Client side electricity utilization potential safety hazard state multi-class mode identification method

By monitoring the voltage, current and phase relationships, establishing a power consumption abnormality determination model and loop impedance matrix, identifying the abnormal electricity consumption behavior of users in low-voltage distribution networks, solving the problem of frequent equipment abnormal movements and large number of users in low-voltage distribution networks. , improves the accuracy of identification results.

CN118980875APending Publication Date: 2024-11-19INNER MONGOLIA ELECTRIC POWER (GRP) CO LTD ALXA POWER SUPPLY BRANCH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411051398.1
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-01
Publication Date
2024-11-19

AI Technical Summary

Technical Problem

The equipment and lines in the low-voltage distribution network are frequently abnormal, the number of users is large, and the power supply network is complex, making it difficult to effectively draw and measure network graphics and difficult to identify users' abnormal power usage behavior.

Method used

By monitoring the user's voltage, current and phase relationship between voltage and current, establish a power consumption abnormality determination model and loop impedance matrix, build a binary linear regression model, calculate the suspicion coefficient, and identify abnormal power consumption behavior.

Benefits of technology

The construction of a low-voltage distribution network topology model is improved based on existing resources, which improves the accuracy of users' abnormal power use identification and multi-faceted data support for identification results, and improves the accuracy and reliability of identification results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118980875A_ABST
    Figure CN118980875A_ABST
Patent Text Reader

Abstract

The invention discloses a client-side electricity utilization potential safety hazard state multi-class mode identification method, which comprises the following steps of: analyzing electricity utilization behaviors of a user according to a power formula metered by each measuring element, and monitoring voltage, current and a phase relation between the voltage and the current of the user by using an electric energy meter. According to the method, the low-voltage power distribution network topology model based on the loop impedance is constructed on the basis of existing resources configured for the low-voltage power distribution network, and the constructed loop impedance model is reasonably simplified on the basis of understanding the actual situation and engineering, so that the reliability of the low-voltage power distribution network is improved. A demand-side real-time voltage data matrix and a demand-side real-time current data matrix are acquired through a data acquisition system, historical voltage and current data of the residential user safe power utilization diagnosis terminal are recorded, user data of a demand side are updated in real time based on the constructed data matrix, abnormal power utilization of a user is identified from multiple aspects of data, and the user safety power utilization diagnosis method is improved. And the accuracy of the identification result is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of electrical safety and pattern recognition, and in particular to a multi-category pattern recognition method for hidden danger states of electrical safety at a customer side. Background Art

[0002] Abnormal electricity use is the most harmful abnormal electricity use to power supply companies. For example, electricity theft is a user who secretly steals electricity during the use of electricity to avoid or underpay electricity bills. Since electricity is intangible and has a very fast transmission speed, it cannot be stored and is consumed as soon as it is stolen. Therefore, electricity theft cannot be proved by the stolen electricity, but can only be identified by behavior. Abnormal electricity use is the most harmful abnormal electricity use to power supply companies. In order to improve the power supply reliability of the power grid, operation status assessment is one of the reliable means to predict and analyze the future operation status of the power grid, screen the weak links of the power grid, and eliminate the invisible faults of the power grid in the bud.

[0003] However, at present, due to the frequent changes of equipment and lines in low-voltage distribution networks, as well as the large number of users, complex power supply networks, and uneven quality of power supply lines, it is difficult to effectively draw and measure network graphics of low-voltage distribution networks. Therefore, it is necessary to develop a multi-category pattern recognition method for customer-side power safety hazard status to solve the problem of identifying abnormal power consumption behavior of users. Summary of the invention

[0004] To this end, the present invention provides a multi-category pattern recognition method for the status of hidden dangers in power consumption on the customer side, in order to solve the problem that there are frequent abnormal changes in equipment and lines in the low-voltage distribution network, and the characteristics of a large number of users, a complex power supply network, and uneven quality of power supply lines, which makes it difficult to effectively draw and measure the network diagram of the low-voltage distribution network. Therefore, it is necessary to develop a multi-category pattern recognition method for the status of hidden dangers in power consumption on the customer side to solve the problem of identifying abnormal power consumption behavior of users. In order to achieve the above purpose, the present invention provides the following technical solutions: According to the first aspect of the present invention, a multi-category pattern recognition method for the status of hidden dangers in power consumption on the customer side includes:

[0005] Step 1: Analyze the user's electricity consumption behavior based on the power formula measured by each measuring element, and use an electric energy meter to monitor the user's voltage, current, and the phase relationship between voltage and current;

[0006] Step 2: Establish an abnormal power consumption judgment model to judge abnormal power consumption behavior;

[0007] Step 3: Construct a loop impedance matrix to evaluate the grid operation status, including:

[0008] Define loop impedance and build a loop impedance model;

[0009] A binary linear regression model was constructed to modify the loop impedance model;

[0010] Determine the loop confrontation calculation results;

[0011] Step 4: Establish an intelligent recognition model for abnormal electricity consumption behavior to identify abnormal electricity consumption by users. The intelligent recognition model for abnormal electricity consumption behavior includes:

[0012] The calculation formula of suspicion coefficient is: suspicion coefficient = ∑ power consumption abnormality score * abnormal number of times;

[0013] Among them, the power consumption anomaly score: the score corresponding to the power consumption anomaly model. The higher the probability of judging it as power theft after the power consumption anomaly occurs, the higher the score. The number of anomalies: refers to the total number of times the element types in various power consumption anomaly models appear in the user's continuous monitoring in one day.

[0014] Furthermore, the power formula is specifically expressed as:

[0015]

[0016] Among them, U is the voltage value of the measuring element, I is the current value of the measuring element, is the phase angle between U and I.

[0017] Furthermore, the power consumption abnormality determination method includes:

[0018] Abnormal monthly line loss of parallel lines: when the pure line loss rate is greater than or equal to 5% and the data type is parallel line subtotal and the line loss electricity is greater than 20000, the parallel line loss rate is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than or equal to 30000 and the data type is parallel line subtotal, the parallel line loss rate is greater than 7%;

[0019] Abnormal line loss in monthly line: When the line loss rate is greater than or equal to 5% and the line loss electricity is greater than 20000 and the data type is single line, the line loss rate of the single line is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than 30000 and the data type is single line, the line loss rate of the single line is greater than 7%;

[0020] Abnormal daily line loss: when the line loss rate is >=5% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >5%; when the power supply - power sales volume is >2000 and the line loss fluctuation rate is >=50% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the electricity quantity difference is >2000 and the line loss fluctuation rate is >50%; the line loss rate is >=7% and the power supply and sales volume difference is >1000 and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >7%.

[0021] Furthermore, the power consumption abnormality determination method further includes:

[0022] Three-phase three-wire current unbalance: When the meter type = three-phase three-wire and the current unbalance rate > 30 and the current phase A > = 0.5 or the current phase C > = 0.5, it is the three-phase three-wire current unbalance rate; when the meter type = three-phase three-wire and the electric unbalance rate > 50 and the current phase A > 0.1 and the current phase A < 0.2 or the current phase C > 0.1 and the current phase C < 0.2, it is the three-phase three-wire micro-current unbalance rate;

[0023] Power factor abnormality: When the line loss rate is greater than or equal to 5% and the line loss power is greater than 20,000 and the data type is single line, the power factor is lower than 0.5; when the line loss rate is greater than or equal to 7% and the line loss power is greater than 30,000 and the data type is single line, the three-phase three-wire power factor is 1;

[0024] The proportion of electricity purchased by the exchange meter is abnormal: the electricity difference proportion is greater than 50 and the electricity difference is greater than 50000.

[0025] Furthermore, the power consumption abnormality determination method further includes:

[0026] Three-phase four-wire current abnormality: When the meter type = three-phase four-wire and the current unbalance rate is >=50 and the current phase A is >=3 or the current phase B is >=3 or the current phase C is >=3 and the current phase A is >0 and the current phase B is >0 and the current phase C is >0 and the line loss difference range is <-800 or the line loss difference range is >800 and the statistical line loss power is <-800 or the statistical line loss power is >800, that is, the three-phase four-wire current is unbalanced;

[0027] The three-phase current difference ratio is abnormal: when the current phase A = 0 or the current phase B = 0 or the current phase C = 0) and the meter type = three-phase four-wire and the line loss difference range <-800 or the line loss difference range >800 and the statistical line loss power <-800 or the statistical line loss power >800, the three-phase four-wire current is abnormal;

[0028] When the current difference ratio of phase A is >50 and the power difference ratio is >30 and the current of phase A is >0.5 or the current of phase B is >0.5 and the current difference ratio of phase B is >50 and the power difference ratio is >30 or the current of phase C is >0.5 and the current difference ratio of phase C is >50 and the power difference ratio is >30, the current difference ratio is greater than 50.

[0029] Furthermore, the definition of user loop impedance in the substation can be expressed as the equivalent impedance of the loop including equipment from the substation distribution transformer low-voltage bus to the user meter.

[0030] Furthermore, according to the KVL law, the loop voltage equation of the loop impedance model is:

[0031]

[0032] Among them, considering any time, the magnitude of the line impedance is much smaller than the equivalent impedance of the user load, so the angle deviation of the user voltage at different times caused by the line impedance is ignored, and it is assumed that the voltage phase angle at t1 and t2 is 0, then the above formula is equivalent to:

[0033]

[0034] According to the above formula, the impedance constraint judgment conditions of each time section are constructed, and the equivalent hypothesis testing model is constructed, that is:

[0035]

[0036] Furthermore, the binary linear regression model is shown as follows:

[0037] y i =βx1 i +γx2 i +ε i

[0038] Add the impedance deviation Z to the above formula and rewrite it into a binary linear regression model:

[0039]

[0040] Furthermore, the determination of the loop confrontation calculation result specifically includes:

[0041] Establish screening rules for voltage, current and power factor at different times: For the user's power factor η(t)=cos(φ t ), it is required to satisfy η(t)∈[0.9,1];

[0042] For the user's current value I(t), in order to exclude the user's abnormal power consumption, the current needs to set a lower limit, which ensures that I(t) ≥ 1A;

[0043] The user's voltage U(t) is required to be within the range of -10% to +7% of the rated voltage, that is,

[0044] V(t)∈[198V,236V].

[0045] The present invention has the following advantages:

[0046] Based on the existing resources already equipped in the low-voltage distribution network, a low-voltage distribution network topology model based on loop impedance was constructed. After understanding the actual situation and the project, the constructed loop impedance model was reasonably simplified. The real-time voltage data array and the real-time current data array on the demand side were obtained through the data acquisition system. The voltage and current historical data of the residential user's safe electricity use diagnosis terminal were recorded. The user data on the demand side was updated in real time based on the constructed data array, and abnormal electricity use of users was identified from multiple aspects of data, which improved the accuracy of the identification results. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flowchart of a method for multi-category pattern recognition of potential safety hazards of electricity use at a client side provided in some embodiments of the present invention.

[0048] Figure 2 An equivalent circuit diagram of a network topology in a customer-side power safety hazard status multi-category pattern recognition method provided in some embodiments of the present invention.

[0049] Figure 3 A flowchart for executing impedance solution in a multi-category pattern recognition method for potential safety hazards of electricity use at a customer side is provided in some embodiments of the present invention. DETAILED DESCRIPTION

[0050] The following is a description of the implementation of the present invention by specific embodiments. People familiar with the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] like Figures 1 to 3 As shown, a multi-category pattern recognition method for customer-side power safety hazard status in an embodiment of the first aspect of the present invention includes:

[0052] Step 1: Analyze the user's electricity consumption behavior based on the power formula measured by each measuring element, and use an electric energy meter to monitor the user's voltage, current, and the phase relationship between voltage and current;

[0053] Step 2: Establish an abnormal power consumption judgment model to judge abnormal power consumption behavior;

[0054] Step 3: Construct a loop impedance matrix to evaluate the grid operation status, including:

[0055] Define loop impedance and build a loop impedance model;

[0056] A binary linear regression model was constructed to modify the loop impedance model;

[0057] Determine the loop confrontation calculation results;

[0058] Step 4: Establish an intelligent recognition model for abnormal electricity consumption behavior to identify abnormal electricity consumption by users. The intelligent recognition model for abnormal electricity consumption behavior includes:

[0059] The calculation formula of suspicion coefficient is: suspicion coefficient = ∑ power consumption abnormality score * abnormal number of times;

[0060] Among them, the power consumption anomaly score: the score corresponding to the power consumption anomaly model. The higher the probability of judging it as power theft after the power consumption anomaly occurs, the higher the score. The number of anomalies: refers to the total number of times the element types in various power consumption anomaly models appear in the user's continuous monitoring in one day.

[0061] In this embodiment, by studying the multi-category pattern recognition method of the power safety hazard status on the customer side, a typical feature library of power safety hazards is constructed. A survey is conducted on the power safety hazard phenomena recorded by various metering devices in the province, such as electric energy meters, meter boxes, acquisition terminals, and transformers. Relying on the MDS data statistics system, an intelligent fault analysis and processing algorithm and model are established to summarize the power safety hazard phenomena and causes of various metering devices in the province, and to associate the corresponding power safety conditions of various metering devices in the region. By classifying and analyzing the environmental conditions, resident categories, meter age, manufacturers, safety types, etc. in different regions, typical cases are generated. It is planned to use artificial intelligence algorithms to establish an equipment operation status assessment model database, select input and output characteristic parameters, establish an equipment operation status assessment model, and generate a typical case expert database. Based on the existing resources already equipped in the low-voltage distribution network, a low-voltage distribution network topology model based on loop impedance is constructed. After understanding the actual situation and the project, the constructed loop impedance model is reasonably simplified. The real-time voltage data array and the real-time current data array on the demand side are obtained through the data acquisition system. The voltage and current historical data of the residential user's safe electricity use diagnosis terminal are recorded, and the user data on the demand side is updated in real time based on the constructed data array. The abnormal electricity use of users is identified from multiple aspects of data, which improves the accuracy of the identification results.

[0062] The power formula is specifically expressed as:

[0063]

[0064] Among them, U is the voltage value of the measuring element, I is the current value of the measuring element, is the phase angle between U and I.

[0065] The methods for determining abnormal power consumption include:

[0066] Abnormal monthly line loss of parallel lines: when the pure line loss rate is greater than or equal to 5% and the data type is parallel line subtotal and the line loss electricity is greater than 20000, the parallel line loss rate is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than or equal to 30000 and the data type is parallel line subtotal, the parallel line loss rate is greater than 7%;

[0067] Abnormal line loss in monthly line: When the line loss rate is greater than or equal to 5% and the line loss electricity is greater than 20000 and the data type is single line, the line loss rate of the single line is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than 30000 and the data type is single line, the line loss rate of the single line is greater than 7%;

[0068] Abnormal daily line loss: when the line loss rate is >=5% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >5%; when the power supply - power sales volume is >2000 and the line loss fluctuation rate is >=50% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the electricity quantity difference is >2000 and the line loss fluctuation rate is >50%; the line loss rate is >=7% and the power supply and sales volume difference is >1000 and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >7%.

[0069] The power consumption abnormality determination method also includes:

[0070] Three-phase three-wire current unbalance: When the meter type = three-phase three-wire and the current unbalance rate > 30 and the current phase A > = 0.5 or the current phase C > = 0.5, it is the three-phase three-wire current unbalance rate; when the meter type = three-phase three-wire and the electric unbalance rate > 50 and the current phase A > 0.1 and the current phase A < 0.2 or the current phase C > 0.1 and the current phase C < 0.2, it is the three-phase three-wire micro-current unbalance rate;

[0071] Power factor abnormality: When the line loss rate is greater than or equal to 5% and the line loss power is greater than 20,000 and the data type is single line, the power factor is lower than 0.5; when the line loss rate is greater than or equal to 7% and the line loss power is greater than 30,000 and the data type is single line, the three-phase three-wire power factor is 1;

[0072] The proportion of electricity purchased by the exchange meter is abnormal: the electricity difference proportion is greater than 50 and the electricity difference is greater than 50000.

[0073] The power consumption abnormality determination method also includes:

[0074] Three-phase four-wire current abnormality: When the meter type = three-phase four-wire and the current unbalance rate is >=50 and the current phase A is >=3 or the current phase B is >=3 or the current phase C is >=3 and the current phase A is >0 and the current phase B is >0 and the current phase C is >0 and the line loss difference range is <-800 or the line loss difference range is >800 and the statistical line loss power is <-800 or the statistical line loss power is >800, that is, the three-phase four-wire current is unbalanced;

[0075] The three-phase current difference ratio is abnormal: when the current phase A = 0 or the current phase B = 0 or the current phase C = 0) and the meter type = three-phase four-wire and the line loss difference range <-800 or the line loss difference range >800 and the statistical line loss power <-800 or the statistical line loss power >800, the three-phase four-wire current is abnormal;

[0076] When the current difference ratio of phase A is greater than 50 and the power difference ratio is greater than 30 and the current of phase A is greater than 0.5 or the current of phase B is greater than 0.5 and the current difference ratio of phase B is greater than 50 and the power difference ratio is greater than 30 or the current of phase C is greater than 0.5 and the current difference ratio of phase C is greater than 50 and the power difference ratio is greater than 30, that is, the current difference ratio is greater than 50;

[0077] In this embodiment, the basic idea of ​​analyzing electricity usage behavior is discussed based on the basics of electrical engineering. Based on this idea, common means of electricity theft and main methods of anti-electricity theft are introduced in detail. At the same time, based on this idea, an abnormal electricity usage model and an intelligent identification model for electricity theft are designed, and a user leakage detection method is designed. From the power formula, we can know that the size of the power measured by our electric energy meter is related to the three variables of voltage, current, and the phase relationship between voltage and current. The electric energy generated by the user is equal to the product of power and time, so the data measured by the electric energy meter is related to voltage, current, the phase relationship between voltage and current, and time. To analyze whether the user's electricity usage is normal, we only need to monitor and analyze whether the user's voltage, current, and the phase relationship between voltage and current are normal;

[0078] Among them, if its power consumption is normal, then we can summarize the following basic characteristics: the measured voltage value is relatively stable and basically unchanged; the measured current changes at any time, but the three phases are basically balanced; the phase angle will not change suddenly, and the power factor changes within a certain range; the user's daily line loss and monthly line loss fluctuate within a certain range;

[0079] The characteristics of abnormal power consumption behavior can be analyzed from the records of the energy meter. Regardless of whether the energy meter is stopped, reversed, or slow, the user's power consumption will be recorded. We can analyze whether the user has abnormal behavior by analyzing the user's power consumption behavior, that is, by analyzing the changes in the user's voltage, current, power, power factor, line loss, etc. The user's voltage, current, power, power factor, line loss, etc. cannot be constant. Slight changes must exist, and this slight change has nothing to do with abnormal power consumption. However, if the amplitude of the change exceeds a certain value, we can determine it as abnormal. Through power consumption abnormality analysis, we can determine what kind of abnormal situation exists on the user side. Based on the principles of electrical engineering, combined with the actual power consumption in this area and previous power consumption abnormality cases, we designed the power consumption abnormality model and judgment conditions. These judgment conditions can be modified in the system, providing opportunities for future modifications and improvements. For voltage abnormalities, since the user's voltage is basically unchanged, the voltage can be judged as being lower than a certain threshold. For current abnormalities, since the current changes at any time but the three phases are basically balanced, the three-phase current is unbalanced or the single phase is equal to zero for judgment. At the same time, the difference between meter sampling and crossover is increased. Usually, the meter is damaged in abnormal electricity consumption, and the abnormal difference between the two can be used as a basis for judging whether there is abnormal behavior. For power factor abnormality, since dedicated transformer users must perform power factor assessment, users with power factors lower than the specified value will be punished, so the power of dedicated transformer users is basically greater than 0.8. Therefore, we use the power factor lower than a certain threshold to make a judgment. For power abnormality, since the power changes at any time, we use the difference ratio between meter sampling and cross-check to make a judgment. For line loss abnormality, since the user line loss varies within a certain range, the urban network line loss rate is generally lower than 5%, and the rural network line loss rate is generally lower than or greater than 7%. We use the line loss greater than a certain threshold to make a judgment. We design two variables: monthly line loss and daily line loss. Since there are two types of users: dedicated selection and parallel line, for non-dedicated line users, we use parallel line loss to make a judgment.

[0080] The definition of user loop impedance in the substation area can be expressed as the equivalent impedance of the loop from the substation distribution transformer low-voltage bus to the user meter including the equipment.

[0081] Furthermore, according to the KVL law, the loop voltage equation of the loop impedance model is:

[0082]

[0083] Among them, considering any time, the magnitude of the line impedance is much smaller than the equivalent impedance of the user load, so the angle deviation of the user voltage at different times caused by the line impedance is ignored, and it is assumed that the voltage phase angle at t1 and t2 is 0, then the above formula is equivalent to:

[0084]

[0085] According to the above formula, the impedance constraint judgment conditions of each time section are constructed, and the equivalent hypothesis testing model is constructed, that is:

[0086]

[0087] The binary linear regression model is shown below:

[0088] y i =βx1 i +γx2 i +ε i Add the impedance deviation Z to the above formula and rewrite it into a binary linear regression model:

[0089]

[0090] The determination of the loop confrontation calculation results specifically includes:

[0091] Establish screening rules for voltage, current and power factor at different times: For the user's power factor η(t)=cos(φ t ), it is required to satisfy η(t)∈[0.9,1];

[0092] For the user's current value I(t), in order to exclude the user's abnormal power consumption, the current needs to set a lower limit, which ensures that I(t) ≥ 1A;

[0093] The user's voltage U(t) is required to be within the range of -10% to +7% of the rated voltage, that is,

[0094] V(t)∈[198V,236V];

[0095] In this embodiment, the low-voltage substation follows the mesh radiation structure of the distribution system, and the high-voltage side of the substation adopts a triangle connection method. On the one hand, it prevents harmonics from being transmitted to the upper level, and on the other hand, it reduces the three-phase voltage distortion on the low-voltage side caused by the three-phase imbalance on the low-voltage side; the low-voltage side adopts the method of direct grounding of the neutral point. The low-voltage side collects electrical energy by the busbar and supplies power to the outgoing line. In the actual scenario, the low-voltage side of the distribution transformer is connected to the outgoing line cabinet, and the connection to the outgoing line is completed in the low-voltage outgoing line cabinet;

[0096] Among them, for the identification of electricity theft, we considered the relationship between electricity theft and abnormal electricity usage. If there is electricity theft, the user must have abnormal electricity usage, and in most cases there is more than one abnormal electricity usage, and the abnormal electricity usage is continuous. According to this principle, we design an intelligent electricity theft identification model, that is, to comprehensively measure the types and frequency of abnormal electricity usage behaviors of users to determine whether there is electricity theft. Here we simplify the calculation of the user's "suspicion coefficient". Any user can only be a "suspect" before being investigated and confirmed as an electricity thief. We set different scores for each abnormal electricity usage model of the user. The cumulative value of the product of the number of times the user has an abnormal electricity usage model in a day and the corresponding score of the model is the user's "suspicion coefficient";

[0097] At the same time, the execution order of impedance solution is as follows: first, obtain the electricity consumption data of users for multiple days, and perform data screening according to the designed data screening rules such as voltage, current and power factor. Secondly, select the parameters that meet the screening rules to form the parameter sequence required for regression model solution; then introduce the assumptions, and further use the binary linear regression model to solve the resistance reactance; finally, determine whether the calculation results meet the regression error judgment and impedance constraint judgment, and determine whether to re-screen the data or obtain a reasonable loop resistance and reactance of the substation.

Claims

1. A multi-category pattern recognition method for customer-side power safety hazard status, It is characterized in that include: Step 1: Analyze the user's electricity consumption behavior based on the power formula measured by each measuring element, and use an electric energy meter to monitor the user's voltage, current, and the phase relationship between voltage and current; Step 2: Establish an abnormal power consumption judgment model to judge abnormal power consumption behavior; Step 3: Construct a loop impedance matrix to evaluate the grid operation status, including: Define loop impedance and build a loop impedance model; A binary linear regression model was constructed to modify the loop impedance model; Determine the loop confrontation calculation results; Step 4: Establish an intelligent recognition model for abnormal electricity consumption behavior to identify abnormal electricity consumption by users. The intelligent recognition model for abnormal electricity consumption behavior includes: The calculation formula of suspicion coefficient is: suspicion coefficient = ∑ power consumption abnormality score * abnormal number of times; Among them, the power consumption anomaly score: the score corresponding to the power consumption anomaly model. The higher the probability of judging it as power theft after the power consumption anomaly occurs, the higher the score. The number of anomalies: refers to the total number of times the element types in various power consumption anomaly models appear in the user's continuous monitoring in one day.

2. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The power formula is specifically expressed as: Among them, U is the voltage value of the measuring element, I is the current value of the measuring element, is the phase angle between U and I.

3. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The methods for determining abnormal power consumption include: Abnormal monthly line loss of parallel lines: when the pure line loss rate is greater than or equal to 5% and the data type is parallel line subtotal and the line loss electricity is greater than 20000, the parallel line loss rate is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than or equal to 30000 and the data type is parallel line subtotal, the parallel line loss rate is greater than 7%; Abnormal line loss in monthly line: When the line loss rate is greater than or equal to 5% and the line loss electricity is greater than 20000 and the data type is single line, the line loss rate of the single line is greater than 5%; when the line loss rate is greater than or equal to 7% and the line loss electricity is greater than 30000 and the data type is single line, the line loss rate of the single line is greater than 7%; Abnormal daily line loss: when the line loss rate is >=5% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >5%; when the power supply - power sales volume is >2000 and the line loss fluctuation rate is >=50% and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the electricity quantity difference is >2000 and the line loss fluctuation rate is >50%; the line loss rate is >=7% and the power supply and sales volume difference is >1000 and the ratio of electricity quantity difference to line loss is <0.3 and the line loss type is line-line loss, that is, the line loss rate is >7%.

4. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The power consumption abnormality determination method also includes: Three-phase three-wire current unbalance: When the meter type = three-phase three-wire and the current unbalance rate > 30 and the current phase A > = 0.5 or the current phase C > = 0.5, it is the three-phase three-wire current unbalance rate; when the meter type = three-phase three-wire and the electric unbalance rate > 50 and the current phase A > 0.1 and the current phase A < 0.2 or the current phase C > 0.1 and the current phase C < 0.2, it is the three-phase three-wire micro-current unbalance rate; Power factor abnormality: When the line loss rate is greater than or equal to 5% and the line loss power is greater than 20,000 and the data type is single line, the power factor is lower than 0.5; when the line loss rate is greater than or equal to 7% and the line loss power is greater than 30,000 and the data type is single line, the three-phase three-wire power factor is 1; The proportion of electricity purchased by the exchange meter is abnormal: the electricity difference proportion is greater than 50 and the electricity difference is greater than 50000.

5. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The power consumption abnormality determination method also includes: Three-phase four-wire current abnormality: When the meter type = three-phase four-wire and the current unbalance rate is >=50 and the current phase A is >=3 or the current phase B is >=3 or the current phase C is >=3 and the current phase A is >0 and the current phase B is >0 and the current phase C is >0 and the line loss difference range is <-800 or the line loss difference range is >800 and the statistical line loss power is <-800 or the statistical line loss power is >800, that is, the three-phase four-wire current is unbalanced; The three-phase current difference ratio is abnormal: when the current phase A = 0 or the current phase B = 0 or the current phase C = 0) and the meter type = three-phase four-wire and the line loss difference range <-800 or the line loss difference range >800 and the statistical line loss power <-800 or the statistical line loss power >800, the three-phase four-wire current is abnormal; When the current difference ratio of phase A is >50 and the power difference ratio is >30 and the current of phase A is >0.5 or the current of phase B is >0.5 and the current difference ratio of phase B is >50 and the power difference ratio is >30 or the current of phase C is >0.5 and the current difference ratio of phase C is >50 and the power difference ratio is >30, the current difference ratio is greater than 50.

6. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The definition of user loop impedance in the substation area can be expressed as the equivalent impedance of the loop from the substation distribution transformer low-voltage bus to the user meter including the equipment.

7. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The loop impedance model is based on the KVL law, and its loop voltage equation is: Among them, considering any time, the magnitude of the line impedance is much smaller than the equivalent impedance of the user load, so the angle deviation of the user voltage at different times caused by the line impedance is ignored, and it is assumed that the voltage phase angle at t1 and t2 is 0, then the above formula is equivalent to: According to the above formula, the impedance constraint judgment conditions of each time section are constructed, and the equivalent hypothesis testing model is constructed, that is:

8. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The binary linear regression model is shown below: and i =βxl i +γx2 i +ε i Add the impedance deviation Z to the above formula and rewrite it into a binary linear regression model:

9. According to claim 1, a multi-category pattern recognition method for hidden dangers of power safety at the customer side, It is characterized in that The determination of the loop confrontation calculation results specifically includes: Establish screening rules for voltage, current and power factor at different times: For the user's power factor η(t)=cos(φ t ), it is required to satisfy η(t)∈[0.9,1]; For the user's current value I(t), in order to exclude the user's abnormal power consumption, the current needs to set a lower limit, which ensures that I(t) ≥ 1A; The user's voltage U(t) is required to be within the range of -10% to +7% of the rated voltage, that is, V(t)∈[198V,236V].