A method for identifying electricity theft in metering boxes based on impedance spectrum characteristics
By acquiring the instantaneous phase-triggered orthogonal frequency division multiplexing detection sequence of the power frequency voltage inside the metering box, extracting complex channel state information and calculating impedance spectrum characteristics, the problems of noise interference and false alarms in metering box anti-theft identification are solved, and more accurate anti-theft identification is achieved.
Patent Information
- Application Number
- CN202611070004.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-17
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies are susceptible to noise interference in the identification of electricity theft in metering boxes, and relying solely on impedance spectrum characteristics makes it difficult to distinguish between legitimate electrical loads and externally connected wires, leading to frequent false alarms.
By acquiring the instantaneous phase of the power frequency voltage inside the metering box, an orthogonal frequency division multiplexing detection sequence is triggered within a specific phase window. Complex channel state information is extracted, and the normalized impedance spectrum amplitude sequence and residual group delay sequence are calculated. These are then compared with historical benchmarks to solve for amplitude and delay abrupt change characteristics. An orthogonal feature space is constructed for classification and decision-making.
It improves the accuracy of electricity theft detection, reduces the frequency of false alarms, and can effectively distinguish between legitimate load start-up and shutdown and physical line changes such as bypassing the meter box to connect long wires, thus enhancing the meter box's ability to detect electricity theft in complex environments.
Smart Images

Figure CN123116040A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of power distribution network monitoring technology, specifically to a method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics. Background Technology
[0002] The metering boxes in distribution substations are responsible for metering and settling electricity billing on the user side. Therefore, preventing electricity theft in these boxes is a crucial aspect of controlling distribution network line losses. Since the impedance spectrum characteristics of a distribution network objectively reflect the broadband response of physical lines at different frequencies, and the specific numerical distribution is directly related to the external network topology and internal load attributes, metering box anti-theft identification based on impedance spectrum characteristics has become an important technological direction in the field of power grid condition monitoring.
[0003] Currently, conventional impedance spectrum application solutions mainly rely on the communication module inside the metering box to send broadband detection signals to the power line. The analysis is carried out by statistically analyzing the average impedance amplitude fluctuation of the line in a specific frequency band. When the impedance amplitude drops or exceeds the set benchmark, it is determined that there is a wiring abnormality in the line and a verification prompt is generated.
[0004] This method has limitations in practical applications. Existing solutions often acquire detection signals at random time points, and the collected impedance spectrum is easily distorted by broadband noise from the power grid. Furthermore, relying solely on impedance amplitude characteristics makes it difficult to distinguish legitimate local high-power load starts and stops from external electricity theft. In addition, current technology lacks the ability to calculate high-frequency phase information of transmitted signals and has not yet established a mapping relationship between the physical length of the conductor and the signal group delay. This results in insufficient perception of topological changes caused by connecting concealed long conductors around the metering box, easily misjudging load fluctuations caused by normal electricity use as genuine electricity theft.
[0005] Therefore, this invention proposes a method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics to address the shortcomings of existing technologies. Summary of the Invention
[0006] To address the shortcomings of existing technologies, this invention provides a method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics. This method solves the technical problems of existing technologies, which are susceptible to noise interference when obtaining metering box operating parameters and cannot distinguish between legitimate electrical loads and externally connected wires by relying solely on impedance spectrum characteristics, thus leading to false alarms in actual electricity theft prevention identification results.
[0007] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying anti-theft electricity in metering boxes based on impedance spectrum characteristics, comprising: The instantaneous phase of the power frequency voltage inside the metering box is acquired, and when the instantaneous phase enters the preset zero-point phase window or peak phase window, an orthogonal frequency division multiplexing detection sequence is triggered. Extract the complex channel state information of the orthogonal frequency division multiplexing probe sequence under the corresponding phase window; The normalized impedance spectrum amplitude sequence is calculated based on the complex channel state information; the phase angle of the complex channel state information is subjected to phase dewinding and clock drift removal processing to extract the residual group delay sequence characterizing the high-frequency topological features; The normalized impedance spectrum amplitude sequence is compared with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum abrupt change features; the residual group delay sequence is compared with the acquired pre-stored historical reference group delay sequence to extract delay abrupt change features. The dispersion calculated between the zero-point phase window and the peak phase window by combining the amplitude spectrum mutation feature and the delay mutation feature are used to perform orthogonal feature classification, and an anti-theft alarm command is output based on the classification result.
[0008] Preferably, the step of acquiring the instantaneous phase of the power frequency voltage inside the metering box, and triggering the transmission of an orthogonal frequency division multiplexing detection sequence when the instantaneous phase enters a preset zero-point phase window or peak phase window includes: The power frequency voltage is extracted and the instantaneous phase is calculated using a voltage zero-crossing detection circuit; The zero-point phase window and the peak-point phase window are constructed based on the phase tolerance width; When the instantaneous phase enters the zero-point phase window or the peak phase window, a highest preemption priority command is issued. Based on the highest preemption priority instruction, a carrier listening mechanism is masked and the orthogonal frequency division multiplexing detection sequence containing multiple subcarriers is sent.
[0009] Preferably, the step of extracting the power frequency voltage and calculating the instantaneous phase through the voltage zero-crossing detection circuit includes: When the power frequency voltage crosses the zero potential, an edge pulse is output. The zero-crossing direction is distinguished according to the edge pulse, and a reference phase corresponding to the zero-crossing node is set. The time of the most recent occurrence of the edge pulse is recorded using a clock timer as the zero-crossing moment; Obtain the current real-time running time and calculate the time difference between the current time and the zero-crossing time; Divide the time difference by the preset power frequency cycle to obtain the ratio, and multiply the ratio by 360 degrees to obtain the multiplication result; After adding the multiplication result to the reference phase, a modulo operation is performed on 360 degrees to calculate the instantaneous phase.
[0010] Preferably, the step of extracting the complex channel state information of the orthogonal frequency division multiplexing probe sequence under the corresponding phase window includes: Receive the orthogonal frequency division multiplexing probe sequence and extract the complex channel state information sequence under the corresponding phase window from the physical layer equalizer register; Separate the in-phase and quadrature components corresponding to each subcarrier in the complex channel state information sequence; A discrete complex transmission function is constructed using the in-phase component and the quadrature component corresponding to each subcarrier, and the discrete complex transmission function is used as the complex channel state information.
[0011] Preferably, the step of calculating the normalized impedance spectrum magnitude sequence based on the complex channel state information includes: Extract the in-phase component and the quadrature component corresponding to each subcarrier in the discrete complex transmission function; The square operation is performed on the in-phase component and the quadrature component respectively. The square value of the in-phase component and the square value of the quadrature component are summed and then the square root is taken to generate the absolute amplitude of the discrete complex transfer function on each subcarrier. The positive bias constant is added to the absolute amplitude to generate a sum, and the reciprocal of the sum is taken to generate the normalized impedance spectrum amplitude corresponding to each subcarrier. The normalized impedance spectrum amplitudes corresponding to each subcarrier are arranged in order of frequency from low to high to form the normalized impedance spectrum amplitude sequence.
[0012] Preferably, the step of performing phase dewinding and clock drift removal processing on the phase angle of the complex channel state information to extract the residual group delay sequence characterizing the high-frequency topological features includes: Calculate the ratio of the orthogonal component to the in-phase component and perform an arctangent operation to generate a wrapping phase angle sequence; Calculate the phase difference between adjacent subcarriers in each subcarrier of the wrapped phase angle sequence, perform jump compensation on the phase difference based on positive jump threshold and negative jump threshold, and convert the wrapped phase angle sequence into an unwound phase angle sequence; The linear trend phase angle component is extracted by polynomial fitting of the unwound phase angle sequence, and the linear trend phase angle component is subtracted from the unwound phase angle sequence item by item to obtain the residual phase angle sequence. The discrete derivatives of the adjacent subcarrier frequencies in the residual phase angle sequence are calculated, and the calculated discrete derivative values are used to construct the residual group delay sequence.
[0013] Preferably, the step of comparing the normalized impedance spectrum amplitude sequence with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum abrupt change features includes: The normalized impedance spectrum amplitude sequence is subtracted point by point from the acquired pre-stored historical reference impedance spectrum amplitude sequence, and the absolute value is taken to generate an amplitude absolute difference sequence. When the input data corresponds to the zero-point phase window, the sequence of absolute difference in amplitude is traversed and the point of maximum difference is extracted, which is defined as the zero-point amplitude abrupt change peak within the same monitoring period; When the input data corresponds to the peak phase window, the absolute difference sequence of amplitude is traversed and the maximum difference point is extracted, which is defined as the peak amplitude sudden change peak within the same monitoring period; When the zero-point amplitude mutation peak is greater than the preset zero-point amplitude abnormality threshold or the peak amplitude mutation peak is greater than the preset peak amplitude abnormality threshold, the subcarrier index containing the one with the larger value between the zero-point amplitude mutation peak and the peak amplitude mutation peak is locked as the target subcarrier index. By combining the pre-configured broadband carrier communication start frequency and the frequency interval between adjacent subcarriers, the center frequency corresponding to the target subcarrier index is calculated. The combination of the zero-point amplitude mutation peak, the peak amplitude mutation peak, and the corresponding center frequency within the same monitoring period is used as the amplitude spectrum mutation feature.
[0014] Preferably, the step of comparing the residual group delay sequence with the acquired pre-stored historical benchmark group delay sequence to extract delay mutation features includes: The residual group delay sequence is subtracted from the pre-stored historical benchmark group delay sequence item by item to generate a group delay difference sequence. The target subcarrier index is used to determine the group delay mapping index, and the value corresponding to the group delay mapping index in the group delay difference sequence is read and defined as the group delay mutation value under the current phase window; The average value of the two group delay mutation values calculated under the zero-point phase window and the peak phase window within the same monitoring period is defined as the final group delay mutation value, and the final group delay mutation value is used as the delay mutation feature.
[0015] Preferably, the step of performing orthogonal feature classification by combining the dispersion calculated from the amplitude spectrum change feature between the zero-point phase window and the peak phase window with the delay change feature, and outputting an anti-theft alarm command based on the classification result includes: The zero-point amplitude mutation peak and the peak amplitude mutation peak are extracted from the amplitude spectrum mutation characteristics; Calculate the absolute difference between the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak; Compare the numerical values of the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak, and extract the maximum value between the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak as the denominator constraint term; Divide the absolute difference by the denominator constraint term to calculate the dispersion. Using the dispersion as the first coordinate axis variable and the final group delay mutation value as the second coordinate axis variable, an orthogonal feature space is constructed. The orthogonal feature space is substituted with a discreteness threshold and a delay mutation threshold to perform orthogonal feature classification, and the anti-theft alarm command is output based on the classification result.
[0016] Preferably, the step of performing orthogonal feature classification by substituting a discreteness threshold and a delay mutation threshold into the orthogonal feature space, and outputting the anti-theft alarm command based on the classification result includes: When the dispersion is determined to be greater than or equal to the dispersion threshold, it is classified as a legitimate nonlinear load start-up and shutdown event, and it is determined that there is no electricity theft behavior on the current network. When it is determined that the dispersion is less than the dispersion threshold and the final group delay mutation value is less than or equal to the delay mutation threshold, it is classified as a local in-situ legal pure resistive load start-up and shutdown event, and it is determined that there is no electricity theft behavior in the current network. When it is determined that the dispersion is less than the dispersion threshold and the final group delay mutation value is greater than the delay mutation threshold, it is classified as an event of bypassing the metering device to connect a passive long wire, and the anti-theft identification alarm command is generated and output.
[0017] This invention provides a method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics. It has the following advantages: 1. This invention triggers an orthogonal frequency division multiplexing detection sequence through a specific phase window of the power frequency voltage, and calculates the normalized impedance spectrum amplitude sequence based on complex channel information. This reduces the impact of random phase sampling and random backoff delay on impedance spectrum feature extraction, thereby improving the stability of the underlying impedance spectrum features during extraction. This provides data support for analyzing physical topology changes and improves the accuracy of metering boxes in anti-theft identification in complex environments from the signal source.
[0018] 2. This invention performs unwinding and clock drift removal processing on the channel phase angle, extracts the residual group delay sequence and compares it with historical benchmarks, and uses the change in group delay to characterize the change in physical reflection path, which makes up for the limitations of simply relying on impedance spectrum amplitude monitoring. This allows for the determination of whether a physical line change has occurred by connecting a long wire around the metering box, effectively making up for the lack of spatial topology state determination in the anti-theft electricity identification scheme.
[0019] 3. This invention constructs an orthogonal feature space to perform classification decisions by combining the discreteness of amplitude mutations calculated under different phase windows and the group delay mutation results. This distinguishes the performance of legitimate load start-up and shutdown and unauthorized wire connections at the impedance spectrum characteristic level, thereby eliminating the judgment interference caused by conventional electricity use behavior and purely resistive load access. This reduces the frequency of false alarms in the actual operation of the metering box and improves the objectivity of the anti-theft identification alarm results output. Attached Figure Description
[0020] Figure 1 This is a diagram illustrating the architecture of the metering box anti-electricity theft identification system based on impedance spectrum characteristics according to the present invention. Figure 2 The flowchart of the meter box anti-electricity theft identification method based on impedance spectrum characteristics of the present invention is shown below; Figure 3 This is a schematic diagram of voltage phase calculation and signal triggering according to the present invention; Figure 4 This is a schematic diagram of complex channel state information extraction according to the present invention; Figure 5 This is a schematic diagram illustrating the impedance amplitude and group delay calculation of the present invention; Figure 6 This is a schematic diagram of mutation feature comparison and frequency point locking in the present invention; Figure 7 This is a schematic diagram of the orthogonal feature classification and alarm output of the present invention.
[0021] Among them, 100 is the phase triggering module; 200 is the information extraction module; 300 is the feature cleaning module; 400 is the mutation extraction module; and 500 is the joint decision module. Detailed Implementation
[0022] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. 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 skilled in the art without creative effort are within the scope of protection of the present invention.
[0023] Reference Figure 1 The present invention provides a meter box anti-theft identification system based on impedance spectrum characteristics, including a phase triggering module 100, an information extraction module 200, a feature cleaning module 300, a mutation extraction module 400, and a joint decision module 500.
[0024] The anti-electricity theft detection system is deployed inside the metering box in the distribution transformer area and is built based on the microprocessor and broadband carrier communication module integrated on the main control board of the metering box. The microprocessor is connected to the broadband carrier communication module through the underlying bus, providing a platform for data acquisition, processing, and transmission scheduling for the phase triggering module 100, information extraction module 200, feature cleaning module 300, mutation extraction module 400, and joint decision module 500. As a preferred embodiment, the broadband carrier communication module is an orthogonal frequency division multiplexing carrier communication module based on the high-speed power line communication (HPLC) standard. This broadband carrier communication module integrates a physical layer baseband chip and analog front-end circuitry, enabling it to multiplex the power network medium for high-frequency signal modulation and demodulation and channel state information extraction.
[0025] The input terminal of the phase trigger module 100 is connected to the microprocessor. The phase trigger module 100 is used to acquire the instantaneous phase of the power frequency voltage in the metering box, and trigger the transmission of the orthogonal frequency division multiplexing detection sequence when the instantaneous phase enters the preset zero-point phase window or peak phase window.
[0026] The information extraction module 200 is connected to the phase triggering module 100 and the broadband carrier communication module respectively. The information extraction module 200 is used to extract the complex channel state information of the orthogonal frequency division multiplexing detection sequence under the corresponding phase window.
[0027] The input of the feature cleaning module 300 is connected to the output of the information extraction module 200. The feature cleaning module 300 is used to calculate the normalized impedance spectrum amplitude sequence based on the complex channel state information, and to perform phase dewinding and clock drift removal processing on the phase angle of the complex channel state information to extract the residual group delay sequence characterizing the high-frequency topological features.
[0028] The mutation extraction module 400 is connected to the output of the feature cleaning module 300 and the microprocessor, respectively. The mutation extraction module 400 is used to compare the normalized impedance spectrum amplitude sequence with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum mutation features; and to compare the residual group delay sequence with the acquired pre-stored historical reference group delay sequence to extract delay mutation features.
[0029] The joint decision module 500 is connected to the output of the mutation extraction module 400. The joint decision module 500 is used to perform orthogonal feature classification by combining the discreteness of the amplitude spectrum mutation feature between the zero-point phase window and the peak phase window and the delay mutation feature, and outputs an anti-theft identification alarm command according to the classification result.
[0030] Reference Figure 2 This invention provides a method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics. This method is executed by the aforementioned electricity theft prevention identification system and includes the following steps: S1, acquire the instantaneous phase of the power frequency voltage in the metering box, and trigger the transmission of the orthogonal frequency division multiplexing detection sequence when the instantaneous phase enters the preset zero-point phase window or peak phase window.
[0031] S2, extract the complex channel state information of the orthogonal frequency division multiplexing probe sequence under the corresponding phase window.
[0032] S3 calculates the normalized impedance spectrum amplitude sequence based on the complex channel state information, and performs phase dewinding and clock drift removal processing on the phase angle of the complex channel state information to extract the residual group delay sequence characterizing the high-frequency topological features.
[0033] S4. The normalized impedance spectrum amplitude sequence is compared with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum abrupt change features; and the residual group delay sequence is compared with the acquired pre-stored historical reference group delay sequence to extract delay abrupt change features.
[0034] S5 combines the amplitude spectrum mutation features with the discreteness calculated between the zero-point phase window and the peak phase window, and the delay mutation features to perform orthogonal feature classification, and outputs an anti-theft alarm command based on the classification results.
[0035] The following section explains each step in the method for identifying electricity theft, using a specific process as an example.
[0036] Reference Figure 3 In this embodiment, step S1 includes the following sub-steps.
[0037] S101 extracts the power frequency voltage and calculates the instantaneous phase through the voltage zero-crossing detection circuit.
[0038] The impedance characteristics of nonlinear loads in the power grid change periodically and dynamically with the instantaneous amplitude of AC voltage. In order to extract comparable stable characteristics under a fixed voltage condition, a phase synchronization mechanism needs to be introduced.
[0039] In this embodiment, the input terminal of the phase trigger module 100 is connected to a microprocessor, which receives a hardware interrupt signal from the voltage zero-crossing detection circuit. When the power frequency AC voltage crosses zero potential from the negative half-cycle into the positive half-cycle, or from the positive half-cycle into the negative half-cycle, the voltage zero-crossing detection circuit outputs an edge pulse to the microprocessor. The microprocessor distinguishes the zero-crossing direction based on the polarity of the edge pulse or the direction of change of adjacent sampled voltages; when the power frequency AC voltage crosses zero potential from the negative half-cycle into the positive half-cycle, the reference phase corresponding to the zero-crossing node is set to 0°; when the power frequency AC voltage crosses zero potential from the positive half-cycle into the negative half-cycle, the reference phase corresponding to the zero-crossing node is set to 180°.
[0040] For the specific hardware construction of this zero-crossing voltage detection circuit, those skilled in the art can use conventional topologies based on optocoupler isolation or voltage comparators for design. The specific circuit connection relationship is a well-known technology in this field and will not be described in detail here.
[0041] To obtain accurate phase data, the microprocessor uses an internal high-frequency clock timer to record the time of the most recent edge pulse occurrence, defining this time as the zero-crossing moment. At the same time, the microprocessor obtains the current real-time running time. Based on this, combined with a pre-set power frequency cycle And the reference phase corresponding to the zero-crossing node is denoted as The microprocessor calculates the current time. Corresponding instantaneous phase This instantaneous phase The specific formula for calculation is as follows: ; As a preferred method, the power frequency period in the above calculation formula It is usually determined based on the nominal operating frequency of the distribution network. For example, in a 50 Hz AC power grid environment, this power frequency cycle... The timeframe is set to 20 milliseconds. The instantaneous phase calculated from this is... The instantaneous phase increases periodically within the range of 0° to 360°. The value of 2 can effectively reflect the true phase node of the current power grid voltage.
[0042] S102, constructing zero-point phase window and peak phase window based on phase tolerance width.
[0043] Considering the slight physical execution delay during microprocessor instruction scheduling and the power-on preparation of the broadband carrier communication module, it is difficult to complete data transmission immediately at the ideal absolute zero or absolute peak point. By setting a phase interval as the trigger boundary, the error caused by the aforementioned physical execution delay can be covered.
[0044] Specifically, the microprocessor is internally configured with a fixed phase tolerance width. The phase tolerance width Based on the pre-measured execution statistics of the microprocessor from receiving the interrupt instruction to completing the writing of the underlying bus data, and combined with the power frequency, the angle value is converted and the value range is set to 1° to 5°.
[0045] Based on the set phase tolerance width The microprocessor extends the phase tolerance width to both sides, centered at the 0° and 180° nodes. The range of the zero-point phase window is used to construct the zero-point phase window. The physical operating range representing the grid voltage amplitude close to zero volts is expressed mathematically as follows: ; Similarly, the microprocessor constructs peak phase windows near the 90° and 270° nodes. The peak phase window The physical operating range representing the maximum positive or negative voltage amplitude of the power grid is expressed mathematically as follows: ; S103: When the instantaneous phase enters the zero-point phase window or the peak phase window, issue the highest preemption priority command.
[0046] The phase triggering module 100 will calculate the instantaneous phase in real time. With the constructed zero-point phase window and peak phase window Perform boundary comparison. Once satisfied... or The condition indicates that the current grid voltage has reached the specific phase range required for impedance feature extraction. At this point, the microprocessor writes specific register configuration values to the wideband carrier communication module via the underlying bus, thereby issuing the highest preemption priority instruction. Conversely, when it is determined that the condition is not met... And not satisfied When the condition is met, it means that the current grid voltage has not yet reached the required specific phase range. The microprocessor will not trigger the register write operation and will return to continue to obtain the current real-time running time in order to perform the next round of phase calculation and boundary comparison.
[0047] The highest priority preemption instruction is a communication scheduling command used to adjust the execution order of the media access control layer queue inside the broadband carrier communication module, and to promote the current orthogonal frequency division multiplexing probe sequence transmission task to the highest priority level, so as to reduce the occupation of bus and transmission resources by other background polling read tasks and ensure subsequent synchronous transmission.
[0048] S104, based on the highest preemption priority command, the carrier listening mechanism is masked and an orthogonal frequency division multiplexing probe sequence containing multiple subcarriers is sent.
[0049] In conventional communication network environments, broadband carrier communication modules typically employ a collision avoidance carrier sense multiple access (CSMA) mechanism. This mechanism involves the broadband carrier communication module performing a channel idle detection on the power line before transmitting data to determine if other communication signals exist on the power line. When other communication signals or background noise interference are detected on the power line, the broadband carrier communication module enters a random backoff period. This random backoff time may cause the actual transmission time to deviate from the target phase window, affecting phase synchronization sampling.
[0050] Therefore, in this embodiment, when the media access control layer of the broadband carrier communication module receives the highest preemption priority instruction, it will disable the internal channel idle detection by modifying the underlying register configuration, thereby shielding the carrier sense multiple access mechanism with collision avoidance at the hardware execution level.
[0051] After bypassing the carrier sensing mechanism, the physical layer baseband processing circuit of the broadband carrier communication module maps the pre-reserved test data to the frequency domain and generates a physical layer waveform containing multiple subcarriers via inverse discrete Fourier transform. This physical layer waveform is the orthogonal frequency division multiplexing (OFDM) probe sequence. This OFDM probe sequence is composed of multiple subcarriers with mutually orthogonal frequencies superimposed, and can excite the impedance response of the power distribution network over a wide frequency band.
[0052] Subsequently, the analog front-end circuit of the broadband carrier communication module couples the orthogonal frequency division multiplexing (OFDM) probe sequence to the power line inside the metering box for transmission. This operation reduces interference from random backoff delays, enabling the high-frequency probe signal to be injected into the power grid at specific transient moments when the voltage amplitude crosses zero or reaches its peak value.
[0053] Reference Figure 4 In this embodiment, step S2 includes the following sub-steps.
[0054] S201, receive the orthogonal frequency division multiplexing probe sequence, and extract the complex channel state information sequence under the corresponding phase window from the physical layer equalizer register.
[0055] After the orthogonal frequency division multiplexing (OFDM) detection sequence is transmitted or reflected through the physical topology of the distribution network where the metering box is located, its amplitude will attenuate and its phase will shift due to line impedance, parasitic capacitance, and inductance. When the analog front end of the broadband carrier communication module receives the signal, its internal physical layer baseband chip executes channel estimation logic to compensate for the signal distortion. In this embodiment, the information extraction module 200 locks the specific reception time slot of the OFDM detection sequence within the zero-point phase window or the peak phase window based on the time synchronization stamp transmitted by the phase triggering module 100.
[0056] Subsequently, the information extraction module 200 reads the physical layer equalizer register inside the broadband carrier communication module via the microprocessor's direct memory access bus. This physical layer equalizer register is a static random access memory (SRAM) inside the broadband carrier communication module used to store the frequency domain channel estimation response results. By reading the register's address segment, the information extraction module 200 obtains a one-dimensional array composed of the frequency domain response values of each subcarrier arranged in ascending order of subcarrier index, and uses this one-dimensional array as a sequence of complex channel state information. This complex channel state information sequence Includes the configured The physical response vector of each effective subcarrier.
[0057] As a preferred approach, the number of effective subcarriers The configuration is based on the standard frequency band of the broadband carrier communication protocol, and the value typically ranges from 100 to 500. For the specific hardware addressing and reading methods of this physical layer equalizer register, those skilled in the art can refer to the register mapping manual of the corresponding broadband carrier chip to write the bus driver. Its underlying driver protocol is well-known in the field and will not be elaborated upon here.
[0058] S202, separate the in-phase component and quadrature component corresponding to each subcarrier in the complex channel state information sequence.
[0059] The extracted complex channel state information sequence The information extraction module 200 arranges the complex channel state information sequence sequentially in hardware registers in either complex interleaved or fixed-point fractional binary two's complement format. This allows the subsequent feature cleaning module 300 to independently process the amplitude and phase. The components are then broken down.
[0060] The information extraction module iterates through the complex channel state information sequence 200 times. This corresponds to the storage data unit for each subcarrier. The subcarrier's index is set to... ,in The set of values is The information extraction module 200 extracts the information according to the high-low byte mapping rule. Each subcarrier's data unit is split into real data and imaginary data.
[0061] In orthogonal frequency division multiplexing (OFDM) modulation, the real part data represents the signal projection of the OFDM detection sequence in phase with the local oscillator, and is denoted as the in-phase component corresponding to the subcarrier. The imaginary part data represents the signal projection of the orthogonal frequency division multiplexing detector sequence and the local oscillator, and is denoted as the orthogonal component corresponding to the subcarrier. The in-phase component With this orthogonal component They collectively preserved the physical information of energy loss and time delay of each subcarrier during transmission.
[0062] S203, construct a discrete complex transmission function using the in-phase component and the quadrature component corresponding to each subcarrier, and use the discrete complex transmission function as the complex channel state information.
[0063] After completing the component separation of the effective subcarriers across the entire frequency band, the information extraction module 200 establishes a continuous array of floating-point numbers in the microprocessor's cache space. This is achieved by introducing an imaginary unit. For each subcarrier, the sequence number index The information extraction module 200 will separate the in-phase component. With this orthogonal component By combining mathematical formulas, a complex expression for a single subcarrier can be constructed. .
[0064] Subsequently, according to the physical arrangement of subcarrier frequencies from low to high, the information extraction module 200 extracts all... The complex expressions of each subcarrier are combined sequentially to generate a discrete complex transfer function in the frequency domain. The discrete complex transfer function The specific expression is: ; in, To meet The imaginary unit.
[0065] The discrete complex transfer function Presented as a set of discrete complex sequences varying with frequency, it retains the amplitude-frequency and phase-frequency response characteristics of the high-frequency detection signal passing through the metering box topology network under the zero-point phase window or peak phase window. Based on this data structure, the information extraction module 200 will construct the generated discrete complex transfer function. This complex channel state information is passed to the next stage.
[0066] Reference Figure 5 In this embodiment, step S3 includes the following sub-steps.
[0067] S301, calculate the normalized impedance spectrum amplitude sequence based on the complex channel state information.
[0068] To obtain amplitude characteristics less affected by random fluctuations in device communication gain, the feature cleaning module 300 performs amplitude calculations on the complex channel state information. In this embodiment, the feature cleaning module 300 extracts the complex channel state information, namely the discrete complex transfer function constructed upstream. In-phase components corresponding to each subcarrier Orthogonal components .
[0069] The feature cleaning module 300 respectively cleans the in-phase component With this orthogonal component Perform the squaring operation on the in-phase component. The square value and the orthogonal component The square roots of the summation of the squares are then taken. This yields the absolute amplitude of the discrete complex transfer function on the corresponding subcarrier. The absolute amplitude The mathematical expression is: ; In actual power line broadband carrier communication, this absolute amplitude It reflects the transmission admittance characteristics of high-frequency channels in the distribution network.
[0070] Since anti-electricity theft detection focuses more on analyzing line impedance changes, and communication receivers generally have multiplicative errors introduced by automatic gain amplifiers, this feature cleaning module 300 further analyzes the absolute amplitude. Take the reciprocal to generate the normalized impedance spectrum amplitude for each subcarrier. To avoid this absolute amplitude When the value approaches zero, it causes division operations to malfunction. As a preferred approach, a very small positive bias constant is introduced into the denominator. The positive bias constant The value range is 1×10 -6 Up to 1×10 -3 The preferred value is 1×10 -4 The normalized impedance spectrum amplitude The specific calculation formula is as follows: ; The aforementioned reciprocal operation is used to convert the amplitude response corresponding to the complex channel state information into a relative impedance characterization value, and, combined with subsequent historical benchmark comparison, to reduce the linear scaling interference caused by the front-end analog circuit amplifier. This normalized impedance spectrum amplitude is derived from the amplitude of the spectrum of all effective subcarriers. The set of data arranged in ascending order of frequency in the frequency domain constitutes the normalized impedance spectrum amplitude sequence. ,in, The normalized impedance spectrum amplitude sequence physically represents the discrete representation of the equivalent impedance of the entire distribution topology as a function of frequency within the frequency band of the high-frequency detection signal.
[0071] S302, perform phase dewinding processing on the phase angle of the complex channel state information.
[0072] Besides amplitude characteristics, the phase changes caused by the propagation and reflection of high-frequency signals in the medium are important parameters reflecting the branch distances of the distribution network topology. The feature cleaning module 300 calculates the orthogonal components corresponding to each subcarrier in the discrete complex transfer function. With the in-phase component The ratio is calculated, and the arctangent of the ratio is performed.
[0073] To avoid this in-phase component The computational anomaly caused by a value of zero is typically addressed by those skilled in the art through programming implementations using an arctangent function that supports four-quadrant output. Through this arctangent operation, the feature cleaning module 300 calculates the phase angle of the complex channel state information, which is specifically represented in the data as a generated packaged phase angle sequence. Its theoretical expression is: ; Because the range of the standard arctangent function is limited to... Within a given interval, when the cumulative phase angle of actual physical propagation exceeds the boundary of that interval, the calculated phase value will change. The abrupt change disrupts the physical continuity of the phase frequency characteristics. Therefore, the feature cleaning module 300 cleans the packaged phase angle sequence... The phase difference between adjacent subcarriers is smoothed and compensated.
[0074] The feature cleaning module 300 calculates the phase difference between adjacent subcarriers. When it determines that the phase difference is greater than a set positive transition threshold, it subtracts the phase difference of the current and subsequent subcarriers. When it is determined that the phase difference is less than the set negative transition threshold, the phase of the current and subsequent subcarriers is increased. When it is determined that the phase difference is less than or equal to the positive transition threshold and greater than or equal to the negative transition threshold, the current phase angle value remains unchanged.
[0075] The positive transition threshold is typically set to π, and the negative transition threshold is typically set to... After the above traversal judgment and jump compensation operations, the wrapped phase data is converted into continuous phase data, thereby generating a continuous unwinding phase angle sequence. .
[0076] S303, perform clock drift removal processing on the phase angle of the complex channel state information, and extract the residual group delay sequence that characterizes the high-frequency topological features.
[0077] Although this unwinding phase angle sequence Continuity was restored, but the broadband carrier communication module experienced crystal oscillator frequency deviation and sampling time offset during transmission and reception. These asynchronous factors introduce linear error interference proportional to the frequency across the entire phase-frequency curve. To isolate this linear error interference, the feature cleaning module 300 cleans the unwound phase angle sequence. Perform polynomial fitting.
[0078] In this embodiment, the feature cleaning module 300 calculates the unwinding phase angle sequence using the least squares method. The linear slope and intercept vary with frequency, from which the linear trend phase angle component is extracted. Subsequently, the unwinding phase angle sequence was... Subtract the phase angle component of the linear trend item by item The residual phase angle sequence after removing crystal oscillator clock drift interference is obtained. : ; In physics, group delay is defined as the derivative of the phase angle with respect to the diagonal frequency, representing the propagation time of the signal envelope in the medium. In discrete data systems, this feature cleaning module 300 processes the residual phase angle sequence... Find the discrete derivatives of the frequencies of adjacent subcarriers. Assume the frequency interval between adjacent subcarriers is . This frequency interval The frequency interval is determined by the quotient of the standard bandwidth of the broadband carrier communication protocol and the total number of subcarriers. Set to 24.4 kHz. To avoid array out-of-bounds errors, for the index... From 1 to ( The discrete derivative formula executed by the feature cleaning module 300 within the calculation interval (for the number of effective subcarriers) is as follows: ; Discrete derivative values calculated from each subcarrier position The sequence formed by these two groups is the residual group delay sequence. ,in, .
[0079] The residual group delay sequence removes the global background bias caused by asynchronous transmission and reception, and in a physical sense can characterize the microscopic reflection time delay of high-frequency signals caused by the local impedance change of the metering box and its external wires, providing spatial state characteristics for subsequent illegal topology identification.
[0080] Reference Figure 6 In this embodiment, step S4 includes the following sub-steps.
[0081] S401, compare the normalized impedance spectrum amplitude sequence with the acquired pre-stored historical reference impedance spectrum amplitude sequence, extract the amplitude spectrum abrupt change features and lock the frequency domain center point.
[0082] Before performing data differential comparison, the mutation extraction module 400 reads the frequency domain impedance baseline data in the factory-configured state without abnormalities from the non-volatile memory of the microprocessor through internal addressing instructions, and uses it as the acquired pre-stored historical reference impedance spectrum amplitude sequence.
[0083] As a preferred method, the process of obtaining the pre-stored historical reference impedance spectrum amplitude sequence is as follows: Based on the standard internal wiring impedance parameters and parasitic parameters measured before the metering box of this model leaves the factory, the theoretical impedance amplitude corresponding to each subcarrier in the entire frequency band is calculated, and the calculated one-dimensional theoretical impedance data is pre-programmed into the non-volatile memory of the microprocessor, serving as the pre-stored historical reference impedance spectrum amplitude sequence. For example, when the application environment is a metering box with a purely resistive matching standard, the historical reference impedance spectrum amplitude corresponding to each subcarrier is uniformly set to a normalization constant of 1.0.
[0084] After acquiring the baseline data, the mutation extraction module 400 performs a difference operation. In this embodiment, the mutation extraction module 400 compares the normalized impedance spectrum amplitude sequence currently output by the feature cleaning module 300 with the acquired pre-stored historical reference impedance spectrum amplitude sequence, and records the acquired pre-stored historical reference impedance spectrum amplitude sequence as... ,in, The Middle The historical reference impedance spectral amplitude corresponding to each subcarrier is denoted as For each subcarrier index The mutation extraction module 400 performs point-by-point subtraction and takes the absolute value. Thus, the mutation extraction module 400 generates a sequence of absolute amplitude differences to characterize the degree of frequency domain impedance deviation. The calculation formula is as follows: ; The equivalent impedance exhibited by a nonlinear load at the zero-crossing point and the peak point of the AC voltage differs significantly. To extract this nonlinear feature, the abrupt change extraction module 400 performs extreme value extraction in conjunction with the current sampling phase window type, and determines the zero-point amplitude abrupt change peak and the peak amplitude abrupt change peak according to the phase window type corresponding to the input data.
[0085] In this embodiment, the mutation extraction module 400 executes the following branch decision logic: when the input data corresponds to the zero-point phase window, the mutation extraction module 400 traverses the absolute difference sequence of amplitudes and extracts the maximum difference point, defining it as the zero-point amplitude mutation peak. It is temporarily stored in the current system cache, waiting for the peak phase window data within the same monitoring period.
[0086] Similarly, when the input data corresponds to the peak phase window, the abrupt change extraction module iterates through the absolute difference sequence of amplitudes 400 times and extracts the point of maximum difference, defining it as the peak amplitude abrupt change peak. .
[0087] Acquiring paired zero-point amplitude abrupt change peaks within the same monitoring period With the peak amplitude abrupt change peak Then, the mutation extraction module 400 performs a joint threshold determination: when the zero-point amplitude mutation peak is determined... The peak amplitude is less than or equal to the set zero-point amplitude anomaly threshold, and at the same time, the peak amplitude changes abruptly. If the peak amplitude is less than or equal to the set abnormal threshold, it is determined that the current topology has not changed abnormally. The output of features to subsequent stages is stopped, the current system's cached computational data is cleared, and the anti-theft detection system is triggered to return to step S1 to wait for the next instantaneous phase trigger of the grid voltage, thus entering the next anti-theft monitoring cycle. If it is determined that at least one of the two values is greater than the corresponding abnormal amplitude threshold, an impedance change event is confirmed, and the subcarrier index containing the larger value is locked as the target subcarrier index. .
[0088] As a preferred approach, the aforementioned zero-point amplitude anomaly threshold and peak amplitude anomaly threshold are determined based on the statistical standard deviation of amplitude difference data collected and calculated multiple times under legal conditions during the initial operation of the metering box. The values are typically 3 to 5 times this statistical standard deviation to avoid interference from white noise in the conventional power line environment. For example, considering the normalized data dimensions, the zero-point amplitude anomaly threshold is specifically set to 0.3, and the peak amplitude anomaly threshold is specifically set to 0.5.
[0089] After locking onto the aforementioned abrupt change peak, and combining it with the pre-configured broadband carrier communication start frequency... and the frequency spacing between adjacent subcarriers Calculate the physical frequency point corresponding to the current phase window and record this physical frequency point as the center frequency. As a preferred method, the starting frequency of this broadband carrier communication The center frequency can be set according to the selected carrier frequency band standard, for example, set to 2 MHz. The calculation formula is: ; The zero-point amplitude abrupt change peak, the peak amplitude abrupt change peak, and the corresponding center frequency together reflect the specific intensity and frequency domain location of the impedance anomaly. The abrupt change extraction module 400 extracts the zero-point amplitude abrupt change peaks generated within the same monitoring period. Peak amplitude sudden change peak and the corresponding center frequency The data are combined, and the combination of the three is used as the characteristic of the amplitude spectrum mutation.
[0090] S402, compare the residual group delay sequence with the acquired pre-stored historical benchmark group delay sequence, and extract the delay mutation feature.
[0091] After locking onto the core frequency point of impedance anomaly in the frequency domain, the mutation extraction module 400 further quantifies the change in physical spatial distance caused by illegal branch lines and performs comparison and extraction operations of spatial delay characteristics. The mutation extraction module 400 reads the spatial delay baseline data under the abnormal state that was fixed at the factory configuration from the non-volatile memory of the microprocessor and uses it as the pre-stored historical reference group delay sequence.
[0092] As a preferred method, the acquisition process of the pre-stored historical reference group delay sequence is as follows: Based on the factory design length of the physical conductors inside the metering box and the theoretical propagation speed of high-frequency signals in the conductor medium, the theoretical inherent group delay without external illegal branches is calculated. This theoretical inherent group delay is then allocated to each subcarrier node to generate a constant baseline array, which is pre-stored in the non-volatile memory of the microprocessor, thus serving as the pre-stored historical reference group delay sequence. For example, in an ideal state without external physical branches, the historical reference group delay corresponding to each subcarrier is uniformly set to 0 nanoseconds.
[0093] The mutation extraction module 400 compares the real-time acquired residual group delay sequence with the pre-stored historical benchmark group delay sequence (denoted as...). ,in, The Middle The historical reference group delay value corresponding to each subcarrier is denoted as The comparison is performed by subtracting each corresponding subcarrier item from the above-mentioned subtraction operation. Through the above subtraction operation, the mutation extraction module 400 generates a group delay difference value sequence that filters out the inherent reflection delay of the legitimate network. Its mathematical expression is: ; Considering that full-band broadband background noise may cause computational interference to the time delay data in low signal-to-noise ratio bands, the mutation extraction module 400 utilizes the target subcarrier index under the current phase window recorded in step S401. Determine the group delay mapping index ;when Less than hour, ;when equal hour, Subsequently, the mutation extraction module 400 extracts the group delay difference sequence. Targeted mapping extraction is performed to improve spatial resolution.
[0094] Specifically, the mutation extraction module 400 directly reads the group delay difference sequence. The group delay mapping index The corresponding value, that is The obtained data is defined as the group delay mutation value under the current phase window. This group delay mutation value quantitatively characterizes the delay time of the microscopic reflected signal caused by the illegal connection of external conductors. Considering that the spatial delay caused by passive physical wiring is independent of the power frequency voltage phase, in order to unify the data dimension of subsequent orthogonal classification, the mutation extraction module 400 takes the average of the two group delay mutation values calculated under the zero-point phase window and the peak phase window within the same monitoring period, defines it as the final group delay mutation value, and uses this group delay mutation value as the only delay mutation feature output to the subsequent stage. The mutation extraction module 400 uses this delay mutation feature together with the frequency domain feature obtained in step S401 to form a complete feature dataset for subsequent identification logic to call.
[0095] Reference Figure 7 In this embodiment, step S5 includes the following sub-steps.
[0096] S501, using the amplitude spectrum abrupt change characteristics, calculates the dispersion between the zero-point phase window and the peak phase window.
[0097] To quantify the degree of nonlinear change in the external load with respect to the AC voltage phase, it is necessary to calculate the differences in amplitude abrupt changes. The joint decision module 500 acquires the amplitude spectrum abrupt change characteristics output from the preceding steps and extracts the zero-point amplitude abrupt change peak corresponding to the zero-point phase window. and the corresponding peak amplitude abrupt change peak of the peak phase window In this embodiment, the joint decision module 500 calculates the peak amplitude abrupt change peak. With the zero-point amplitude abrupt peak The absolute difference between them.
[0098] Because the reference impedance and line attenuation levels differ between different power distribution branches, the aforementioned absolute difference is easily affected by transmission gain drift. To eliminate the proportional error caused by this attenuation, the joint decision module 500 compares the peak amplitude abrupt change peak. With the zero-point amplitude abrupt peak The maximum value between the two values is extracted as the denominator constraint term. Since the extraction of abrupt peaks requires exceeding a set anomaly threshold, this denominator constraint term is always greater than zero, preventing computational anomalies caused by a zero denominator. The joint decision module 500 divides the calculated absolute difference by this denominator constraint term, thereby calculating the dispersion. This dispersion The specific calculation formula is as follows: ; The calculated dispersion This is a dimensionless parameter characterizing state changes. At the physical level, this dispersion characterizes the impedance fluctuations of an unknown connected device causing impedance anomalies under different phases of AC power. Typically, legitimate nonlinear loads exhibit significant differences in conduction state between zero-crossing and peak points, resulting in higher dispersion values; while passive short-circuited conductors or purely resistive loads show relatively constant impedance across the entire voltage phase, leading to lower dispersion values.
[0099] S502, combine the dispersion and the delay mutation feature to perform orthogonal feature classification, and output the anti-theft alarm command based on the classification result.
[0100] Obtain this dispersion that characterizes the impedance fluctuation properties. Subsequently, the joint decision module 500 retrieves the group delay mutation value from the delay mutation feature. The dispersion and the group delay abrupt change value reflect the topological state from two independent dimensions: the degree of nonlinearity of AC impedance and the reflection distance in the transmission space. Together, they constitute an orthogonal feature space for classification. The joint decision module 500, within this orthogonal feature space, applies the dispersion... The group's time-delay mutation value is used as the first coordinate axis variable. As the second coordinate axis variable, coordinate quadrant division is performed to classify this orthogonal feature.
[0101] To avoid misjudgments caused by single-dimensional feature determination, the joint decision module 500 incorporates the discreteness... Compared with the delayed mutation value of this group The orthogonal feature classification is performed by substituting the values into a preset threshold judgment logic. In this embodiment, the preset threshold judgment logic is configured with a discrete threshold. With time-delayed mutation threshold .
[0102] As a preferred approach, this dispersion threshold The threshold is set based on the statistical lower limit of the dispersion of a known legal nonlinear load, for example, the dispersion threshold is set between 0.4 and 0.6; the delay mutation threshold is set accordingly. The delay drift of the physical circuit group is determined based on the upper limit of the delay drift caused by normal ambient temperature fluctuations. For example, the threshold for this delay change is set to 100 nanoseconds to filter out environmental interference.
[0103] Based on the above threshold configuration, the joint decision module 500 executes the following classification and determination logic: When the dispersion is determined Greater than or equal to the dispersion threshold This indicates that the source causing the impedance change has phase-dependent and nonlinear conduction characteristics. The joint decision module 500 classifies it as a legitimate nonlinear load start-stop event and determines that there is no electricity theft on the current network.
[0104] When the dispersion is determined Less than this dispersion threshold And the delayed mutation value of this group Less than or equal to the delay threshold This indicates that the unknown access device exhibits linear impedance and has not created any new spatial conductor reflection paths outside the original topology. The joint decision module 500 classifies this as a local, in-situ, legitimate purely resistive load start-stop event, determining that there is no electricity theft on the current network.
[0105] When the dispersion is determined Less than this dispersion threshold And at the same time, the delayed mutation value of this group Greater than the delay threshold This indicates that the source of the anomaly not only exhibits stable impedance across the entire phase but also elongates the physical reflection path of the high-frequency detection signal. This state matches the spatial topology of a passive long wire bypassing the metering device.
[0106] Based on the calculation results of the above classification and judgment logic, when the feature combination points to a passive long conductor, the joint decision module 500 generates and outputs the anti-theft alarm command to the peripheral control unit. This anti-theft alarm command includes the abnormal occurrence time, abnormal frequency parameters, and a logical classification identifier. The abnormal frequency parameters are the center frequency extracted by the joint decision module 500 from the amplitude spectrum mutation feature. This alarm command triggers the microprocessor or broadband carrier communication module to execute corresponding protection actions or data reporting actions. This classification process, combining electromagnetic and spatial physical parameters to identify abnormal topologies, can improve the reliability of metering status monitoring.
[0107] To further illustrate the implementation process and technical effects of this invention, the following description is provided in conjunction with specific application scenarios. The specific numerical values and scenario parameters described below are only used to explain the implementation principle of this invention and do not limit the scope of protection of this invention.
[0108] The nominal operating frequency of the distribution network is 50 Hz, corresponding to the power frequency cycle. Configured to 20 milliseconds. Wideband carrier communication start frequency. Configured at 2000 kHz, with adjacent subcarrier frequency spacing Configured at 24.4 kHz, number of effective subcarriers The configuration is set to 100. The historical reference impedance spectrum amplitude value corresponding to each subcarrier in the pre-stored historical reference impedance spectrum amplitude sequence retrieved from the microprocessor's non-volatile memory is denoted as... All are configured to a constant of 1.0; in the pre-stored historical reference group delay sequence, the delay value corresponding to each subcarrier is recorded as... All parameters are configured to 0 nanoseconds. The physical topology scenario is as follows: an external person bypasses the metering device and connects an additional long passive wire.
[0109] The microprocessor acquires the instantaneous phase of the power frequency voltage in the metering box. When the instantaneous phase enters the preset zero-point phase window, the microprocessor issues the highest preemption priority instruction to trigger the broadband carrier communication module to send the orthogonal frequency division multiplexing detection sequence.
[0110] The information extraction module 200 receives the orthogonal frequency division multiplexing (OFDM) probe sequence and extracts the complex channel state information under the corresponding phase window. Indexed by sequence number. Taking a subcarrier as an example, separate the in-phase component corresponding to that subcarrier. 0.5, orthogonal components The value is 0.5. The information extraction module 200 constructs a complex expression for a single subcarrier and generates the discrete complex transmission function corresponding to that point. .
[0111] The feature cleaning module 300 performs squaring operations on the in-phase component and the quadrature component respectively, sums them, and then performs square root operations to generate the absolute amplitude on the corresponding subcarrier. The mathematical expression is: ; Introducing a positive bias constant The feature cleaning module 300 further calculates the normalized impedance spectrum amplitude. The calculation formula is as follows: ; The mutation extraction module 400 performs point-by-point subtraction between the normalized impedance spectrum amplitude sequence and the acquired pre-stored historical reference impedance spectrum amplitude sequence, and takes the absolute value. The generated absolute difference sequence of amplitudes numerical value for: ; The mutation extraction module iterates through the absolute difference sequence of amplitudes 400 times and extracts the point of maximum difference. Data comparison shows that this point of maximum difference occurs at... At that point, its value is 0.4140. The mutation extraction module 400 defines it as a zero-point amplitude mutation peak. The data is temporarily stored in the current system cache, awaiting peak phase window data within the same monitoring period. Simultaneously, the feature cleaning module 300 processes the phase angle to extract the residual group delay sequence, and the mutation extraction module 400 reads the calculated group delay mutation value under the current phase window at that point, which is 150 nanoseconds.
[0112] Within the same monitoring period, when the instantaneous phase enters the preset peak phase window, the broadband carrier communication module triggers the transmission of the orthogonal frequency division multiplexing (OFDM) probe sequence again. The information extraction module 200 separates the data within the peak phase window. In-phase component corresponding to subcarrier The value is 0.495, and the orthogonal component is... The value is 0.495. The feature cleaning module 300 calculates the absolute amplitude of the corresponding subcarrier. The normalized impedance spectrum amplitude is 0.7000. The value is 1.4283. The absolute difference in amplitude generated by the mutation extraction module 400 is compared with... It is 0.4283.
[0113] The mutation extraction module iterates through the absolute difference sequence of amplitudes 400 times and extracts the point of maximum difference. Data comparison shows that this point of maximum difference also appears in... At that point, its value is 0.4283. The mutation extraction module 400 defines it as the peak amplitude mutation peak. The group delay abrupt change value at this point under the current phase window is 140 nanoseconds.
[0114] The mutation extraction module 400 performs joint threshold determination, with the zero-point amplitude anomaly threshold configured at 0.3 and the peak amplitude anomaly threshold configured at 0.5. Due to the zero-point amplitude mutation peak... (0.4140) is greater than the set zero-point amplitude anomaly threshold (0.3), satisfying the condition that at least one of them is greater than the corresponding amplitude anomaly threshold, confirming the existence of an impedance abrupt change event. Comparing the two values, the peak amplitude abrupt change peak... (0.4283) is the larger value, and the mutation extraction module 400 locks the subcarrier index containing this value as the target subcarrier index. .
[0115] Solve for the corresponding center frequency The calculation formula is: kilohertz.
[0116] The mutation extraction module 400 will extract the generated zero-point amplitude mutation peak. Peak amplitude sudden change peak and the corresponding center frequency The data is combined to represent the abrupt change characteristics of the amplitude spectrum. Simultaneously, the abrupt change extraction module 400 determines the group delay mapping index. The average of the two group delay abrupt change values (150 nanoseconds and 140 nanoseconds) calculated under the zero-point phase window and peak phase window within the same monitoring period, which is 145 nanoseconds, is defined as the final group delay abrupt change value. This group delay abrupt change value is then used as the sole delay abrupt change feature output to the subsequent stage. .
[0117] The joint decision module 500 uses this amplitude spectrum abrupt change characteristic to calculate the dispersion. The specific calculation formula is as follows: ; The joint decision module 500 retrieves the group delay mutation value from the delay mutation feature. (145 nanoseconds), substituted into the preset threshold judgment logic for orthogonal feature classification. Discreteness threshold. Configured to 0.5, delay mutation threshold The configuration is set to 100 nanoseconds.
[0118] The joint decision module 500 determines the dispersion. (0.0334) is less than the dispersion threshold. (0.5), and at the same time, the delayed mutation value of this group (145 nanoseconds) is greater than the delay threshold. (100 nanoseconds). Based on the calculation results of the above classification and judgment logic, the feature combination points to the passive long wire. The joint decision module 500 generates and outputs the anti-theft identification alarm command to the peripheral control unit. Among them, the abnormal frequency parameter included in the anti-theft identification alarm command is the center frequency of 3000.4 kHz extracted from the amplitude spectrum change feature.
[0119] By combining the above workflow and data comparison, the system effectively processes and extracts impedance spectrum features, and through multi-dimensional parameter determination, outputs the identification result corresponding to the target topology, thus realizing the anti-theft identification of the metering box.
[0120] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.
Claims
1. A method for identifying electricity theft prevention in metering boxes based on impedance spectrum characteristics, characterized in that, The method includes: The instantaneous phase of the power frequency voltage inside the metering box is acquired, and when the instantaneous phase enters the preset zero-point phase window or peak phase window, an orthogonal frequency division multiplexing detection sequence is triggered. Extract the complex channel state information of the orthogonal frequency division multiplexing probe sequence under the corresponding phase window; The normalized impedance spectrum amplitude sequence is calculated based on the complex channel state information; the phase angle of the complex channel state information is subjected to phase dewinding and clock drift removal processing to extract the residual group delay sequence characterizing the high-frequency topological features; The normalized impedance spectrum amplitude sequence is compared with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum abrupt change features; the residual group delay sequence is compared with the acquired pre-stored historical reference group delay sequence to extract delay abrupt change features. The dispersion calculated between the zero-point phase window and the peak phase window by combining the amplitude spectrum mutation feature and the delay mutation feature are used to perform orthogonal feature classification, and an anti-theft alarm command is output based on the classification result.
2. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 1, characterized in that, The step of acquiring the instantaneous phase of the power frequency voltage inside the metering box, and triggering the transmission of an orthogonal frequency division multiplexing detection sequence when the instantaneous phase enters a preset zero-point phase window or peak phase window includes: The power frequency voltage is extracted and the instantaneous phase is calculated using a voltage zero-crossing detection circuit; The zero-point phase window and the peak-point phase window are constructed based on the phase tolerance width; When the instantaneous phase enters the zero-point phase window or the peak phase window, a highest preemption priority command is issued. Based on the highest preemption priority instruction, a carrier listening mechanism is masked and the orthogonal frequency division multiplexing detection sequence containing multiple subcarriers is sent.
3. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 2, characterized in that, The steps of extracting the power frequency voltage and calculating the instantaneous phase through the voltage zero-crossing detection circuit include: When the power frequency voltage crosses the zero potential, an edge pulse is output. The zero-crossing direction is distinguished according to the edge pulse, and a reference phase corresponding to the zero-crossing node is set. The time of the most recent occurrence of the edge pulse is recorded using a clock timer as the zero-crossing moment; Obtain the current real-time running time and calculate the time difference between the current time and the zero-crossing time; Divide the time difference by the preset power frequency cycle to obtain the ratio, and multiply the ratio by 360 degrees to obtain the multiplication result; After adding the multiplication result to the reference phase, a modulo operation is performed on 360 degrees to calculate the instantaneous phase.
4. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 1, characterized in that, The step of extracting the complex channel state information of the orthogonal frequency division multiplexing probe sequence under the corresponding phase window includes: Receive the orthogonal frequency division multiplexing probe sequence and extract the complex channel state information sequence under the corresponding phase window from the physical layer equalizer register; Separate the in-phase and quadrature components corresponding to each subcarrier in the complex channel state information sequence; A discrete complex transmission function is constructed using the in-phase component and the quadrature component corresponding to each subcarrier, and the discrete complex transmission function is used as the complex channel state information.
5. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 4, characterized in that, The step of calculating the normalized impedance spectrum magnitude sequence based on the complex channel state information includes: Extract the in-phase component and the quadrature component corresponding to each subcarrier in the discrete complex transmission function; The square operation is performed on the in-phase component and the quadrature component respectively. The square value of the in-phase component and the square value of the quadrature component are summed and then the square root is taken to generate the absolute amplitude of the discrete complex transfer function on each subcarrier. The positive bias constant is added to the absolute amplitude to generate a sum, and the reciprocal of the sum is taken to generate the normalized impedance spectrum amplitude corresponding to each subcarrier. The normalized impedance spectrum amplitudes corresponding to each subcarrier are arranged in order of frequency from low to high to form the normalized impedance spectrum amplitude sequence.
6. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 4, characterized in that, The steps of performing phase dewinding and clock drift removal processing on the phase angle of the complex channel state information to extract the residual group delay sequence characterizing the high-frequency topological features include: Calculate the ratio of the orthogonal component to the in-phase component and perform an arctangent operation to generate a wrapping phase angle sequence; Calculate the phase difference between adjacent subcarriers in each subcarrier of the wrapped phase angle sequence, perform jump compensation on the phase difference based on positive jump threshold and negative jump threshold, and convert the wrapped phase angle sequence into an unwound phase angle sequence; The linear trend phase angle component is extracted by polynomial fitting of the unwound phase angle sequence, and the linear trend phase angle component is subtracted from the unwound phase angle sequence item by item to obtain the residual phase angle sequence. The discrete derivatives of the adjacent subcarrier frequencies in the residual phase angle sequence are calculated, and the calculated discrete derivative values are used to construct the residual group delay sequence.
7. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 1, characterized in that, The step of comparing the normalized impedance spectrum amplitude sequence with the acquired pre-stored historical reference impedance spectrum amplitude sequence to extract amplitude spectrum abrupt change features includes: The normalized impedance spectrum amplitude sequence is subtracted point by point from the acquired pre-stored historical reference impedance spectrum amplitude sequence, and the absolute value is taken to generate an amplitude absolute difference sequence. When the input data corresponds to the zero-point phase window, the sequence of absolute difference in amplitude is traversed and the point of maximum difference is extracted, which is defined as the zero-point amplitude abrupt change peak within the same monitoring period; When the input data corresponds to the peak phase window, the absolute difference sequence of amplitude is traversed and the maximum difference point is extracted, which is defined as the peak amplitude sudden change peak within the same monitoring period; When the zero-point amplitude mutation peak is greater than the preset zero-point amplitude abnormality threshold or the peak amplitude mutation peak is greater than the preset peak amplitude abnormality threshold, the subcarrier index containing the one with the larger value between the zero-point amplitude mutation peak and the peak amplitude mutation peak is locked as the target subcarrier index. By combining the pre-configured broadband carrier communication start frequency and the frequency interval between adjacent subcarriers, the center frequency corresponding to the target subcarrier index is calculated. The combination of the zero-point amplitude mutation peak, the peak amplitude mutation peak, and the corresponding center frequency within the same monitoring period is used as the amplitude spectrum mutation feature.
8. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 7, characterized in that, The step of comparing the residual group delay sequence with the acquired pre-stored historical benchmark group delay sequence to extract delay mutation features includes: The residual group delay sequence is subtracted from the pre-stored historical benchmark group delay sequence item by item to generate a group delay difference sequence. The target subcarrier index is used to determine the group delay mapping index, and the value corresponding to the group delay mapping index in the group delay difference sequence is read and defined as the group delay mutation value under the current phase window; The average value of the two group delay mutation values calculated under the zero-point phase window and the peak phase window within the same monitoring period is defined as the final group delay mutation value, and the final group delay mutation value is used as the delay mutation feature.
9. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 8, characterized in that, The step of performing orthogonal feature classification by combining the dispersion calculated from the amplitude spectrum change feature between the zero-point phase window and the peak phase window with the delay change feature, and outputting an anti-theft alarm command based on the classification result includes: The zero-point amplitude mutation peak and the peak amplitude mutation peak are extracted from the amplitude spectrum mutation characteristics; Calculate the absolute difference between the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak; Compare the numerical values of the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak, and extract the maximum value between the peak amplitude abrupt change peak and the zero-point amplitude abrupt change peak as the denominator constraint term; Divide the absolute difference by the denominator constraint term to calculate the dispersion. Using the dispersion as the first coordinate axis variable and the final group delay mutation value as the second coordinate axis variable, an orthogonal feature space is constructed. The orthogonal feature space is substituted with a discreteness threshold and a delay mutation threshold to perform orthogonal feature classification, and the anti-theft alarm command is output based on the classification result.
10. The method for identifying anti-theft electricity boxes based on impedance spectrum characteristics according to claim 9, characterized in that, The step of performing orthogonal feature classification by substituting a discreteness threshold and a delay mutation threshold into the orthogonal feature space, and outputting the anti-theft alarm command based on the classification result includes: When the dispersion is determined to be greater than or equal to the dispersion threshold, it is classified as a legitimate nonlinear load start-up and shutdown event, and it is determined that there is no electricity theft behavior on the current network. When it is determined that the dispersion is less than the dispersion threshold and the final group delay mutation value is less than or equal to the delay mutation threshold, it is classified as a local in-situ legal pure resistive load start-up and shutdown event, and it is determined that there is no electricity theft behavior in the current network. When it is determined that the dispersion is less than the dispersion threshold and the final group delay mutation value is greater than the delay mutation threshold, it is classified as an event of bypassing the metering device to connect a passive long wire, and the anti-theft identification alarm command is generated and output.
