CN-121984870-A - Low-voltage station line topology identification method and system based on waveform coherence

CN121984870ACN 121984870 ACN121984870 ACN 121984870ACN-121984870-A

Abstract

The invention relates to a low-voltage transformer area line topology identification method and system based on waveform coherence. The method comprises the steps of constructing a precise moment network covering an intelligent gateway and each monitoring terminal, injecting characteristic current into a power grid by the monitoring terminals, collecting line background noise by the monitoring terminals to generate reverse offset signals to preprocess original current signals, performing secondary filtering through an FIR band-pass filter, constructing four-dimensional characteristic vectors, introducing attention mechanisms, determining effective signals by combining an edge collaborative checking mechanism, generating reference checking signals, feeding back to the monitoring terminals to perform parameter calibration, identifying fault nodes of a low-voltage station line topological structure, and performing topology information complementation. By constructing a precise moment network, implementing dynamic characteristic current injection and combining multidimensional coherence waveform identification and edge collaborative verification, the automatic and high-precision identification of the low-voltage distribution network topology can be effectively realized.

Inventors

LI JUNCHEN
WANG YIPING
WANG DEJUN
GE JIAWEI

Assignees

佳源科技股份有限公司
云起智能(南京)科技有限公司

Dates

Publication Date: 20260505
Application Date: 20251209

Claims (10)

1. The method for identifying the circuit topology of the low-voltage transformer area based on waveform coherence is characterized by comprising the following steps: The intelligent gateway issues accurate time setting commands to each monitoring terminal in an HPLC communication mode, the monitoring terminals receive time setting values and record the time setting values, and an accurate time network covering the intelligent gateway and each monitoring terminal is constructed; under the accurate moment network, the intelligent gateway dynamically selects an injection frequency point by carrying out frequency spectrum analysis in real time, determines characteristic current waveform parameters and synchronizes the characteristic current waveform parameters to each monitoring terminal; the monitoring terminal acquires an original current signal, acquires a line background noise generation reverse offset signal to preprocess the original current signal, and carries out secondary filtering on the preprocessed signal through an FIR band-pass filter to obtain a filtered signal; Based on the signals after filtering, extracting characteristic parameters from a time domain, a frequency domain, a phase domain and an energy domain, constructing a four-dimensional characteristic vector, introducing an attention mechanism to dynamically allocate characteristic weights, generating an in-phase reference signal and a quadrature reference signal, and determining an effective signal by combining an edge collaborative checking mechanism; The intelligent gateway generates a reference check signal based on historical data and feeds back the reference check signal to the monitoring terminal after receiving the identification result of the monitoring terminal, and the monitoring terminal performs parameter calibration based on the reference check signal, identifies a fault node of a low-voltage station line topological structure and performs topology information completion.
2. The method for identifying the low-voltage transformer area circuit topology based on waveform coherence according to claim 1, wherein the intelligent gateway issues accurate time setting commands to each monitoring terminal in an HPLC communication mode, the monitoring terminal receives and records the time setting values, and a precise time network covering the intelligent gateway and each monitoring terminal is constructed, and the method comprises the following steps: The intelligent gateway issues accurate time setting commands to each monitoring terminal in an HPLC communication mode, the monitoring terminals receive and record time setting values, and time setting accuracy error control is carried out based on the time setting values; under the control of time setting precision errors, calculating clock deviation values of all monitoring terminals relative to the intelligent gateway; Based on the clock deviation value, the intelligent gateway carries out synchronous time service on each monitoring terminal in a mode of combining broadcasting and on-demand at a designated moment, so that all the monitoring terminals are synchronously calibrated to a target accurate moment, and the construction of a network covering the accurate moment of the intelligent gateway and each monitoring terminal is completed.
3. The method for identifying the low-voltage transformer area line topology based on waveform coherence according to claim 1, wherein the intelligent gateway dynamically selects injection frequency points by performing frequency spectrum analysis in real time, determines characteristic current waveform parameters and synchronizes the characteristic current waveform parameters to each monitoring terminal, and the monitoring terminal injects characteristic current into a power grid based on the characteristic current waveform parameters, and the method comprises the following steps: the intelligent gateway collects line background noise, and selects a frequency point with the lowest noise power spectrum density from a plurality of preset candidate frequency points as an injection frequency point; the intelligent gateway judges the load type through the monitored line current effective value and the monitored change rate, and matches the corresponding characteristic current waveform parameters according to the load type; and the intelligent gateway synchronizes the injection frequency point and the characteristic current waveform parameters to all monitoring terminals through an HPLC channel, and injects the characteristic current after the power frequency voltage crosses the zero point.
4. The method for identifying the circuit topology of the low-voltage transformer area based on waveform coherence as set forth in claim 1, wherein the monitoring terminal collects an original current signal, collects a reverse cancellation signal generated by a background noise of the circuit to preprocess the original current signal, and secondarily filters the preprocessed signal through an FIR band-pass filter to obtain a filtered signal, comprising: The monitoring terminal acquires an original current signal, preprocesses the original current signal through self-adaptive noise cancellation, filters broadband interference, and obtains a preprocessed signal; and configuring an FIR software band-pass filter, performing secondary filtering on the preprocessed signals through the FIR band-pass filter, and synchronizing the characteristic current injection with the precision of the power frequency voltage zero crossing point under the accurate time network to obtain the filtered signals.
5. The method for identifying a low voltage transformer area line topology based on waveform coherence of claim 1, further comprising: After four-dimensional feature vectors comprising a time domain, a frequency domain, a phase domain and an energy domain are constructed, each feature weight is dynamically distributed through an attention mechanism based on a real-time scene, and feature matching scores are calculated.
6. The method for identifying a low voltage transformer area circuit topology based on waveform coherence of claim 5, wherein generating an in-phase reference signal, a quadrature reference signal, and determining an effective signal in combination with an edge cooperative verification mechanism comprises: generating an in-phase reference signal and a quadrature reference signal, calculating an in-phase branch amplitude according to the in-phase reference signal, and calculating a quadrature branch amplitude according to the quadrature reference signal; calculating according to the in-phase branch amplitude and the quadrature branch amplitude to obtain a synthesized signal amplitude; adopting a reinforcement learning algorithm, and dynamically adjusting a threshold range according to the signal-to-noise ratio change; And carrying out signal validity judgment according to the feature matching score, the threshold range and the synthesized signal amplitude, and determining a valid signal.
7. The method of low voltage transformer area line topology identification based on waveform coherence of claim 6, further comprising: capturing sampling time corresponding to the maximum value of the synthesized signal amplitude through a sliding window; comparing the sampling time with zero crossing time on the power frequency voltage, and calculating a time difference; And determining the initial signal direction of the effective signal based on the time difference, and correcting through a random forest classification model to obtain the signal direction of the effective signal.
8. The method for identifying the low-voltage transformer area line topology based on waveform coherence according to claim 1, wherein the monitoring terminal performs parameter calibration based on the reference check signal, and identifies a fault node of the low-voltage transformer area line topology structure, comprising: the monitoring terminal compares the self-monitoring data with the reference check signal, and if the deviation is larger than a reference threshold value, the monitoring terminal is marked as a fault node; And if the monitoring terminal does not recognize the effective signal, sending a re-recognition request to the intelligent gateway, and if the number of times that the monitoring terminal does not recognize the effective signal is larger than the target number of times, marking the monitoring terminal as a fault node.
9. The method for identifying a low voltage transformer area line topology based on waveform coherence of claim 1, further comprising: The monitoring terminal monitors the line current in real time, and if the line load suddenly changes, the monitoring terminal triggers the quick re-identification to update the line topology structure of the low-voltage transformer area; Based on the updated low-voltage station area line topology structure, the intelligent gateway complements fault area topology information based on adjacent node signal coverage.
10. The low-voltage station line topology identification system based on waveform coherence is characterized by comprising a monitoring terminal and an intelligent gateway, wherein the monitoring terminal is in connection communication with the intelligent gateway, and the system comprises the following components: The intelligent gateway issues accurate time setting commands to each monitoring terminal in an HPLC communication mode, the monitoring terminals receive time setting values and record the time setting values, and an accurate time network covering the intelligent gateway and each monitoring terminal is constructed; under the accurate moment network, the intelligent gateway dynamically selects an injection frequency point by carrying out frequency spectrum analysis in real time, determines characteristic current waveform parameters and synchronizes the characteristic current waveform parameters to each monitoring terminal; The monitoring terminal collects original current signals, collects line background noise to generate reverse offset signals, preprocesses the original current signals, and secondarily filters the preprocessed signals through the FIR band-pass filter to obtain filtered signals; The monitoring terminal extracts characteristic parameters from a time domain, a frequency domain, a phase domain and an energy domain based on the signals after the filtering processing, constructs a four-dimensional characteristic vector, introduces a attention mechanism to dynamically allocate characteristic weights, generates an in-phase reference signal and a quadrature reference signal, and combines an edge collaborative checking mechanism to determine an effective signal; The intelligent gateway generates a reference check signal based on historical data and feeds back the reference check signal to the monitoring terminal after receiving the identification result of the monitoring terminal, and the monitoring terminal performs parameter calibration based on the reference check signal, identifies a fault node of a low-voltage station line topological structure and performs topology information completion.

Description

Low-voltage station line topology identification method and system based on waveform coherence Technical Field The invention relates to the technical field of low-voltage power line topology identification, in particular to a low-voltage station area line topology identification method and system based on waveform coherence. Background In recent years, the power industry in China realizes rapid development and construction. As a key component of the power system, the low-voltage distribution network not only bears important responsibilities of peak regulation and frequency modulation, voltage level stability maintenance and the like for guaranteeing safe and stable operation of the power system, but also promotes the core link of the power network of China to develop in an integrated, coordinated and intelligent direction. With the continuous deep construction of the electric power Internet of things and the intelligent power grid, a large number of electric vehicles, distributed photovoltaic devices, energy storage devices and the like are connected into the power grid, so that the low-voltage power distribution network is characterized by multi-source isomerism and dynamic fluctuation, and higher technical requirements on the identification precision, response speed and anti-interference capability of the topological structure are provided. At present, the low-voltage distribution network topology in China mainly has two main characteristics that the development process is relatively slow, part of the areas still depend on the traditional manual account, the network structure is complex and variable, and the scenes of user capacity increase, line transformation and the like are frequent, so that the topology information update is delayed. The traditional low-voltage distribution network topology structure identification means mainly depend on the partition switching identification and manual recording modes. On the one hand, the method can interrupt the power supply of a user, the influence duration of single switching-off is about 10-30 minutes, inconvenience is brought to the life and business power consumption of residents, and on the other hand, the manual operation is easily influenced by subjective factors and environmental conditions, so that the topology information leakage rate is high, and the operation and maintenance requirements of the intelligent power grid are difficult to meet. Therefore, a topology identification technology based on a characteristic signal injection method appears, and the core principle is that by monitoring and analyzing current signals in a power grid, extracting characteristic information in the current signals and combining the prior knowledge of the power grid, the accurate identification of the power grid topology structure is realized, and the topology identification technology has the advantages of high identification precision, strong environmental adaptability and the like, and is theoretically suitable for topology identification scenes of various complex low-voltage power networks. However, in the practical application process, the method has obvious limitations that firstly, the operation equipment in a partial area (such as an industrial load intensive area and a new energy collection area) can generate strong electromagnetic noise in a characteristic signal frequency band, the noise intensity can reach 3-5 times of that of a characteristic signal, the characteristic signal is submerged, the recognition success rate is lower than 60%, secondly, the dynamic adaptation capability is lacking, when the load of a station area fluctuates or the access state of equipment changes, signal parameters need to be debugged again, the operation and maintenance cost is higher, thirdly, single signal judgment is easily influenced by accidental interference, the misjudgment rate is higher, fourthly, the prior art is more dependent on single-dimension recognition logic of 'double-phase-sensitive detection+fixed threshold amplitude judgment', interference and dynamic change under complex scenes are difficult to cope, and the accuracy and reliability of topology recognition are further limited. Therefore, when the characteristic signal injection method is used for topology identification, the noise intensity is high, the dynamic adaptation capability is lacked, the identification logic is single, and the problems of low accuracy and low practicability of the topology identification are often caused. Disclosure of Invention Based on the above, in order to solve the above technical problems, a method and a system for identifying a low-voltage transformer area circuit topology based on waveform coherence are provided, so that accuracy and practicability of topology identification can be improved. A low voltage transformer area line topology identification method based on waveform coherence, the method comprising: The intelligent gateway issues accurate time setting commands to ea