Method for checking relationship between electric energy metering box and transformer substation based on voltage curve matching

By extracting high-frequency voltage fingerprints from the electricity metering box and combining them with active coding ripple matching, the problems of misjudgment and crosstalk in the identification of household transformer relationships in low-voltage distribution networks are solved, achieving high-precision and low-cost topology verification.

CN122859946BActive Publication Date: 2026-11-10HENAN YUHE ELECTRIC CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202611342905.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-09-01
Publication Date
2026-11-10
Estimated Expiration
2046-09-01

AI Technical Summary

Technical Problem

Existing methods for identifying the relationship between households and transformers in low-voltage distribution networks fail when the macroscopic voltage curves overlap in adjacent transformer areas with similar structures and moderate loads. Furthermore, crosstalk between power line carriers across transformer areas leads to data corruption. Traditional active signal injection methods are costly and negatively impact power quality.

Method used

Using the electricity meter box as the smallest verification unit, the high-frequency fingerprint voltage curve at the zero-crossing point of the power frequency voltage is extracted. Combined with passive waveform similarity and active coding ripple voltage curve matching, pseudo-random binary coding modulation is performed using the built-in switching power supply of the meter to achieve coherent demodulation and multi-dimensional decision-making of the high-frequency fingerprint and establish physical affiliation.

Benefits of technology

It effectively distinguishes between homogeneous impedance and power consumption behavior, resists high-frequency crosstalk, reduces engineering implementation costs, protects power quality, and achieves high-precision topology verification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122859946B_ABST
    Figure CN122859946B_ABST
Patent Text Reader

Abstract

The application discloses a kind of electric energy metering box house relation verification method based on voltage curve matching, it is related to electric energy metering box technical field.Pass through intercepting ac voltage power frequency zero-crossing point fragment extraction high-frequency micro-distortion characteristics, generate high-frequency fingerprint voltage curve and carry out normality similarity matching;When passive matching result falls into impedance convergence fuzzy uncertain interval, dynamic trigger active coding matching mechanism;By issuing instruction control the switch power module built-in in specific meter in metering box, execute high-speed load switching based on pseudo-random binary sequence, under the premise of not affecting power grid total harmonic distortion rate, weak active coding physical ripple is excited;Active coherence coefficient is obtained by cooperating digital coherent demodulation calculation, and carries out joint decision.The present application overcomes the traditional macroscopic voltage matching failure and the logic file misjudgment problem caused by carrier crosstalk, realizes the high confidence of station area topology in limit working condition under the condition of no external hardware cost.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of electricity metering box technology, and more specifically to a method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching. Background Technology

[0002] Low-voltage distribution area topology files, which map the physical connections between distribution transformers and the electricity meters of the users they supply in a low-voltage distribution network, serve as the data foundation for advanced application analysis in modern smart grids and the Internet of Things (IoT) for distribution networks. Operations such as synchronous line loss analysis, distributed photovoltaic (PV) absorption assessment, three-phase imbalance management, and low-voltage fault diagnosis all heavily rely on 100% accurate transformer-user relationship files. However, due to historical construction reasons, frequent equipment changes, and distribution area load cutovers, the logical files in the electricity marketing system often deviate from the actual physical connections on-site.

[0003] Currently, the widely used automatic verification technology for user-transformer relationships in the industry mainly relies on a passive macroscopic voltage curve similarity matching method. Its basic principle is as follows: Using distribution area concentrators and smart meters, the effective voltage values ​​of each node are collected periodically (typically with a time resolution of 15 minutes or 5 minutes), forming a macroscopic steady-state voltage time series curve containing tens to hundreds of data points. Then, algorithms such as Pearson correlation coefficient, Euclidean distance, or dynamic time warping (DTW) are used to calculate the similarity between the voltage curve of the user under test and the low-voltage bus voltage curves of each candidate transformer. The distribution area with the highest similarity is determined to belong to that user.

[0004] While the aforementioned methods have achieved some success in most conventional scenarios, they still exhibit significant fundamental flaws when facing complex distribution network conditions. First, in scenarios such as high-density urban residential areas or large industrial parks, multiple distribution transformers operate in parallel, and their lines are intertwined. User electricity consumption behavior in these areas is highly homogeneous (e.g., simultaneous increases in nighttime lighting and air conditioning loads). According to basic circuit laws, under the premise that the voltage sources of the upstream power grid are of the same frequency and phase, the impedance of the power supply lines in each distribution area is similar, and the load fluctuation trend is consistent, the macroscopic voltage RMS curves of different users under adjacent distribution areas exhibit severe geometric overlap. Extensive engineering measurements show that the Pearson correlation coefficient of voltage curves in different distribution areas under such conditions often ranges from 0.97 to 0.99, causing matching algorithms based on macroscopic steady-state characteristics to fall into recognition dead zones, significantly increasing the misjudgment rate. Second, current low-voltage distribution networks widely employ broadband power line carrier (HPLC) technology for data communication. HPLC signals exhibit strong electromagnetic coupling and cross-distribution crosstalk within shared trench cables or dense distribution room spaces. Because HPLC employs a complex routing and networking protocol, crosstalk can cause meters in transformer area A to establish stable communication links with concentrators in transformer area B, creating logical attribution at the application layer. Traditional verification systems struggle to distinguish between this purely protocol-level logical connectivity and the actual conductor-level physical connectivity, leading to cross-referencing and confusion in transformer area records. To address this issue, some existing technologies propose active signal injection methods. This involves installing dedicated signal generators at the low-voltage outgoing terminals or branch boxes of distribution transformers to forcibly inject non-power frequency interharmonic currents (such as dual-frequency signals of a specific frequency) or broadband high-frequency pulses into the power grid. While this method improves identification accuracy, it is costly to implement, requiring significant additional hardware investment. Furthermore, the continuous injection of high-power signals can increase the total harmonic distortion (THD) of the local power grid, negatively impacting power quality and hindering the implementation of routine, fully automated deployment.

[0005] Therefore, it is necessary to propose a method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching to solve the above problems. Summary of the Invention

[0006] The purpose of this invention is to solve the problem of macroscopic voltage curve overlap failure caused by the similar structure and smooth load of adjacent transformer substations in the existing low-voltage distribution network household transformer relationship identification method, as well as the problem of file disorder caused by cross-transformer interference of power line carrier.

[0007] To achieve the above objectives, the present invention specifically adopts the following technical solution:

[0008] This invention provides a method for verifying the relationship between a power metering box and a transformer based on voltage curve matching. The method uses the power metering box as the smallest verification unit and includes the following steps:

[0009] Step S100: Obtain the instantaneous waveforms of the power frequency voltage at the reference point of the low-voltage side of the low-voltage transformer and at each user's meter in the power metering box of the user to be verified, and accurately locate the zero-crossing point of the voltage for each power frequency cycle; extract time-domain segments centered on each zero-crossing point, and superimpose and average the segments of multiple consecutive power frequency cycles to generate a zero-crossing distortion characteristic waveform; perform high-pass filtering on the zero-crossing distortion characteristic waveform to filter out the power frequency fundamental wave and low-order system harmonics, and extract the high-frequency fingerprint voltage curve;

[0010] Step S200: Using the electricity metering box as a unit, calculate the passive waveform similarity coefficient between the high-frequency fingerprint voltage curve of each user's meter in the box and the high-frequency fingerprint voltage curve of the transformer reference point; when the passive waveform similarity coefficient falls into the preset fuzzy uncertainty range, dynamically trigger the active coding ripple voltage curve matching mechanism.

[0011] Step S300: The intelligent terminal of the distribution area sends a detection command to the power metering box, which is the target ambiguity area, designating at least one user meter in the box as an active coding injection source; the internal switching power supply module of the active coding injection source is controlled to perform high-speed load switching according to a preset pseudo-random binary coding sequence, thereby generating an active coding ripple voltage curve on the incoming bus voltage of the power metering box.

[0012] Step S400: The transformer low-voltage side reference point master meter and each user meter synchronously record the voltage waveform curve within the modulation time window, and perform coherent correlation operation between the recorded voltage curve and the pseudo-random binary encoded sequence through a digital matched filter, demodulate and extract their respective active encoded voltage feature vectors, and calculate the active coherence coefficient.

[0013] Step S500: Based on the passive waveform similarity coefficient and the active coherence coefficient, perform multi-dimensional joint decision-making to determine and confirm the physical affiliation between the power metering box and the corresponding transformer substation.

[0014] The physical principle of this invention is as follows: First, due to the objective differences in the internal leakage inductance, winding distributed capacitance, and nonlinear characteristics of the power electronic loads connected to the bus side of different distribution transformers, and due to the nonlinear switching of the diode rectifier bridge, these differences will excite a unique microscopic transient high-frequency resonant response near the zero-crossing point of the AC power frequency. By extracting the microscopic fingerprint of this zero-crossing period, the macroscopically overlapping voltage curves can be effectively separated in the microscopic time-frequency domain. Second, when the microscopic high-frequency fingerprint cannot provide a deterministic conclusion, the active charge modulation with extremely low energy is performed using the built-in switching power supply of the smart meter. This process is essentially superimposing a real physical voltage drop that obeys Kirchhoff's laws onto the loop impedance. Disconnected switches or out-of-phase buses cannot conduct such baseband ripples, therefore this method can penetrate communication protocol crosstalk caused by electromagnetic radiation, transforming the judgment of physical connectivity into a coherent demodulation process for a specific encoded digital signal.

[0015] Further, in step S100, the sampling rate of the instantaneous power frequency voltage waveform is not less than 6.4kHz; the time window width of the time domain segment truncated with each voltage zero-crossing point as the center is 0.2ms to 1.0ms before and after the zero-crossing point; and the number of continuous power frequency cycles used in the superposition averaging denoising is not less than 10. These parameter configurations ensure that, while satisfying the Nyquist theorem, high-frequency glitches are fully captured, and background white noise is effectively suppressed through ensemble averaging, thereby improving the signal-to-noise ratio of subsequent matching.

[0016] Furthermore, in step S100, the cutoff frequency of the high-pass filter is set to 500Hz, and the frequency components of the extracted high-frequency fingerprint voltage curve are concentrated between 500Hz and 3.2kHz. Filtering out frequency components below 500Hz can eliminate the energy masking effect of the 50Hz fundamental wave and major low-frequency system harmonics (such as the 3rd, 5th, and 7th harmonics), highlighting the weak high-frequency characteristics closely related to the physical structure of the transformer substation.

[0017] Furthermore, in step S200, the lower threshold of the preset fuzzy uncertainty interval is 0.65, and the upper threshold is 0.90.

[0018] Furthermore, in step S200, if the passive waveform similarity coefficient between more than 80% of the user meters in the energy metering box and the high-frequency fingerprint voltage curve of the transformer reference point is greater than the upper limit threshold of 0.90, then it is directly determined that the energy metering box as a whole belongs to the transformer area, and the active coded ripple voltage curve matching mechanism is not triggered. Through this adaptive exit mechanism, more than 90% of the topology verification in the system can be completed in passive mode without any signal injection, which greatly reduces the computational load of the system.

[0019] Furthermore, in step S300, the high-speed load switching duration of the pseudo-random binary encoded sequence is 50ms to 200ms, preferably 80ms to 160ms, and more preferably 100ms to 140ms. This time window selection ensures that the matched filter has sufficient integration length to obtain processing gain, while avoiding the accumulation of thermal stress in the internal power supply components of the energy meter due to frequent switching over long periods.

[0020] Furthermore, in step S400, when the transformer low-voltage side reference point master meter and each user meter record the voltage waveform curve within the modulation time window, the synchronous recording sampling rate is not less than 12.8kHz, and the network-wide time synchronization accuracy is maintained by the broadband power line carrier network clock. Microsecond-level time synchronization ensures that the transmitted and received sequences are strictly phase-aligned during coherent computation to obtain the maximum cross-correlation peak value.

[0021] Furthermore, in step S500, the determination rule for the multi-dimensional joint decision is as follows: if and only if the active coherence coefficient is greater than or equal to 0.95, and the active coded voltage feature vectors extracted from the demodulation of each user's meter in the energy metering box show consistent origin, then the energy metering box is finally confirmed to belong to the corresponding transformer substation. This consistent origin verification not only enhances the robustness of the decision but also effectively detects abnormal physical topologies such as unauthorized connections of individual meters within the same metering box.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] (1) It effectively overcomes the limitations of the macroscopic voltage RMS curve matching method. By introducing the extraction and matching of high-frequency fingerprints of power frequency zero crossing points, the system is able to distinguish between homogeneous impedance and homogeneous power consumption behavior.

[0024] (2) It has the ability to resist high-frequency spatial crosstalk and realizes absolute verification of physical connectivity. The baseband pulse modulation is performed using the meter's own switching power supply. The resulting physical ripple cannot pass through the disconnected electrical node, which fundamentally corrects the topology file disorder caused by the inverted HPLC communication logic.

[0025] (3) Significantly reduced engineering implementation costs and protected power quality. This invention does not require the installation of any high-power external signal injection equipment in the distribution network, and directly reuses the existing hardware of the meter; the generated millivolt-level ripples are within the background noise tolerance of the power grid itself, completely avoiding the problem of excessive local total harmonic distortion caused by the traditional active method, and achieving the effect of obtaining high-precision topology rights with zero additional modification cost. Attached Figure Description

[0026] Figure 1The overall flowchart of the method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching provided by the present invention is shown below.

[0027] Figure 2 This is a schematic diagram illustrating the signal processing principle of zero-crossing distortion feature waveform truncation, superposition denoising, and high-pass filtering to extract high-frequency fingerprint voltage curves in an embodiment of the present invention.

[0028] Figure 3 This is a comparison diagram of the coded ripple waveform generated by the controlled modulation of the switching power supply in this embodiment of the invention, and the coherent matched filter output results on the target transformer and non-target transformer sides.

[0029] Figure 4 This is a comparison chart showing the similarity between the traditional macroscopic voltage RMS curve and the high-frequency fingerprint voltage curve of this invention across different station ranges. Detailed Implementation

[0030] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0031] Combination Figure 1 The overall flowchart of the present invention shown shows that after the process is started, it first enters the normalized passive detection stage (step S100), extracts high-frequency fingerprints and calculates similarity (step S200). Figure 1 The decision branch regarding whether the waveform falls within the fuzzy range: When the passive waveform similarity is extremely high or low, the membership relationship is directly confirmed along the negative branch without triggering the active mechanism; only when the similarity falls within the fuzzy range... When the fuzzy and uncertain interval is reached, the process proceeds downward along the branch, precisely triggering the active encoding matching mechanism in step S300 and subsequent coherent demodulation (step S400), and finally completing the joint decision (step S500).

[0032] Part 1: System-level Underlying Architecture Design and Core Algorithms

[0033] The hardware upon which this invention is based mainly includes intelligent distribution transformer terminals (TTUs) in the distribution room of the transformer substation, concentrators / master meters, and smart energy meters distributed in the user-side energy metering boxes. All of these terminals have built-in analog-to-digital converters (ADCs) and microcontrollers / digital signal processors (MCUs / DSPs).

[0034] 1. Hardware implementation and algorithm derivation of passive fingerprint extraction in step S100

[0035] The system requires the ADC to perform high-speed sampling of the AC voltage waveform, with the sampling rate set to [value missing]. And above. Because the real voltage waveform contains background noise, relying solely on sign changes to determine the zero-crossing point results in significant errors. This embodiment employs a linear interpolation method with anti-jitter mechanism to achieve sub-sampling rate level zero-crossing point location. Assuming two consecutive sampling points... and If voltage polarity reverses between them, the precise zero-crossing time will be... The calculation formula is:

[0036]

[0037] Combination Figure 2 The diagram shown illustrates the principle of high-frequency fingerprint voltage curve signal processing, used to locate... Centered on the origin, the time window is defined as follows: (Total width) The instantaneous waveform of the power grid. In order to effectively suppress random white power grid noise that follows a normal distribution, continuous waveforms are extracted. One power frequency cycle ( In-phase time domain segment The denoised zero-crossing distortion characteristic waveform is obtained by ensemble averaging. :

[0038]

[0039] The superimposed waveform still contains a significant amount of low-frequency fundamental energy. The transfer function of a Butterworth high-pass filter based on the infinite impulse response (IIR) is designed within a microcontroller. The cutoff frequency is set to By using difference equations Filtering, thoroughly filtering out The fundamental frequency and lower-order system harmonics were used to extract the high-frequency fingerprint voltage curve.

[0040] 2. Passive waveform similarity calculation and adaptive triggering logic in step S200

[0041] Based on the high-frequency fingerprint voltage curve extracted in step S100, the microcontroller (MCU) calls a waveform similarity calculation algorithm to calculate the passive waveform similarity coefficient between each user's meter in the energy metering box and the transformer reference point master meter. The microcontroller has internal decision logic that compares the calculation result with a preset threshold. If the similarity coefficient falls within the fuzzy uncertainty range of [0.65, 0.90], the MCU outputs a trigger signal to adaptively initiate the subsequent active coded ripple voltage curve matching mechanism; if it is not within this range, the decision logic is completed directly, thereby achieving optimal allocation of computing system resources at the underlying level and avoiding invalid active injection.

[0042] 3. The physical injection mechanism of the active encoding ripple in step S300 and the PRBS generator

[0043] Smart meters contain a switching power supply (SMPS), typically employing a flyback converter topology. In step S300, the MCU controls the enable state of the SMPS's PWM chip, or controls the switching of the DC-side virtual load, thereby altering the input current instantaneously drawn by the primary coil. Let the equivalent internal resistance of the busbar system be... Tiny changes in current can induce superimposed ripple at the input terminal:

[0044]

[0045] When the modulation signal uses a pseudo-random binary sequence (PRBS, denoted as...) When the value is 0 or 1, the signal on the incoming bus will present the baseband envelope corresponding to the sequence.

[0046] 4. Digital matched filter coherent demodulation in step S400

[0047] The transformer low-voltage side reference point summary table is no less than High sampling rate synchronous recording. The receiver uses a digital matched filter whose impulse response is the time-reversed conjugate of the transmitted sequence. Discrete convolution calculations are performed to locate the signal.

[0048]

[0049] When the sequences are fully aligned, the correlation spike is output to characterize the active coherence coefficient.

[0050] Combination Figure 3 The comparison diagram of matched filter output results shown illustrates this. Area A represents the PRBS encoded signal from the transmitting end. Area B represents the output from the target transformer side. Due to the presence of physical conduction wires, the integrator accumulates real charge energy, resulting in a sharp main correlation peak and a high active coherence coefficient. Area C is the non-target transformer side. Although spatial carrier crosstalk exists, the physical circuit breaker isolates the low-frequency ripple. The matched filter cannot extract an effective sequence, exhibiting a white noise distribution, and the active coherence coefficient is low. .

[0051] 5. Step S500: Multidimensional Joint Decision-Making and Physical Affiliation Confirmation

[0052] After coherent demodulation, the edge computing module or digital signal processor (DSP) within the transformer substation (TTU) executes a multi-dimensional joint decision-making algorithm. The processor comprehensively retrieves the passive waveform similarity coefficients cached in step S200, as well as the active coherence coefficients and active encoded voltage feature vectors obtained in step S400. Based on underlying logic thresholds, if and only if the active coherence coefficients reach or exceed a preset threshold of 0.95, and the feature vector verification shows consistent origin, the processor determines that the physical electrical loop is closed, outputs a determination of ownership of the power metering box belonging to the corresponding transformer substation, and completes the closed-loop verification.

[0053] Part Two: Multi-Dimensional Operating Condition Implementation Examples and Parameter Verification

[0054] Example 1: Joint Validation under Conditions of Severe Macroscopic Convergence and Severe HPLC Crosstalk

[0055] Scenario Description: In a certain town area, transformer #1 and transformer #2 operate in parallel, with a 100-meter-long low-voltage main cable laid in a shared trench. Meter box B01 is physically powered by transformer #1 and contains 6 smart meters. The load in this area is stable and homogeneous, and crosstalk in the shared trench during carrier communication is severe, causing the records of some meters in B01 to repeatedly jump in the main station.

[0056] Execution steps and data processing:

[0057] Step S100: Meter No. 1 and transformers No. 1 and No. 2 in B01 Initiate high-frequency sampling. Extract both the preceding and following sections. The segment, an average of 15 consecutive cycles, is... High-frequency fingerprint voltage curves are obtained using high-pass filtering.

[0058] Step S200: Calculate the similarity coefficient between the high-frequency fingerprint of meter #1 and transformer #1. The similarity coefficient with variable #2 The preset fuzzy range is... If the value falls within this range, adaptive trigger step S300 is initiated.

[0059] Step S300: Select table 1 as the injection source. Generate a 7-bit PRBS code (1011001). Set the high-speed load switching duration to... Meter 1's switching power supply generates an amplitude of approximately... The faint coded ripple.

[0060] Step S400: Terminals #1 and #2 are synchronized with the HPLC beacon, using... Record the waveform and perform digital matched filtering.

[0061] Step S500: Calculate the active coherence coefficient for terminal #1. ;2# terminal calculation obtained According to the judgment rules The system determined that B01 has 100% physical connectivity and belongs to transformer #1, successfully eliminating the crosstalk logic error.

[0062] Example 2: Parameter Boundary Support Verification and Extreme Interference Test

[0063] To fully support the numerical range of the claims, three units are connected on the bus side. Variable frequency water pumps simulate strong interference environments.

[0064] Test Group A (Lower Bound Verification): Zero-crossing window set to both the front and back PRBS switching time window set to The narrower window captured near-zero high-frequency glitches. Short-time modulation limits the disturbance. Under strong background noise, the active coherence coefficient of the demodulated target station area is 0.953. Accurately determine the rights and interests. Prove the enforceability of the lower limit endpoint value of the claim.

[0065] Test Group B (Upper Limit Verification): Zero-crossing window set to both front and back PRBS switching time window set to . Modulation provides a longer integral gain, significantly improving anti-interference capability. The active coherence coefficient reaches 0.991. Temperature rise monitoring shows that the switching module temperature rises by approximately 1.5℃, within the safety tolerance range. This demonstrates the superiority of the upper limit value.

[0066] Example 3: Fast Exit Mechanism When Passive Features Have High Significance

[0067] Scenario Description: In a rural transformer substation, the power supply radius and line impedance of transformers #3 and #4 differ significantly. Execution Steps: Perform S100 and S200 operations on the 8 meters in metering box C01 to be calibrated. Calculations show that the 7 meters in C01 account for approximately... The high-frequency fingerprint similarity with that of variant #3 is concentrated in The similarity between C01 and transformer #3 is greater than the upper limit of the fuzzy region (0.90). The similarity with transformer #4 is less than 0.40. The algorithm triggers a rapid determination: directly confirming that C01 belongs to transformer #3, terminating the verification, and no longer triggering the active encoding mechanism in step S300. This demonstrates the adaptive energy-saving optimization of this method.

[0068] Example 4: Detecting unauthorized and haphazard connections inside the box using the principle of consistency of origin

[0069] Six electricity meters are logged in metering box D01. S300 selects meter 1 for modulation. During S400 demodulation, the waveforms of all meters in the box are extracted and cross-correlation is calculated. Under normal circumstances, the meters are connected in parallel to the same busbar, and the demodulation vectors should be nearly identical. Data shows that the characteristic vector of meter 5 has a significantly delayed phase and a greatly attenuated amplitude, with the cross-correlation coefficient dropping to 0.51. Based on transmission line theory, attenuation corresponds to redundant impedance. The system determines that meter 5 is not connected to the same busbar in the metering box, indicating unauthorized cross-box wiring, triggering an inspection work order. This demonstrates its practical value in identifying microscopic abnormal wiring conditions.

[0070] Comparative Example 1: Comparison between traditional macroscopic voltage RMS curve matching and the present invention

[0071] Combination Figure 4 The similarity comparison chart shown indicates that area A represents the traditional macroscopic voltage RMS curve. Within 24 hours, influenced by a gradual load, the reference curve of transformer #1 highly overlaps with the curve of the meter under test. The Pearson correlation coefficients are... and The difference is only It falls into a recognition dead zone and is misled by carrier crosstalk. Region B represents the high-frequency fingerprint voltage curve of this invention. At the microsecond level after filtering out the fundamental wave, due to subtle differences in physical impedance, the passive waveform similarity decreases to the true value. It accurately falls into the fuzzy range, thus successfully triggering subsequent proactive investigations.

[0072] Performance data testing and effect comparison summary

[0073] The test data of the above embodiments and comparative examples under the same topology and interference are summarized in Table 1.

[0074] Table 1: Comparison of Core Performance Test Data between Examples and Comparative Examples

[0075] Test group Core mechanisms and parameters Household transformer topology discrimination accuracy Cross-regional carrier crosstalk suppression capability Impact of power quality on grid supply Comparative Example 1 No active modulation (macroscopic RMS matching) 68.4% (trapped in the homogenized impedance blind zone) Extremely poor, completely misled by false messages. No impact Example 2 (Group A) Narrow window 0.2ms / Short modulation 50ms 96.2% Good (low-frequency physical energy prevents it from exceeding the limit) Extremely small, transient fluctuations of a few millivolts. Example 1 (Preferred) Moderate 0.5ms / Golden Window 120ms 100.0% Extremely strong (the main peak of the relevant output is extremely significant). Unnoticeable and does not affect THD levels. Example 2 (Group B) Wide window 1.0ms long modulation 200ms 100.0% Extremely strong resistance to interference from high-power frequency inverters. Minimal heat generation; power supply heat is within safe limits.

[0076] As shown in Table 1, by constructing a collaborative system of passive microscopic initial screening and active controlled ripple point removal, this invention successfully increases the topology confirmation rate of the household transformer under extreme interference conditions to 100% without increasing the cost of external hardware and under the constraint of zero increase in power quality THD.

[0077] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. The scope of patent protection of the present invention shall be determined by the claims. Similarly, any equivalent structural changes made based on the content of the present invention's specification shall also be included within the scope of protection of the present invention.

Claims

1. A method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching, characterized in that, The method uses the electricity metering box as the smallest verification unit and includes the following steps: Step S100: Obtain the instantaneous waveforms of the power frequency voltage at the reference point of the low-voltage side of the low-voltage transformer and at each user's meter in the power metering box of the user to be verified, and accurately locate the zero-crossing point of the voltage for each power frequency cycle; extract time-domain segments centered on each zero-crossing point, and superimpose and average the segments of multiple consecutive power frequency cycles to generate a zero-crossing distortion characteristic waveform; perform high-pass filtering on the zero-crossing distortion characteristic waveform to filter out the power frequency fundamental wave and low-order system harmonics, and extract the high-frequency fingerprint voltage curve; Step S200: Using the electricity metering box as a unit, calculate the passive waveform similarity coefficient between the high-frequency fingerprint voltage curve of each user's meter in the box and the high-frequency fingerprint voltage curve of the transformer reference point; when the passive waveform similarity coefficient falls into the preset fuzzy uncertainty range, dynamically trigger the active coding ripple voltage curve matching mechanism. Step S300: The intelligent terminal of the distribution area sends a detection command to the power metering box, which is the target ambiguity area, designating at least one user meter in the box as an active coding injection source; the internal switching power supply module of the active coding injection source is controlled to perform high-speed load switching according to a preset pseudo-random binary coding sequence, thereby generating an active coding ripple voltage curve on the incoming bus voltage of the power metering box. Step S400: The transformer low-voltage side reference point master meter and each user meter synchronously record the voltage waveform curve within the modulation time window, and perform coherent correlation operation between the recorded voltage curve and the pseudo-random binary encoded sequence through a digital matched filter, demodulate and extract their respective active encoded voltage feature vectors, and calculate the active coherence coefficient. Step S500: Based on the passive waveform similarity coefficient and the active coherence coefficient, perform multi-dimensional joint decision-making to determine and confirm the physical affiliation between the power metering box and the corresponding transformer substation.

2. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S100, the sampling rate of the instantaneous power frequency voltage waveform is not less than 6.4kHz; the time window width of the time domain segment truncated with each voltage zero-crossing point as the center is 0.2ms to 1.0ms before and after the zero-crossing point; and the number of continuous power frequency cycles used in the superposition averaging denoising is not less than 10.

3. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S100, the cutoff frequency of the high-pass filter is set to 500Hz, and the frequency components of the extracted high-frequency fingerprint voltage curve are concentrated between 500Hz and 3.2kHz.

4. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S200, the lower threshold of the preset fuzzy uncertainty interval is 0.65 and the upper threshold is 0.

90.

5. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 4, characterized in that, In step S200, if the passive waveform similarity coefficient between more than 80% of the user meters in the power metering box and the high-frequency fingerprint voltage curve of the transformer reference point is greater than the upper limit threshold of 0.90, then it is directly determined that the power metering box as a whole belongs to the transformer area, and the active coded ripple voltage curve matching mechanism is not triggered.

6. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S300, the high-speed load switching duration of the pseudo-random binary encoded sequence is 50ms to 200ms.

7. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S400, when the transformer low-voltage side reference point master meter and each user meter record the voltage waveform curve within the modulation time window, the synchronous recording sampling rate is not less than 12.8kHz, and the network-wide time synchronization accuracy is maintained by the broadband power line carrier network clock.

8. The method for verifying the relationship between electricity metering boxes and transformers based on voltage curve matching according to claim 1, characterized in that, In step S500, the determination rule for the multidimensional joint decision is: if and only if the active coherence coefficient is greater than or equal to 0.95, and the active encoded voltage feature vectors extracted by demodulation of each user's meter in the power metering box show the same source consistency, then the power metering box is finally confirmed to belong to the corresponding transformer area.

Citation Information

Patent Citations

  • Automatic identification method and system for topological relation of transformer area

    CN113659714A

  • Low-voltage transformer area missort identification method and system

    CN122174173A