A power metering box anti-stealing electricity monitoring method and system

Abstract

The present application relates to the field of electricity larceny prevention monitoring, and particularly relates to an electricity larceny prevention monitoring method and system for an electric energy metering box, which comprises the following steps: synchronously collecting voltage and current original high-frequency sequences and obtaining three-phase total active power within a set basic time window; intercepting a target analysis sequence from the voltage sequence and obtaining multiple sliding analysis sequences by sliding the current sequence within a time delay range, and determining the voltage time sequence delay degree of the current window through time domain correlation comparison; constructing a delay degree historical time sequence based on a sliding step, constructing a discrete probability distribution and calculating a current response rigidity degree; calculating the current response rigidity degree and three-phase active power imbalance degree calculated based on the three-phase total active power to obtain a box electricity larceny confidence degree; and finally outputting an electricity larceny prevention monitoring alarm result through threshold judgment. The present application effectively strips the data boundary between normal load fluctuation and electricity larceny behavior, and improves the monitoring accuracy.

Description

Technical Field

[0001] This invention relates to the field of electricity theft monitoring. In particular, it relates to a method and system for monitoring electricity theft from electricity metering boxes. Background Technology

[0002] With the application of power big data technology in the distribution network, analyzing abnormal electricity consumption behavior using electrical sequence data collected from electricity metering boxes has become a technical means to achieve refined management and anti-theft monitoring of low-voltage distribution networks. In processing massive amounts of user-side data, identifying electricity theft by mining the abnormal characteristics of electrical parameters such as three-phase voltage and current is a current research direction in this field.

[0003] For example, Chinese patent application CN114881808B discloses a method for accurate identification of electricity theft based on big data and an anti-theft system. The method uses intelligent collection of electricity consumption data and multi-dimensional indicators such as the calculated imbalance rate, power factor and contract capacity ratio to evaluate and predict the suspicion coefficient of electricity theft. It also sets personalized static indicator warning ranges for different types of users.

[0004] However, this technology still primarily relies on steady-state sequence components and static indicators such as voltage and current amplitudes, or three-phase imbalance rates derived from these amplitudes, to construct anti-theft criteria. In the actual operating conditions of low-voltage distribution networks, modern users are heavily connected to single-phase high-power loads and nonlinear rectifier loads. These loads exhibit randomness and asymmetry in the time domain, and their normal operational fluctuations themselves induce high-frequency three-phase imbalance phenomena at the electricity metering box. Typically, the changes in three-phase voltage and current caused by load fluctuations are random and dynamic; however, due to the connection of abnormal impedances, electricity theft leads to a fixed temporal sequence and response relationship between the voltage of a specific phase and the currents of the other phases. Because the aforementioned technology cannot extract this inter-phase temporal correspondence, normal electricity consumption and abnormal electricity theft data are severely confused, resulting in high false alarm rates, high false negative rates, and low accuracy in anti-theft monitoring. Summary of the Invention

[0005] To address the problem of low accuracy in anti-theft monitoring caused by the inability to extract the timing correspondence between phases in existing technologies, this invention provides solutions in the following aspects.

[0006] In the first aspect, a method for monitoring electricity theft in an electricity metering box includes:

[0007] Within the set basic time window, the original high-frequency voltage sequence and the original high-frequency current sequence of the power metering box are synchronously collected, and the total three-phase active power of the power metering box is obtained.

[0008] The target analysis sequence is extracted from the original high-frequency voltage sequence, and the original high-frequency current sequence is slid within a set time delay range to obtain multiple sliding analysis sequences; the target analysis sequence is compared with each of the sliding analysis sequences in the time domain to determine the voltage timing delay degree of the current base time window.

[0009] The voltage timing delay is obtained by sliding along the time axis to acquire multiple consecutive historical basic time windows, and a historical time series of delay is constructed. A discrete probability distribution is constructed based on the occurrence frequency of each different voltage timing delay value in the historical time series of delay, and the current response stiffness is calculated based on the information entropy of the discrete probability distribution.

[0010] The three-phase active power imbalance is calculated based on the total three-phase active power. The current response stiffness is then combined with the three-phase active power imbalance to obtain the confidence level of electricity theft in the enclosure.

[0011] Determine whether the confidence level of the electricity theft in the enclosure is greater than a preset alarm threshold. If it is, output the anti-electricity theft monitoring alarm result.

[0012] Preferably, the method for obtaining the target analysis sequence includes:

[0013] The edge buffer of the symmetrical region at the beginning and end of the original high-frequency voltage sequence is removed, and the retained central segment is taken as the target analysis sequence; wherein, the span of the edge buffer is determined according to the preset maximum number of discrete time delay search points.

[0014] Preferably, the step of sliding the original high-frequency current sequence within a set time delay range to obtain multiple sliding analysis sequences includes:

[0015] Let the delay factor be any discrete integer within the delay range, and slide point by point within the delay range;

[0016] Whenever the time delay factor slides to a certain discrete integer, a sequence of the corresponding length is extracted from the original high-frequency current sequence to obtain a sliding analysis sequence. All of the multiple sliding analysis sequences are obtained by traversing through the sequence.

[0017] Preferably, determining the voltage timing delay of the current base time window includes:

[0018] The target analysis sequence is paired with each of the sliding analysis sequences point by point, and the discrete mutual information value is calculated for each of the time delay factor states.

[0019] By iterating through all the aforementioned delay factors, the delay factor corresponding to the point where the discrete mutual information value reaches its maximum value is determined as the voltage timing delay degree of the current base time window.

[0020] Preferably, the historical time series of the delay includes:

[0021] A first-in-first-out (FIFO) buffer queue with a preset length is established, using the same length as the base time window as the sliding step. Whenever the next sliding step is entered and the voltage timing delay of the current base time window is output, the value is stored in the FIFO buffer queue. When the amount of stored data reaches the preset length, the earliest stored voltage timing delay in the FIFO buffer queue is removed. All voltage timing delays currently retained in the FIFO buffer queue are used as the historical time series of the delay.

[0022] Preferably, the method for obtaining the current response stiffness includes:

[0023] The frequency of occurrence of each different voltage timing delay value in the historical time series of the delay is statistically analyzed to construct the corresponding discrete probability distribution;

[0024] The statistical information entropy of the discrete probability distribution is calculated, and the reciprocal of the statistical information entropy is taken as the independent variable. The current response stiffness is then calculated by constructing a negative exponential function.

[0025] Preferably, the method for obtaining the confidence level of electricity theft in the enclosure includes:

[0026] Within a set sliding period, the three-phase active power imbalance is calculated based on the collected total three-phase active power; the current response stiffness is multiplied by the three-phase active power imbalance, and the product result is used as a negative exponent term of the natural logarithm for nonlinear mapping calculation to obtain the confidence level of electricity theft in the electricity metering box.

[0027] Secondly, an anti-theft monitoring system for an electricity metering box includes a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned anti-theft monitoring method for an electricity metering box is implemented.

[0028] The present invention has the following effects:

[0029] 1. Traditional anti-theft technologies mainly rely on static indicators, which are highly susceptible to interference from nonlinear high-power loads connected to modern user sides under low-voltage distribution network conditions, resulting in severe data overlap and confusion. This invention, by introducing an overlapping sliding window of historical continuous observations, calculates the current response rigidity, which can perfectly filter out transient jump noise caused by random switching of legitimate loads, and realize the identification of abnormal physical topology tampering behavior inside the metering box;

[0030] 2. To address the confusion caused by the long-term stable operation of high-power symmetrical loads or the continuous operation of nonlinear purely resistive loads, this invention performs a product and negative exponential mapping on the current response stiffness and the three-phase active power imbalance under the same frequency time reference to calculate the enclosure electricity theft confidence. The product term increases only when the current response stiffness and the three-phase active power imbalance increase simultaneously, causing the final calculated enclosure electricity theft confidence to approach its maximum value. Once either of these indicators approaches 0, the product term becomes 0, causing the final output enclosure electricity theft confidence to approach 0, thereby eliminating the confusion caused by benign symmetrical constant loads and effectively avoiding false alarms. Attached Figure Description

[0031] Figure 1 This is a flowchart of steps S1-S5 in an embodiment of the present invention for a method of monitoring electricity theft prevention in an electricity metering box.

[0032] Figure 2 This is a structural block diagram of an anti-theft monitoring system for an electricity metering box according to an embodiment of the present invention. Detailed Implementation

[0033] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0034] Reference Figure 1 A method for monitoring electricity theft prevention in an electricity metering box includes steps S1-S5, as detailed below:

[0035] S1: Within the set basic time window, synchronously collect the original high-frequency voltage sequence and the original high-frequency current sequence of the power metering box, and obtain the total three-phase active power of the power metering box.

[0036] It should be noted that when nonlinear high-power load switching occurs in a low-voltage distribution network, the transmission of transient harmonics and micro-distortions between phases is random and transient. Once an abnormal theft impedance is forcibly connected inside the energy metering box (such as cross-phase connection or change in the physical topology of the current shunt), the micro-voltage transient waveform distortion of one phase will inevitably be transmitted to another phase through this fixed physical path, inducing a current response. Furthermore, this micro-transmission exhibits a rigid time delay in the time domain. Therefore, this embodiment performs high-frequency synchronous sampling of the voltage and current of a specific phase at a microscale.

[0037] In one specific embodiment, a high-frequency synchronous acquisition chip or a wide-area synchronous measurement unit (PMU) installed inside the power metering box is used to strictly synchronize the first-phase voltage and second-phase current of the user-side distribution network with the same clock frequency. In the continuously acquired electrical data stream, the original high-frequency sampling sequence within the current sliding cycle is extracted according to a set basic time window. The original high-frequency sampling sequence includes the corresponding original high-frequency voltage sequence within the basic time window. and the original high-frequency sequence of current ,Right now: ; ,in, This represents the total number of sampling points within the basic time window.

[0038] As an optional implementation method, high frequency sampling rate It can be preset as Basic time window It can be set to Under these parameters, the original high-frequency sequence of the voltage and the original high-frequency sequence of current Total number of sampling points All point.

[0039] S2: Extract the target analysis sequence from the original high-frequency voltage sequence, and slide the original high-frequency current sequence within the set time delay range to obtain multiple sliding analysis sequences; perform time-domain correlation comparison processing between the target analysis sequence and each sliding analysis sequence to determine the voltage timing delay degree of the current base time window;

[0040] It should be noted that under low-voltage distribution network conditions, the three-phase electrical sequence data collected at the side end exhibits nonlinearity and dynamic transients. With the increasing number of nonlinear loads and single-phase high-power loads connected to modern users, the harmonic injection, reactive power fluctuations, and transient switching generated during the operation of these devices can cause frequent and severe three-phase voltage distortion and current surges. Existing technologies rely solely on the magnitude of voltage and current amplitudes or macroscopic imbalance indicators for electricity theft detection. However, under the aforementioned load fluctuations, the instantaneous drops or surges in voltage and current caused by normal electricity consumption are almost identical in amplitude characteristics to illegal electricity theft. If amplitude alone cannot distinguish between benign fluctuations and malicious tampering, this further leads to high false alarm and high false negative rates in anti-theft systems.

[0041] It should be further noted that the changes in three-phase voltage and current caused by load fluctuations are usually random and dynamic; however, electricity theft, due to the connection of abnormal impedance, results in a fixed temporal sequence and response relationship between the voltage of a specific phase and the currents of the other phases. A microscopic voltage transient waveform distortion occurring in one phase will inevitably be propagated to another phase through this fixed path with extremely high time stability, inducing a current response. Furthermore, this propagation exhibits an indelible physical time delay in the time domain. Therefore, this step utilizes waveforms sampled at high frequency and synchronously, and performs a sliding misalignment comparison of voltage distortion and current abrupt changes in the time domain to provide a basis for subsequent identification of physical topology changes.

[0042] In one embodiment, the target analysis sequence is obtained by: from a sequence of length [missing information]. The original high-frequency sequence of voltage In the process, the edge buffers of the symmetrical regions at the beginning and end are removed, and the retained central segment is used as the fixed target analysis sequence. :

[0043]

[0044] In the formula, This is the preset maximum number of search points for discrete time delay. At this point, the target analysis sequence... Total length for The span of this edge buffer is determined based on the preset maximum number of discrete time delay search points. It is used to reserve a sufficient edge blocking buffer band for time delay traversal search when performing time delay sliding on the original high-frequency current sequence, to prevent exceeding the array boundary during sliding truncation, thereby ensuring the boundary safety of the micro-feature extraction algorithm.

[0045] The process of obtaining multiple sliding analysis sequences from the original high-frequency sequence of the sliding current within a set time delay range includes: setting the time delay factor... As any discrete integer within the time delay range, within the symmetric integer interval defined by the number of search points for the maximum discrete time delay. The inner step size is The point-by-point traversal sliding. Whenever this delay factor... When sliding to a specific discrete integer, from the original high-frequency sequence of the current Extracting a sequence of the corresponding length from the middle yields a sliding analysis sequence. :

[0046]

[0047] All sliding analysis sequences are obtained by traversing the sequence, regardless of the time delay factor. Take any value within the interval, and extract each subsequence. The lengths are all fixed and locked to The point is within the original array boundary.

[0048] For example, as an optional implementation, the AC power frequency of my country's low-voltage distribution network is used as... The synchronous high-frequency sampling rate is set to Taking the operating condition as an example, the time step between each sampling point for To cover a complete AC power frequency cycle. All possible electromagnetic conduction delays within the range, with a maximum number of discrete delay search points set. Calculations show that the maximum search delay span is exactly: .

[0049] The methods for obtaining the target analysis sequence include: from a sequence of length... The original high-frequency sequence of voltage In the middle, remove the first and last ones. The edge buffer of each point is used to extract the central segment as the fixed target analysis sequence. : .

[0050] At this point, the fixed target analysis sequence Total length for The span of this edge buffer is determined based on the preset maximum number of discrete time delay search points. It is used to reserve sufficient edge blocking buffer bands for time delay traversal search when performing time delay sliding on the original high-frequency current sequence, so as to prevent exceeding the array boundary during sliding truncation.

[0051] Meanwhile, time delay factor exist The sliding analysis sequence is truncated within a specific integer range, and each time the delay factor slides to a certain integer, a specific sliding analysis sequence is extracted. The corresponding expression is: .

[0052] Methods for determining the voltage timing delay within the current base time window include: analyzing the fixed target sequence. With each current sliding analysis sequence Point-by-point pairing and joint input are fed into the discrete mutual information operator to calculate the discrete mutual information value under each time delay factor state. The formula for calculating discrete mutual information is as follows:

[0053]

[0054] In the formula, Analyze the target sequence Values And sliding analysis sequence Values The joint probability distribution of ; Analyze the target sequence The marginal probability distribution; For sliding analysis sequences The marginal probability distribution.

[0055] Traverse all delay factors This will make the discrete mutual information value The time delay factor corresponding to the point where the maximum value is reached is determined as the voltage timing delay degree of the current base time window. The voltage timing delay represents the discrete physical response time required for the electromagnetic anomaly of a specific phase voltage to be transmitted to another phase through the internal topological path and induce a sudden current change after a microscopic distortion occurs in that phase voltage. A larger voltage timing delay, or a more pronounced random fluctuation, indicates a looser dynamic relationship between the two phases, consistent with the random interaction pattern of normal loads. A smaller voltage timing delay, and one that tends to be a constant on the time axis, indicates a tighter topological connection, consistent with the physical conduction characteristics after forced access to a power theft branch.

[0056] S3: Obtain the voltage timing delay degree of multiple consecutive basic time windows based on the time axis sliding, and construct a historical time series of delay degree; construct a discrete probability distribution based on the occurrence frequency of each different voltage timing delay degree value in the historical time series of delay degree, and calculate the current response stiffness based on the information entropy of the discrete probability distribution;

[0057] It should be noted that in actual low-voltage distribution network operation, the load on the user side is constantly switching operating conditions. This nonlinear operating condition generates a large number of random transient currents, which causes severe random drift and glitches in the voltage timing delay extracted in the previous steps during continuous observation on the time axis. This drastic change in operating conditions leads to a decrease in the confidence of micro-delay monitoring, resulting in the inability to distinguish between transient jumps caused by load switching and continuous rigid alignment caused by electricity theft. Therefore, this step filters out transient jump noise caused by legitimate random load switching by analyzing the dynamic law of voltage timing delay during its time evolution.

[0058] In one specific embodiment, the construction method of the historical time series of latency includes: to balance the statistical sample size of data and the real-time performance of alarms, this embodiment adopts an overlapping sliding window mechanism. The system establishes a buffer queue on the time axis with a fixed sliding step size equal to the length of the base time window. Each time the next sliding step is entered, the voltage timing latency of the latest base time window is output. At that time, the time-series rigidity analysis module stores the value in a file with a preset length. The earliest voltage timing delay data stored in the FIFO buffer queue is simultaneously removed. At each time step, the historical time series of delays containing continuous historical spans is directly extracted from the FIFO buffer queue.

[0059]

[0060] As an optional implementation, the preset length of the first-in-first-out (FIFO) cache queue. for Because the sliding step size is perfectly aligned with the basic acquisition window length, subsequent steps... The queue can then perform an adaptive iteration, thereby increasing the output frequency of the current response stiffness to [value missing]. This avoids the long monitoring blind spots and alarm delays that occur when using a non-overlapping sliding method, ensuring the ability to capture instantaneous abnormal power consumption behavior.

[0061] The methods for obtaining the current response stiffness include: statistically analyzing the frequency of different voltage timing delay values ​​in the historical time series of the delay, and calculating the frequency of each value in the time series. The probability of appearing in each window To construct the corresponding discrete probability distribution, the statistical information entropy of this discrete probability distribution is calculated, and its reciprocal is taken as the independent variable. The dimensionless current response stiffness is then calculated by constructing a negative exponential power function. The calculation formula is as follows:

[0062]

[0063] In the formula, For the historical time series of this delay, the first... A variety of different voltage timing delay values; This represents the probability of this value occurring throughout the entire historical time series of the delay. This represents the total number of different voltage timing delay values ​​that actually occur in the historical time series of this delay.

[0064] Specifically, when the voltage timing delay value is highly concentrated in a single voltage timing delay value within the historical time series of that delay, the calculated statistical information entropy approaches [the expected value]. This makes the overall current response stiffness The value of is smoothly monotonically approaching the maximum value. This indicates that the cross-phase timing response within this time period possesses extremely high temporal structure stability and rigidity; conversely, when normal power consumption leads to an extremely scattered distribution of voltage timing delay values, the statistical information entropy increases accordingly. Due to the constraint of the negative exponential power function, The value will decrease smoothly and monotonically and approach the minimum value. This not only widens the gap between normal electricity fluctuations and abnormal electricity theft and tampering, but also completely eliminates the problem of mathematical logic collapse caused by a zero denominator.

[0065] The current response stiffness characterizes the temporal structure stability of the response delay of the phase current to the micro-distortion of the voltage during the time evolution process. The larger the value, the more fixed the time sequence of the current response is, and it does not drift with the load switching conditions. This indicates that the power meter box is highly suspected of electricity theft. The smaller the value, the more scattered the voltage time delay value is, which is a benign manifestation of normal load fluctuation.

[0066] S4: The three-phase active power imbalance is calculated based on the total three-phase active power. The current response stiffness and the three-phase active power imbalance are then combined to calculate the confidence level of electricity theft in the enclosure.

[0067] It should be noted that, under low-voltage distribution network conditions, although the current response rigidity constructed through the aforementioned steps can effectively filter out transient noise caused by random switching conditions, the data characteristics at the electricity metering box side still pose a risk of confusion under specific legal electricity consumption conditions. For example, when a high-power symmetrical load operates stably for a long time on the user side, or when a large nonlinear purely resistive load operates continuously without interruption, the internal physical topology of such loads remains completely constant over a long period, which will also induce highly stable micro-time-series correlation responses between cross-phase data. This stable characteristic generated by continuous stable operating conditions has a very high similarity to the fixed foreign object topology introduced by electricity theft in the time domain, resulting in the problem that, when relying solely on the current response rigidity indicator, there is still residual overlap between normal and abnormal electricity consumption areas in the feature space. Therefore, this step constructs a box-based electricity theft confidence level to separate the data boundary between normal continuous stable loads and electricity theft, thereby improving the accuracy of anti-theft monitoring.

[0068] In one embodiment, the method for obtaining the confidence level of electricity theft in the metering box includes: each time the next sliding step is entered, the total three-phase active power collected by the electricity metering box in the current sliding cycle is obtained. The three-phase active power imbalance is calculated based on the obtained total three-phase active power. The specific calculation process is as follows: first, the average value of the three-phase active power is obtained; then, the absolute deviation between the active power of each of the three phases and the average value of the three-phase active power is calculated, and the maximum absolute deviation value is found; the maximum absolute deviation value is divided by the average value of the three-phase active power to obtain the three-phase active power imbalance.

[0069] It should be noted that the three-phase active power imbalance is used to characterize the asymmetric residual loss of three-phase energy caused by abnormal physical branch connections within the electricity metering box. Under normal power consumption conditions, although the three-phase power fluctuates, its distribution is relatively balanced, and the imbalance is low. However, under abnormal topology changes such as single-phase electricity theft or cross-phase external connections, the power of a specific phase will suffer severe losses or drastic fluctuations, leading to a significant increase in the three-phase active power imbalance.

[0070] Then, the current response stiffness Three-phase active power imbalance The product is performed, and the result is used as the negative exponent of the natural logarithm for nonlinear mapping calculation to obtain the confidence level of electricity theft in the electricity metering box. The calculation formula is as follows:

[0071]

[0072] In the formula, The confidence level for electricity theft from the electricity metering box is given, and its value range is an open interval. ; The current response stiffness corresponding to the historical time series of the delay obtained from the aforementioned steps; The three-phase active power imbalance of the energy metering box is calculated within the current sliding cycle. Represented by natural constant An exponential function with base 1.

[0073] The range of the confidence level for electricity theft from the enclosure is as follows: This represents the comprehensive confidence probability of determining that the electricity metering box contains an abnormal electricity theft topology. Specifically, when both the current response stiffness and the three-phase active power imbalance increase simultaneously, the product term increases, causing the confidence level to rapidly increase. Approach; as long as any one of them is close The overall confidence level will tend to be This effectively avoids false alarms. The confidence level of the box-type electricity theft is getting closer to... This indicates a very high degree of certainty that the meter box is involved in electricity theft; the closer it gets to... This indicates that the system is operating normally.

[0074] S5: Determine whether the confidence level of electricity theft in the enclosure is greater than the preset alarm threshold. If it is, output the anti-electricity theft monitoring alarm result.

[0075] In one specific embodiment, the system will output the confidence level of electricity theft from the enclosure in real time. A hard comparison is performed with a preset alarm threshold. As an optional implementation, the preset alarm threshold... Set as If the confidence level of electricity theft in the metering box exceeds the preset alarm threshold, the system will immediately output the anti-theft monitoring alarm result for the metering box and trigger the electricity theft alarm command. If the confidence level of electricity theft in the metering box does not exceed the threshold, it indicates that the system is operating normally and no alarm result will be output.

[0076] This invention also provides an anti-theft monitoring system for electricity metering boxes. For example... Figure 2 As shown, the system includes a processor and a memory. The memory stores computer program instructions, which, when executed by the processor, implement a method for monitoring electricity theft from an electricity metering box according to the first aspect of the present invention. The system also includes other components well-known to those skilled in the art, such as a communication bus and a communication interface. Their configuration and functions are known in the art and will not be described further here.

[0077] It should be noted that those skilled in the art can make various modifications and improvements without departing from the inventive concept, and these all fall within the scope of protection of this invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for monitoring electricity theft prevention in an electricity metering box, characterized in that, include: Within the set basic time window, the original high-frequency voltage sequence and the original high-frequency current sequence of the power metering box are synchronously collected, and the total three-phase active power of the power metering box is obtained. The target analysis sequence is extracted from the original high-frequency voltage sequence, and the original high-frequency current sequence is slid within a set time delay range to obtain multiple sliding analysis sequences; the target analysis sequence is compared with each of the sliding analysis sequences in the time domain to determine the voltage timing delay degree of the current base time window. The voltage timing delay is obtained by sliding along the time axis to acquire multiple consecutive historical basic time windows, and a historical time series of delay is constructed. A discrete probability distribution is constructed based on the occurrence frequency of each different voltage timing delay value in the historical time series of delay, and the current response stiffness is calculated based on the information entropy of the discrete probability distribution. The three-phase active power imbalance is calculated based on the total three-phase active power. The current response stiffness is then combined with the three-phase active power imbalance to obtain the confidence level of electricity theft in the enclosure. Determine whether the confidence level of the electricity theft in the enclosure is greater than a preset alarm threshold. If it is, output the anti-electricity theft monitoring alarm result.

2. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The methods for obtaining the target analysis sequence include: The edge buffer of the symmetrical region at the beginning and end of the original high-frequency voltage sequence is removed, and the retained central segment is taken as the target analysis sequence; wherein, the span of the edge buffer is determined according to the preset maximum number of discrete time delay search points.

3. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The step of sliding the original high-frequency current sequence within a set time delay range to obtain multiple sliding analysis sequences includes: Let the delay factor be any discrete integer within the delay range, and slide point by point within the delay range; Whenever the time delay factor slides to a certain discrete integer, a sequence of the corresponding length is extracted from the original high-frequency sequence of the current to obtain a sliding analysis sequence. All of the multiple sliding analysis sequences are obtained by traversing through the sequence.

4. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The determination of the voltage timing delay degree of the current base time window includes: The target analysis sequence is paired with each of the sliding analysis sequences point by point, and the discrete mutual information value is calculated for each of the time delay factor states. By iterating through all the aforementioned delay factors, the delay factor corresponding to the point where the discrete mutual information value reaches its maximum value is determined as the voltage timing delay degree of the current base time window.

5. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The historical time series of the delay includes: A first-in-first-out (FIFO) buffer queue with a preset length is established, using the same length as the base time window as the sliding step. Whenever the next sliding step is entered and the voltage timing delay of the current base time window is output, the value is stored in the FIFO buffer queue. When the amount of stored data reaches the preset length, the earliest stored voltage timing delay in the FIFO buffer queue is removed. All voltage timing delays currently retained in the FIFO buffer queue are used as the historical time series of the delay.

6. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The method for obtaining the current response stiffness includes: The frequency of occurrence of each different voltage timing delay value in the historical time series of the delay is statistically analyzed to construct the corresponding discrete probability distribution; The statistical information entropy of the discrete probability distribution is calculated, and the reciprocal of the statistical information entropy is taken as the independent variable. The current response stiffness is then calculated by constructing a negative exponential function.

7. The method for monitoring electricity theft prevention in an electricity metering box according to claim 1, characterized in that, The methods for obtaining the confidence level of electricity theft in the enclosure include: Within a set sliding period, the three-phase active power imbalance is calculated based on the collected total three-phase active power; the current response stiffness is multiplied by the three-phase active power imbalance, and the product result is used as a negative exponent term of the natural logarithm for nonlinear mapping calculation to obtain the confidence level of electricity theft in the electricity metering box.

8. A monitoring system for preventing electricity theft in an electricity metering box, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement a method for monitoring electricity theft from an electricity metering box according to any one of claims 1-7.

Images