A method and system for identifying the topology of a charging pile in a transformer area based on two-factor electrical fingerprints

Abstract

The application provides a transformer area charging pile topology identification method and system based on a double-factor electrical fingerprint, and belongs to the technical field of power distribution network planning and operation, which comprises the following steps: synchronously collecting voltage and current waveforms, and extracting steady-state active power and specific harmonic current proportion; calculating a characteristic deviation degree, and when the characteristic deviation degree is less than a dynamic matching deviation threshold, determining a load corresponding to an access time as a candidate charging pile; obtaining an active power difference value of a node and adjacent upstream nodes, and confirming an effective charging pile according to a consistency verification result; identifying an access phase, and according to a power conservation relationship, combining a topology connection sequence, and positioning a specific branch line step by step; recording each identified charging pile and corresponding access phase and attribution branch, and generating a transformer area charging pile topology distribution map. The application solves the problem that the existing topology identification method is single in characteristics, is easily disturbed by a background, and leads to inaccurate identification of a charging pile access position and phase, and cannot be finely positioned to a branch line.

Description

Technical Field

[0001] This invention relates to the field of power distribution network planning and operation technology, specifically to a method and system for topology identification of charging piles in transformer substations based on two-factor electrical fingerprints. Background Technology

[0002] With the rapid growth in the number of electric vehicles, a large number of charging piles are being connected to low-voltage distribution transformer areas, leading to increasingly complex load structures in these areas. To ensure the safe and economical operation of these transformer areas and support orderly charging dispatch, power grid companies need to accurately determine the specific connection location, branch line, and phase of each charging pile within the transformer area—that is, the complete topology of the charging piles.

[0003] Currently, topology identification of low-voltage distribution areas mainly relies on manual surveys and ledger records, with maintenance personnel checking line routes and connections on-site. This method is inefficient and prone to discrepancies between records and reality due to line modifications and temporary connections. Another method, topology identification based on power line carrier communication, involves deploying carrier signal transceivers on both the transformer and user sides, judging equipment connections based on signal attenuation and transmission path characteristics. However, when there are many branches in the distribution area and significant load variations, the carrier signal is susceptible to interference and cross-distribution crosstalk, significantly reducing identification accuracy. In charging pile access scenarios, single-type features are easily affected by fluctuations in the background load of the distribution area and overlapping features of similar equipment, leading to missed identifications or misidentification of ordinary nonlinear loads as charging piles. This fails to meet the precision required for precise branch line location in the refined management of large-scale charging pile access.

[0004] In summary, how to overcome the shortcomings of existing topology identification methods in charging pile access scenarios, such as single feature identification, susceptibility to background interference, and difficulty in accurately locating branch lines, and to realize a charging pile topology identification method with strong anti-interference ability and precise positioning, is an urgent problem to be solved in this field. Summary of the Invention

[0005] This invention provides a method and system for topology identification of charging piles in transformer substations based on two-factor electrical fingerprints. This method addresses the problems of existing topology identification methods, which suffer from inaccurate identification of charging pile access locations and phases and are susceptible to background interference due to their single feature and inability to accurately locate branch lines.

[0006] In view of the above problems, the present invention provides a method and system for topology identification of charging piles in transformer substations based on two-factor electrical fingerprints.

[0007] In a first aspect, the present invention provides a method for topology identification of charging piles in a transformer substation based on two-factor electrical fingerprinting, comprising: Voltage and current waveforms are simultaneously collected on the low-voltage side of the transformer and at each branch node in the distribution area, and the steady-state active power and the proportion of specific harmonic current at the connection time are extracted to form a two-factor electrical fingerprint. Calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile. Obtain the active power difference between the node where the candidate charging pile is located and its adjacent upstream node within the same time window. Based on the consistency verification result between the active power difference and the steady-state active power, confirm the valid charging pile. Identify the access phase of the valid charging pile, and based on the power conservation relationship between the node and its upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level. Record each identified charging pile, its corresponding access phase, and its affiliated branch to generate a topology distribution map of charging piles in the area.

[0008] Secondly, the present invention provides a topology identification system for charging piles in a transformer substation based on two-factor electrical fingerprinting, comprising: The two-factor fingerprint extraction module is used to synchronously collect voltage and current waveforms on the low-voltage side of the transformer and at each branch node, and extract the steady-state active power and the proportion of specific harmonic current at the connection time to form a two-factor electrical fingerprint. The candidate charging pile determination module is used to calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile. The active power consistency verification module is used to obtain the active power difference between the node where the candidate charging pile is located and the adjacent upstream node within the same time window, and to confirm the valid charging pile based on the consistency verification result of the active power difference and the steady-state active power. The phase and branch positioning module is used to identify the access phase of the valid charging pile and, based on the power conservation relationship between the node and the upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level. The topology map generation module is used to record each identified charging pile and its corresponding access phase and affiliated branch, and generate a topology map of the charging piles in the area.

[0009] One or more technical solutions provided in this invention have at least the following technical effects or advantages: The technical solution of this invention first synchronously collects three-phase voltage and current waveforms on the low-voltage side of the transformer and at each branch node. After anti-aliasing filtering and synchronous analog-to-digital conversion, the current waveform is aligned based on the voltage zero-crossing point, and real-time sliding window Fourier analysis is performed to continuously output the active power sequence and the third and fifth harmonic current content sequences. When the active power increment exceeds the power change threshold and stabilizes within the typical power range of the slow charging pile within multiple consecutive power frequency cycles, the connection time is marked. After transient avoidance, the steady-state active power and the ratio difference of each harmonic current are extracted to form a two-factor electrical fingerprint. A multi-dimensional feature label integrating power level and harmonic spectrum characteristics is established for each connected load, effectively distinguishing charging piles from background loads such as purely resistive loads and rotating motors, providing highly recognizable feature inputs for subsequent identification.

[0010] Furthermore, the real-time extracted two-factor electrical fingerprint is compared with a pre-constructed slow-charging pile feature benchmark. The normalized relative deviation of each dimension is calculated and aggregated into a feature deviation degree using the sum of squares and square roots. Simultaneously, a dynamic matching deviation threshold is constructed using the fluctuation within the access background window and the offline tolerance upper limit through root mean square calculation. When the feature deviation degree is less than the dynamic matching deviation threshold, the load is identified as a candidate charging pile. Through multi-dimensional feature quantitative matching and adaptive threshold adjustment, the matching conditions are automatically relaxed to capture real charging piles when the background fluctuates, while strict standards are maintained to exclude non-charging pile loads when the background is stable, achieving a balance between high capture rate and low false positive rate.

[0011] Furthermore, the active power increments of the candidate charging pile node and its adjacent upstream node within the same time window are obtained, the difference between the two is calculated, and compared with the steady-state active power to obtain the power conservation deviation rate. When the deviation rate is less than the preset power conservation error threshold, the candidate charging pile is confirmed as a valid charging pile. Upstream and downstream verification is performed using the power conservation principle of the distribution network to effectively eliminate false power surge events caused by load switching transmission from adjacent nodes or measurement disturbances, ensuring that all charging pile events relied upon for subsequent location have genuine local access evidence.

[0012] Furthermore, by comparing the three-phase power increments of the node where the effective charging pile is located, the connected phase is identified. Based on this, starting from the current node, the system traverses downstream level by level along the topology connection sequence. Combining the power conservation check between parent and child nodes, when the difference between the sum of the power increments of the parent node and the child node exceeds the power change threshold, it is confirmed that the charging pile is directly connected to the current node, and the assigned branch line is output. This refines the location of the charging pile from the node level to the specific branch line level, while clearly defining the connected phase, achieving precise location of phase-node-branch line, and eliminating the positioning ambiguity caused by insufficient branch penetration in traditional methods.

[0013] Finally, a tree-like topology structure for the charging station area is established, and the charging pile records for each node are initialized. The two-factor electrical fingerprint of each valid charging pile is used as an identifier and stored along with the access phase in the corresponding node. After automatically deduplicating duplicate records at the same node through fingerprint similarity comparison, the records are merged to generate a charging pile topology distribution map for the area. Furthermore, the real-time power of the recorded charging piles is continuously monitored. Upon detecting a power drop and confirming disconnection, the topology map is automatically updated. This results in a visualized charging pile topology distribution map, allowing maintenance personnel to intuitively grasp the precise location and online status of each charging pile, and maintaining the real-time accuracy of the topology information through a dynamic update mechanism.

[0014] In summary, the technical solution of this invention enhances the distinction between charging piles and background loads by using a two-factor fingerprint composed of the difference between steady-state active power and harmonic current ratios; improves the robustness of judgment under complex operating conditions by using a background adaptive dynamic threshold; eliminates the interference of false events by using upstream and downstream power conservation verification; and achieves precise positioning from transformer outgoing lines to terminal branches by using step-by-step topology penetration. The final output dynamic topology distribution map can provide real-time, reliable, and precise topology data for services such as orderly charging scheduling in distribution areas, three-phase imbalance management, line overload early warning, and business expansion applications. It effectively solves the problem that existing topology identification methods are prone to inaccurate identification of charging pile access location and phase, and cannot precisely locate branch lines due to single features and susceptibility to background interference. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the method for topology identification of charging piles in transformer substations based on two-factor electrical fingerprints provided by the present invention.

[0016] Figure 2 This is a schematic diagram of the topology distribution of charging piles in the charging pile topology identification method based on two-factor electrical fingerprint provided by the present invention.

[0017] Figure 3 This is a schematic diagram of the topology identification system for charging piles in a transformer substation based on two-factor electrical fingerprints provided by the present invention.

[0018] In the attached diagram, the labels representing each component are as follows: The module includes a two-factor fingerprint extraction module 11, a candidate charging pile determination module 12, an active power consistency verification module 13, a phase and branch location module 14, and a topology distribution map generation module 15. Detailed Implementation

[0019] This invention provides a method and system for topology identification of charging piles in transformer substations based on two-factor electrical fingerprints. This solves the problems of existing topology identification methods, which suffer from inaccurate identification of charging pile access location and phase, and inability to accurately locate branch lines due to single features and susceptibility to background interference.

[0020] It should be noted that the terms "comprising" and "having" are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to these processes, methods, products, or devices.

[0021] Example 1, as Figure 1 As shown, this invention provides a method for topology identification of charging piles in a transformer substation based on two-factor electrical fingerprints. The method includes: S100: Simultaneously collect voltage and current waveforms on the low-voltage side of the transformer and at each branch node, and extract the steady-state active power and the proportion of specific harmonic current at the connection time to form a two-factor electrical fingerprint.

[0022] This step involves deploying synchronous sampling devices at the low-voltage outgoing line side of the transformer in the distribution area and at the nodes of each branch line to collect voltage and current waveform data in real time. When a load connection event is detected, the active power value of the load after entering steady state is extracted from the waveform at the time of the event trigger, and the percentage of the amplitude of a preset specific harmonic in its current waveform relative to the fundamental frequency amplitude is calculated. The steady-state active power and this set of harmonic current ratios are combined to form a two-factor electrical fingerprint.

[0023] Step S100 in the method provided by the present invention includes: Voltage signals are acquired by voltage transformers connected in parallel to the low-voltage side of the transformer and each branch node, and current signals at the same node are acquired by through-type current transformers. The voltage and current signals are three-phase voltage and three-phase current signals. After performing anti-aliasing low-pass filtering on the collected voltage and current signals, synchronous analog-to-digital conversion is performed to obtain the three-phase voltage sampling sequence and the three-phase current sampling sequence. The power frequency zero-crossing phase point of the three-phase voltage sampling sequence is detected and used as the starting boundary to perform full-cycle truncation and alignment of the three-phase current sampling sequence to generate a current waveform; Real-time sliding window Fourier analysis is performed on the synchronously acquired voltage and current waveforms to continuously output the active power sequence and the third and fifth harmonic current content sequences. When the increment of the active power sequence in multiple consecutive power frequency cycles exceeds the power change threshold in the characteristic benchmark of the slow charging pile, and the active power after the increment is stable in the typical power range of the slow charging pile, the start time of the increment is marked as the access time. The average active power over N consecutive power frequency cycles after the access time is delayed by one start-up transient avoidance window is taken as the steady-state active power, wherein the start-up transient avoidance window is M power frequency cycles, and N and M are both positive integers; The third harmonic current content rate and the fifth harmonic current content rate in the steady-state window after the access time are respectively subtracted from the corresponding harmonic current content rate in the background window before the access time to obtain the third harmonic current ratio difference and the fifth harmonic current ratio difference. The steady-state window is N consecutive power frequency cycles after the access time and after the start transient avoidance window is delayed, and the background window is K consecutive power frequency cycles before the access time, where K is a positive integer. The steady-state active power, the ratio difference of the third harmonic current, and the ratio difference of the fifth harmonic current are combined into a three-dimensional vector, which serves as the two-factor electrical fingerprint.

[0024] In this step, three-phase voltage signals are first collected on each phase of the low-voltage side of the transformer and each branch node through parallel voltage transformers, and three-phase current signals at the same node are collected at the same time through through-type current transformers, so as to realize the synchronous acquisition of voltage and current at the same node and in the same phase.

[0025] For example, at the incoming line of branch box No. 1 in transformer area B, through-type current transformers are installed on the A-phase, B-phase, and C-phase busbars of the node respectively, and voltage transformers are connected in parallel to obtain the voltage and current signals of the three phases of the node synchronously.

[0026] Secondly, the acquired three-phase voltage signals and three-phase current signals are subjected to anti-aliasing low-pass filtering to remove high-frequency noise and potential aliasing components. The filtered six signals are then sent to a synchronous sampling analog-to-digital converter to generate a three-phase voltage sampling sequence and a three-phase current sampling sequence.

[0027] Anti-aliasing low-pass filtering refers to performing low-pass filtering on the analog signal before analog-to-digital conversion to eliminate or attenuate frequency components in the signal that are higher than the Nyquist frequency, i.e., half of the sampling frequency. It is a noise reduction method in the existing technology.

[0028] A synchronous sampling analog-to-digital converter (ADC) is an analog-to-digital converter with multi-channel synchronous sampling, holding, and conversion capabilities. All input channels share the same sampling clock signal and trigger source, enabling the analog signals of each channel to be sampled and held at the same time, and then quantization and encoding conversion to be completed sequentially or in parallel. This ensures strict alignment of multiple signals on the time axis and eliminates sampling time deviations between channels.

[0029] For example, the six analog signals, including phase A voltage and phase A current, collected at the inlet of branch box No. 1, are first passed through a low-pass filter with a cutoff frequency of 2kHz, and then sent to a six-channel synchronous analog-to-digital converter chip to output the digital sampling sequence of three-phase voltage and three-phase current at a sampling rate of 12.8kHz.

[0030] Next, the zero-crossing phase point of each phase in the three-phase voltage sampling sequence is detected. Using this zero-crossing point as the time reference boundary, the current sampling sequence of the same phase is truncated for an entire period to ensure that the current waveform and the voltage waveform are strictly aligned in phase, thereby generating a current waveform that can be used for harmonic analysis.

[0031] In this context, the zero-crossing phase point of the power frequency is the instantaneous phase moment in an AC power system when the instantaneous value of a voltage or current signal changes from negative to positive, or from positive to negative, and crosses zero. This step uses this as the starting boundary to perform full-cycle truncation and alignment of the current waveform, thereby establishing a time reference that is strictly synchronized with the phase of the grid voltage, ensuring that the starting phase of each truncation of the current waveform data window remains consistent.

[0032] For example, if the 1000th sampling point in the voltage sampling sequence of phase A is detected, which crosses the zero-crossing phase point of the power frequency from negative to positive, 256 sampling points are extracted from the zero-crossing phase point of the power frequency as the starting boundary, corresponding to one 50Hz power frequency cycle, and the current waveform data of one complete cycle of phase A is generated.

[0033] Next, real-time sliding window Fourier analysis is performed on the aligned voltage and current waveforms. The window function slides once for each power frequency cycle, and the active power value of the node, as well as the third harmonic current content rate and the fifth harmonic current content rate, are continuously output to form the active power sequence and the harmonic content rate sequence.

[0034] Real-time sliding window Fourier analysis is a method for dynamic spectrum calculation of continuously sampled time-series signals. The process is as follows: a fixed-length time window is defined, typically an integer number of power frequency cycles. The signal within the window is extracted and subjected to a discrete Fourier transform to obtain the spectrum information within that window, including the amplitude and phase of the fundamental frequency and each harmonic. Then, the window is slid backward along the time axis by a fixed step size, typically one power frequency cycle or half a cycle, and the Fourier transform is performed again on the signal within the new window. This process is repeated continuously.

[0035] For example, on the A-phase monitoring terminal at the incoming line of branch box No. 1, the time window length for real-time sliding window Fourier transform analysis is fixed at 10 complete power frequency cycles. For a 50Hz power grid, one power frequency cycle is 20ms, so the time window length is 10 × 20ms = 200ms, meaning that 200ms of continuous sampled data is extracted each time for Fourier transform. The sliding window step size is set to one power frequency cycle, meaning that the time window slides forward one step along the time axis every 20ms, and there is a 180ms overlap in time between adjacent analysis data windows.

[0036] For example, in the first analysis window: voltage and current data from 1ms to 200ms are captured. After Fourier transform, the average A-phase active power within 200ms is calculated to be 5.82kW, the third harmonic current content is 18.1%, and the fifth harmonic current content is 9.0%. These three values ​​are used as the first feature data point. In the second analysis window: the window slides forward 20ms, capturing data from 21ms to 220ms. The active power is calculated to be 5.80kW, the third harmonic current content is 17.9%, and the fifth harmonic current content is 9.1%, used as the second feature data point. In the third analysis window: the window slides forward another 20ms, capturing data from 41ms to 240ms. The active power is calculated to be 5.79kW, the third harmonic current content is 18.2%, and the fifth harmonic current content is 8.9%, used as the third feature data point. Following this rule, a new data point is output every 20ms, continuously recording data. Over time, three characteristic sequence curves are formed: the active power sequence of phase A, the third harmonic current content rate sequence, and the fifth harmonic current content rate sequence.

[0037] Furthermore, the active power sequence is monitored in real time. When the increase in active power over multiple consecutive power frequency cycles exceeds the preset power change threshold in the slow charging pile characteristic benchmark, and the active power value after the increase remains stable within the typical power range of the slow charging pile for at least five power frequency cycles, the time corresponding to the first cycle of the increase is marked as the access time.

[0038] In this step, the characteristic benchmark of the slow charging pile is obtained by repeatedly starting charging and synchronously collecting the electrical quantities of the nodes for various mainstream models of slow charging piles under the rated voltage fluctuation range of ±10% and the charging power corresponding to different battery states of charge. The steady-state active power, the third harmonic current ratio difference, and the fifth harmonic current ratio difference extracted each time are used as fingerprint samples. After removing outlier samples, the mean of each dimension is calculated to obtain the benchmark steady-state active power, benchmark third harmonic current ratio difference, and benchmark fifth harmonic current ratio difference. These are stored together with the power change threshold, the typical power range of the slow charging pile, and the number of transient avoidance cycles.

[0039] Specifically, the slow charging pile characteristic benchmark is a multi-dimensional reference template, pre-built offline, used to determine online whether the connected load is a candidate charging pile. The construction process first selects various mainstream models on the market, such as slow charging piles with rated power levels of 3.3kW and 7kW, and conducts charging start-up tests under the following operating conditions: Voltage fluctuation condition: within ±10% of the rated voltage of 220V, i.e., 198V to 242V, with a fixed step size, such as 2V, to adjust the voltage of the test area, covering the upper and lower boundaries of voltage deviation and the rated point; Battery state of charge condition: charging is triggered when the electric vehicle battery connected to the charging pile is in different states of charge, such as SOC of 20%, 40%, 60%, and 80%, to cover the charging power range corresponding to different power absorption characteristics during the constant current charging stage.

[0040] Furthermore, in each test, electrical quantities of the charging pile access node were simultaneously collected, and fingerprint feature values ​​in three dimensions were extracted: steady-state active power, i.e., the average active power value during the power stabilization phase after access (unit: kW); third harmonic current ratio difference, i.e., the difference between the steady-state third harmonic current content rate after access and the background harmonic content rate before access; and fifth harmonic current ratio difference, i.e., the difference between the steady-state fifth harmonic current content rate after access and the background harmonic content rate before access. The purpose of the above difference processing is to eliminate the interference of background harmonics in the transformer area on the uniqueness of the fingerprint, so that the feature benchmark reflects the net harmonic increment characteristics introduced by the charging pile itself.

[0041] Furthermore, for the fingerprint sample sets collected under all operating conditions, outlier samples are removed using statistical methods, such as removing abnormal data points that exceed three times the standard deviation. Then, the arithmetic mean of the effective samples in the three dimensions, the benchmark steady-state active power, the benchmark third harmonic current ratio difference, and the benchmark fifth harmonic current ratio difference are calculated respectively, which together constitute the electrical fingerprint benchmark vector of the slow charging pile.

[0042] Finally, the following features of the slow charging pile are stored: Power change threshold, which is the minimum power increment used to determine whether a power surge is a charging pile access event. This can be set to 80% of the lower limit of the rated power; for example, the threshold for a 3.3kW slow charging pile is set to 2.6kW. Typical power range of the slow charging pile, which defines the reasonable fluctuation range of steady-state power, such as [3.0kW, 7.5kW], covering the steady-state power range of mainstream 3.3kW and 7kW slow charging piles under voltage fluctuations. Transient avoidance cycle number: the number of cycles that need to be avoided after the charging pile starts due to transient processes such as capacitor charging and control handshake. For example, it can be set to 5 power frequency cycles. During this period, fingerprint features are not extracted; extraction is performed after the power enters a stable state.

[0043] The above parameters, together with the benchmark vector, constitute a complete characteristic benchmark for slow charging piles, which is stored in the local memory of the monitoring terminal or the database of the main station in the distribution area for use in the online matching process.

[0044] The requirement that the increase in active power exceed a threshold for P consecutive power frequency cycles is to avoid misjudgments caused by instantaneous impact loads, such as surge currents that typically last less than 1-2 power frequency cycles during motor startup. A value of P that is too small will result in weak anti-interference capability, while a value that is too large will delay the calibration of the connection time. Generally, a value of 3 to 5 is optimal for P.

[0045] For example, the number of power frequency cycles is set to P, where P is 3. That is, if the increase in active power exceeds the power change threshold within 3 consecutive power frequency cycles, and the power after the increase falls into the typical power range, it is determined to be a valid access event. The power change threshold is the threshold stored in the slow charging pile characteristic benchmark, such as 2.6kW. The typical power range of the slow charging pile is the range stored in the slow charging pile characteristic benchmark, such as [3.0kW, 7.5kW].

[0046] For example, at the A-phase monitoring point of branch box 1 in transformer area B, the active power sequence shows that the power is 0.5kW in cycle T0, jumps to 3.8kW in cycle T1, 5.6kW in cycle T2, and 5.8kW in cycle T3. Since the increments for three consecutive cycles all exceed the 2.6kW threshold, and the stable power from 5.6kW to 5.8kW falls within the range of [3.0kW, 7.5kW], the starting time of cycle T1 is marked as the connection time.

[0047] Furthermore, the average active power over N consecutive power frequency cycles after a delay of one startup transient avoidance window is taken as the steady-state active power. The startup transient avoidance window M is the number of transient avoidance cycles stored in the slow charging pile characteristic benchmark. For example, M = 5 power frequency cycles. During the first 5 cycles after the charging pile starts, transient processes such as control handshake, relay activation, and output capacitor charging occur, and the power is not yet stable, requiring avoidance.

[0048] The steady-state calculation window N is the length of the data window used to calculate the steady-state active power and the steady-state harmonic content. It is recommended to take a value between 5 and 20 to ensure that the obtained data is a stable average value. N is a positive integer and is set to N = 10 power frequency cycles. That is, after avoiding transients, the average value of active power over 10 consecutive cycles is taken as the steady-state active power to smooth out measurement noise and small fluctuations.

[0049] For example, if the access time is marked at the start of cycle T1, then skip the 5 transient avoidance cycles from T1 to T5, take the active power data of 10 cycles from T6 to T15, calculate their arithmetic mean as the steady-state active power, and obtain the steady-state active power as 5.82kW.

[0050] Furthermore, the average values ​​of the third harmonic current content rate and the fifth harmonic current content rate within the steady-state window after the access time are taken, and then the average values ​​of the corresponding harmonic current content rates within the background window before the access time are subtracted from each of them to obtain the third harmonic current ratio difference and the fifth harmonic current ratio difference.

[0051] The steady-state window and steady-state active power extraction share the same data window, which is the N consecutive cycles after a delay of M cycles from the access time. The background window K is the length of the data window used to calculate the background harmonic content before access. A value between 5 and 20 is recommended to ensure smooth and stable background harmonics without introducing excessive irrelevant background information, and K should be a positive integer. For example, K = 10 power frequency cycles can be set. The 10 consecutive cycles before the access time are used as the background window to calculate the background harmonic level before access.

[0052] For example, with the access time as the starting point of cycle T1, the background window covers 10 cycles from T(-9) to T0 before T1. The calculated average background third harmonic current content is 2.2%, and the average fifth harmonic current content is 1.1%. The steady-state window covers 10 cycles from T6 to T15, and the calculated average steady-state third harmonic current content is 18.4%, and the average fifth harmonic current content is 9.3%. Therefore, the difference in the proportion of the third harmonic current is 18.4% - 2.2% = 16.2%, and the difference in the proportion of the fifth harmonic current is 9.3% - 1.1% = 8.2%.

[0053] Finally, the three feature values ​​extracted in the above three steps are combined into a three-dimensional vector as the two-factor electrical fingerprint of the access load. For example, the final two-factor electrical fingerprint vector of the access event is [5.82kW, 16.2%, 8.2%].

[0054] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0055] In summary, compared with traditional single-feature recognition methods, this step effectively enhances the feature differentiation between charging piles and background loads in the transformer area, such as purely resistive lighting loads and rotating motor loads, significantly improving the accuracy of charging pile access events and reducing missed detections and misjudgments caused by feature overlap. At the same time, through precise locking of access time and synchronous extraction of steady-state features, it achieves rapid capture and multi-dimensional feature solidification of charging pile access behavior.

[0056] S200: Calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile.

[0057] This step involves matching the extracted two-factor electrical fingerprint of the access load with a pre-built feature benchmark for slow-charging piles. The feature deviation between the two-factor electrical fingerprint vector and the feature benchmark vector of the slow-charging pile is calculated, and this deviation is compared with a dynamic matching deviation threshold. If the deviation is less than the threshold, the electrical characteristics of the access load are determined to be highly consistent with the slow-charging pile characteristics, and it is marked as a candidate charging pile; if the deviation is greater than or equal to the threshold, it is excluded.

[0058] Step S200 in the method provided by the present invention includes: Extract the steady-state active power, the third harmonic current ratio difference, and the fifth harmonic current ratio difference from the two-factor electrical fingerprint; Extract the benchmark steady-state active power, the benchmark third harmonic current ratio difference, and the benchmark fifth harmonic current ratio difference from the aforementioned slow charging pile characteristic benchmarks; Based on the steady-state active power and the reference steady-state active power, the relative deviation of active power is determined, wherein the relative deviation of active power increases monotonically with the absolute value of the difference between the steady-state active power and the reference steady-state active power. Based on the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current, the relative deviation of the third harmonic is determined, wherein the relative deviation of the third harmonic increases monotonically with the absolute value of the difference between the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current. Based on the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference, the fifth harmonic relative deviation is determined, wherein the fifth harmonic relative deviation increases monotonically with the absolute value of the difference between the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference. The characteristic deviation is obtained by comprehensively calculating the relative deviation of active power, the relative deviation of the third harmonic, and the relative deviation of the fifth harmonic.

[0059] In this step, the feature values ​​of three dimensions in the two-factor electrical fingerprint three-dimensional vector—steady-state active power, third harmonic current ratio difference, and fifth harmonic current ratio difference—are first used as real-time inputs for deviation calculation. For example, if an access event is detected in phase A of branch box 1 in transformer area B, the extracted two-factor electrical fingerprint vector is [5.82kW, 16.2%, 8.2%]. From this, steady-state active power = 5.82kW, third harmonic current ratio difference = 16.2%, and fifth harmonic current ratio difference = 8.2%.

[0060] Secondly, from the pre-built and stored slow charging pile feature benchmark data package, read the three-dimensional components corresponding to the benchmark vector: benchmark steady-state active power, benchmark third harmonic current ratio difference, and benchmark fifth harmonic current ratio difference.

[0061] For example, from the slow charging pile characteristic benchmarks stored locally on the monitoring terminal of area B, the benchmark steady-state active power is read as 5.80kW, the benchmark third harmonic current ratio difference is 16.5%, and the benchmark fifth harmonic current ratio difference is 8.0%.

[0062] Next, calculate the relative deviations between the real-time values ​​and the reference values ​​for each of the three dimensions. The definitions of the relative deviations for each dimension include: the active power relative deviation is based on the absolute value of the difference between steady-state active power and the reference steady-state active power, and the active power relative deviation increases monotonically with the absolute value. It can be expressed as a normalized ratio, i.e., active power relative deviation = |steady-state active power - reference steady-state active power| / reference steady-state active power. For example, active power relative deviation = |5.82 - 5.80| / 5.80 ≈ 0.0034.

[0063] The relative deviation of the third harmonic is based on the absolute value of the difference between the proportional difference of the third harmonic current and the proportional difference of the reference third harmonic current | proportional difference of third harmonic current - proportional difference of reference third harmonic current|. The relative deviation of the third harmonic increases monotonically with the absolute value. For example, the relative deviation of the third harmonic = |16.2% - 16.5%| / 16.5% = 0.01818.

[0064] The relative deviation of the fifth harmonic is based on the absolute value of the difference between the proportional difference of the fifth harmonic current and the proportional difference of the reference fifth harmonic current, |Fifth Harmonic Current Proportional Difference - Reference Fifth Harmonic Current Proportional Difference|. The relative deviation of the fifth harmonic increases monotonically with the absolute value. For example, the relative deviation of the fifth harmonic = |8.2% - 8.0%| / 8.0% = 0.02500.

[0065] Finally, the normalized relative deviations of the three dimensions are combined by calculating the sum of squares and then taking the square root to obtain the feature deviation. For example, feature deviation = ≈0.03110, which means the feature deviation is approximately 3.11%.

[0066] After obtaining the feature deviation, when the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile. The steps for determining the dynamic matching deviation threshold include: Within the background window before the access time, extract the active power, third harmonic current content, and fifth harmonic current content of multiple power frequency cycles. The difference between the maximum and minimum values ​​of the active power is calculated as the active background fluctuation amount, the difference between the maximum and minimum values ​​of the third harmonic current content rate is calculated as the third harmonic background fluctuation amount, and the difference between the maximum and minimum values ​​of the fifth harmonic current content rate is calculated as the fifth harmonic background fluctuation amount. The active power tolerance upper limit corresponding to the benchmark steady-state active power, the first tolerance upper limit corresponding to the benchmark third harmonic current ratio difference, and the second tolerance upper limit corresponding to the benchmark fifth harmonic current ratio difference are obtained from the slow charging pile characteristic benchmark. The steps for obtaining the active power tolerance upper limit, the first tolerance upper limit, and the second tolerance upper limit include: In the process of constructing the characteristic benchmark of the slow charging pile, the standard deviation of the steady-state active power sample set after removing outliers is calculated, and the calculated standard deviation is multiplied by the tolerance multiple in the characteristic benchmark of the slow charging pile to obtain the upper limit of the active power tolerance. The standard deviation of the third harmonic current ratio difference sample set after removing outliers is calculated. The calculated standard deviation is multiplied by the tolerance multiple in the slow charging pile characteristic benchmark to obtain the first tolerance upper limit. The standard deviation of the fifth harmonic current ratio difference sample set after removing outliers is calculated, and the calculated standard deviation is multiplied by the tolerance multiple to obtain the second tolerance upper limit; The active power tolerance upper limit, the first tolerance upper limit, and the second tolerance upper limit are associated and stored in the slow charging pile characteristic benchmark; The active power background fluctuation is calculated by performing a root mean square operation on the active power tolerance upper limit to obtain the active power deviation boundary. The third harmonic background fluctuation is calculated by performing a root mean square operation on the first tolerance upper limit to obtain the third harmonic deviation boundary. The fifth harmonic background fluctuation is calculated by performing a root mean square operation on the second tolerance upper limit to obtain the fifth harmonic deviation boundary. The active power deviation boundary, the third harmonic deviation boundary, and the fifth harmonic deviation boundary are comprehensively converged and calculated to obtain the dynamic matching deviation threshold.

[0067] Specifically, within the background window K before the access time, the active power, the third harmonic current content rate, and the fifth harmonic current content rate are extracted cycle by cycle. The maximum and minimum values ​​in each sequence are found, and the difference between the maximum and minimum values ​​is calculated as the background fluctuation amount.

[0068] For example, in phase A of branch box 1 in transformer area B, the connection time is the starting point of cycle T1, and the background window covers 10 cycles from T(-9) to T0. The active power sequence extracted cycle by cycle is [0.48, 0.52, 0.51, 0.49, 0.53, 0.50, 0.47, 0.55, 0.52, 0.50] kW, with a maximum value of 0.55 kW and a minimum value of 0.47 kW. The active power background fluctuation is 0.55 - 0.47 = 0.08 kW. Similarly, the maximum value of the third harmonic current content sequence is 3.1%, and the minimum value is 2.0%. The third harmonic background fluctuation is 3.1% - 2.0% = 1.1%. The fifth harmonic background fluctuation is calculated to be 0.5%.

[0069] Furthermore, the pre-calculated and stored active power tolerance upper limit, first tolerance upper limit, and second tolerance upper limit are read from the slow charging pile characteristic benchmark. The process of constructing the tolerance upper limit is as follows: when the slow charging pile characteristic benchmark is constructed offline, the standard deviation σ is calculated for each dimension of the sample set after removing outliers, and then the standard deviation is multiplied by the preset tolerance multiple.

[0070] The recommended range for the tolerance factor is 2 to 3, with a default value of 3. A value of 3 times the standard deviation means that, assuming the samples follow a normal distribution, approximately 99.7% of the deviations from the fingerprint samples of normal charging stations will fall within this tolerance range.

[0071] For example, when constructing the characteristic benchmark for slow charging piles, the standard deviation of the steady-state active power sample set after removing outliers is 0.15kW. Using a tolerance factor of 3, the upper limit of the active power tolerance is 3 × 0.15 = 0.45kW. The standard deviation of the third harmonic current ratio difference sample set is 1.2%, so the first upper limit of the tolerance is 3 × 1.2 = 3.6%. The standard deviation of the fifth harmonic current ratio difference sample set is 0.8%, so the second upper limit of the tolerance is 3 × 0.8 = 2.4%. These three upper limits of tolerance are associated and stored in the slow charging pile characteristic benchmark, and are retrieved from it during this matching process.

[0072] Furthermore, the obtained background fluctuations in each dimension are compared with the corresponding dimension tolerance upper limits using root mean square (RMS) calculations to obtain the deviation boundaries between the combined offline sample statistical tolerance and online background fluctuations. These boundaries are then normalized by dividing by the baseline value of the corresponding dimension, converting them into dimensionless relative deviation boundaries. For example, after normalization, the active power deviation boundary = / 5.80≈0.0788; Third harmonic deviation boundary = / 16.5≈0.2279; Fifth harmonic deviation boundary = / 8.0≈0.3063.

[0073] Finally, the normalized relative deviation boundaries of the three dimensions are combined and converged using the original sum of squares and square roots to obtain the dynamic matching deviation threshold. For example, the dynamic matching deviation threshold = The value is approximately 0.3898, meaning the dynamic matching deviation threshold is approximately 38.98%. The characteristic deviation is compared to the dynamic matching deviation threshold. If the characteristic deviation is less than the threshold, the connected load is considered a candidate charging pile; if the characteristic deviation is greater than or equal to the threshold, it is excluded. For example, if the characteristic deviation of 3.11% is less than the dynamic matching deviation threshold of 38.98%, the connected load is considered a candidate charging pile.

[0074] The above thresholds are dynamic thresholds. Since the background fluctuation is calculated online in real time at the time of access, when the background load of the transformer area fluctuates drastically, the background fluctuation increases. After the root mean square calculation, the deviation boundaries of each dimension increase accordingly, which automatically widens the matching threshold and avoids the false exclusion of real charging piles due to background disturbances. When the background of the transformer area is stable, the background fluctuation approaches zero, and the matching threshold converges to the basic leniency determined by the offline tolerance upper limit, maintaining the ability to exclude non-charging pile loads.

[0075] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0076] In summary, this step achieves accurate initial screening of charging pile access events by quantitatively calculating the deviation between the real-time extracted two-factor electrical fingerprint and the slow charging pile feature benchmark, and comparing it with the dynamically adjusted matching deviation threshold that adapts to the background fluctuations of the charging station area.

[0077] S300: Obtain the active power difference between the node where the candidate charging pile is located and its adjacent upstream node within the same time window, and confirm the valid charging pile based on the consistency verification result of the active power difference and the steady-state active power.

[0078] This step obtains the change in active power within the same time window at the node where the candidate charging pile is located and its adjacent upstream nodes. The difference in active power between the upstream node and the node where the candidate charging pile is located within this window is calculated and compared with the steady-state active power of the candidate charging pile. If the two are consistent within a certain tolerance range, the candidate charging pile is confirmed as a valid charging pile; if there is a significant deviation, it is judged as a false event and removed.

[0079] Step S300 in the method provided by the present invention includes: Extract the steady-state active power from the two-factor electrical fingerprint; The average active power of the node where the candidate charging pile is located within the background window before the access time is read and used as the node background power. The average active power of the node where the candidate charging pile is located within the steady-state window after the access time is read as the node steady-state power; The node power increment is obtained by subtracting the node background power from the node steady-state power. The average active power of the adjacent upstream nodes within the background window is read as the upstream background power; The average active power of the adjacent upstream nodes within the steady-state window is read as the upstream steady-state power; Subtracting the upstream background power from the upstream steady-state power yields the upstream power increment; Calculate the absolute value of the difference between the upstream power increment and the node power increment, and use it as the active power difference; The ratio of the active power difference to the steady-state active power is calculated as the power conservation deviation rate. Obtain the power conservation error threshold, wherein the power conservation error threshold is a preset ratio of the reference steady-state active power in the characteristic reference of the slow charging pile; When the power conservation deviation rate is less than the power conservation error threshold, it is confirmed that the active power difference is consistent with the steady-state active power, and the candidate charging pile is determined to be a valid charging pile.

[0080] In this step, the steady-state active power of the candidate charging pile is first extracted from the two-factor electrical fingerprint. Simultaneously, the average active power of the node containing the candidate charging pile within the background window K before the access time is read as the node's background power, and the average active power within the steady-state window N after the access time is read as the node's steady-state power.

[0081] For example, in phase A of branch box 1 in transformer substation B, a candidate charging pile was identified. The steady-state active power extracted from the two-factor electrical fingerprint of this event is 5.82 kW. The background window K = 10 cycles, and the node background power is the average active power of the 10 cycles before connection, which is 0.50 kW. The steady-state window N = 10 cycles, and the node steady-state power is the average active power of the steady-state window after connection, which is 6.32 kW.

[0082] Secondly, calculate the node power increment as node steady-state power minus node background power. Simultaneously, read the average active power of adjacent upstream nodes within the same background window as the upstream background power, and the average active power within the same steady-state window as the upstream steady-state power, and calculate the upstream power increment as upstream steady-state power minus upstream background power.

[0083] For example, the adjacent upstream node of branch box No. 1 is the low-voltage side outgoing line of the transformer in the distribution area. The average active power of the upstream node within the same background window T(-9) to T0 is read as 12.30kW, and the average active power within the same steady-state window T6 to T15 is read as 18.15kW. The node power increment is calculated as 6.32 - 0.50 = 5.82kW, and the upstream power increment is calculated as 18.15 - 12.30 = 5.85kW.

[0084] Next, calculate the absolute value of the difference between the upstream power increment and the node power increment as the active power difference, i.e., active power difference = |upstream power increment - node power increment|. The ratio of the active power difference to the steady-state active power is used as the power conservation deviation rate. For example, active power difference = |5.85 - 5.82| = 0.03 kW. Power conservation deviation rate = 0.03 / 5.82 ≈ 0.00515, or approximately 0.52%.

[0085] Finally, the power conservation error threshold is obtained. When the power conservation deviation rate is less than the power conservation error threshold, it is confirmed that the active power difference is consistent with the steady-state active power, and the candidate charging pile is determined to be a valid charging pile.

[0086] The power conservation error threshold is a preset percentage of the benchmark steady-state active power in the slow-charging pile characteristic benchmark. It is used to tolerate power differences between upstream and downstream caused by measurement device accuracy errors, line losses, and minor fluctuations in background load. A value that is too small (below 1%) may lead to the false exclusion of genuine charging piles due to normal measurement noise; a value that is too large (exceeding 5%) may loosen the filtering capability for false events. Therefore, the preset percentage range is recommended to be 1% to 5%, with a default value of 3% being preferred.

[0087] For example, if the baseline steady-state active power is 5.80kW, and a preset ratio of 3% is taken, then the power conservation error threshold is 3%. If the power conservation deviation rate is 0.52% < the power conservation error threshold of 3%, and the active power difference is consistent with the steady-state active power, then the candidate charging pile is confirmed as a valid charging pile.

[0088] In another possible embodiment, the node power increment is 5.80kW, the upstream power increment is 2.10kW, the active power difference is |2.10-5.80|=3.70kW, and the power conservation deviation rate is 3.70 / 5.80≈63.8%>3% of the power conservation error threshold. Therefore, it is determined that the power increment of the node is not caused by the load connection of this node. If it may be caused by the load switching of other downstream branches transmitted to the upstream node through the line, or if the measurement is abnormal, the candidate charging pile is eliminated.

[0089] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0090] In summary, this step achieves secondary verification of the initial screening results and elimination of false events by obtaining the change in active power between the node where the candidate charging pile is located and the adjacent upstream node within the same time window and calculating the difference, and then verifying the consistency of the difference with the steady-state active power.

[0091] S400: Identify the access phase of the valid charging pile, and based on the power conservation relationship between the node and the upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level.

[0092] This step performs phase identification and branch line location for valid charging piles. First, on the three-phase line of the node where the valid charging pile is located, the change in active power of each phase before and after the connection time is obtained. The phase with the largest power increase exceeding the set proportion of other phases is identified as the connection phase of the charging pile. Then, starting from the low-voltage side of the transformer, each branch node is traversed from top to bottom along the transformer topology. Based on the power conservation relationship satisfied by the parent node and each child node within the same time window, the branch path where the valid charging pile is located is locked. When there is only a single branch downstream of a branch node and the power increase matches the steady-state active power of the charging pile, that branch is identified as the specific branch line to which the valid charging pile belongs.

[0093] Step S400 in the method provided by the present invention includes: For the three-phase voltage sampling sequence and three-phase current sampling sequence of the node where the effective charging pile is located, calculate the average active power of each phase in the steady state window after the access time, and subtract the average active power of each phase in the background window before the access time to obtain the power increment of each phase. The phase with the largest power increment among all phases that exceeds the power change threshold in the slow charging pile characteristic benchmark is taken as the access phase of the effective charging pile; Taking the node where the effective charging pile is located as the current node, read the average total active power of the current node within the steady-state window and the average total active power within the background window; The power increment at the current node is obtained by subtracting the average background total active power from the average steady-state total active power. Obtain all direct downstream child nodes corresponding to the current node in the topology connection order of all branch nodes in the transformer area, read the average active power of each direct downstream child node in the steady state window and the average active power in the background window, and calculate the power increment of each child node. When there are two or more direct downstream sub-nodes whose sub-node power increments are all greater than the power change threshold, the difference between the largest sub-node power increment and the second largest sub-node power increment is compared. If the difference is less than the power change threshold, the downstream positioning is terminated and the current node is taken as the branch node to which the effective charging pile belongs. Otherwise, when the absolute value of the difference between the current node power increment and the sum of the power increments of all directly downstream child nodes is less than the power change threshold, the directly downstream child node with the largest power increment is selected as the new current node, the current node power increment is recalculated, and the operation of obtaining all directly downstream child nodes corresponding to the current node is returned. When the absolute value of the difference between the current node's power increment and the sum of the power increments of all directly downstream child nodes is greater than or equal to the power change threshold, the effective charging pile is confirmed to be directly connected to the current node, and the branch line corresponding to the current node is output as the specific branch line to which the effective charging pile belongs.

[0094] In this step, the three-phase voltage sampling sequence and three-phase current sampling sequence of the node where the effective charging pile is located are first used to calculate the average active power of each phase within the steady-state window N after the access time. The average active power of each phase within the background window K before the access time is then subtracted to obtain the power increment of each phase. The phase with the largest power increment that exceeds the power change threshold in the slow charging pile characteristic benchmark is taken as the access phase of the effective charging pile.

[0095] For example, at branch box node 1 in area B, the average active power of each phase of the confirmed valid charging pile within the steady-state window is 6.35kW for phase A, 1.20kW for phase B, and 0.80kW for phase C. Within the background window, the average active power of each phase is 0.50kW for phase A, 1.18kW for phase B, and 0.79kW for phase C. The power increments for each phase are calculated as follows: Phase A power increment = 6.35 - 0.50 = 5.85kW, Phase B power increment = 1.20 - 1.18 = 0.02kW, and Phase C power increment = 0.80 - 0.79 = 0.01kW. Phase A has the largest power increment and exceeds the power change threshold of 2.6kW in the slow charging pile characteristic benchmark; therefore, the charging pile is determined to be connected to phase A.

[0096] Secondly, taking the node where the effective charging pile is located as the current node, read the average total active power of this node within the steady-state window N and the average total active power within the background window K, and calculate the current node power increment = average total active power within the steady-state window N - average total active power within the background window K. Here, the average total active power is the sum of the average active power of each phase.

[0097] For example, the average total active power of branch box node 1 within the steady-state window is 6.35kW + 1.20kW + 0.80kW = 8.35kW, and the average total active power within the background window is 0.50kW + 1.18kW + 0.79kW = 2.47kW. Therefore, the power increment of the current node is 8.35 - 2.47 = 5.88kW.

[0098] Next, based on the topological connection order of all branch nodes in the transformer area, obtain all direct downstream child nodes corresponding to the current node. For each direct downstream child node, read its average active power within the same steady-state window N and the average active power within the background window K, and calculate the power increment of each child node.

[0099] For example, in transformer area B, there are three direct downstream child nodes of branch box node 1: branch box 2, branch box 3, and branch box 4. The average active power of the three child nodes is read within the steady-state window and the background window, and the power increments are calculated as follows: power increment of child node 2 = 0.05kW, power increment of child node 3 = 5.80kW, and power increment of child node 4 = 0.03kW.

[0100] Finally, topology positioning logic is executed based on the power increment distribution of child nodes. When there are two or more directly downstream child nodes whose child node power increments are all greater than the power change threshold, the difference between the largest and second largest child node power increments is compared. If the difference is less than the power change threshold, it means that multiple branches downstream of the current node have loads connected simultaneously, and the branch to which the charging pile belongs cannot be uniquely determined. In this case, downstream positioning is terminated, and the current node is taken as the branch node to which the valid charging pile belongs.

[0101] For example, assuming the power increment of sub-node 3 is 5.80kW, and the power increment of sub-node 2 also suddenly becomes 5.50kW, both exceeding the 2.6kW threshold, and the difference between the two is 0.30kW, which is less than 2.6kW, then downstream positioning stops, and branch box 1 is taken as the belonging branch node.

[0102] If only one sub-node has a significant power increase, or if the difference between the largest and second-largest power increases among multiple sub-nodes exceeds the power change threshold, a power conservation check is initiated: If the absolute value of the difference between the current node's power increase and the sum of the power increases of all directly downstream sub-nodes is less than the power change threshold, it indicates that the charging pile is located in a downstream branch. In this case, the directly downstream sub-node with the largest power increase is selected as the new current node, and the above steps are repeated.

[0103] For example, the power increment of sub-node 2 is 0.05kW, the power increment of sub-node 3 is 5.80kW, and the power increment of sub-node 4 is 0.03kW. The difference between the largest and second largest is 5.80 - 0.05 = 5.75kW, which exceeds the 2.6kW threshold, triggering a power conservation check. The sum of the power increments of the sub-nodes is 0.05 + 5.80 + 0.03 = 5.88kW, and the difference between this and the current node's power increment of 5.88kW is |5.88 - 5.88| = 0kW, which is less than 2.6kW. Therefore, branch box 3 is selected as the new current node, and the above steps are repeated to continue downstream positioning.

[0104] If the absolute value of the difference is greater than or equal to the power change threshold, it means that the power increment of the current node is greater than the sum of the power increments of all known child nodes, and the charging pile is directly connected to the current node instead of the downstream child node. In this case, the branch line corresponding to the current node is output as the specific branch line to which the valid charging pile belongs.

[0105] For example, when tracing to branch box 3 as the current node, its direct downstream sub-nodes are distribution boxes 3-1 and 3-2. The calculated power increment of sub-node 3-1 is 5.78kW, and the power increment of sub-node 3-2 is 0.02kW, with a sum of sub-node power increments of 5.80kW. The current node power increment of branch box 3 is also 5.80kW. The difference |5.80 - 5.80| = 0kW < 2.6kW, so distribution box 3-1 is selected as the new current node. Continuing to trace to distribution box 3-1, there are no downstream sub-nodes, or the power increments of all sub-nodes are zero, with a sum of sub-node power increments of 0kW. The current node power increment is 5.78kW, and the difference |5.78 - 0| = 5.78kW > 2.6kW. Therefore, it is confirmed that the charging pile is directly connected to distribution box 3-1, and distribution box 3-1 is identified as the specific branch line to which this valid charging pile belongs.

[0106] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0107] In summary, this step accurately identifies the connected phase of a valid charging pile by comparing the power increments of each phase before and after the connection time, eliminating the ambiguity of knowing only the node but not the phase, and providing accurate phase identification basis for single-phase overload early warning and three-phase imbalance management.

[0108] S500: Records each identified charging pile and its corresponding access phase and affiliated branch, generating a topology distribution map of charging piles in the area.

[0109] This step involves uniformly recording and archiving each valid charging pile identified and confirmed in the previous steps, along with its access phase and affiliated branch line information, and generating a charging pile topology distribution map for the transformer substation. Specifically, this includes: creating a file entry for each identified charging pile, containing fields such as charging pile identifier, access phase, affiliated branch node name or number, transformer substation name, and identification timestamp; aggregating the file information of all charging piles by transformer substation, combining it with the existing distribution network topology data of the substation, and marking each charging pile at the end of its designated branch line on the topology single-line diagram, distinguishing access points with different colors or line types, ultimately generating a visualized charging pile topology distribution map for the transformer substation.

[0110] Step S500 in the method provided by the present invention includes: Establish a transformer substation topology with the low-voltage side of the transformer substation as the root and each branch node as a subtree, and initialize the charging pile records corresponding to each branch node to be empty; For each load identified as a valid charging pile, obtain the two-factor electrical fingerprint, access phase, and specific branch line corresponding to the valid charging pile; The dual-factor electrical fingerprint is used as the charging pile identifier and stored together with the access phase in the charging pile record of the corresponding node of the specific branch line to which it belongs; When any node records multiple charging pile identifiers, calculate the feature deviation between each pair of the two-factor electrical fingerprints corresponding to each charging pile identifier. If the feature deviation is less than the dynamic matching deviation threshold, retain the charging pile identifier and access phase recorded once and discard duplicate records. Traverse all branch nodes in the transformer substation area, merge the charging pile records of each node with the transformer substation topology, and generate a transformer substation charging pile topology distribution map.

[0111] In this step, a hierarchical tree topology is first established with the low-voltage side of the transformer substation as the root node and each branch node, including branch boxes and distribution boxes, as subtree nodes, based on the known connections of the transformer substation's power distribution network. An empty charging pile record list is created for each branch node to store the identified charging pile information later.

[0112] For example, in transformer substation B, the low-voltage side outgoing line of the transformer is taken as the root node, and branch boxes No. 1, No. 5, etc. are connected to it as first-level child nodes. Branch boxes No. 1 are further connected to branch boxes No. 2, No. 3, and No. 4 as second-level child nodes, and so on, to construct a complete tree topology. The charging pile records of each branch node are initialized as an empty list.

[0113] Secondly, for each valid charging pile that has been confirmed and whose phase identification and branch location have been completed, obtain its corresponding two-factor electrical fingerprint vector, access phase, and the name of the specific branch line node to which it belongs.

[0114] For example, in transformer substation B, the confirmed valid charging pile EV01 has a dual-factor electrical fingerprint of [5.82kW, 16.2%, 8.2%], is connected to phase A, and its specific branch line is located at distribution box 3-1. These three pieces of information are used as the charging pile's identification file.

[0115] Furthermore, using a two-factor electrical fingerprint as the unique identifier for the charging pile, the fingerprint vector and the access phase are stored together in the charging pile record list of the corresponding node of the branch line.

[0116] For example, the two-factor electrical fingerprint [5.82kW, 16.2%, 8.2%] of charging pile EV01 is used as an identifier, along with the connected phase A, and stored in the charging pile record list of distribution box node 3-1. At this time, the charging pile record of distribution box 3-1 is updated from an empty list to contain one record: [Identifier=(5.82kW, 16.2%, 8.2%), Phase=A].

[0117] Next, when multiple charging pile identifiers exist in the charging pile record list of any node, the feature deviation between each pair of the two-factor electrical fingerprints corresponding to each identifier is calculated. If the feature deviation between two fingerprints is less than the dynamic matching deviation threshold, it is determined to be a duplicate record of the same charging pile. The charging pile identifier and access phase of one of the records are retained, and the duplicate record is discarded.

[0118] For example, suppose that a new record is subsequently added to the charging pile record list of distribution box node 3-1: [Identifier=(5.78kW,16.4%,8.1%), Phase=Phase A]. Calculate the characteristic deviation between the two record fingerprints: Active power relative deviation = |5.82-5.78| / 5.80≈0.0069, Third harmonic relative deviation = |16.2-16.4| / 16.5≈0.0121, Fifth harmonic relative deviation = |8.2-8.1| / 8.0≈0.0125. Characteristic deviation = ≈0.0187, or 1.87%. If the dynamic matching deviation threshold is 38.98%, then 1.87% < 38.98%, which is considered a duplicate identification of the same charging pile. The first record is retained, and the subsequent record is discarded.

[0119] In another possible embodiment, if the fingerprint of another record is [3.50kW, 8.0%, 5.0%], the calculated feature deviation may exceed the dynamic matching deviation threshold, then it is determined to be a different charging pile, and both records are retained.

[0120] Finally, iterate through all branch nodes in the transformer substation area and merge the charging pile record list of each node with the established transformer substation topology. For example... Figure 2 As shown, on the topology single-line diagram, each node with a charging pile record is labeled with the corresponding charging pile identifier and the access phase, and each phase is distinguished by different colors or marks to generate a visual topology distribution map of charging piles in the transformer area.

[0121] For example, traversing all nodes in transformer area B, node 3-1 records charging pile EV01, node 3-2 records charging pile EV04 (phase A), node 2 has no record, node 4 records charging pile EV02 (phase B), and node 5 records charging pile EV03 (phase C). This information is merged into the topology to generate a topology diagram: EV01 (phase A) is labeled below node 3-1, EV04 (phase A) below node 3-2, EV02 (phase B) below node 4, and EV03 (phase C) below node 5, with phase A marked in red, phase B in yellow, and phase C in blue. For example... Figure 2 As shown, the final output shows the complete charging pile topology distribution map of area B.

[0122] In this step, after traversing all branch nodes of the transformer area, merging the charging pile records of each node with the transformer area topology, and generating a transformer area charging pile topology distribution map, the following steps are also included: For each recorded valid charging pile in the topology distribution map of the charging pile area, the three-phase voltage signal and three-phase current signal of the node to which the valid charging pile belongs are continuously collected, and the total active power is calculated by sliding window to obtain the real-time total active power sequence. When the decrease in the real-time total active power sequence over multiple consecutive power frequency cycles exceeds the power change threshold in the slow charging pile characteristic benchmark, and the decreased real-time total active power is less than the lower limit of the typical power range of the slow charging pile, a power drop event is marked. After a delay of one start-up transient avoidance window at the time of the power drop event, the average real-time total active power within a continuous power frequency cycle is taken as the steady-state power after the power cut-off. When the steady-state power after the removal is less than the lower limit of the typical power range of the slow charging pile, the removal of the effective charging pile is confirmed. The charging pile identifier and access phase of the effective charging pile are removed from the charging pile record of the node to which it belongs. Based on the removed charging pile record and the topology of the transformer area, a new transformer area charging pile topology distribution map is generated.

[0123] Specifically, for each valid charging pile recorded in the topology distribution map of the charging piles in the distribution area, the three-phase voltage signal and three-phase current signal of the node to which it belongs are continuously collected. The total active power, i.e. the sum of the three-phase active power, is calculated cycle by cycle in a sliding window manner to form a real-time total active power sequence.

[0124] For example, in distribution box B, node 3-1 records charging pile EV01, which is phase A, with a fingerprint of [5.82kW, 16.2%, 8.2%]. Three-phase voltage and current are continuously collected for this node, and the analysis window slides every 20ms to continuously calculate and output the total active power value, forming a real-time total active power sequence.

[0125] Secondly, the total active power sequence is monitored in real time. When the decrease in the sequence over multiple consecutive power frequency cycles exceeds the power change threshold in the characteristic benchmark of the slow charging pile, and the real-time total active power after the decrease is less than the lower limit of the typical power range of the slow charging pile, the moment when the decrease begins is marked as the power drop event moment.

[0126] The number of consecutive power frequency cycles is consistent with the access detection, taking P power frequency cycles, with a recommended default value of P=3. The power change threshold is taken from the threshold stored in the slow charging pile characteristic benchmark, for example, 2.6kW. The lower limit of the typical power range of a slow charging pile is, for example, 3.0kW. If the power decreases below this lower limit, it indicates that the charging pile may have stopped charging, rather than a normal fluctuation in charging power.

[0127] For example, the real-time total active power sequence of distribution box node 3-1 decreased from 6.30kW to 0.45kW in three consecutive cycles. The reduction of 5.85kW exceeded the 2.6kW threshold, and the reduced 0.45kW was less than the lower limit of the typical power range of 3.0kW. The starting time of the decrease was marked as the power drop event.

[0128] Secondly, after a delay of one transient avoidance window M at the power drop event marker, the average real-time total active power over N consecutive power frequency cycles is taken as the steady-state power after the power cut-off. The transient avoidance window M is taken as the number of transient avoidance cycles stored in the slow charging pile characteristic benchmark, with a recommended default value of M=5. The steady-state calculation window N is taken as the recommended default value of N=10.

[0129] For example, skip M=5 transient avoidance cycles after the power drop event, and take the active power data for 10 cycles from the power drop event time + 6 cycles to the power drop event time + 15 cycles. Calculate the average value to obtain a steady-state power of 0.45kW after the cutoff.

[0130] Finally, when the steady-state power after disconnection is less than the lower limit of the typical power range of the slow charging pile, the effective charging pile is confirmed to have been disconnected. The charging pile identifier and access phase of the charging pile are removed from the charging pile record list of the node to which it belongs. Based on the removed charging pile records and the transformer substation topology, a new charging pile topology distribution map for the transformer substation is generated.

[0131] For example, if the steady-state power after disconnection is 0.45kW < 3.0kW, it confirms that charging pile EV01 has been disconnected. Remove the entry for EV01 from the charging pile record list of distribution box node 3-1. Re-traverse all branch nodes of transformer area B, merge the updated charging pile records and topology structure, and generate a new transformer area charging pile topology distribution map. At this time, EV01 will no longer be marked below distribution box 3-1. If there are no other charging pile records at this node, it will be restored to an empty state.

[0132] It should be noted that the above values ​​are for illustrative purposes only and do not constitute a limitation on the present invention.

[0133] In summary, this step ensures the uniqueness and accuracy of the charging pile files by structurally archiving the two-factor electrical fingerprint, access phase, and affiliated branch line of each identified charging pile, and automatically deduplicating duplicate records of the same node based on fingerprint similarity comparison.

[0134] In summary, this invention achieves fully automatic and accurate identification of the charging pile access location and phase in the transformer substation through a complete technical route including two-factor electrical fingerprint extraction, multi-dimensional feature matching for initial screening, power conservation verification, phase and branch hierarchical positioning, and topology graph generation and dynamic maintenance.

[0135] Example 2, as Figure 3 As shown, this invention provides a topology identification system for charging piles in a transformer substation based on two-factor electrical fingerprinting. The system includes: The two-factor fingerprint extraction module 11 is used to synchronously collect voltage and current waveforms on the low-voltage side of the transformer and at each branch node, and extract the steady-state active power and the proportion of specific harmonic current at the connection time to form a two-factor electrical fingerprint.

[0136] Among them, voltage and current waveforms are synchronously collected on the low-voltage side of the transformer and at each branch node, and the steady-state active power and the proportion of specific harmonic currents at the time of connection are extracted to form a two-factor electrical fingerprint, including: Voltage signals are acquired by voltage transformers connected in parallel to the low-voltage side of the transformer and each branch node, and current signals at the same node are acquired by through-type current transformers. The voltage and current signals are three-phase voltage and three-phase current signals. After performing anti-aliasing low-pass filtering on the collected voltage and current signals, synchronous analog-to-digital conversion is performed to obtain the three-phase voltage sampling sequence and the three-phase current sampling sequence. The power frequency zero-crossing phase point of the three-phase voltage sampling sequence is detected and used as the starting boundary to perform full-cycle truncation and alignment of the three-phase current sampling sequence to generate a current waveform; Real-time sliding window Fourier analysis is performed on the synchronously acquired voltage and current waveforms to continuously output the active power sequence and the third and fifth harmonic current content sequences. When the increment of the active power sequence in multiple consecutive power frequency cycles exceeds the power change threshold in the characteristic benchmark of the slow charging pile, and the active power after the increment is stable in the typical power range of the slow charging pile, the start time of the increment is marked as the access time. The average active power over N consecutive power frequency cycles after the access time is delayed by one start-up transient avoidance window is taken as the steady-state active power, wherein the start-up transient avoidance window is M power frequency cycles, and N and M are both positive integers; The third harmonic current content rate and the fifth harmonic current content rate in the steady-state window after the access time are respectively subtracted from the corresponding harmonic current content rate in the background window before the access time to obtain the third harmonic current ratio difference and the fifth harmonic current ratio difference. The steady-state window is N consecutive power frequency cycles after the access time and after the start transient avoidance window is delayed, and the background window is K consecutive power frequency cycles before the access time, where K is a positive integer. The steady-state active power, the ratio difference of the third harmonic current, and the ratio difference of the fifth harmonic current are combined into a three-dimensional vector, which serves as the two-factor electrical fingerprint.

[0137] The characteristic benchmark of the slow charging pile is obtained by repeatedly starting charging and synchronously collecting the electrical quantities of the nodes for various mainstream models of slow charging piles under the rated voltage fluctuation range of ±10% and the charging power corresponding to different battery states of charge. The steady-state active power, the ratio difference of the third harmonic current and the ratio difference of the fifth harmonic current extracted each time are used as fingerprint samples. After removing outliers, the mean of each dimension is calculated to obtain the benchmark steady-state active power, the benchmark ratio difference of the third harmonic current and the benchmark ratio difference of the fifth harmonic current. These are stored together with the power change threshold, the typical power range of the slow charging pile and the number of transient avoidance cycles.

[0138] The candidate charging pile determination module 12 is used to calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile.

[0139] The calculation of the feature deviation between the two-factor electrical fingerprint and the slow charging pile feature benchmark includes: Extract the steady-state active power, the third harmonic current ratio difference, and the fifth harmonic current ratio difference from the two-factor electrical fingerprint; Extract the benchmark steady-state active power, the benchmark third harmonic current ratio difference, and the benchmark fifth harmonic current ratio difference from the aforementioned slow charging pile characteristic benchmarks; Based on the steady-state active power and the reference steady-state active power, the relative deviation of active power is determined, wherein the relative deviation of active power increases monotonically with the absolute value of the difference between the steady-state active power and the reference steady-state active power. Based on the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current, the relative deviation of the third harmonic is determined, wherein the relative deviation of the third harmonic increases monotonically with the absolute value of the difference between the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current. Based on the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference, the fifth harmonic relative deviation is determined, wherein the fifth harmonic relative deviation increases monotonically with the absolute value of the difference between the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference. The characteristic deviation is obtained by comprehensively calculating the relative deviation of active power, the relative deviation of the third harmonic, and the relative deviation of the fifth harmonic.

[0140] The calculation of the feature deviation between the two-factor electrical fingerprint and the slow charging pile feature benchmark includes: Extract the steady-state active power, the third harmonic current ratio difference, and the fifth harmonic current ratio difference from the two-factor electrical fingerprint; Extract the benchmark steady-state active power, the benchmark third harmonic current ratio difference, and the benchmark fifth harmonic current ratio difference from the aforementioned slow charging pile characteristic benchmarks; Based on the steady-state active power and the reference steady-state active power, the relative deviation of active power is determined, wherein the relative deviation of active power increases monotonically with the absolute value of the difference between the steady-state active power and the reference steady-state active power. Based on the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current, the relative deviation of the third harmonic is determined, wherein the relative deviation of the third harmonic increases monotonically with the absolute value of the difference between the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current. Based on the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference, the fifth harmonic relative deviation is determined, wherein the fifth harmonic relative deviation increases monotonically with the absolute value of the difference between the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference. The characteristic deviation is obtained by comprehensively calculating the relative deviation of active power, the relative deviation of the third harmonic, and the relative deviation of the fifth harmonic.

[0141] The active power consistency verification module 13 is used to obtain the active power difference between the node where the candidate charging pile is located and the adjacent upstream node within the same time window, and to confirm the valid charging pile based on the consistency verification result of the active power difference and the steady-state active power.

[0142] Specifically, the active power difference between the node where the candidate charging pile is located and its adjacent upstream node within the same time window is obtained. Based on the consistency verification result between the active power difference and the steady-state active power, valid charging piles are confirmed, including: Extract the steady-state active power from the two-factor electrical fingerprint; The average active power of the node where the candidate charging pile is located within the background window before the access time is read and used as the node background power. The average active power of the node where the candidate charging pile is located within the steady-state window after the access time is read as the node steady-state power; The node power increment is obtained by subtracting the node background power from the node steady-state power. The average active power of the adjacent upstream nodes within the background window is read as the upstream background power; The average active power of the adjacent upstream nodes within the steady-state window is read as the upstream steady-state power; Subtracting the upstream background power from the upstream steady-state power yields the upstream power increment; Calculate the absolute value of the difference between the upstream power increment and the node power increment, and use it as the active power difference; The ratio of the active power difference to the steady-state active power is calculated as the power conservation deviation rate. Obtain the power conservation error threshold, wherein the power conservation error threshold is a preset ratio of the reference steady-state active power in the characteristic reference of the slow charging pile; When the power conservation deviation rate is less than the power conservation error threshold, it is confirmed that the active power difference is consistent with the steady-state active power, and the candidate charging pile is determined to be a valid charging pile.

[0143] The phase and branch positioning module 14 is used to identify the access phase of the valid charging pile, and based on the power conservation relationship between the node and the upstream and downstream nodes, combined with the topological connection sequence of each branch node, to locate the specific branch line to which the valid charging pile belongs level by level.

[0144] This includes identifying the access phase of the valid charging pile, and based on the power conservation relationship between the node and its upstream and downstream nodes, combined with the topological connection sequence of each branch node, locating the specific branch line to which the valid charging pile belongs level by level, including: For the three-phase voltage sampling sequence and three-phase current sampling sequence of the node where the effective charging pile is located, calculate the average active power of each phase in the steady state window after the access time, and subtract the average active power of each phase in the background window before the access time to obtain the power increment of each phase. The phase with the largest power increment among all phases that exceeds the power change threshold in the slow charging pile characteristic benchmark is taken as the access phase of the effective charging pile; Taking the node where the effective charging pile is located as the current node, read the average total active power of the current node within the steady-state window and the average total active power within the background window; The power increment at the current node is obtained by subtracting the average background total active power from the average steady-state total active power. Obtain all direct downstream child nodes corresponding to the current node in the topology connection order of all branch nodes in the transformer area, read the average active power of each direct downstream child node in the steady state window and the average active power in the background window, and calculate the power increment of each child node. When there are two or more direct downstream sub-nodes whose sub-node power increments are all greater than the power change threshold, the difference between the largest sub-node power increment and the second largest sub-node power increment is compared. If the difference is less than the power change threshold, the downstream positioning is terminated and the current node is taken as the branch node to which the effective charging pile belongs. Otherwise, when the absolute value of the difference between the current node power increment and the sum of the power increments of all directly downstream child nodes is less than the power change threshold, the directly downstream child node with the largest power increment is selected as the new current node, the current node power increment is recalculated, and the operation of obtaining all directly downstream child nodes corresponding to the current node is returned. When the absolute value of the difference between the current node's power increment and the sum of the power increments of all directly downstream child nodes is greater than or equal to the power change threshold, the effective charging pile is confirmed to be directly connected to the current node, and the branch line corresponding to the current node is output as the specific branch line to which the effective charging pile belongs.

[0145] The topology distribution map generation module 15 is used to record each identified charging pile and its corresponding access phase and affiliated branch to generate a topology distribution map of the charging piles in the transformer area.

[0146] This involves recording each identified charging pile, its corresponding access phase, and its affiliated branch to generate a charging pile topology distribution map for the area, including: Establish a transformer substation topology with the low-voltage side of the transformer substation as the root and each branch node as a subtree, and initialize the charging pile records corresponding to each branch node to be empty; For each load identified as a valid charging pile, obtain the two-factor electrical fingerprint, access phase, and specific branch line corresponding to the valid charging pile; The dual-factor electrical fingerprint is used as the charging pile identifier and stored together with the access phase in the charging pile record of the corresponding node of the specific branch line to which it belongs; When any node records multiple charging pile identifiers, calculate the feature deviation between each pair of the two-factor electrical fingerprints corresponding to each charging pile identifier. If the feature deviation is less than the dynamic matching deviation threshold, retain the charging pile identifier and access phase recorded once and discard duplicate records. Traverse all branch nodes in the transformer substation area, merge the charging pile records of each node with the transformer substation topology, and generate a transformer substation charging pile topology distribution map.

[0147] This process, which involves traversing all branch nodes of the transformer substation, merging the charging pile records of each node with the substation topology to generate a charging pile topology distribution map of the substation, also includes: For each recorded valid charging pile in the topology distribution map of the charging pile area, the three-phase voltage signal and three-phase current signal of the node to which the valid charging pile belongs are continuously collected, and the total active power is calculated by sliding window to obtain the real-time total active power sequence. When the decrease in the real-time total active power sequence over multiple consecutive power frequency cycles exceeds the power change threshold in the slow charging pile characteristic benchmark, and the decreased real-time total active power is less than the lower limit of the typical power range of the slow charging pile, a power drop event is marked. After a delay of one start-up transient avoidance window at the time of the power drop event, the average real-time total active power within a continuous power frequency cycle is taken as the steady-state power after the power cut-off. When the steady-state power after the removal is less than the lower limit of the typical power range of the slow charging pile, the removal of the effective charging pile is confirmed. The charging pile identifier and access phase of the effective charging pile are removed from the charging pile record of the node to which it belongs. Based on the removed charging pile record and the topology of the transformer area, a new transformer area charging pile topology distribution map is generated.

[0148] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0149] This specification and accompanying drawings are merely illustrative examples of the invention and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of the invention. Clearly, those skilled in the art can make various alterations and modifications to the invention without departing from its scope. Therefore, if such modifications and modifications fall within the scope of the invention and its equivalents, the invention is intended to include these modifications and modifications.

Claims

1. A method for topological identification of charging piles in a transformer substation based on two-factor electrical fingerprinting, characterized in that, The method includes: Voltage and current waveforms are simultaneously collected on the low-voltage side of the transformer and at each branch node in the distribution area, and the steady-state active power and the proportion of specific harmonic current at the connection time are extracted to form a two-factor electrical fingerprint. Calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile. Obtain the active power difference between the node where the candidate charging pile is located and its adjacent upstream node within the same time window. Based on the consistency verification result between the active power difference and the steady-state active power, confirm the valid charging pile. Identify the access phase of the valid charging pile, and based on the power conservation relationship between the node and its upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level. Record each identified charging pile, its corresponding access phase, and its affiliated branch to generate a topology distribution map of charging piles in the area.

2. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 1, characterized in that, Voltage and current waveforms are simultaneously collected on the low-voltage side of the transformer and at each branch node in the distribution area. The steady-state active power and the proportion of specific harmonic currents at the moment of connection are extracted to form a two-factor electrical fingerprint, including: Voltage signals are acquired by voltage transformers connected in parallel to the low-voltage side of the transformer and each branch node, and current signals at the same node are acquired by through-type current transformers. The voltage and current signals are three-phase voltage and three-phase current signals. After performing anti-aliasing low-pass filtering on the collected voltage and current signals, synchronous analog-to-digital conversion is performed to obtain the three-phase voltage sampling sequence and the three-phase current sampling sequence. The power frequency zero-crossing phase point of the three-phase voltage sampling sequence is detected and used as the starting boundary to perform full-cycle truncation and alignment of the three-phase current sampling sequence to generate a current waveform; Real-time sliding window Fourier analysis is performed on the synchronously acquired voltage and current waveforms to continuously output the active power sequence and the third and fifth harmonic current content sequences. When the increment of the active power sequence in multiple consecutive power frequency cycles exceeds the power change threshold in the characteristic benchmark of the slow charging pile, and the active power after the increment is stable in the typical power range of the slow charging pile, the start time of the increment is marked as the access time. The average active power over N consecutive power frequency cycles after the access time is delayed by one start-up transient avoidance window is taken as the steady-state active power, wherein the start-up transient avoidance window is M power frequency cycles, and N and M are both positive integers; The third harmonic current content rate and the fifth harmonic current content rate in the steady-state window after the access time are respectively subtracted from the corresponding harmonic current content rate in the background window before the access time to obtain the third harmonic current ratio difference and the fifth harmonic current ratio difference. The steady-state window is N consecutive power frequency cycles after the access time and after the start transient avoidance window is delayed, and the background window is K consecutive power frequency cycles before the access time, where K is a positive integer. The steady-state active power, the ratio difference of the third harmonic current, and the ratio difference of the fifth harmonic current are combined into a three-dimensional vector, which serves as the two-factor electrical fingerprint.

3. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 2, characterized in that, The characteristic benchmark of the slow charging pile is obtained by repeatedly starting charging and synchronously collecting node electrical quantities of various mainstream slow charging piles under the rated voltage fluctuation range of ±10% and the charging power corresponding to different battery states of charge. The steady-state active power, the third harmonic current ratio difference and the fifth harmonic current ratio difference extracted each time are used as fingerprint samples. After removing outlier samples, the mean of each dimension is calculated to obtain the benchmark steady-state active power, benchmark third harmonic current ratio difference and benchmark fifth harmonic current ratio difference. These are stored together with the power change threshold, the typical power range of the slow charging pile and the number of transient avoidance cycles.

4. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 1, characterized in that, Calculating the feature deviation between the two-factor electrical fingerprint and the slow charging pile feature benchmark includes: Extract the steady-state active power, the third harmonic current ratio difference, and the fifth harmonic current ratio difference from the two-factor electrical fingerprint; Extract the benchmark steady-state active power, the benchmark third harmonic current ratio difference, and the benchmark fifth harmonic current ratio difference from the aforementioned slow charging pile characteristic benchmarks; Based on the steady-state active power and the reference steady-state active power, the relative deviation of active power is determined, wherein the relative deviation of active power increases monotonically with the absolute value of the difference between the steady-state active power and the reference steady-state active power. Based on the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current, the relative deviation of the third harmonic is determined, wherein the relative deviation of the third harmonic increases monotonically with the absolute value of the difference between the ratio difference of the third harmonic current and the ratio difference of the reference third harmonic current. Based on the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference, the fifth harmonic relative deviation is determined, wherein the fifth harmonic relative deviation increases monotonically with the absolute value of the difference between the fifth harmonic current ratio difference and the reference fifth harmonic current ratio difference. The characteristic deviation is obtained by comprehensively calculating the relative deviation of active power, the relative deviation of the third harmonic, and the relative deviation of the fifth harmonic.

5. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 1, characterized in that, The steps for determining the dynamic matching deviation threshold include: Within the background window before the access time, extract the active power, third harmonic current content, and fifth harmonic current content of multiple power frequency cycles. The difference between the maximum and minimum values ​​of the active power is calculated as the active background fluctuation amount, the difference between the maximum and minimum values ​​of the third harmonic current content rate is calculated as the third harmonic background fluctuation amount, and the difference between the maximum and minimum values ​​of the fifth harmonic current content rate is calculated as the fifth harmonic background fluctuation amount. The active power tolerance upper limit corresponding to the benchmark steady-state active power, the first tolerance upper limit corresponding to the benchmark third harmonic current ratio difference, and the second tolerance upper limit corresponding to the benchmark fifth harmonic current ratio difference are obtained from the slow charging pile characteristic benchmark. The steps for obtaining the active power tolerance upper limit, the first tolerance upper limit, and the second tolerance upper limit include: In the process of constructing the characteristic benchmark of the slow charging pile, the standard deviation of the steady-state active power sample set after removing outliers is calculated, and the calculated standard deviation is multiplied by the tolerance multiple in the characteristic benchmark of the slow charging pile to obtain the upper limit of the active power tolerance. The standard deviation of the third harmonic current ratio difference sample set after removing outliers is calculated. The calculated standard deviation is multiplied by the tolerance multiple in the slow charging pile characteristic benchmark to obtain the first tolerance upper limit. The standard deviation of the fifth harmonic current ratio difference sample set after removing outliers is calculated, and the calculated standard deviation is multiplied by the tolerance multiple to obtain the second tolerance upper limit; The active power tolerance upper limit, the first tolerance upper limit, and the second tolerance upper limit are associated and stored in the slow charging pile characteristic benchmark; The active power background fluctuation is calculated by performing a root mean square operation on the active power tolerance upper limit to obtain the active power deviation boundary. The third harmonic background fluctuation is calculated by performing a root mean square operation on the first tolerance upper limit to obtain the third harmonic deviation boundary. The fifth harmonic background fluctuation is calculated by performing a root mean square operation on the second tolerance upper limit to obtain the fifth harmonic deviation boundary. The active power deviation boundary, the third harmonic deviation boundary, and the fifth harmonic deviation boundary are comprehensively converged and calculated to obtain the dynamic matching deviation threshold.

6. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprints according to claim 1, characterized in that, Obtain the active power difference between the node where the candidate charging pile is located and its adjacent upstream node within the same time window. Based on the consistency verification result of the active power difference and the steady-state active power, confirm the valid charging pile, including: Extract the steady-state active power from the two-factor electrical fingerprint; The average active power of the node where the candidate charging pile is located within the background window before the access time is read and used as the node background power. The average active power of the node where the candidate charging pile is located within the steady-state window after the access time is read as the node steady-state power; The node power increment is obtained by subtracting the node background power from the node steady-state power. The average active power of the adjacent upstream nodes within the background window is read as the upstream background power; The average active power of the adjacent upstream nodes within the steady-state window is read as the upstream steady-state power; Subtracting the upstream background power from the upstream steady-state power yields the upstream power increment; Calculate the absolute value of the difference between the upstream power increment and the node power increment, and use it as the active power difference; The ratio of the active power difference to the steady-state active power is calculated as the power conservation deviation rate. Obtain the power conservation error threshold, wherein the power conservation error threshold is a preset ratio of the reference steady-state active power in the characteristic reference of the slow charging pile; When the power conservation deviation rate is less than the power conservation error threshold, it is confirmed that the active power difference is consistent with the steady-state active power, and the candidate charging pile is determined to be a valid charging pile.

7. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 1, characterized in that, Identify the access phase of the valid charging pile, and based on the power conservation relationship between the node and its upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level, including: For the three-phase voltage sampling sequence and three-phase current sampling sequence of the node where the effective charging pile is located, calculate the average active power of each phase in the steady state window after the access time, and subtract the average active power of each phase in the background window before the access time to obtain the power increment of each phase. The phase with the largest power increment among all phases that exceeds the power change threshold in the slow charging pile characteristic benchmark is taken as the access phase of the effective charging pile; Taking the node where the effective charging pile is located as the current node, read the average total active power of the current node within the steady-state window and the average total active power within the background window; The power increment at the current node is obtained by subtracting the average background total active power from the average steady-state total active power. Obtain all direct downstream child nodes corresponding to the current node in the topology connection order of all branch nodes in the transformer area, read the average active power of each direct downstream child node in the steady state window and the average active power in the background window, and calculate the power increment of each child node. When there are two or more direct downstream sub-nodes whose sub-node power increments are all greater than the power change threshold, the difference between the largest sub-node power increment and the second largest sub-node power increment is compared. If the difference is less than the power change threshold, the downstream positioning is terminated and the current node is taken as the branch node to which the effective charging pile belongs. Otherwise, when the absolute value of the difference between the current node power increment and the sum of the power increments of all directly downstream child nodes is less than the power change threshold, the directly downstream child node with the largest power increment is selected as the new current node, the current node power increment is recalculated, and the operation of obtaining all directly downstream child nodes corresponding to the current node is returned. When the absolute value of the difference between the current node's power increment and the sum of the power increments of all directly downstream child nodes is greater than or equal to the power change threshold, the effective charging pile is confirmed to be directly connected to the current node, and the branch line corresponding to the current node is output as the specific branch line to which the effective charging pile belongs.

8. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 1, characterized in that, Record each identified charging pile, its corresponding access phase, and its affiliated branch to generate a charging pile topology distribution map for the area, including: Establish a transformer substation topology with the low-voltage side of the transformer substation as the root and each branch node as a subtree, and initialize the charging pile records corresponding to each branch node to be empty; For each load identified as a valid charging pile, obtain the two-factor electrical fingerprint, access phase, and specific branch line corresponding to the valid charging pile; The dual-factor electrical fingerprint is used as the charging pile identifier and stored together with the access phase in the charging pile record of the corresponding node of the specific branch line to which it belongs; When any node records multiple charging pile identifiers, calculate the feature deviation between each pair of the two-factor electrical fingerprints corresponding to each charging pile identifier. If the feature deviation is less than the dynamic matching deviation threshold, retain the charging pile identifier and access phase recorded once and discard duplicate records. Traverse all branch nodes in the transformer substation area, merge the charging pile records of each node with the transformer substation topology, and generate a transformer substation charging pile topology distribution map.

9. The method for topological identification of charging piles in a distribution area based on two-factor electrical fingerprinting according to claim 8, characterized in that, After traversing all branch nodes of the transformer substation, merging the charging pile records of each node with the transformer substation topology to generate a transformer substation charging pile topology distribution map, the process also includes: For each recorded valid charging pile in the topology distribution map of the charging pile area, the three-phase voltage signal and three-phase current signal of the node to which the valid charging pile belongs are continuously collected, and the total active power is calculated by sliding window to obtain the real-time total active power sequence. When the decrease in the real-time total active power sequence over multiple consecutive power frequency cycles exceeds the power change threshold in the slow charging pile characteristic benchmark, and the decreased real-time total active power is less than the lower limit of the typical power range of the slow charging pile, a power drop event is marked. After a delay of one start-up transient avoidance window at the time of the power drop event, the average real-time total active power within a continuous power frequency cycle is taken as the steady-state power after the power cut-off. When the steady-state power after the removal is less than the lower limit of the typical power range of the slow charging pile, the removal of the effective charging pile is confirmed. The charging pile identifier and access phase of the effective charging pile are removed from the charging pile record of the node to which it belongs. Based on the removed charging pile record and the topology of the transformer area, a new transformer area charging pile topology distribution map is generated.

10. A topology identification system for charging piles in a transformer substation based on two-factor electrical fingerprinting, characterized in that, The system is used to execute the two-factor electrical fingerprint-based charging pile topology identification method according to any one of claims 1 to 9, the system comprising: The two-factor fingerprint extraction module is used to synchronously collect voltage and current waveforms on the low-voltage side of the transformer and at each branch node, and extract the steady-state active power and the proportion of specific harmonic current at the connection time to form a two-factor electrical fingerprint. The candidate charging pile determination module is used to calculate the feature deviation between the dual-factor electrical fingerprint and the slow charging pile feature benchmark. When the feature deviation is less than the dynamic matching deviation threshold, the load corresponding to the access time is determined as a candidate charging pile. The active power consistency verification module is used to obtain the active power difference between the node where the candidate charging pile is located and the adjacent upstream node within the same time window, and to confirm the valid charging pile based on the consistency verification result of the active power difference and the steady-state active power. The phase and branch positioning module is used to identify the access phase of the valid charging pile and, based on the power conservation relationship between the node and the upstream and downstream nodes, combined with the topological connection sequence of each branch node, locate the specific branch line to which the valid charging pile belongs level by level. The topology map generation module is used to record each identified charging pile and its corresponding access phase and affiliated branch, and generate a topology map of the charging piles in the area.

Images