CN112184477A

CN112184477A - Clustering and PQUI recognition algorithm-based electric quantity supplementing method

Info

Publication number: CN112184477A
Application number: CN202010866383.6A
Authority: CN
Inventors: 侯素颖; 裘炜浩; 王建波; 金挺超; 蔡慧; 郁春雷; 包锦辉; 陈嘉
Original assignee: State Grid Zhejiang Electric Power Co Ltd; China Jiliang University; Zhejiang Huayun Information Technology Co Ltd; Taizhou Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Current assignee: State Grid Zhejiang Electric Power Co Ltd; China Jiliang University; Zhejiang Huayun Information Technology Co Ltd; Taizhou Power Supply Co of State Grid Zhejiang Electric Power Co Ltd
Priority date: 2020-08-25
Filing date: 2020-08-25
Publication date: 2021-01-05

Abstract

The invention discloses a cluster and PQUI recognition algorithm-based electric quantity tracing and supplementing method, and relates to the field of power grid operation and maintenance. At present, the electricity stealing capacity cannot be accurately calculated, the calculated electricity stealing capacity is too much different from the actual electricity stealing capacity, and fairness is difficult to maintain between users and power companies. The method comprises the steps of judging electricity stealing time of a user according to load data of the user, determining an electricity stealing method according to the relation between the load data, determining a correction coefficient according to electricity stealing behaviors, and calculating the compensation electric quantity according to the correction coefficient; and a clustering algorithm and/or a PQUI algorithm are/is adopted when the electricity stealing days are calculated. According to the technical scheme, the clustering algorithm and the PQUI recognition algorithm are adopted to calculate the corresponding electricity stealing amount, the calculation matching is good, the accuracy is high, the acquired data can meet the requirements, the compensation amount of the electricity stealing users can be accurately, objectively and quickly calculated, the compensation amount is consistent with the actually generated electricity, and the electricity utilization fairness is improved.

Description

Clustering and PQUI recognition algorithm-based electric quantity supplementing method

Technical Field

The invention relates to the field of power grid operation and maintenance, in particular to a method for supplementing electric quantity based on clustering and a PQUI recognition algorithm.

Background

The commensurate increase in power production and consumption levels has greatly pushed the development of electricity management technologies, but with the attendant increased problems of electricity theft. The problem of electricity stealing not only brings about the safety problem of electricity utilization, but also leads to the fairness problem among users and the direct economic loss of power supply enterprises. Nowadays, most of the recollection electric quantity is estimated by directly depending on manual analysis, but the method is too flexible to cause insufficient persuasion and is difficult to convince. Therefore, a reasonable method is provided to find out a reasonable electricity stealing interval from the historical load data and correctly calculate the compensation electricity quantity.

The electricity stealing time of the user can be found out more reasonably and accurately by combining various electricity stealing time confirmation methods and the electricity consumption curve of the user. Because the number of users is large, the electricity consumption requirements are different, and electricity stealing methods are various, the electricity stealing time judging method cannot accurately judge the electricity stealing time. Therefore, various methods are required.

After the electricity stealing time is determined, the electricity stealing methods need to be confirmed, due to the fact that the electricity stealing methods are different, correction coefficients of all users are different, electricity stealing electric quantity cannot be calculated accurately, the calculated electricity stealing electric quantity is too much different from the actual electricity stealing electric quantity, and fairness is difficult to maintain between users and power companies.

Disclosure of Invention

The technical problem to be solved and the technical task are to perfect and improve the prior technical scheme and provide a method for supplementing electric quantity based on clustering and a PQUI recognition algorithm so as to achieve the aim of accurately calculating the supplemented electric quantity. Therefore, the invention adopts the following technical scheme.

A cluster and PQUI recognition algorithm-based electric quantity compensation method is implemented by adopting an electric energy metering device remote monitoring and diagnosing system, wherein the electric energy metering device remote monitoring and diagnosing system comprises an electric energy information acquisition system, a data collection and analysis server and a database server; the electricity utilization information acquisition system acquires electricity utilization data and events of power consumers in real time; the data collection and analysis server analyzes and calculates electricity stealing behaviors and corresponding compensation electric quantity according to the acquired data; the database server stores electricity consumption data, events, analysis and judgment thresholds, an electricity stealing behavior feature library and correction coefficients under each electricity stealing behavior, and the data collection and analysis server comprises the following steps when analyzing and calculating the compensation electric quantity:

1) judging whether the electricity stealing amount depends on the time length; if not, calculating the difference between the on-site copy electric quantity and the final settlement electric quantity of the power supply enterprise directly; if yes, entering the next step;

2) calculating the number of electricity stealing days, and judging whether the number of electricity stealing days can be determined; if yes, directly calculating the electricity stealing amount according to the determined days; if not, calculating the electricity stealing amount by preset days; the number of electricity stealing days is calculated by adopting a clustering algorithm and/or a PQUI algorithm;

3) identifying electricity stealing behaviors; acquiring power utilization data of a suspected electricity stealing user; comparing the electricity stealing behavior feature libraries in the corresponding metering modes to match the most probable electricity stealing behaviors;

4) inquiring a database server, comparing with electricity stealing behaviors, judging whether a correction coefficient exists, if so, determining actual daily electric quantity according to the corresponding correction coefficient, and then adding the daily electric quantity within the electricity stealing time to obtain the actual electric quantity; if not, entering the next step;

5) judging whether the daily electric quantity prediction requirement is met; if not, replacing the actual load with the capacity indicated by the charging electric energy meter, then judging the electricity utilization time of one day according to the produced commercial electricity utilization and the domestic electricity utilization of different electricity utilization users, and multiplying the electricity utilization time by the electricity stealing time to obtain electricity stealing amount; if so, predicting the daily electric quantity by using an algorithm, and calculating the electric stealing quantity of the current day of stopping electric stealing by multiplying the daily electric quantity by the electric stealing days;

6) and storing the calculated electricity stealing amount and the calculation basis, and modifying the correction coefficient of the database server when the difference value between the calculated electricity stealing amount and the actual electricity stealing amount exceeds a set threshold value.

As a preferable technical means: when the number of electricity stealing days is calculated in the step 2), the specific steps of adopting a clustering algorithm and/or a PQUI algorithm are as follows;

201) and (3) clustering algorithm: dividing the power consumption into three types, and dividing the power consumption into a label 1 type, a label 2 type, a label 3 type and a label 3 type according to the clustering center of the power consumption; regarding the 3 types of labels, if the time corresponding to the electricity consumption of the 3 types of labels, namely the abnormal time continuously exceeds 15 days, the continuous abnormal time is considered as electricity stealing time;

202) PQUI algorithm: carrying out relevant calculation according to the load of the user and extracting useful features; the three-phase four-wire calculation formula is as follows:

S₁＝U_aI_a+U_bI_b+U_cI_c……(1)

K＝(S₁-S₂)/W……(3)

in the formula of U_a，U_b，U_c，I_a，I_b，I_cThree-phase voltage and three-phase current respectively; w is the electricity consumption; p and Q are respectively active power and reactive power; s₁，S₂Is the apparent power;

for normal users, the electricity consumption condition of the users at ordinary times can not be changed too much, and the fluctuation of the voltage and the current can not be too large, so that the K value can not fluctuate too much and can be in a stable range; when a user has an electricity stealing behaviour, the change in the electricity consumption behaviour directly leads to a change in the relationship between the loads, S₁，S₂The difference between the two values is increased, the electricity consumption W is reduced, and therefore the ratio K is increased; therefore, when the K value is far larger than the K value in normal electricity utilization, the electricity utilization of the day is considered to be abnormal, and if the abnormal K value appears for 15 consecutive days, the user is considered to steal the electricity;

the PQUI method requires electric quantity, voltage, current and power for calculation, so that the PQUI method is not suitable for low-voltage users, and can only calculate special transformer users; clustering only needs to use electric quantity data, so that low-voltage and special transformer users can use the data; therefore, when the user is a low-voltage user, the clustering algorithm is used for judging the electricity stealing time; when the user is a special transformer user, the condition needs to be supplemented; the PQUI recognition algorithm judges whether the power utilization is abnormal or not through K value fluctuation, so that when the fluctuation is small, accurate judgment cannot be carried out; the clustering algorithm directly judges through the continuous time of the lowest class, so that the defect of small fluctuation is avoided, but because of the need of multiple iterations, a large amount of time is needed for calculation; in order to avoid the defects of the two algorithms, a parameter D is added to calculate the fluctuation condition through distance, and then a proper algorithm is selected according to the D; the specific calculation is as follows:

firstly, carrying out normalization processing on data:

in the formula, W_i(i-1, 2,3, …, n) is a set of electrical data of a user, W_minIs the minimum value in the data, W_maxIs the maximum value in the data;

then, the parameter D is calculated:

D＝|W_i ^*-mean(W^*)|/n……(5)

in the formula, W^*Is W_i ^*Sum of all data in (1), mean (W)^*) The average value of the normalized data is obtained, and n is the number of the data;

setting the threshold value to be 0.2, and using a PQUI algorithm when D is more than or equal to 0.2; when D <0.2, a clustering algorithm is used.

Through a large amount of experimental calculation and comparison, the threshold value is 0.2, the calculation result is more accurate, and the calculation result is most matched with the actual electricity stealing time.

As a preferable technical means: the electric energy metering device remote monitoring and diagnosing system also comprises an electricity stealing behavior characteristic library establishing module, and comprises the following steps when the electricity stealing behavior characteristic library is established:

A1) acquiring historical electricity stealing data;

A2) processing historical electricity stealing data; obtaining electricity stealing methods, numbering each electricity stealing method, and classifying; the classification of electricity stealing techniques includes: electricity stealing without meters, disconnection of a voltage loop, poor contact of the voltage loop, voltage division of the voltage loop, open circuit of a current loop, short circuit of the current loop, shunt of the current loop, phase-shifting electricity stealing, change of the internal structure of the electricity meter, damage of the electricity meter by large current or mechanical force and external interference;

A3) establishing a corresponding electricity stealing behavior feature library aiming at three metering modes of high supply and high metering, high supply and low metering and low supply and low metering; wherein:

the data of the electricity stealing behavior characteristic library in the high power supply and high metering mode comprise categories, output electricity stealing methods corresponding to the categories, A phase voltage, C phase voltage, A phase current, C phase current and active power; judging a fault phase and an electricity stealing method according to the phase voltage, the phase current and the active power of each phase when judging electricity stealing behavior;

the data of the electricity stealing behavior characteristic library under the high-power supply and low-metering mode comprise categories, output electricity stealing methods corresponding to the categories, A-phase voltage, B-phase voltage, C-phase voltage, A-phase current, B-phase current, C-phase current and active power; judging a fault phase and an electricity stealing method according to the phase voltage, the phase current and the active power of each phase when judging electricity stealing behavior;

the data of the electricity stealing behavior feature library in the low supply and low metering mode comprise categories, output electricity stealing methods and electric quantity corresponding to the categories; and judging the fault phase and the electricity stealing method according to the electric quantity when judging the electricity stealing behavior.

The technical scheme provides a perfect electricity stealing behavior data characteristic library on the basis of three typical electricity utilization types, namely the electricity stealing behavior characteristic library is established aiming at electricity utilization characteristics of high supply and high metering, high supply and low metering and low supply and low metering. On the basis, a perfect matching mechanism is established, the electricity stealing behavior matching is carried out on the user data actually containing the suspicion of electricity stealing, the possible electricity stealing behavior is identified, technical support is provided for electricity stealing verification, the electricity stealing identification efficiency is effectively improved, and the labor cost is reduced.

As a preferable technical means: the electricity stealing methods are classified into electricity stealing methods without meters and comprising the following steps: the voltage transformer is disconnected with a wire Q1, and is connected across a transformer cross-over Q2, and is connected with a bypass cross-over electric energy meter Q3;

the electricity stealing methods classified as the broken line of the voltage loop include: the method comprises the following steps of loosening a fuse Q4 of the TV, breaking a fuse Q5 in a fuse tube, loosening a connecting terminal Q6 of a voltage loop, breaking a wire core Q7 of a voltage loop wire and loosening a voltage connecting piece Q8 of the electric energy meter;

the classification of electricity stealing methods into poor-contact electricity stealing methods of voltage loops comprises the following steps: a voltage connecting piece Q9 of the electric energy meter is unscrewed, a connecting terminal Q10 of a voltage loop is unscrewed, and a low-voltage fuse Q11 of the TV is unscrewed;

the electricity stealing method classified into voltage division of the voltage loop comprises the following steps: a resistor Q12 is connected in series in a secondary circuit of the TV, a zero line on the wire inlet side of the single-phase meter is broken, and a resistor voltage drop Q13 is connected in series between the outlet wire and the ground;

the electricity stealing methods classified as open-circuit current circuit electricity stealing methods include: loosening a TA secondary outgoing line terminal Q14, artificially manufacturing a poor contact fault Q15 of a connecting terminal in a TA secondary circuit, breaking a wire core Q16 of a current circuit lead, and breaking a zero line and electricity stealing Q17;

the electricity stealing methods classified as short circuit of current loop include: the current terminal Q18 of the short circuit electric energy meter, the terminal row Q19 in the short circuit current loop, and the short circuit TA primary side or secondary side Q20;

the electricity stealing methods classified into current loop shunting include: replacing TA Q21 with different transformation ratios, changing a secondary tap Q22 of a tapped TA and changing the number of primary turns Q23 of a core-through TA;

the electricity stealing methods classified into phase-shift electricity stealing methods include: the phase line and the zero line of the single-phase meter are interchanged, and the ground wire is used as the zero line Q24; exchanging an inlet/outlet line Q25 on the primary side of the TA; exchanging a homonymous terminal Q26 on the secondary side of the TA; exchanging an incoming and outgoing line Q27 of a current terminal of the electric energy meter; exchanging TA to a phase Q28 of a connecting line of the electric energy meter; swapping the polarity Q29 of TV primary or secondary; exchanging the TV to a phase Q30 of a power meter connecting line; phase-shifting Q31 with special inductor or capacitor;

the electricity stealing techniques classified into those for changing the internal structure of the electricity meter include: reducing the number of current coil turns Q32; performing resistance spot welding on the manganin, and cutting off a manganin signal wire Q33; the current sampling loop is connected in parallel and in series with a resistor Q34; replacing the voltage sampling loop grading sampling resistor Q35; a voltage coil series resistor and other electronic elements to divide the voltage Q36; a copper wire hook short circuit Q37; implanting a remote control shunt Q38;

the electricity stealing techniques classified as electricity stealing techniques that damage electricity meters with large current or mechanical force include: burning out the current coil Q39 by overload current, impacting the electric meter Q40 by electric power of short-circuit current, and damaging the electric meter Q41 by mechanical external force;

the electricity stealing techniques classified as external interference include: the high-voltage pulse electricity stealing system comprises a strong magnetic interference electricity stealing Q42, a high-frequency interference electricity stealing Q43, a high-voltage pulse electricity stealing Q44 and a short-circuit metering box inlet and outlet wire Q45.

As a preferable technical means: the characteristic library of electricity stealing behavior under the high-supply and high-metering mode is as follows:

the electricity stealing behavior characteristic library under the high-power supply and low-metering mode is as follows:

the electricity stealing behavior characteristic library under the low-supply and low-metering mode is as follows:

。

as a preferable technical means: the electricity stealing behavior judgment comprises electricity stealing behavior identification in a high supply and low metering mode, electricity stealing behavior identification in a high supply and high metering mode and electricity stealing behavior identification in a low supply and low metering mode;

the electricity stealing behavior identification under the high power supply and low metering mode comprises the following steps:

3101) inputting electricity utilization data of suspected electricity stealing users;

3102) judging A, B, C whether the phase current voltage has a value; if not, determining that the judgment cannot be carried out, and finishing; if yes, entering the next step;

3103) judging whether the three-phase voltage and the three-phase current are both close to 0; if yes, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 9 types of electricity stealing reasons, and ending; if not, entering the next step;

3104) judging whether the three-phase current has one phase or multiple phases close to 0; if yes, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 1 type of electricity stealing reason and an electricity stealing phase with the output current close to 0, and ending; if not, entering the next step;

3105) judging whether the three-phase voltage has one phase or multiple phases close to 0; if yes, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting the type 2 electricity stealing reasons and the electricity stealing phase with the output voltage close to 0, and ending; if not, entering the next step;

3106) summoning the split-phase power factor;

3107) judging whether all the three-phase power factors are normal or not; if not, go to step 2110; if yes, entering the next step;

3108) judging whether the three-phase voltages are all larger than or close to 220V; if not, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 4 types of electricity stealing reasons and electricity stealing phases with undersize output voltage, and ending; if yes, entering the next step;

3109) judging whether the three-phase currents are equal, if not, comparing the three-phase currents with the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 3 types of electricity stealing reasons and electricity stealing phases with smaller output currents, and ending; if yes, outputting 3 types of electricity stealing reasons and electricity stealing phases with smaller current, and ending;

3110) judging whether the daily electric quantity of the user is 0 or not; if yes, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 6 or 8 types of electricity stealing reasons and electricity stealing phases with abnormal power factors, and ending; if not, entering the next step;

3111) judging whether the three-phase current has one phase or multiple phases smaller than 0; if yes, comparing the electricity stealing behavior feature library in the high-power supply and low-metering mode, outputting 5 types of electricity stealing reasons and electricity stealing phases with negative current, and ending; if not, outputting 7 types of electricity stealing reasons and electricity stealing phases with negative voltage, and ending;

the electricity stealing behavior identification under the high power supply and high metering mode comprises the following steps:

3201) inputting electricity data of suspected electricity stealing users

3202) Judging A, C whether the phase current and voltage have values, if not, judging that the judgment cannot be carried out and finishing;

3203) judging whether A, C phase voltage and current are both close to 0; if yes, comparing the electricity stealing behavior feature library in the high-supply high-count metering mode, outputting 9 types of electricity stealing reasons, and ending; if not, entering the next step;

3204) determining A, C whether the phase voltage has one or more phases approaching 0; if yes, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 1 type electricity stealing reason and an electricity stealing phase with the voltage close to 0, and ending; if not, entering the next step;

3205) determining A, C whether the phase current has one or more phases that are close to 0; if yes, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, and outputting the type 2 electricity stealing reasons and the electricity stealing phases with the power supply close to 0; if not, entering the next step;

3206) summoning the split-phase power factor;

3207) judging whether A, C-phase power factors are all normal; if not, go to step 2210; if yes, entering the next step;

3208) determining A, C whether the phase voltages are both greater than or near 220V; if not, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 3 types of electricity stealing reasons and electricity stealing phases with undersize voltage, and ending; if yes, entering the next step;

3209) judging A, C whether the phase currents are equal; if yes, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 4 types of electricity stealing reasons and A, C two phases of electricity stealing phases, and ending; if not, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 4 types of electricity stealing reasons and electricity stealing phases with smaller current, and ending;

3210) determining A, C whether the phase voltage is less than 0; if yes, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 7 types of electricity stealing reasons and electricity stealing phases with the voltage less than 0, and ending; if not, entering the next step;

3211) judging A, C whether the phase current is less than 0; if yes, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting 5 types of electricity stealing reasons and electricity stealing phases with the current less than 0, and ending; if not, entering the next step;

3212) judging whether the daily electric quantity of the user is 0; if yes, comparing the electricity stealing behavior feature library in the high-supply high-metering mode, outputting 6 or 8 types of electricity stealing reasons and an electricity stealing phase AC or AB phase, and ending; if not, comparing the electricity stealing behavior feature library in the high power supply and high metering mode, outputting two phases of 8 types of electricity stealing reasons and electricity stealing phases BC, and ending;

the electricity stealing behavior identification under the low supply and low metering mode comprises the following steps:

3301) inputting electricity data of suspected electricity stealing users

3302) Judging whether the daily electric quantity of the user has a value or not; if not, determining that the judgment cannot be carried out, and finishing; if yes, entering the next step;

3303) judging whether the daily electric quantity of the user is close to 0; if yes, comparing the electricity stealing behavior feature library in the low-supply and low-metering mode, outputting the 1-class electricity stealing reason, and ending; if not, entering the next step;

3304) judging whether the daily electric quantity of the user is negative or not; if yes, comparing the electricity stealing behavior feature library in the low-supply and low-metering mode, outputting the 2 types of electricity stealing reasons, and ending; if not, comparing the electricity stealing behavior feature library in the low-supply and low-metering mode, outputting 3 types of electricity stealing reasons, and ending.

As a preferable technical means: the database server stores the electricity stealing behaviors and correction coefficients under each electricity stealing behavior, wherein the correction coefficients comprise a private-transformer high-supply and high-supply user correction coefficient, a private-transformer high-supply and low-supply user correction coefficient and a low-voltage user correction coefficient;

the user correction factors for the high supply and high supply of the special transformer are shown in the following table:

the private high and low user correction factors are shown in the following table:

the low voltage user correction factor is shown in the following table:

。

has the advantages that: the technical scheme solves the defects that the current electric quantity compensation can only be estimated, the accuracy is low, the persuasion is insufficient, and a large amount of manpower is consumed. According to the technical scheme, the clustering algorithm and the PQUI recognition algorithm are adopted to calculate the corresponding electricity stealing amount, the calculation matching is good, the accuracy is high, the acquired data can meet the requirements, the compensation amount of the electricity stealing users can be accurately, objectively and quickly calculated, the compensation amount is consistent with the actually generated electricity, and the electricity utilization fairness is improved.

Drawings

FIG. 1 is a flow chart of the present invention.

Fig. 2 is a line graph of daily electricity usage by a low voltage customer.

FIG. 3 is a graph of the results of the user's power usage through a clustering algorithm.

FIG. 4 is a line graph of daily power usage by a specific power consumer.

FIG. 5 is a proprietary user PQUI result diagram.

Detailed Description

The technical scheme of the invention is further explained in detail by combining the drawings in the specification.

As shown in fig. 1, a cluster and PQUI recognition algorithm-based electric quantity compensation method is implemented by using an electric energy metering device remote monitoring and diagnosing system, wherein the electric energy metering device remote monitoring and diagnosing system comprises an electric energy information acquisition system, a data collection and analysis server and a database server; the electricity utilization information acquisition system acquires electricity utilization data and events of power consumers in real time; the data collection and analysis server analyzes and calculates electricity stealing behaviors and corresponding compensation electric quantity according to the acquired data; the database server stores electricity consumption data, events, analysis and judgment thresholds, an electricity stealing behavior feature library and correction coefficients under each electricity stealing behavior, and the data collection and analysis server comprises the following steps when analyzing and calculating the compensation electric quantity:

2) judging whether the number of electricity stealing days can be determined; if yes, directly calculating according to the determined days; if not, calculating by preset days;

When the number of electricity stealing days is determined in the step 2), a clustering algorithm and/or a PQUI algorithm are/is adopted;

201) when the number of electricity stealing days is calculated in the step 2), the specific steps of adopting a clustering algorithm and/or a PQUI algorithm are as follows;

S₁＝U_aI_a+U_bI_b+U_cI_c……(1)

K＝(S₁-S₂)/W……(3)

for normal users, the electricity consumption condition of the users at ordinary times can not be changed too much, and the fluctuation of the voltage and the current can not be too large, so that the K value can not fluctuate too much and can be in a stable range; when a user has an electricity stealing behaviour, the change in the electricity consumption behaviour directly leads to a change in the relationship between the loads, S₁，S₂Decrease, difference of bothThe value becomes larger, the used amount W decreases, and therefore the ratio K becomes larger; therefore, when the K value is far larger than the K value in normal electricity utilization, the electricity utilization of the day is considered to be abnormal, and if the abnormal K value appears for 15 consecutive days, the user is considered to steal the electricity;

firstly, carrying out normalization processing on data:

then, the parameter D is calculated:

D＝|W_i ^*-mean(W^*)|/n……(5)

The electric energy metering device remote monitoring and diagnosing system also comprises an electricity stealing behavior characteristic library establishing module, and comprises the following steps when the electricity stealing behavior characteristic library is established:

A1) acquiring historical electricity stealing data;

The electricity stealing methods are classified into electricity stealing methods without meters and comprising the following steps: the voltage transformer is disconnected with a wire Q1, and is connected across a transformer cross-over Q2, and is connected with a bypass cross-over electric energy meter Q3;

The electricity stealing behavior is expressed in tabular form as:

watch 1 stealing electricity behavior watch

The characteristic library of electricity stealing behavior under the high-supply and high-metering mode is as follows:

the electricity stealing behavior judgment comprises electricity stealing behavior identification in a high supply and low metering mode, electricity stealing behavior identification in a high supply and high metering mode and electricity stealing behavior identification in a low supply and low metering mode;

3106) summoning the split-phase power factor;

3201) inputting electricity data of suspected electricity stealing users

3206) summoning the split-phase power factor;

3301) inputting electricity data of suspected electricity stealing users

The database server stores the electricity stealing behaviors and correction coefficients under each electricity stealing behavior, wherein the correction coefficients comprise a private-transformer high-supply and high-supply user correction coefficient, a private-transformer high-supply and low-supply user correction coefficient and a low-voltage user correction coefficient;

the user correction factors for the high supply and high supply of the special transformer are shown in the following table 2:

TABLE 2 SPECIFIC CHANGE HIGH SUPPLY-HIGH CALCULATION USER CORRECTION COEFFICIENCY TABLE

The private high-and low-user correction coefficients are shown in table 3:

TABLE 3 correction coefficient tables for exclusive HDR/LF users

The low voltage user correction factor is shown in table 4:

table 4 low voltage user correction coefficient table

The technical solution is further illustrated by the following specific examples:

calculation process of certain low-voltage user compensation electric quantity based on clustering algorithm for judging electricity stealing time

As shown in fig. 2, electricity data of a low-voltage user is obtained, a line graph of daily electricity consumption as shown in fig. 2 is generated, and a result schematic diagram after clustering processing as shown in fig. 3 is generated.

The lowest class is numbered 2345789222526353637383940414243444546..... 674675676677678679680681704808810811813 (1 for 2017.01.01; 820 for 2018.12.31)

The analysis shows that: the first over 15 days of low power usage is from time 2017.02.04 and over 60 days, as shown in the box of FIG. 2. The electricity stealing time is directly judged.

As a result: day 2017.02.04 is the time when electricity theft began.

And (3) calculating the compensation electric quantity: considering that the electricity stealing time of the user is 2 months and 4 days in 2017, the electricity quantity of the user is close to 0 during the electricity stealing period, so that the electricity stealing method can be judged to be one of the following methods: the bypass type electric energy meter is additionally connected with a bypass type electric energy meter (Q3), a wiring terminal (Q9) of a unscrewing voltage loop, a voltage connecting piece (Q10) of the unscrewing electric energy meter, a current terminal (Q18) of the short-circuit electric energy meter, a copper wire hook short circuit (Q37) and the like. Therefore, in the electric quantity compensation process, the capacity (rated voltage and rated current) indicated by the rated current value of the charging electric energy meter is determined to replace the actual load (xKW), the electricity stealing days are 717 days (2017.02.04-2019.01.23), the domestic electricity consumption of the residential users is calculated according to 6 hours, and the compensation electric quantity is xKW, 6, 717 4296xKWh

Actual compensation of electric quantity:

the daily electric quantity is: 3.7 × 6 ═ 22.2kwh

And (3) totalizing the compensation electric quantity: 22.2 × 180 ═ 3996kwh

Supplementing the electricity charge: 3996 x 0.538 x 2149.85 yuan

Default electricity usage: (3996 × 0.538) × 3 ═ 6449.54 yuan

Adding additional electric charge: 2149.85+ 6449.54-8599.39 yuan

Second, a calculation process of compensation electric quantity of a specific transformer user for judging electricity stealing time based on PQUI algorithm

Acquiring power consumption data of a certain special transformer user, and generating a line graph of daily power consumption of the certain special transformer user as shown in fig. 4; and as shown in FIG. 5, a result diagram of the user PQUI determination method.

Judging the electricity stealing time: as can be seen from fig. 4, the power consumption is abnormal during the period from 17 th 6 to 3 rd 2016 (as indicated by the box in fig. 4). As can be seen from fig. 5, the ratio increases (as shown in the box of fig. 5) during the same time, and the ratio is far higher than the normal value for 50 consecutive days, so that the power consumption is considered abnormal during this time. The electricity utilization abnormality in the time is considered in both aspects. Thus, electricity stealing time is considered to be 6 months and 17 days to 2016 and 8 months and 3 days.

Identification of electricity stealing skills: as can be seen from fig. 4, since the amount of electricity used is decreased 2/3 during the period of electricity stealing, it is assumed that the user uses a method of reducing or cutting off the current of two phases thereof, which corresponds to Q14-Q20 in table 1 and category 1 in table 3.

Electric quantity compensation: when the two-phase current is reduced, the correction coefficient is 3, so the compensation electric quantity is:

W_pursuing＝W_{Theft-proof device}*3-W_{Theft-proof device}＝178212*3-178212＝296416(W)

Actual compensation results:

the electricity to be measured is 3UI Cos phi T

＝3*225.5*612.9*0.98*1128/1000

458345 degrees

Less electricity metering: 458345 + 161929 degrees.

The method for supplementing electric quantity based on clustering and PQUI recognition algorithm shown in FIG. 1 is a specific embodiment of the present invention, already embodies the essential features and progress of the present invention, and can make equivalent modifications in shape, structure, etc. according to the practical use requirements, and is within the scope of protection of the present invention.

Claims

1. A cluster and PQUI recognition algorithm-based electric quantity tracing and supplementing method is characterized by comprising the following steps: the remote monitoring and diagnosing system is implemented by adopting an electric energy metering device, and comprises an electricity utilization information acquisition system, a data collection and analysis server and a database server; the electricity utilization information acquisition system acquires electricity utilization data and events of power consumers in real time; the data collection and analysis server analyzes and calculates electricity stealing behaviors and corresponding compensation electric quantity according to the acquired data; the database server stores electricity consumption data, events, analysis and judgment thresholds, an electricity stealing behavior feature library and correction coefficients under each electricity stealing behavior, and the data collection and analysis server comprises the following steps when analyzing and calculating the compensation electric quantity:

2. The method of claim 1 for supplementing electric power based on clustering and PQUI recognition algorithm, wherein: when the number of electricity stealing days is calculated in the step 2), the specific steps of adopting a clustering algorithm and/or a PQUI algorithm are as follows;

S₁＝U_aI_a+U_bI_b+U_cI_c……(1)

K＝(S₁-S₂)/W……(3)

in the formula of U_a,U_b,U_c,I_a,I_b,I_cThree-phase voltage and three-phase current respectively; w is the electricity consumption; p and Q are respectivelyActive power and reactive power; s₁,S₂Is the apparent power;

for normal users, the electricity consumption condition of the users at ordinary times can not be changed too much, and the fluctuation of the voltage and the current can not be too large, so that the K value can not fluctuate too much and can be in a stable range; when a user has an electricity stealing behaviour, the change in the electricity consumption behaviour directly leads to a change in the relationship between the loads, S₁,S₂The difference between the two values is increased, the electricity consumption W is reduced, and therefore the ratio K is increased; therefore, when the K value is far larger than the K value in normal electricity utilization, the electricity utilization of the day is considered to be abnormal, and if the abnormal K value appears for 15 consecutive days, the user is considered to steal the electricity;

firstly, carrying out normalization processing on data:

then, the parameter D is calculated:

D＝|W_i ^*-mean(W^*)|/n……(5)

setting the threshold value to be 0.2, and using a PQUI algorithm when D ≧ 0.2; when D <0.2, a clustering algorithm is used.

3. The method of claim 2 for supplementing electric power based on clustering and PQUI recognition algorithm, wherein: the electric energy metering device remote monitoring and diagnosing system also comprises an electricity stealing behavior characteristic library establishing module, and comprises the following steps when the electricity stealing behavior characteristic library is established:

A1) acquiring historical electricity stealing data;

4. The method of claim 3 for supplementing electric quantity based on clustering and PQUI recognition algorithm, wherein: the electricity stealing methods are classified into electricity stealing methods without meters and comprising the following steps: the voltage transformer is disconnected with a wire Q1, and is connected across a transformer cross-over Q2, and is connected with a bypass cross-over electric energy meter Q3;

5. The method of claim 4 for supplementing electric quantity based on clustering and PQUI recognition algorithm, wherein:

。

6. the method of claim 5 for supplementing electric power based on clustering and PQUI recognition algorithm, wherein: the electricity stealing behavior judgment comprises electricity stealing behavior identification in a high supply and low metering mode, electricity stealing behavior identification in a high supply and high metering mode and electricity stealing behavior identification in a low supply and low metering mode;

3106) summoning the split-phase power factor;

3201) inputting electricity data of suspected electricity stealing users

3206) summoning the split-phase power factor;

3301) inputting electricity data of suspected electricity stealing users

7. The method of claim 5 for supplementing electric power based on clustering and PQUI recognition algorithm, wherein: the database server stores the electricity stealing behaviors and correction coefficients under each electricity stealing behavior, wherein the correction coefficients comprise a private-transformer high-supply and high-supply user correction coefficient, a private-transformer high-supply and low-supply user correction coefficient and a low-voltage user correction coefficient;

the low voltage user correction factor is shown in the following table: