiplication operation between the equivalent harmonic impedance of the power distribution line and the command compensation current, and combines this with the local harmonic voltage prediction deviation calculation equation to output the local harmonic voltage prediction deviation. Specifically: ; In the formula, Indicates discrete-time index as The first time Prediction deviation of sub-local harmonic voltage; Indicates discrete-time index as The first time Secondary command compensation current; Indicates discrete-time index as The first time Equivalent harmonic impedance of secondary distribution lines.

[0063] The local harmonic voltage prediction deviation characterizes the distortion increment at the distribution line connection end after the actual output AC current of the main circuit inverter topology. Before the control board sends pulse width modulation signals to the gates of multiple sets of insulated gate bipolar transistors, the microprocessor completes the numerical calculation of the distortion increment in advance. The timing of the prediction calculation is earlier than the actual conduction time of the internal devices of the main circuit inverter topology, and the acquired local harmonic voltage prediction deviation is written into the cache register as the comparison data source for subsequent active blocking determination.

[0064] After the control board within the distributed governance unit completes the writing of the local harmonic voltage prediction deviation, the microprocessor retrieves the local harmonic voltage prediction deviation from the cache register. The control board has preset safe operating voltage distortion thresholds and tolerance delay calculation cycles embedded within it. The microprocessor executes numerical comparison instructions to determine the relationship between the local harmonic voltage prediction deviation and the safe operating voltage distortion threshold.

[0065] When it is determined that the local harmonic voltage prediction deviation exceeds the safe operating voltage distortion threshold, and the duration calculated from the first continuous sampling exceeding the limit recorded by the internal clock is longer than the tolerance delay calculation period, the microprocessor determines that there is a high-frequency parallel resonance risk at the power distribution line access point, generates a hardware interrupt signal, and the control board cuts off the pulse width modulation signal transmission channel according to the hardware interrupt signal. The microprocessor generates an active blocking level signal, which is directly transmitted by the control board to the gates of multiple sets of insulated gate bipolar transistors inside the main circuit inverter topology.

[0066] Multiple sets of insulated gate bipolar transistors receive active latching level signals and are forced to switch to the off state. The main circuit inverter topology stops outputting AC current to the distribution line, and the output current is reduced to 0A. The forced off state blocks the high-frequency parallel resonance amplification effect generated by the AC current output by the main circuit inverter topology and the inherent capacitive reactance parameter of the distribution line, thus limiting the voltage waveform distortion rate inside the distribution line.

[0067] When the predicted deviation of the local harmonic voltage is determined to be less than or equal to the safe operating voltage distortion threshold, the control board maintains the original internal signal transmission channel. The microprocessor sends pulse width modulation signals to the gates of multiple sets of insulated gate bipolar transistors inside the main circuit inverter topology, and the main circuit inverter topology outputs the corresponding AC current to the power distribution line according to the instruction compensation current.

[0068] After the control board executes the active interlocking action, the interlocking status feature code is embedded within the control board. The microprocessor reads the node hardware identification code of the target distributed governance unit, concatenates the node hardware identification code with the interlocking status feature code according to the data format, and generates a status update data frame. The control board transmits the status update data frame to the second industrial Ethernet communication module, which then sends the status update data frame to the fiber optic ring network.

[0069] The fiber optic ring network broadcasts the status update data frame to the local communication neighbor set. Other decentralized governance units within the local communication neighbor set receive the status update data frame. The control boards of other decentralized governance units re-trigger the local consistency iterative calculation of the compensation capacity allocation weight based on the blocking status feature code carried in the status update data frame, and reallocate the instruction compensation current while removing the blocking node.

[0070] Unlocked distributed governance units within the local communication neighbor set receive status update data frames containing lockout status feature codes via the second industrial Ethernet communication module. The control board within each unlocked distributed governance unit extracts the node hardware identifier code attached to the status update data frame. The microprocessor within the control board accesses the local capacity allocation status table, executes a matching and retrieval instruction for the node hardware identifier code, identifies the distributed governance unit that sent the lockout status feature code, and deletes the corresponding distributed governance unit's node hardware identifier code and remaining available compensation capacity percentage from the local capacity allocation status table. After data deletion is complete, the microprocessor reads the connectivity topology graph in memory, disconnects the data transmission interaction link corresponding to the lockout node's hardware identifier code, and generates a reduced-order connectivity topology graph.

[0071] The microprocessor restarts the consensus iteration calculation according to the reduced-order connectivity topology graph. The decentralized governance units participating in the consensus iteration calculation are limited to the remaining nodes in the local communication neighbor set that are in a normal communication state. The microprocessor continues to cycle through the compensation capacity allocation weights in the environment of removing locked nodes until the difference magnitude of the cyclic calculation reaches the limit of the preset convergence tolerance threshold, and then stops the cyclic calculation and outputs the target compensation capacity allocation weights for secondary convergence.

[0072] The microprocessor retrieves the rated output current parameters of the main circuit inverter topology and multiplies them with the target compensation capacity allocation weights for secondary convergence to calculate and output the secondary command compensation current. After the secondary command compensation current is generated, the microprocessor again retrieves the equivalent harmonic impedance of the distribution line and performs phasor multiplication to calculate the corrected local harmonic voltage prediction deviation. When it is determined that the corrected local harmonic voltage prediction deviation is less than or equal to the safe operating voltage distortion threshold, the microprocessor determines that the secondary command compensation current meets the safe issuance conditions.

[0073] The unblocked decentralized governance unit within the local communication neighbor set retrieves the time synchronization stamp recorded in memory. The microprocessor has an internal time alignment comparison function that performs a correlation operation with the time synchronization stamp to generate a global synchronization trigger edge signal, which is then combined with the secondary instruction compensation current to generate a secondary pulse width modulation signal.

[0074] The microprocessor has preset trigger value boundaries that limit the execution conditions. When the global synchronization trigger edge signal reaches the trigger value boundary, the control board in the unblocked distributed management unit relies on the global synchronization trigger edge signal to synchronously send the secondary pulse width modulation signal to the gates of multiple sets of insulated gate bipolar transistors inside the main circuit inverter topology.

[0075] Multiple sets of insulated-gate bipolar transistors (IGBTs) switch synchronously on and off according to the duty cycle parameters of the second pulse width modulation (PWM) signal. The main circuit inverter topology of the unblocked distributed control unit synchronously outputs the corresponding AC current to the distribution line. The multi-node closed-loop linkage regulation is completed, fully realizing the dynamic monitoring and sensing of distribution network harmonics and the linkage control of multiple distributed control units.

[0076] Specific application examples: Based on the specific harmonic mitigation nodes at the nonlinear load connection point, the constants and initial parameters set in this invention are as follows: The calculation step size between adjacent sliding time windows is set to 0.02s; the initial fixed step size constant is set to 0.05; the local node connectivity degree of the target distributed governance unit is counted as 3; and the safe operating voltage distortion threshold is set to 5.0V.

[0077] The microprocessor extracts the harmonic amplitude change rate of 150.0 A / s carried by the multicast data packet and reads the internally stored step size mapping adjustment coefficient of 0.001. The microprocessor performs multiplication and addition operations to generate an unconstrained adaptive step size of 0.05 + 0.001 × 150.0 = 0.20. The microprocessor performs a division operation on the sum of constant 1 and the local node connection degree 3 and constant 1 to generate a convergence boundary threshold of 0.250.

[0078] The microprocessor invokes the constrained adaptive step-size mapping equation to perform numerical comparison: ; It was determined that the unconstrained adaptive step size of 0.20 was less than the convergence boundary threshold of 0.250. The microprocessor then used 0.20 as the constrained adaptive step size in subsequent calculations, thus limiting the risk of numerical divergence in the local consistency iterative calculation.

[0079] The microprocessor reads the first value generated by the local consistency iterative calculation. The compensation current for the next instruction is 12.5A, and the synchronous call to the recursive least squares algorithm model continuously outputs the first... The equivalent harmonic impedance of the secondary distribution line is 0.6Ω. The microprocessor is then used to calculate the local harmonic voltage prediction deviation equation: ; The microprocessor within the control board executes a numerical comparison instruction and detects that the local harmonic voltage prediction deviation of 7.5V is greater than the preset safe operating voltage distortion threshold of 5.0V. When the duration calculated from the first continuous sampling exceeding the limit, as recorded by the clock, is longer than the tolerance delay calculation period, the microprocessor immediately generates an active latching level signal, forcing multiple sets of insulated-gate bipolar transistors to switch to the off state, thus blocking the high-frequency parallel resonant amplification effect.

[0080] Actual operation tests were conducted and data were compared. Specific verification results are detailed in the appendix. Figure 7 and attached Figure 8 The experiment compared the operating data of the traditional centralized control mode with those of the present invention under the same power distribution line and background harmonic abrupt change conditions. The experimental data are recorded in Table 1.

[0081] Table 1. Comparison of Operating Data between Traditional Centralized Control Mode and the Invention

[0082] It should be noted that the traditional centralized control mode relies on the master node to directly issue commands and allocate resources. There is no peer-to-peer communication and consistency iteration process between local nodes. Therefore, the evaluation metric of the number of local consistency iteration convergence steps is not applicable in the traditional control mode.

[0083] See attached document Figure 7 And Table 1, Appendix Figure 7 This diagram compares the convergence of the local consistency iterative calculation process for compensating capacity allocation weights. The horizontal axis represents the number of iterations in the loop operation, dimensionless, ranging from 0 to 40. The vertical axis represents the compensating capacity allocation weights, dimensionless, ranging from 0.2 to 0.8. The diagram includes both fixed-step-size and constrained adaptive-step-size iterative trajectories.

[0084] In the initial stage of the cyclic operation iteration from 0 to 10, the fixed step size iteration trajectory does not introduce the harmonic amplitude change rate for dynamic adjustment, resulting in very slow weight convergence and numerical oscillations when approaching the target consensus value. This invention uses a microprocessor to substitute a restricted adaptive step size into the compensation weight consistency iteration state update equation for cyclic operation. The restricted adaptive step size iteration trajectory exhibits a fast and smooth convergence trend. When the cyclic operation iteration reaches 18 times, the difference amplitude rapidly shrinks to within the preset convergence tolerance threshold without numerical oscillations, achieving a rapid balance of the remaining available compensation capacity among multiple decentralized governance units.

[0085] See attached document Figure 8 And Table 1, Appendix Figure 8This diagram illustrates the voltage distortion blocking control during the execution of the prediction deviation calculation task. The horizontal axis represents the discrete-time index, dimensionless, with a value range of 0 to 200. The left vertical axis represents the actual voltage distortion increment of the distribution line, in volts (V), with a value range of 0 to 50; the right vertical axis represents the output current of the main circuit inverter topology, in amperes (A), with a value range of 0 to 20. The diagram includes voltage distortion curves for the traditional control mode, distribution line voltage distortion curves, and main circuit inverter topology output current curves.

[0086] When the discrete-time index reaches around 100, high-frequency background harmonics are injected into the power distribution line. Traditional control mode lacks prediction deviation calculation tasks and underlying blocking logic. The main circuit inverter topology continuously outputs AC current, which leads to an amplification of the voltage distortion increment in the actual power distribution line. The voltage distortion curve of traditional control mode shows that the highest high-frequency parallel resonant overvoltage peak reaches 45.8V.

[0087] Before the control board sends pulse width modulation signals to the gates of multiple sets of insulated gate bipolar transistors, the microprocessor pre-calculates that the predicted deviation of the local harmonic voltage is greater than the safe operating voltage distortion threshold. As can be seen from the output current curve of the main circuit inverter topology, the microprocessor decisively generates an active blocking level signal at discrete time index 115, causing the output current of the main circuit inverter topology to instantly drop to 0A. This forced shutdown state blocks the AC current output path, and the voltage distortion curve of the distribution line is strictly suppressed to the lower end of the 5.2V range throughout the entire operating cycle, avoiding the risk of equipment damage caused by high-frequency parallel resonant amplification effects.

Claims

1. A distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network, characterized in that, include: The distributed harmonic sensing node integrates an edge computing chip and a first industrial Ethernet communication module. The edge computing chip generates parameters for each harmonic component. When the parameters for each harmonic component are greater than a preset distortion threshold, the edge computing chip generates a trigger signal. The first industrial Ethernet communication module encapsulates the trigger signal to generate a multicast data packet. The distributed governance unit includes a main circuit inverter topology with multiple sets of insulated gate bipolar transistors inside, a DC-side capacitor, a second industrial Ethernet communication module, a control board, and a microprocessor. The second industrial Ethernet communication module and the microprocessor are electrically connected to the control board. The distributed harmonic sensing node and the distributed governance unit form a peer-to-peer communication network within the optical fiber ring network. The microprocessor generates a pulse width modulation signal based on the multicast data message to control the on and off of the multiple sets of insulated gate bipolar transistors. The main circuit inverter topology generates an alternating current that is injected into the power distribution line in combination with the DC voltage of the DC side capacitor.

2. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 1, characterized in that, The distributed harmonic sensing node integrates a high-frequency sampling module, which continuously acquires the instantaneous voltage waveform and instantaneous current waveform of the power distribution line and converts them into digital discrete signals. The edge computing chip calls the internally embedded discrete Fourier transform algorithm to calculate the digital discrete signal and generate the parameters of each harmonic component.

3. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 1, characterized in that, During the time period when the control board does not receive the multicast data packet, the microprocessor continuously collects background harmonic voltage and background harmonic current in the power distribution line to generate a data sequence. The microprocessor extracts the numerical fluctuation parameters within the data sequence and compares them with an internally preset excitation dead zone threshold. When the numerical fluctuation parameters are greater than the excitation dead zone threshold, the background harmonic voltage, background harmonic current, and equivalent harmonic impedance of the distribution line are converted into per-unit values ​​for normalization. The normalized per-unit values ​​of the background harmonic voltage and background harmonic current are used as the calculation data source and input into the internally fixed recursive least squares algorithm model for continuous iterative calculation, outputting the equivalent harmonic impedance of the distribution line.

4. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 2, characterized in that, The edge computing chip uses an internally set fixed sliding time window to perform data timing segmentation on the generated harmonic component parameters and extracts the harmonic current amplitude corresponding to the sliding time window. The edge computing chip performs differential calculations on the harmonic current amplitudes corresponding to two adjacent sliding time windows, and calculates and outputs the harmonic amplitude change rate, which characterizes the severity of high-frequency waveform distortion in the power distribution line.

5. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 4, characterized in that, The edge computing chip selects harmonic voltage phasors and harmonic current phasors from the parameters of each harmonic component and extracts the phase angle difference, performs a product operation and outputs the harmonic active power value. The edge computing chip determines the sign of the harmonic active power value and extracts the harmonic active power direction quantity, which represents the propagation direction of the nonlinear power flow inside the power distribution line.

6. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 5, characterized in that, The edge computing chip performs dynamic division of the communication addressing range based on the harmonic active power direction quantity, and retrieves the downstream node address set or the upstream node address set from the distribution network topology mapping table. The edge computing chip selects the media access control address of the subscriber within the set as the target receiver identifier and writes it into the addressing information segment of the multicast data packet to achieve targeted push.

7. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 3, characterized in that, The microprocessor reads the difference between the rated compensation capacity and the current output capacity of the main circuit inverter topology and generates the percentage of remaining available compensation capacity. The control board combines and encapsulates the remaining available compensation capacity percentage with the node hardware identifier code set inside the decentralized governance unit to generate a state initialization data frame. All decentralized governance units within the local communication neighbor set send and receive the state initialization data frame to each other. The microprocessor parses the state initialization data frame, establishes a local capacity allocation state table in memory, and writes the node hardware identifier code and the remaining available compensation capacity percentage into the local capacity allocation state table.

8. The distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 7, characterized in that, The microprocessor generates a restricted adaptive step size based on the local node connectivity degree and the preset harmonic amplitude change rate, extracts the remaining available compensation capacity ratio of adjacent decentralized governance units within the local communication neighbor set, and substitutes the restricted adaptive step size into the compensation weight consistency iterative state update equation to perform loop calculation. When the difference magnitude of the loop operation is less than or equal to the preset convergence tolerance threshold, the microprocessor outputs the convergence target compensation capacity allocation weight, and performs multiplication calculations based on the effective value output current parameters of the main circuit inverter topology under rated operating conditions to generate the instruction compensation current.

9. A distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 8, characterized in that, The microprocessor performs a phasor multiplication operation between the equivalent harmonic impedance of the power distribution line and the command compensation current, and outputs the local harmonic voltage prediction deviation. When it is determined that the local harmonic voltage prediction deviation is greater than the safe operating voltage distortion threshold, and the duration calculated from the first continuous sampling exceeding the limit recorded by the internal clock is longer than the tolerance delay calculation period, the microprocessor generates an active blocking level signal to force the multiple sets of insulated gate bipolar transistors to switch to the off state. After the control board performs the active interlocking action, the microprocessor reads the node hardware identification code of the target distributed governance unit, concatenates the node hardware identification code with the interlocking status feature code in a data format, generates a status update data frame, and the control board delivers the status update data frame to the optical fiber ring network via the second industrial Ethernet communication module.

10. A distributed control and linkage regulation system for dynamic monitoring of harmonics in a power distribution network according to claim 9, characterized in that, The control board in the unlocked distributed governance unit extracts the node hardware identification code attached to the status update data frame. The microprocessor accesses the local capacity allocation status table to execute the node hardware identification code matching and retrieval instruction, locks the distributed governance unit that sent the lockout status feature code, and deletes the node hardware identification code and the remaining available compensation capacity ratio of the corresponding distributed governance unit from the local capacity allocation status table. The microprocessor within the unlocked decentralized governance unit disconnects the data transmission interaction link in the peer-to-peer communication network corresponding to the hardware identifier code of the locked node, generates a reduced-order connectivity topology graph, and restarts the consistency iteration calculation.