A power distribution network operation situation awareness and fault diagnosis system and method

Abstract

The application provides a power distribution network operation situation awareness and fault diagnosis system and method. The system according to the application comprises a perception layer module, a network layer module and a platform layer module connected in sequence. The perception layer module is deployed at a key node of the power distribution network, is used for synchronously collecting electrical quantity data of the power distribution network, performing local anomaly detection and preprocessing on the electrical quantity data, and sending the processed data to the network layer module. The data output end of the network layer module is connected to the data input end of the platform layer module. The network layer module is used for receiving the processed data based on a 5G communication network and transmitting the processed data to the platform layer module. The platform layer module is deployed at a master station side, and its data output end is connected to a visual display device. The platform layer module is used for receiving and converging the processed data, performing operation situation awareness, topology anomaly detection and fault diagnosis positioning on the power distribution network based on the processed data, and outputting a diagnosis result to the visual display device.

Description

Technical Field

[0001] This invention relates to the field of power system automation technology, and in particular to a distribution network operation status perception and fault diagnosis system and method. Background Technology

[0002] As the last mile of the power system, the distribution network's operational status directly affects power supply reliability and power quality. With the large-scale integration of new loads such as distributed photovoltaic power, energy storage, and electric vehicle charging stations, the distribution network exhibits new characteristics such as variable operating modes, significant three-phase imbalance, and rapid fault transients, placing higher demands on situational awareness and fault diagnosis. Currently, distribution network monitoring mainly faces the following problems: 1. Insufficient sensing capability: Traditional distribution network monitoring relies on SCADA systems and fault indicators, which have low sampling rates (second-level) and sparse data, making it difficult to capture millisecond-level fault transient processes. Fault indicators can only report overcurrent information and cannot provide key electrical quantities such as voltage phasors and phase angles, resulting in low fault location accuracy.

[0003] 2. Lagging Topology Identification: Current topology analysis mainly relies on switch status reporting, which has a long update cycle (minutes) and suffers from missed and false alarms regarding switch changes. In scenarios such as frequent switching of distributed power sources and network reconfiguration, the actual topology and system model are severely mismatched, leading to the failure of state estimation and fault diagnosis.

[0004] 3. Time synchronization relies on GPS: Existing PMUs generally use GPS for time synchronization, which poses potential security risks. With the full deployment of the BeiDou system, the power system urgently needs to achieve independent and controllable time synchronization.

[0005] 4. Communication bottlenecks are prominent: The number of distribution network nodes is huge (up to tens of thousands), traditional fiber optic communication has high deployment costs and poor flexibility, and 4G network bandwidth is limited and latency is unstable (50-100ms), making it difficult to meet the real-time data transmission needs of PMU.

[0006] 5. Poor adaptability of diagnostic models: Traditional fault diagnosis methods are based on set value logic or simple threshold judgment, which are difficult to adapt to the complex fault characteristics after the connection of distributed power sources (such as the decrease in fault current amplitude, uncertainty in direction, waveform distortion, etc.), resulting in high false alarm rate and false alarm rate. Summary of the Invention

[0007] The purpose of this invention is to provide a power distribution network operation status perception and fault diagnosis system and method, which aims to solve the above-mentioned problems in the prior art.

[0008] This invention provides a power distribution network operation status awareness and fault diagnosis system, comprising a sensing layer module, a network layer module, and a platform layer module connected in sequence. The sensing layer module is deployed at key nodes of the distribution network. Its electrical quantity acquisition terminal is connected to the primary equipment of the distribution network, and its data output terminal is connected to the data input terminal of the network layer module. The sensing layer module is used to synchronously acquire electrical quantity data of the distribution network, perform local anomaly detection and preprocessing on the electrical quantity data, and send the processed data to the network layer module. The data output terminal of the network layer module is connected to the data input terminal of the platform layer module; the network layer module is used to receive the processed data based on the 5G communication network and transmit the processed data to the platform layer module. The platform layer module is deployed on the main station side, and its data output end is connected to the visualization display device. The platform layer module is used to receive and aggregate the processed data, perform operational status awareness, topology anomaly detection and fault diagnosis and location on the distribution network based on the processed data, and output the diagnosis results to the visualization display device.

[0009] This invention provides a method for power distribution network operation status perception and fault diagnosis, including: The electrical quantity data of the distribution network is collected synchronously by the sensing layer module deployed at key nodes of the distribution network, and the electrical quantity data is subjected to local anomaly detection and preprocessing. The processed data is then sent to the network layer module. The network layer module receives the processed data based on the 5G communication network and transmits the processed data to the platform layer module. The platform layer module deployed on the main station side receives and aggregates the processed data, performs operational status awareness, topology anomaly detection, and fault diagnosis and location on the distribution network based on the processed data, and outputs the diagnostic results to the visualization display device.

[0010] This invention also provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it implements the steps of the above-described distribution network operation status perception and fault diagnosis method.

[0011] This invention also provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by a processor, implements the steps of the above-described power distribution network operation status perception and fault diagnosis method.

[0012] The embodiments of the present invention can include the following beneficial effects: The embodiments of the present invention propose a distribution network operation status and topology anomaly correlation detection and location device based on sensing and communication technologies, through... The deep integration of PMU, 5G communication, BeiDou timing and TodyNet model constructs an integrated "sensing-communication-analysis" power distribution network monitoring system to achieve high-precision real-time sensing, low-latency reliable transmission, autonomous and controllable timing and intelligent diagnosis and positioning. Attached Figure Description

[0013] To more clearly illustrate the technical solutions in one or more embodiments of this specification or in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 This is a schematic diagram of the power distribution network operation status perception and fault diagnosis system according to an embodiment of the present invention; Figure 2 This is an embodiment of the present invention. PMU system installation diagram; Figure 3 This is a schematic diagram of the real-time fault acquisition unit according to an embodiment of the present invention; Figure 4 This is a flowchart of the power distribution network operation status perception and fault diagnosis method according to an embodiment of the present invention. Detailed Implementation

[0015] To enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this specification, and not all of the embodiments. Based on one or more embodiments of this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of this document.

[0016] System Implementation Examples According to embodiments of the present invention, a power distribution network operation status awareness and fault diagnosis system is provided. Figure 1 This is a schematic diagram of the power distribution network operation status perception and fault diagnosis system according to an embodiment of the present invention, as shown below. Figure 1 As shown, the power distribution network operation status perception and fault diagnosis system according to an embodiment of the present invention includes a perception layer module 10, a network layer module 12 and a platform layer module 14 connected in sequence. The sensing layer module 10 is deployed at key nodes of the distribution network. Its electrical quantity acquisition terminal is connected to the primary equipment of the distribution network, and its data output terminal is connected to the data input terminal of the network layer module 12. The sensing layer module 10 is used to synchronously acquire electrical quantity data of the distribution network, perform local anomaly detection and preprocessing on the electrical quantity data, and send the processed data to the network layer module 12. The sensing layer module 10 includes a miniature synchronous phasor measurement device, wherein the miniature synchronous phasor measurement device includes an FPGA submodule, a DSP submodule, a BeiDou timing submodule, a 5G communication module, and a power supply submodule. The first input terminal of the FPGA submodule is connected to the second pulse output terminal of the Beidou time synchronization submodule, the second input terminal of the FPGA submodule is connected to the electrical quantity output terminal of the primary equipment of the power distribution network, and the output terminal of the FPGA submodule is connected to the input terminal of the DSP submodule. The FPGA submodule is used to synchronously collect electrical quantity data of the power distribution network according to the second pulse signal provided by the Beidou time synchronization submodule, and to perform over-limit detection, buffering and preprocessing on the electrical quantity data, and transmit the preprocessed data to the DSP submodule. The FPGA submodule includes a multi-channel synchronous sampling unit, an over-limit detection unit, and a data buffer unit; The input terminal of the multi-channel synchronous sampling unit is connected to the electrical quantity output terminal of the primary equipment of the power distribution network, and the control terminal of the multi-channel synchronous sampling unit is connected to the second pulse output terminal of the Beidou time synchronization submodule. The multi-channel synchronous sampling unit is used to perform synchronous sampling according to the second pulse signal at a configurable sampling rate and output the sampled data to the data buffer unit. The input terminal of the over-limit detection unit is connected to the output terminal of the multi-channel synchronous sampling unit. The over-limit detection unit is used to monitor the instantaneous current value in real time, and generates an interrupt signal when the detected value exceeds a preset threshold. The input terminal of the data buffer unit is connected to the output terminal of the multi-channel synchronous sampling unit and the output terminal of the over-limit detection unit, respectively. The data buffer unit is used to buffer the sampled data and interrupt signal, and output the buffered data to the DSP submodule. The output of the DSP submodule is connected to the input of the 5G communication module. The DSP submodule is used to receive the preprocessed data, calculate phasor data based on the preprocessed data, extract time-domain features and frequency-domain features based on the phasor data, classify the severity of anomalies using an anomaly detection algorithm based on the extracted time-domain features and frequency-domain features, compress the classified data, and transmit the compressed data to the 5G communication module. The DSP submodule includes a phasor calculation unit, a feature extraction unit, an anomaly classification unit, and a data compression unit; The input terminal of the phasor calculation unit is connected to the output terminal of the FPGA submodule. The phasor calculation unit is used to calculate phasor data according to a preset period and generate data frames that conform to the IEEE standard. The input terminal of the feature extraction unit is connected to the output terminal of the phasor calculation unit. The feature extraction unit is used to extract time-domain features and frequency-domain features based on the phasor data. The time-domain features include at least one of the effective value of current, peak factor and waveform distortion rate, and the frequency-domain features include harmonic content. The input of the anomaly classification unit is connected to the output of the feature extraction unit. The anomaly classification unit is used to obtain anomaly feature parameters based on the time-domain features and frequency-domain features, and to classify the severity of anomalies into multiple preset levels according to the anomaly feature parameters. The anomaly classification unit includes: The first-level determination subunit outputs a local recording instruction when the current exceedance is less than the first preset threshold and the duration is less than the first preset duration, storing the time domain features and frequency domain features in the local memory and marking the timestamp. The second-level judgment subunit outputs a waveform recording start command when the current exceeds the limit to the second preset threshold range or the duration reaches the second preset duration range, triggering the FPGA submodule to record the original sampled data and transmit the recorded data to the platform layer module after compression. The third-level judgment subunit outputs an emergency reporting command when the current exceeds the third preset threshold, or the zero-sequence current exceeds the limit, or a voltage drop is detected. The abnormal classification results and related data are reported to the platform layer module in real time, triggering the fault diagnosis process. Wherein, the first preset threshold is less than the lower limit of the range of the second preset threshold, and the upper limit of the range of the second preset threshold is less than the third preset threshold; The input end of the data compression unit is connected to the output end of the anomaly classification unit. The data compression unit is used to compress the data output by the anomaly classification unit and then output it to the 5G communication module. The output terminal of the 5G communication module serves as the data output terminal of the perception layer module and is connected to the data input terminal of the network layer module; the 5G communication module is used to encapsulate the compressed data into data packets and send them to the network layer module through the 5G air interface. The BeiDou timing submodule is used to receive BeiDou satellite signals, generate second pulse signals, and provide them to the FPGA submodule. The power output terminal of the power supply submodule is connected to the power input terminals of the FPGA submodule, the DSP submodule, the Beidou timing submodule, and the 5G communication module, respectively, to provide working power to each submodule.

[0017] The data output terminal of the network layer module 12 is connected to the data input terminal of the platform layer module 14; the network layer module 12 is used to receive the processed data based on the 5G communication network and transmit the processed data to the platform layer module 14. The platform layer module 14 is deployed on the main station side, and its data output end is connected to the visualization display device. The platform layer module 14 is used to receive and aggregate the processed data, perform operational status awareness, topology anomaly detection and fault diagnosis and location on the distribution network based on the processed data, and output the diagnosis results to the visualization display device. The platform layer module 14 includes a phasor data concentrator and a TodyNet analysis platform; The input of the phasor data concentrator is connected to the data output of the network layer module, and the output of the phasor data concentrator is connected to the input of the TodyNet analysis platform. The phasor data concentrator is used to receive multi-source processed data transmitted by the network layer module, sort and align the multi-source processed data according to timestamps, interpolate and complete missing data, and transmit the aligned data to the TodyNet analysis platform. The output of the TodyNet analysis platform is connected to a visualization display device; the TodyNet analysis platform is used to receive aligned data, and based on the aligned data, to perform operational status awareness, topology anomaly detection, and fault diagnosis and location of the power distribution network, and output the diagnostic results to the visualization display device. The TodyNet analysis platform includes a graph construction unit, a graph neural network model unit, and a topology anomaly detection unit. The input of the graph construction unit is connected to the output of the phasor data concentrator, and the output of the graph construction unit is connected to the input of the graph neural network model unit. The graph construction unit is used to construct a distribution network topology graph with each PMU node in the distribution network as a graph node and the feeder connection relationship as an edge, and output the distribution network topology graph to the graph neural network model unit. The output of the graph neural network model unit is connected to the visualization display device; the graph neural network model unit is used to receive the power distribution network topology map, perform fault type classification and fault distance regression on the power distribution network based on the power distribution network topology map, and output the fault diagnosis results to the visualization display device. The input of the topology anomaly detection unit is connected to the output of the phasor data concentrator, and the output of the topology anomaly detection unit is connected to the visualization display device. The topology anomaly detection unit is used to receive the aligned data, detect topology change events based on current waveform similarity and voltage phase angle change rate, and output the topology anomaly detection results to the visualization display device.

[0018] The following describes in detail the above-mentioned technical solutions of the present invention with reference to the specific circumstances of the power distribution network operation status perception and fault diagnosis system in the embodiments of the present invention.

[0019] 1. System Overall Architecture The embodiments of this invention adopt a three-layer architecture design: Sensing layer: Improved miniature synchronous phasor measurement devices deployed at key nodes of the distribution network ( PMU), such as Figure 2 As shown, each The PMU integrates a high-precision sampling module, a BeiDou timing module, a 5G communication module, and an edge computing module to achieve real-time current and voltage acquisition, anomaly detection, and data uploading.

[0020] Network layer: Based on 5G communication network, network slicing technology is used to ensure high-priority, low-latency transmission of PMU data, with an end-to-end average latency of ≤15ms.

[0021] Platform layer: The phasor data concentrator (PDC) and TodyNet analysis platform are deployed on the main station side to realize multi-source data aggregation, fault diagnosis, topology correction and visualization.

[0022] 2. Perception Layer: Improved PMU device (1) Hardware architecture The PMU adopts an FPGA+DSP heterogeneous computing architecture, such as Figure 3 As shown, it specifically includes: A. FPGA module (Xilinx Zynq UltraScale+ series): Performs 8-channel synchronous sampling (three-phase voltage, three-phase current, zero-sequence voltage, zero-sequence current), with a configurable sampling rate of 12.8kHz / 25.6kHz / 51.2kHz; receives 1PPS pulses per second from the BeiDou time synchronization module to achieve high-precision time synchronization, with a timing error ≤1. s; Implement hardware-level over-limit detection with a response time ≤1ms; Complete data caching and preprocessing.

[0023] B. DSP Module (TI TMS320C6678 selected): Based on the sampling data provided by the FPGA, it calculates the phasor once per cycle (20ms), meeting the IEEE C37.118 standard; it implements a multi-scale feature extraction algorithm, including time-domain features (RMS, peak value, waveform distortion rate) and frequency-domain features (harmonic content, interharmonics); it runs a real-time anomaly detection algorithm to classify the severity of anomalies; and it executes a data compression algorithm to reduce the amount of data transmitted.

[0024] C. Beidou Timing Module (Using Hexin Xingtong UM982): Supports Beidou B1 / B3 dual-frequency timing, with a first positioning time of ≤1s during hot start; outputs high-precision 1PPS pulse per second, with timing accuracy better than 20ns (1 Built-in high-stability crystal oscillator (OCXO, stability ≤5×) In the event of BeiDou signal loss, the timekeeping capability is ≥24 hours, and the timekeeping error is ≤1. Supports dual-mode reception of BeiDou / GPS, and automatically switches to backup mode when the primary BeiDou signal fails.

[0025] D. 5G Communication Module (using Fibocom FM150 series): Supports 5G SA / NSA networking modes, downlink speed ≥2Gbps, uplink speed ≥1Gbps; integrates a protocol conversion module to achieve compatibility between IEEE C37.118 data format and 5G network protocol stack; supports network slicing function, and can configure dedicated slices to ensure PMU data transmission priority; has a built-in breakpoint resume mechanism to ensure no data loss in the event of temporary network interruption.

[0026] E. Power Module: Supports dual AC / DC power supply (AC 85-265V / DC 24-48V); built-in backup lithium battery, which can work continuously for ≥2 hours in the event of main power failure; equipped with power status monitoring and reporting functions.

[0027] (2) Anomaly detection This invention proposes a hierarchical real-time current anomaly detection algorithm framework: A. Bottom layer (physical layer) Real-time monitoring of instantaneous values ​​of three-phase current and zero-sequence current, and setting over-limit detection; A dual comparator circuit is used, with a response time of ≤1ms; The detection threshold can be adaptively adjusted based on historical data, thereby avoiding false alarms caused by a fixed threshold.

[0028] B. Middle layer (feature layer) The time-domain characteristics of the current per cycle are calculated: effective value (RMS), peak value, and waveform distortion rate. Calculate the frequency domain features every 10 cycles: the content of each harmonic (2 - 25 times), the content of interharmonics; Based on a sliding window (window length: 5 cycles, sliding step: 1 cycle), calculate the feature change rate to identify abnormal trends.

[0029] C. High layer (logical layer) Based on the features of time - series data, calculate the correlation and similarity of the currents of adjacent nodes to identify abnormal propagation patterns; Introduce the double - current mutation ratio criterion: calculate the ratio of the current mutations upstream and downstream of the fault point: ; When R > K1 and R < K2, it is judged as an in - zone fault; otherwise, it is judged as an out - of - zone disturbance (K1 and K2 are setting coefficients, and K1 = 0.8, K2 = 1.2 can be set); This criterion has good anti - interference ability against non - fault disturbances such as the output fluctuations of photovoltaic power sources. Simulation verification shows that the false - alarm rate is 0 when the photovoltaic output fluctuates by 50%.

[0030] (3)Classification of abnormal severity According to the abnormal characteristic parameters, the abnormal severity is divided into three levels, as shown in Table 1: Table 1 Classification of abnormal levels

[0031] 3. Network layer: 5G communication optimization (1)End - to - end communication architecture The PMU data transmission system based on the 5G network includes: PMU terminal: Integrate a 5G communication module, encapsulate PMU data into UDP / IP packets, and send them through the 5G air interface; 5G base station: Responsible for air - interface resource scheduling and allocate dedicated resources for PMU terminals; 5G core network: Realize the separation of the user plane and the control plane, and provide network slicing management functions; Phasor data concentrator (PDC): Deployed on the master - station side, receive PMU data from each terminal, and perform time alignment and data recombination.

[0032] (2)Network slicing design To ensure the real - time and reliable transmission of PMU data, dedicated network slices can be designed. For example: Create network slice instances in the 5G core network management system, allocate SST (Slice / Service Type)=1 (eMBB type for general services), SST = 2 (URLLC type for PMU services), and SD (Slice Differentiator)=000001; Configure the sliced ​​resource pool: allocate dedicated frequency bands on the radio side (part of the RBs in the n78 band, such as 5MHz of dedicated resources in the 20MHz bandwidth); allocate FlexE channels on the transmission side, with a guaranteed bandwidth of 100Mbps and the highest priority; deploy a dedicated UPF (User Plane Function) on the core network side, and establish a direct N3 interface connection with the gNB to which the PMU terminal belongs; The 5G module of the PMU terminal activates the PDN connection through the AT command "AT+CGDCONT=1, "IPV4V6", "apn", "0.0.0.0", 0, 0, 0, 2", where the last parameter 2 specifies the slice ID to use. Configure QoS flow: Use the AT command "AT+QOS=1,83" to set the QoS flow identifier to 83 (corresponding to URLLC service) to ensure end-to-end latency ≤15ms.

[0033] 4. Platform Layer: TodyNet Analysis Platform 4.1 Phasor Data Concentrator (PDC) The PDC software openPDC is configured as follows: Input Adapter: Configure the UDP input adapter to listen on port 50000, parse IEEE C37.118 frames, and extract timestamps, phasor data, and status words; Time alignment module: It adopts a sliding window mechanism with a window width of 2ms. The received data frames are sorted by timestamp, and missing data frames are filled in using an interpolation algorithm (cubic spline interpolation). Output adapter: Outputs the aligned data to the TodyNet analytics platform via an Apache Kafka message queue. The Kafka topic is configured as pmu_data, with 10 partitions and 2 replicas. Data storage: The raw data is stored in InfluxDB with a retention policy of 30 days. The data sampling frequency is 50 frames / second, each frame contains 8 channels of phasor data, and the daily data volume is approximately 50GB.

[0034] 4.2 TodyNet Model The TodyNet (Topology Dynamics Network) model is based on graph neural networks (GNNs) and spatiotemporal correlation analysis. The specific implementation steps are as follows: A. Graph Construction Module Using each PMU node in the distribution network as a graph node and the feeder connection relationship as an edge, construct the distribution network topology graph G=(V,E); The node feature vector includes: three-phase voltage amplitude, phase angle, current amplitude, phase angle, frequency, harmonic content, etc., totaling 24 dimensions; The edge feature vector includes: line length, impedance parameters, number of historical faults, etc., totaling 5 dimensions; Dynamic graph structure updates: When a topology change is detected (such as a switch shift), the graph structure is reconstructed.

[0035] B. Graph Neural Network Model The GraphSAGE model is used, with the following parameters: Input layer: 24-dimensional node features and 5-dimensional edge features, concatenated before input; Hidden layers: 2 GraphSAGE convolutional layers, 128 neurons per layer, ReLU activation function; Aggregation function: Mean aggregation is used, and the number of neighbor samples is 10; Output layer: The classification head outputs the fault type (single-phase grounding, two-phase short circuit, three-phase short circuit, no fault), and the regression head outputs the fault distance (normalized value 0-1). Loss function: The classification loss uses cross-entropy loss, the regression loss uses mean squared error loss, and the total loss is the weighted sum of the two (weight ratio 1:0.5). Optimizer: Adam, learning rate 0.001, batch size 64, training epochs 200.

[0036] C. Topology Anomaly Detection Module Detecting topological changes based on current waveform similarity and phase change rate: The Pearson correlation coefficient r between the current waveforms of adjacent nodes is calculated in real time, with a calculation window of 10 cycles (200ms). When r suddenly drops from a normal value (≥0.95) to below 0.8, it is marked as a suspected topological change; Simultaneously detect the phase change rate: calculate the rate of change of the node voltage phase angle dθ / dt, and mark it as a suspected topology change when |dθ / dt|>30° / s; Combining the two criteria, when r < 0.8 and |dθ / dt| > 30° / s, it is determined to be a topological change event, and the timestamp and change type (such as switch open or closed) are recorded. Perform time correlation analysis between topology change events and fault events. If the fault occurs within 100ms after the topology change, it is determined to be a fault caused by the topology change.

[0037] D. Visualization Module Visualization is achieved using web front-end technologies: Front-end framework: Vue.js + ECharts + D3.js; Real-time topology diagram: Dynamically draws the distribution network topology based on the force-directed graph of D3.js. The node color indicates the operating status (normal green, alarm yellow, fault red), and the edge thickness indicates the current magnitude. Fault propagation path graph: Based on the fault propagation probability output by GNN, a heatmap of the fault propagation path is drawn; Abnormal Event Timeline: Based on ECharts, it plots multi-channel PMU data curves, marks abnormal event time points, and supports scaling, dragging, and data export; Data Interface: The backend uses the Flask framework, provides a RESTful API, uses JSON as the data format, and supports real-time WebSocket push.

[0038] 5. Topological Anomaly Correlation Detection 5.1 Data Preprocessing Consume PMU data from Kafka; each frame contains a timestamp t and the voltage phasor of node i. Current phasor ; Five consecutive cycles of data (100ms) were filtered using a Butterworth low-pass filter with a cutoff frequency of 100Hz to remove high-frequency noise. Calculate the sequence of RMS current values ​​for each node. and voltage phase angle sequence .

[0039] 5.2 Current Waveform Similarity Calculation Take the current sampling sequence of adjacent nodes i and j and k = 1~256 (256 points in one cycle); Calculate the Pearson correlation coefficient: ; in, This represents the mean of the current sampling sequence at node i. This represents the mean of the current sampling sequence at node j.

[0040] when When the value suddenly drops from above 0.95 to below 0.8, it is marked as a candidate point for topological change, and the time of change is recorded. and range of change .

[0041] 5.3 Calculation of Phase Change Rate Voltage phase angle sequence at node i Perform differential calculations to obtain the angular velocity. Where Δt = 20 ms; Calculate angular acceleration ; when >30° / s and When the value is greater than 10° / s², it is marked as a candidate point for topological change.

[0042] 5.4 Comprehensive Judgment Candidate points that meet the above two conditions are time-aligned. If the time difference is ≤40ms (2 cycles), they are merged into one topology change event. Record detailed information about topology change events: time Node i, type of change (determining whether it is open or closed based on the direction of current change), confidence level (based on the magnitude of the change in r and...) The value is calculated by combining the magnitudes, and the range is 0-1. Topology change events are written to a database table for use in fault correlation analysis.

[0043] 5.5 Fault Correlation Analysis When a fault event is detected (based on the fault level determination in the three levels of anomaly detection), the fault time is recorded. Faulty node f; Query the topology change event table to find All topology change events that occurred within the first 100ms; If a topology change event exists, and the changed node is adjacent to the faulty node (distance ≤ 2 edges), then the fault is determined to be caused by a topology change, and an associated report is generated. The correlation report includes: topology change time, type, and node; failure time, type, and node; correlation confidence level (weighted average based on topology change confidence level and failure detection confidence level). Related reports are pushed to the front-end visualization interface via WebSocket and simultaneously stored in the database.

[0044] Example 1: Single-phase ground fault detection and location The embodiments of the present invention are deployed at nodes A, B, and C of a certain 10kV distribution network. The PMU device is configured with a sampling rate of 25.6 kHz (512 points per cycle).

[0045] At 14:23:45.123 on a certain day, a single-phase ground fault occurred at node B (phase A grounded). The FPGA recorded the fault at the third sampling point (approximately 117) after the fault occurred. s) Upon detecting that the instantaneous value of the A-phase current exceeds the upper limit threshold, a hardware interrupt is triggered; the FPGA immediately starts waveform recording, recording the raw sampled data for 5 seconds before and after the fault. At the beginning of the next cycle (14:23:45.140), DSP cores 1-2 read the FPGA cache data, calculate the phasor data for the two cycles before and after the fault, and find that the effective value of the A-phase current suddenly increases from the normal value of 100A to 650A, and the effective value of the voltage suddenly drops from 10kV to 5.2kV. DSP cores 5-6 execute the anomaly detection algorithm and calculate the ratio of the two current surges. ,in From node A (650A-100A=550A). The fault originates from node C (120A-100A=20A), with R=550 / 20=27.5>1.2, indicating an in-area fault (fault point located at node B). DSP core 7 compresses the waveform data (using DPCM encoding, compression ratio 5.2:1) and generates a fault report (including fault time, type, amplitude, phase angle, etc.), which is then urgently reported via the 5G module. The 5G module is configured with a dedicated slice (SST=2, SD=000001), encapsulates the fault report into a UDP packet, and sends it to the base station via the 5G air interface. The base station uses URLLC scheduling, with an end-to-end latency of 12ms. The PDC receives the fault report at 14:23:45.155. The TodyNet platform consumes Kafka data, and the GraphSAGE model inputs the fault characteristics of node B, outputting a fault type of single-phase grounding and a normalized fault distance value of 0.48 (corresponding to an actual distance of approximately 480 meters). The topology anomaly detection module checks for topology change events within the previous 100ms, determining it to be a line-specific fault. The visualization interface displays node B turning red in real time, and a fault alarm window pops up, showing the fault location as 480 meters downstream of node B.

[0046] Example 2: Anomaly Classification Handling in Distributed Photovoltaic Fluctuation Scenarios A distribution network node D is connected to a distributed photovoltaic (PV) system. At 10:05:30 one day, due to cloud cover, the PV output fluctuated from 100kW to 50kW within 2 seconds (a fluctuation of 50%). Node D's... The PMU detected a decrease in the effective current value from 100A to 50A, and the rate of change was... RMS = 50% > 10%, but not exceeding the over-limit threshold (upper limit 110A, lower limit 90A). The anomaly detection algorithm identified it as an abnormal trend, but the ratio of the two current surges showed R = 0.98 (the upstream and downstream current changes are close), which does not meet the fault conditions within the zone. The anomaly severity level was classified as Level 1 Concern. The FPGA recorded the abnormal time period data (10:05:30.000~10:05:32.000) in the local eMMC, marked with a timestamp, but did not upload it to the main station. The main station TodyNet platform did not receive an alarm, and the visualization interface showed that node D's status was green, but the waveform of this period could be viewed in the historical data query. This example verifies the anti-interference capability of the present invention against photovoltaic fluctuations and avoids false alarms.

[0047] Example 3: Topology Change Detection and Association Analysis In a power distribution network, nodes E and F are connected via switch K during normal operation. At 16:20:10:50 on a certain day, switch K tripped due to a protection mechanism. The connection between nodes E and F... The PMU detected a sudden drop in current waveform similarity r from 0.98 to 0.65, while the voltage phase angle change rate at node E reached 35° / s and at node F reached 32° / s. The topology anomaly detection module determined this to be a topology change event, recording the change time as 16:20:10.500, the change type as switch open, and a confidence level of 0.92. At 16:20:10.550, a single-phase ground fault occurred downstream of node F. The fault correlation analysis module found a topology change event (16:20:10.500) within 100ms before the fault time 16:20:10.550, and since the changing node F was adjacent to the faulty node, it was determined that the fault was caused by a topology change, and a correlation report was generated. The visualization interface displays the topology change timeline, marking the correlation between switch K open and fault F, helping maintenance personnel quickly locate the cause of the fault.

[0048] Method Implementation Examples According to embodiments of the present invention, a method for power distribution network operation status perception and fault diagnosis is provided. Figure 4 This is a flowchart of the power distribution network operation status perception and fault diagnosis method according to an embodiment of the present invention, as follows: Figure 4 As shown, the power distribution network operation status perception and fault diagnosis method according to an embodiment of the present invention specifically includes: Step S401: The electrical quantity data of the distribution network is synchronously collected by the sensing layer module deployed at the key nodes of the distribution network, and the electrical quantity data is subjected to local anomaly detection and preprocessing. The processed data is then sent to the network layer module. Step S402: The network layer module receives the processed data based on the 5G communication network and transmits the processed data to the platform layer module. Step S403: The processed data is received and aggregated by the platform layer module deployed on the main station side. Based on the processed data, the power distribution network is subjected to operation status perception, topology anomaly detection and fault diagnosis and location. The diagnosis results are then output to the visualization display device.

[0049] The embodiments of the present invention are method embodiments corresponding to the system embodiments described above. The specific operations of each step can be understood by referring to the description of the system embodiments, and will not be repeated here.

[0050] In summary, compared with the prior art, the embodiments of the present invention have the following beneficial effects: 1. High-precision sensing capability: through The PMU achieves a configurable sampling rate of 12.8kHz to 51.2kHz, with a hardware over-limit detection response time of ≤1ms, phasor calculation accuracy meeting the IEEE C37.118 standard, amplitude error ≤0.1%, and phase angle error ≤0.01°. 2. Autonomous and controllable time synchronization: Utilizing a BeiDou time synchronization module, the time synchronization accuracy is better than 20ns, the timekeeping capability is ≥24 hours, and the timekeeping error is ≤1. s, completely eliminating dependence on GPS; 3. Low-latency reliable communication: Based on 5G network slicing technology, the end-to-end latency is ≤15ms, the reliability is 99.999%, and it supports breakpoint resumption to ensure that no data is lost; 4. Intelligent diagnosis and localization: Combining the TodyNet model (GraphSAGE graph neural network) with the topology change association algorithm, the fault localization error is ≤50 meters, and the topology change detection accuracy is ≥95%; 5. Anomaly Classification and Handling: A three-level anomaly classification mechanism is adopted, with attention level local records not reported, alarm level records compressed and transmitted, and fault level records reported urgently, effectively reducing the processing burden on the main station. 6. High system integration: The hardware integrates sampling, timing, communication and computing modules, with small size (standard 1U chassis), low power consumption (≤25W), and easy deployment and maintenance.

[0051] Device Example 1 This invention provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the computer program is executed by the processor, it performs the steps described in the method embodiment.

[0052] Device Example 2 This invention provides a computer-readable storage medium storing an information transmission implementation program, which, when executed by a processor, performs the steps described in the method embodiment.

[0053] The computer-readable storage media described in this embodiment include, but are not limited to, ROM, RAM, disk, or optical disk.

[0054] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power distribution network operation status awareness and fault diagnosis system, characterized in that, It includes a perception layer module, a network layer module, and a platform layer module connected in sequence; The sensing layer module is deployed at key nodes of the distribution network. Its electrical quantity acquisition terminal is connected to the primary equipment of the distribution network, and its data output terminal is connected to the data input terminal of the network layer module. The sensing layer module is used to synchronously acquire electrical quantity data of the distribution network, perform local anomaly detection and preprocessing on the electrical quantity data, and send the processed data to the network layer module. The data output terminal of the network layer module is connected to the data input terminal of the platform layer module; the network layer module is used to receive the processed data based on the 5G communication network and transmit the processed data to the platform layer module. The platform layer module is deployed on the main station side, and its data output end is connected to the visualization display device. The platform layer module is used to receive and aggregate the processed data, perform operational status awareness, topology anomaly detection and fault diagnosis and location on the distribution network based on the processed data, and output the diagnosis results to the visualization display device.

2. The system according to claim 1, characterized in that, The perception layer module includes a miniature synchronous phasor measurement device, which includes an FPGA submodule, a DSP submodule, a BeiDou timing submodule, a 5G communication module, and a power supply submodule. The first input terminal of the FPGA submodule is connected to the second pulse output terminal of the Beidou time synchronization submodule, the second input terminal of the FPGA submodule is connected to the electrical quantity output terminal of the primary equipment of the power distribution network, and the output terminal of the FPGA submodule is connected to the input terminal of the DSP submodule. The FPGA submodule is used to synchronously collect electrical quantity data of the power distribution network according to the second pulse signal provided by the Beidou time synchronization submodule, and to perform over-limit detection, buffering and preprocessing on the electrical quantity data, and transmit the preprocessed data to the DSP submodule. The output of the DSP submodule is connected to the input of the 5G communication module. The DSP submodule is used to receive the preprocessed data, calculate phasor data based on the preprocessed data, extract time-domain features and frequency-domain features based on the phasor data, classify the severity of anomalies using an anomaly detection algorithm based on the extracted time-domain features and frequency-domain features, compress the classified data, and transmit the compressed data to the 5G communication module. The output terminal of the 5G communication module serves as the data output terminal of the perception layer module and is connected to the data input terminal of the network layer module; the 5G communication module is used to encapsulate the compressed data into data packets and send them to the network layer module through the 5G air interface. The BeiDou timing submodule is used to receive BeiDou satellite signals, generate second pulse signals, and provide them to the FPGA submodule. The power output terminal of the power supply submodule is connected to the power input terminals of the FPGA submodule, the DSP submodule, the Beidou timing submodule, and the 5G communication module, respectively, to provide working power to each submodule.

3. The system according to claim 2, characterized in that, The FPGA submodule includes a multi-channel synchronous sampling unit, an over-limit detection unit, and a data buffer unit; The input terminal of the multi-channel synchronous sampling unit is connected to the electrical quantity output terminal of the primary equipment of the power distribution network, and the control terminal of the multi-channel synchronous sampling unit is connected to the second pulse output terminal of the Beidou time synchronization submodule. The multi-channel synchronous sampling unit is used to perform synchronous sampling according to the second pulse signal at a configurable sampling rate and output the sampled data to the data buffer unit. The input terminal of the over-limit detection unit is connected to the output terminal of the multi-channel synchronous sampling unit. The over-limit detection unit is used to monitor the instantaneous current value in real time, and generates an interrupt signal when the detected value exceeds a preset threshold. The input terminal of the data buffer unit is connected to the output terminal of the multi-channel synchronous sampling unit and the output terminal of the over-limit detection unit, respectively. The data buffer unit is used to buffer the sampled data and interrupt signal, and output the buffered data to the DSP submodule.

4. The system according to claim 2, characterized in that, The DSP submodule includes a phasor calculation unit, a feature extraction unit, an anomaly classification unit, and a data compression unit; The input terminal of the phasor calculation unit is connected to the output terminal of the FPGA submodule. The phasor calculation unit is used to calculate phasor data according to a preset period and generate data frames that conform to the IEEE standard. The input terminal of the feature extraction unit is connected to the output terminal of the phasor calculation unit. The feature extraction unit is used to extract time-domain features and frequency-domain features based on the phasor data. The time-domain features include at least one of the effective value of current, peak factor and waveform distortion rate, and the frequency-domain features include harmonic content. The input of the anomaly classification unit is connected to the output of the feature extraction unit. The anomaly classification unit is used to obtain anomaly feature parameters based on the time-domain features and frequency-domain features, and to classify the severity of anomalies into multiple preset levels according to the anomaly feature parameters. The input end of the data compression unit is connected to the output end of the anomaly classification unit. The data compression unit is used to compress the data output by the anomaly classification unit and then output it to the 5G communication module.

5. The system according to claim 4, characterized in that, The anomaly classification unit includes: The first-level determination subunit outputs a local recording instruction when the current exceedance is less than the first preset threshold and the duration is less than the first preset duration, storing the time domain features and frequency domain features in the local memory and marking the timestamp. The second-level judgment subunit outputs a waveform recording start command when the current exceeds the limit to the second preset threshold range or the duration reaches the second preset duration range, triggering the FPGA submodule to record the original sampled data and transmit the recorded data to the platform layer module after compression. The third-level judgment subunit outputs an emergency reporting command when the current exceeds the third preset threshold, or the zero-sequence current exceeds the limit, or a voltage drop is detected. The abnormal classification results and related data are reported to the platform layer module in real time, triggering the fault diagnosis process. Wherein, the first preset threshold is less than the lower limit of the range of the second preset threshold, and the upper limit of the range of the second preset threshold is less than the third preset threshold.

6. The system according to claim 1, characterized in that, The platform layer module includes a phasor data concentrator and the TodyNet analysis platform; The input of the phasor data concentrator is connected to the data output of the network layer module, and the output of the phasor data concentrator is connected to the input of the TodyNet analysis platform. The phasor data concentrator is used to receive multi-source processed data transmitted by the network layer module, sort and align the multi-source processed data according to timestamps, interpolate and complete missing data, and transmit the aligned data to the TodyNet analysis platform. The output of the TodyNet analysis platform is connected to a visualization display device. The TodyNet analysis platform is used to receive aligned data and perform operational status awareness, topology anomaly detection, and fault diagnosis and location on the distribution network based on the aligned data, and output the diagnostic results to the visualization display device.

7. The system according to claim 6, characterized in that, The TodyNet analysis platform includes a graph construction unit, a graph neural network model unit, and a topology anomaly detection unit. The input of the graph construction unit is connected to the output of the phasor data concentrator, and the output of the graph construction unit is connected to the input of the graph neural network model unit. The graph construction unit is used to construct a distribution network topology graph with each PMU node in the distribution network as a graph node and the feeder connection relationship as an edge, and output the distribution network topology graph to the graph neural network model unit. The output of the graph neural network model unit is connected to the visualization display device; the graph neural network model unit is used to receive the power distribution network topology map, perform fault type classification and fault distance regression on the power distribution network based on the power distribution network topology map, and output the fault diagnosis results to the visualization display device. The input of the topology anomaly detection unit is connected to the output of the phasor data concentrator, and the output of the topology anomaly detection unit is connected to the visualization display device. The topology anomaly detection unit is used to receive the aligned data, detect topology change events based on current waveform similarity and voltage phase angle change rate, and output the topology anomaly detection results to the visualization display device.

8. A method for power distribution network operation status perception and fault diagnosis, characterized in that, include: The electrical quantity data of the distribution network is collected synchronously by the sensing layer module deployed at key nodes of the distribution network, and the electrical quantity data is subjected to local anomaly detection and preprocessing. The processed data is then sent to the network layer module. The network layer module receives the processed data based on the 5G communication network and transmits the processed data to the platform layer module. The platform layer module deployed on the main station side receives and aggregates the processed data, performs operational status awareness, topology anomaly detection, and fault diagnosis and location on the distribution network based on the processed data, and outputs the diagnostic results to the visualization display device.

9. An electronic device, characterized in that, include: The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the steps of the power distribution network operation status perception and fault diagnosis method as described in claim 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores an information transmission implementation program, which, when executed by a processor, implements the steps of the power distribution network operation status perception and fault diagnosis method as described in claim 8.

Images