Online identification method and system for internal wiring of user based on loading reactive power

Abstract

The application discloses a kind of based on loading reactive power user internal wiring online identification method and system, belong to distribution network operation and maintenance technical field.The system includes host unit and multiple slave unit;Host unit is located in distribution system total incoming line end, include programmable reactive power generation module and signal generation and synchronous control module, for generating and injecting the reactive power characteristic signal based on pseudo-random binary sequence, and broadcast synchronization frame to each slave;Slave unit is installed at each outgoing switch, include current sampling and signal conditioning module and slave DSP, for acquiring the current signal of this branch under microsecond level synchronous time base, and extracting characteristic intensity index by cross-correlation operation;Data fusion and decision module is located in host end, for carrying out connectivity decision according to threshold to each branch, generates switch-load corresponding relationship.The application realizes the fast, accurate identification of user internal wiring relationship under the condition of not power off.

Description

Technical Field

[0001] This invention patent belongs to the field of power distribution network operation and maintenance technology, and in particular to a method and system for online identification of user internal wiring based on applied reactive power. Background Technology

[0002] Currently, traditional load sensing technologies are mainly divided into two categories: invasive and non-invasive. Invasive load monitoring is a direct measurement method that accurately collects electrical data by installing a distributed sensor network on each device under test or power distribution branch. This method offers accurate identification results, reliable data, and extremely high measurement precision. However, this method requires disconnecting the existing wiring for installation, making it unsuitable for critical locations such as hospitals, data centers, and production lines where power outages are not permitted. Non-invasive load monitoring, on the other hand, does not require connection to internal wiring and can perform load detection without interrupting power supply, offering advantages such as low cost and simple installation. Typical methods include those based on VI trajectories, steady-state characteristics, and transient characteristics. However, this type of method heavily relies on advanced identification algorithms, is prone to monitoring blind spots in complex power consumption scenarios, and has limited ability to identify loads in specific states.

[0003] Therefore, accurately identifying the user's internal wiring method and its corresponding load switch, and achieving rapid identification without power interruption, has become a key research focus. In recent years, Digital Signal Processors (DSPs), with their advantages of high flexibility, high precision, and high stability, have gradually replaced traditional analog signal processing methods in numerous applications, providing a new path for load identification. In the load identification and detection process, DSP technology is widely used for tasks such as real-time acquisition of power grid voltage and current signals, dynamic data display, keyboard access, and communication. Leveraging the high-speed signal acquisition capabilities of DSPs and analog signal carrier communication algorithms, efficient data transmission between the signal generator and the slave device is achieved, significantly improving the reliability of the communication process and system stability.

[0004] However, despite the widespread application of DSPs in the efficient identification of switch-load relationships due to their high-speed signal acquisition and processing capabilities, they still face challenges in practical applications. DSPs are susceptible to background noise interference when processing signals, leading to decreased detection accuracy. Simultaneously, all detection units must maintain microsecond-level time synchronization. In long-distance, multi-node low-voltage power distribution environments, synchronization signals are easily affected by transmission delays and noise, causing phase deviations. If synchronization fails, the data from each node cannot be aligned, resulting in the failure of correlation detection of characteristic signals. Therefore, to address the difficulty in quickly locating power switches corresponding to critical loads such as hospital ICUs and operating rooms, there is an urgent need to establish a method for rapidly and accurately identifying the user's internal wiring methods under uninterrupted power supply conditions. Summary of the Invention

[0005] This invention aims to overcome the shortcomings of existing technologies and solve the technical challenge of quickly and accurately identifying the correspondence between switches and loads within a user's internal system under uninterrupted power supply conditions. To this end, this invention provides a method and device for online identification of user internal wiring based on reactive power. This solution integrates the precise controllability of power electronic converters with the high-speed acquisition and processing capabilities of digital signal processors. By injecting characteristic reactive power signals into the user's power distribution system and performing synchronous detection, it achieves rapid and accurate identification of switch-load connections without affecting normal power supply.

[0006] The technical problem solved by this invention patent is achieved through the following technical solution:

[0007] The first aspect of the present invention is to provide a method for online identification of user internal wiring based on reactive power, comprising the following steps:

[0008] Step S1: Connect the host to the main incoming line of the power distribution system of the user under test, install multiple slaves at each outgoing switch to be identified, and realize microsecond-level clock synchronization between the host and each slave through broadcast synchronization frames;

[0009] Step S2: The host generates a reactive power characteristic signal and injects it into the main incoming line of the power distribution system; the reactive power characteristic signal is generated based on a pseudo-random binary sequence and is modulated onto a carrier wave orthogonal to the phase of the grid voltage.

[0010] Step S3: Each slave device synchronously acquires the current signal of its own branch under a unified time base, preprocesses the acquired signal, calculates its cross-correlation function with the local feature template, and obtains the feature intensity index.

[0011] Step S4: Each slave device uploads its characteristic strength index to the host. The host performs connectivity determination on each branch based on the threshold and generates a switch-load correspondence.

[0012] Step S5: The system performs self-verification based on the known topology loop and optimizes the injected signal parameters and threshold.

[0013] Furthermore, clock synchronization in step S1 includes:

[0014] The master broadcasts a synchronization frame containing a timestamp. Each slave calibrates its local clock based on a clock synchronization protocol and synchronizes using a time difference measurement and compensation model.

[0015] Furthermore, the generation of the reactive power characteristic signal in step S2 includes:

[0016] The host DSP runs a feature encoding algorithm to generate a pseudo-random binary sequence, which is then modulated into a reactive current command after pulse shaping filtering. The reactive current is then driven by the controller to drive the PWM inverter to output the reactive current.

[0017] Furthermore, the preprocessing in step S3 includes:

[0018] After anti-aliasing synchronous sampling, removal of fundamental component, and noise reduction of the acquired current signal, the normalized cross-correlation function between the current signal and the local feature template is calculated, the peak value is taken as the feature intensity index, and the corresponding time delay is recorded.

[0019] Furthermore, the threshold generation method in step S4 is as follows:

[0020] Given a false alarm probability, a decision threshold is generated by combining background measurement data when there is no signal injection.

[0021] Furthermore, the connectivity decision condition in step S4 is:

[0022] If the characteristic strength index of a branch exceeds the threshold and its delay deviates from the estimated delay of the main path within the allowable range, then the branch is determined to be connected to the main incoming line; otherwise, it is determined to be disconnected.

[0023] Furthermore, the self-verification in step S5 includes:

[0024] The identification process is repeated in a test loop with a known topology to construct a confusion matrix, calculate recall and precision, and optimize signal amplitude, feature sequence length, and threshold factor.

[0025] A second aspect of the present invention is to provide an online identification system for user internal wiring based on reactive power, for implementing the above-described method, comprising:

[0026] The main unit, located at the main incoming line end of the power distribution system, includes:

[0027] A programmable reactive power generation module is used to generate and inject reactive characteristic signals;

[0028] The signal generation and synchronization control module is used to generate characteristic signal encoding and modulation instructions, and broadcast synchronization frames;

[0029] Slave units are installed at each outgoing switch, and each slave unit includes:

[0030] Current sampling and signal conditioning module;

[0031] The slave DSP is used for synchronous acquisition, feature extraction, and cross-correlation calculation;

[0032] The data fusion and decision module, located on the host side, is used to aggregate the feature strength data of each slave device and generate topological relationships.

[0033] Furthermore, the programmable reactive power generation module is a PWM inverter based on a fully controlled power electronic converter, and is equipped with an LCL filter for outputting reactive current.

[0034] Furthermore, the slave unit is installed at the outgoing switch using a clamp-like structure.

[0035] The technical architecture of this invention revolves around the core idea of ​​active signal injection and distributed synchronous detection.

[0036] First, at the signal excitation and host end, the present invention generates and injects identifiable feature signals in the following manner:

[0037] 1. The programmable reactive power generation module serves as the core excitation source. This module is built on a fully controlled power electronic converter (such as an IGBT bridge). Its key feature is its ability to receive precise commands from the control terminal and dynamically generate reactive current with programmable control over amplitude, frequency, and waveform, thus serving as a highly recognizable characteristic signal in large and complex power grids.

[0038] 2. The signal generation and synchronization control module constitutes the core of the system. This module uses a high-performance digital signal processor (DSP) as the control center. On the one hand, it runs a characteristic signal encoding algorithm to generate a unique instruction sequence; on the other hand, it drives the power electronic converter through an on-chip pulse width modulation (PWM) unit to convert digital instructions into physical reactive power that can be injected into the power grid in real time and accurately.

[0039] The key point of this invention is that the injected reactive power characteristic signal has minimal impact on the active power operation of the system, achieving true interference-free operation; at the same time, through the precise control of the DSP, the signal has strong uniqueness and high signal-to-noise ratio characteristics, laying the foundation for reliable detection in complex power distribution environments.

[0040] Secondly, at the signal detection and identification end (distributed slave), this invention achieves accurate positioning through collaborative processing:

[0041] 1. The core decision-making unit of the system is the signal detection and processing module distributed in each outgoing circuit. Each module includes a high-precision current sensor, a signal conditioning circuit, and a slave DSP.

[0042] 2. Upon receiving the master synchronization command, each slave DSP starts its high-speed ADC to synchronously acquire the current waveform data of its branch.

[0043] 3. The slave DSP calls the built-in real-time signal processing algorithm (such as digital phase-locked amplification or cross-correlation detection) to compare the acquired current signal with the locally reproduced characteristic signal reference copy. This algorithm can effectively suppress background noise and interference, and only branches that actually flow through the characteristic reactive current will produce significant cross-correlation peaks or characteristic intensity outputs.

[0044] The data fusion and relationship determination module (located on the host) aggregates the feature intensity data uploaded by each slave device and makes intelligent judgments based on preset thresholds. Branches with feature intensity exceeding the threshold are identified as target load circuits connected to the main incoming line. The system ultimately generates and outputs a clear switch number-load correspondence diagram to complete the wiring identification.

[0045] The advantages and positive effects of this invention are:

[0046] 1. The core advantage of this invention lies in its active signal injection-distributed synchronous detection technology architecture, which fundamentally solves the long-standing contradiction in traditional invasive and non-invasive identification methods: low accuracy with minimal disturbance and the need for power outages to achieve high accuracy. This system uses characteristic reactive power signals as the detection carrier, achieving non-invasive online identification of switch-load connections in large-scale power consumption scenarios with minimal impact on normal load operation. By integrating the precise controllability of power electronic converters with the high-speed acquisition and processing capabilities of digital signal processors, this device can quickly and accurately identify user internal wiring relationships based on unique signal characteristics, thus maintaining high reliability and high identification accuracy even in complex large-scale power consumption scenarios.

[0047] 2. This invention innovatively combines the controllability of power electronic converters with the high-precision processing capabilities of digital signal processors. It generates highly distinctive reactive power characteristic signals through programmability and effectively suppresses the effects of background noise and line crosstalk by utilizing high-performance synchronization mechanisms and advanced signal processing algorithms. This design, combining strong characteristics with strong anti-interference capabilities, enables the system to achieve microsecond-level synchronization and high signal-to-noise ratio detection even in long-distance, multi-node low-voltage power distribution environments, significantly improving the recognition success rate under complex power grid conditions.

[0048] 3. This invention's system achieves rapid and accurate identification of user internal wiring relationships through a collaborative mechanism between the host and distributed slave devices. This method completely overcomes the inherent defects of traditional load identification methods, such as low identification accuracy, slow response speed, or the need for power outages. It ensures that the confirmation process of switch-load correspondence is both fast and accurate, without affecting the normal operation of large equipment systems. This capability significantly improves the response efficiency of load identification in large and complex scenarios.

[0049] 4. The system of this invention has a simple structure and is easy to operate. The detection unit can adopt a non-intrusive clamp design, which can be quickly installed without power interruption; the identification results can be displayed locally and support remote uploading, which makes it easy for maintenance personnel or emergency command centers to grasp the power distribution topology in real time, providing key technical support for achieving rapid and accurate emergency load management, and effectively ensuring the power supply reliability of important locations. Attached Figure Description

[0050] Figure 1 The diagram shows the structure of the user internal wiring online identification system based on reactive power according to an embodiment of the present invention; wherein, 100 is the master unit, 200-1 is the first slave unit, 200-2 is the second slave unit, 200-3 is the third slave unit, 300 is the data fusion and decision module, and 400 is the high-speed communication bus.

[0051] Figure 2 A flowchart illustrating the online identification method for user internal wiring based on reactive power provided in an embodiment of the present invention;

[0052] Figure 3 This is a schematic diagram of the tree topology of the user power distribution system used in the embodiments of the present invention. Detailed Implementation

[0053] The present application will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention. Only the parts relevant to the present invention are shown in the drawings, and the embodiments and features described herein can be combined with each other unless otherwise specified.

[0054] Example 1

[0055] This embodiment provides an online identification system for user internal wiring based on reactive power, such as... Figure 1 As shown, it includes:

[0056] The main unit is located at the main incoming line end of the power distribution system. The main unit further includes:

[0057] Programmable reactive power generation module: This module is based on a fully controlled power electronic converter, preferably a PWM inverter composed of IGBTs, and equipped with an LCL filter. This module is used to receive control commands and dynamically generate reactive current with programmable amplitude, frequency, and waveform control, as a characteristic signal injected into the power grid.

[0058] Signal Generation and Synchronization Control Module: This module uses a high-performance digital signal processor (DSP) as its control center. On one hand, the DSP runs a characteristic signal encoding algorithm to generate a unique instruction sequence; on the other hand, it drives the programmable reactive power generation module through an on-chip pulse width modulation (PWM) unit to convert digital instructions into reactive power that can be injected into the power grid in real time. Furthermore, this module is also responsible for broadcasting synchronization frames containing high-precision timestamps to all slave devices via a high-speed communication bus.

[0059] Multiple slave units are installed non-intrusively (preferably clamp-on structure) at each of the outgoing line switches to be identified. Each slave unit includes:

[0060] Current sampling and signal conditioning module: Includes a high-precision current sensor and signal conditioning circuit, used to collect the current signal of this branch and perform preprocessing.

[0061] The slave DSP, upon receiving the synchronization command from the master, activates the high-speed ADC to synchronously acquire current waveform data under a unified time base. The slave DSP incorporates real-time signal processing algorithms (such as digital lock-in amplification or cross-correlation detection algorithms) for feature extraction of the acquired signals.

[0062] The data fusion and decision module, located on the host side, is used to aggregate the feature intensity data uploaded by each slave device, make intelligent decisions based on preset thresholds, and finally generate and output a topology map of "switch number-load correspondence".

[0063] Example 2

[0064] This embodiment provides a method for online identification of user internal wiring based on reactive power, applied to the system described in Embodiment 1. The method is as follows: Figure 2 As shown, it includes the following steps:

[0065] Step S1: Connect the master unit (signal injection unit) to the main incoming line of the power distribution system of the user under test, and install multiple slave units (signal detection units) in a non-intrusive clamp-on structure at each outgoing switch to be identified. The master unit broadcasts a synchronization frame containing a high-precision timestamp to all slave units via a high-speed communication bus. All slave units receive this frame and calibrate their local clocks based on a precision clock synchronization protocol to ensure that the entire system enters a unified time base. To achieve microsecond-level synchronization, the following time difference measurement and compensation model is adopted:

[0066] Assume the host is in absolute time Send synchronization frame, the first k Each slave machine at local time The frame has been received. Ignoring transmission jitter, the channel transmission delay... It can be modeled as:

[0067] (1)

[0068] The slave device calibrates its local clock skew accordingly and applies compensation in subsequent samples to ensure the first... n The global synchronization timestamp for each sampling point is:

[0069] (2)

[0070] in, The absolute start time after synchronization. The sampling period is For the first k The remaining clock offset after compensation for each slave device (target value) |< ).

[0071] Step S2: This step is completed collaboratively by the host's signal generation and synchronization control module and the programmable reactive power generation module.

[0072] (1) Feature encoding and baseband signal generation: The host's built-in DSP runs a feature encoding algorithm to generate a discrete baseband feature sequence. The sequence employs a pseudo-random binary sequence (PRBS) with sharp autocorrelation properties, whose autocorrelation function... satisfy:

[0073] (3)

[0074] in, The sequence period length, For discrete-time indexes of the sequence, This is a time shift. This characteristic is beneficial for efficiently extracting signals from noise at the receiver through correlation operations.

[0075] (2) Carrier modulation and reactive current command synthesis: the baseband characteristic sequence The reactive current reference command is generated by applying it to a carrier wave orthogonal to the grid voltage phase using linear modulation. :

[0076] (4)

[0077] in, The amplitude of the injected signal (usually set to the rated load current) 1%-5% of To set the coefficients, Power grid frequency, phase offset Ensure that pure reactive current is generated. for The continuous form, For pulse shaping filters used for spectrum shaping (such as raised cosine filters). This represents the convolution operation. t It is a continuous-time variable.

[0078] (3) Closed-loop control and power injection: The DSP drives the PWM inverter through digital control algorithms (such as proportional resonant controllers) to make the output current Precise tracking Discrete-domain transfer function of the controller It can be designed as:

[0079] (5)

[0080] in, For proportional and resonant gain, To control the cycle, z These are complex variables in the discrete domain. The inverter, using an LCL filter composed of inductor L and capacitor C, converts the modulated PWM wave into a smooth waveform. Injected into the power grid.

[0081] Step S3: Distributed synchronization signal detection, feature extraction and enhancement processing, each slave unit executes this step independently and in parallel.

[0082] (1) Anti-aliasing synchronous sampling: Under the unified time base drive, each slave device samples the branch current at the Nyquist frequency (≥2 times the highest frequency component of the signal). Perform synchronous sampling to obtain a discrete sequence. :

[0083] (6)

[0084] in, This is additive background noise (including load harmonics, switching noise, etc.). i k For the first k The current in the branch circuit, This represents the discrete-time sampling point number.

[0085] (2) Adaptive noise reduction and preprocessing: for Preprocessing is performed, including removing the power frequency fundamental component and using adaptive noise reduction based on wavelet thresholding or Wiener filtering to improve the signal-to-noise ratio (SNR).

[0086] (3) Cross-correlation feature extraction and intensity calculation: The core is to calculate the preprocessed signal. With local feature templates (and Normalized cross-correlation function (corresponding to) :

[0087] (7)

[0088] in, s For reference signal, This is the time offset. For the normalized cross-correlation function, search The peak value is used to obtain the characteristic strength index of the branch. and corresponding delay :

[0089] (8)

[0090] Step S4: Data fusion, intelligent decision-making, and topology reconstruction; all slave devices will use feature strength indicators. and latency Uploaded to the host for centralized fusion and judgment.

[0091] (1) Adaptive threshold generation: The host generates thresholds based on the feature strength indices reported by all slaves. Data, combined with background measurements under no-signal-injection conditions, dynamically generates decision thresholds. One approach is based on the Neyman-Pearson criterion, given a false alarm probability... Determine the threshold:

[0092] (9)

[0093] in, It is the decision threshold. It is the inverse cumulative distribution function. It is the cumulative probability corresponding to the quantile. It is a characteristic intensity index when no signal is injected. The mean, It is a characteristic intensity index when no signal is injected. The variance describes the degree of fluctuation in background noise.

[0094] (2) Logical judgment: The host executes the traversal judgment algorithm.

[0095] H1 Hypothesis (Branch Connectivity): If and ( To estimate the main path delay, If the deviation is within acceptable limits, then the branch is determined to be connected to the main incoming line.

[0096] H0 assumption (branch not connected): If If the connection is not established, it is considered disconnected.

[0097] (3) Result Generation and Uncertainty Assessment: The host synthesizes all decision results and generates a deterministic list of outgoing switch-load correspondences. Simultaneously, the confidence level of each decision can be calculated. ,For example:

[0098] (10)

[0099] in, This is the sensitivity coefficient. Confidence information is output along with the topology results to provide a quantitative basis for subsequent decision-making.

[0100] Step S5: System Performance Self-Verification and Parameter Optimization. To improve the long-term reliability and adaptability of the system, a self-verification and optimization process is implemented. Steps S2 to S4 are repeated in a test loop with a known real topology. By comparing the identification results with the real topology, a confusion matrix is ​​constructed, and key performance indicators are calculated.

[0101] (11)

[0102] (12)

[0103] in, TP For true examples, it represents the number of samples that are actually positive and correctly identified as positive. TN True negative examples represent the number of samples that are actually negative and have been correctly identified as negative. FP False positives represent the number of samples that are actually negative but are incorrectly identified as positive. FN False negatives represent the number of samples that are actually positive but are incorrectly identified as negative. Acc Accuracy is the proportion of all correctly classified samples out of the total sample. Recall Recall rate is the proportion of samples that are actually positive that are correctly identified.

[0104] Based on these metrics, feedback is used to optimize key parameters, such as signal amplitude. Length of feature sequence Decision threshold factor This allows for online adaptive improvement of system performance. The optimization objective can be expressed as: under constraints... ≤ Maximize recall rate Recall ,in For false alarm rate, To maximize the false alarm rate.

[0105] Figure 1The structural block diagram of the user internal wiring online identification system based on reactive power provided for the implementation of the present invention is shown below. The host unit 100 is connected to the first slave unit 200-1, the second slave unit 200-2, the third slave unit 200-3 and the data fusion and decision module 300 respectively through the high-speed communication bus 400.

[0106] Figure 2 The flowchart of the online identification method for user internal wiring based on reactive power provided for the implementation of this invention mainly includes five sub-processes: system initialization and synchronization, feature signal generation and input, parallel execution of each slave, data fusion and topology reconstruction, and system performance self-verification and optimization. The system initialization and synchronization sub-process initializes the system and completes the time base setting; the feature signal generation and input sub-process's main function is: the entire system's incoming terminals are connected to the host, and under a second-level time base, the host issues a reactive current command, which is then controlled by the inverter to generate reactive power; the parallel execution of each slave sub-process synchronously collects current, performs signal processing and noise reduction, then calculates the correlation function and extracts the feature intensity, and uploads the extracted feature intensity to the host; the data fusion and topology reconstruction sub-process receives slave data, generates a decision threshold by analyzing the feature intensity, and then determines whether the branch is connected or disconnected; the system performance self-verification and optimization sub-process repeats steps S2 to S4, constructs a confusion matrix, calculates key performance indicators, and completes system performance self-verification and optimization.

[0107] Figure 3 This is a schematic diagram of the tree topology of the user power distribution system used in this embodiment of the invention. The topology, from top to bottom, is as follows: the hospital serves as the overall application scenario, under which there are power distribution cabinet 1 and power distribution cabinet 2. Power distribution cabinet 2 further leads out multiple branches such as power distribution cabinet 2-branch 1, power distribution cabinet 2-branch 2, and power distribution cabinet 2-branch 3. Power distribution cabinet 1 further leads out power distribution cabinet 1-branch 1, forming a branch layer. The ends of each branch are connected to specific socket node 1, specific socket node 2, etc., forming a load node layer.

[0108] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several modifications and improvements can be made without departing from the inventive concept, and these all fall within the protection scope of the present invention.

Claims

1. A method for online identification of user internal wiring based on applied reactive power, characterized in that, Includes the following steps: Step S1: Connect the host to the main incoming line of the power distribution system of the user under test, install multiple slaves at each outgoing switch to be identified, and realize microsecond-level clock synchronization between the host and each slave through broadcast synchronization frames; Step S2: The host generates a reactive power characteristic signal and injects it into the main incoming line of the power distribution system; the reactive power characteristic signal is generated based on a pseudo-random binary sequence and is modulated onto a carrier wave orthogonal to the phase of the grid voltage. Step S3: Each slave device synchronously acquires the current signal of its own branch under a unified time base, preprocesses the acquired signal, calculates its cross-correlation function with the local feature template, and obtains the feature intensity index. Step S4: Each slave device uploads its characteristic strength index to the host. The host performs connectivity determination on each branch based on the threshold and generates a switch-load correspondence. Step S5: The system performs self-verification based on the known topology loop and optimizes the injected signal parameters and threshold.

2. The method according to claim 1, characterized in that, The clock synchronization in step S1 includes: The master broadcasts a synchronization frame containing a timestamp. Each slave calibrates its local clock based on a clock synchronization protocol and synchronizes using a time difference measurement and compensation model.

3. The method according to claim 1, characterized in that, The generation of reactive power characteristic signals in step S2 includes: The host DSP runs a feature encoding algorithm to generate a pseudo-random binary sequence, which is then modulated into a reactive current command after pulse shaping filtering. The reactive current is then driven by the controller to drive the PWM inverter to output the reactive current.

4. The method according to claim 1, characterized in that, The preprocessing in step S3 includes: After anti-aliasing synchronous sampling, removal of fundamental component, and noise reduction of the acquired current signal, the normalized cross-correlation function between the current signal and the local feature template is calculated, the peak value is taken as the feature intensity index, and the corresponding time delay is recorded.

5. The method according to claim 1, characterized in that, The threshold generation method in step S4 is as follows: Given a false alarm probability, a decision threshold is generated by combining background measurement data when there is no signal injection.

6. The method according to claim 1, characterized in that, The connectivity determination condition in step S4 is: If the characteristic strength index of a branch exceeds the threshold and its delay deviates from the estimated delay of the main path within the allowable range, then the branch is determined to be connected to the main incoming line; otherwise, it is determined to be disconnected.

7. The method according to claim 1, characterized in that, The self-verification in step S5 includes: The identification process is repeated in a test loop with a known topology to construct a confusion matrix, calculate recall and precision, and optimize signal amplitude, feature sequence length, and threshold factor.

8. A user internal wiring online identification system based on applied reactive power, used to implement the method described in any one of claims 1 to 7, characterized in that, include: The main unit, located at the main incoming line end of the power distribution system, includes: A programmable reactive power generation module is used to generate and inject reactive characteristic signals; The signal generation and synchronization control module is used to generate characteristic signal encoding and modulation instructions, and broadcast synchronization frames; Slave units are installed at each outgoing switch, and each slave unit includes: Current sampling and signal conditioning module; The slave DSP is used for synchronous acquisition, feature extraction, and cross-correlation calculation; The data fusion and decision module, located on the host side, is used to aggregate the feature strength data of each slave device and generate topological relationships.

9. The system according to claim 8, characterized in that, The programmable reactive power generation module is a PWM inverter based on a fully controlled power electronic converter, and is equipped with an LCL filter for outputting reactive current.

10. The system according to claim 8, characterized in that, The slave unit is installed at the outgoing switch using a clamp-like structure.

Images