CN114519514B

CN114519514B - Low-voltage transformer area reasonable line loss value measuring and calculating method, system and computer equipment

Info

Publication number: CN114519514B
Application number: CN202210103803.4A
Authority: CN
Inventors: 姜磊; 张曼; 程绪敏; 杨钊; 崔在京
Original assignee: Brilliant Data Analytics Inc
Current assignee: Brilliant Data Analytics Inc
Priority date: 2022-01-27
Filing date: 2022-01-27
Publication date: 2023-09-19
Anticipated expiration: 2042-01-27
Also published as: CN114519514A

Abstract

The invention belongs to the technical field of electric power, and discloses a method, a system and computer equipment for calculating reasonable line loss measurement values of a low-voltage transformer area, wherein the method comprises the following steps: acquiring original data and preprocessing; extracting multidimensional characteristic indexes including capacity type basic indexes, operation type indexes, platform type indexes and electric quantity type indexes for each platform area according to the preprocessed data, and constructing a platform area line loss potential characteristic index system; calculating mutual information between each dimension characteristic index and the line loss rate of the transformer area, realizing importance sorting of the characteristic indexes of the transformer area, and screening the important characteristic indexes to form the line loss characteristic indexes of the transformer area by combining the correlation among the dimension characteristic indexes; constructing a low-voltage area classification model, and classifying areas; screening various typical samples of the areas for model training; and respectively constructing a reasonable line loss calculation model of each type of the transformer area. The invention classifies the areas and builds a corresponding reasonable line loss calculation model of the areas, and can accurately calculate the theoretical line loss of the areas.

Description

Low-voltage transformer area reasonable line loss value measuring and calculating method, system and computer equipment

Technical Field

The invention belongs to the technical field of electric power, and particularly relates to a low-voltage station reasonable line loss measurement value calculation method, a system and computer equipment based on a big data algorithm.

Background

In the aspect of power station area management, the line loss rate is a comprehensive reflection of the economic operation management level of the power grid and the economic benefit of a power supply enterprise, and the line loss can be divided into technical line loss (theoretical line loss) and management line loss. At present, unified line loss rate indexes are adopted to evaluate the line loss of the transformer area, and overall management of the line loss is basically realized. With the improvement of line loss management level in recent years, the requirement of the lean degree of line loss management is higher and higher, and the original evaluation mode of line loss value 'one-cut' is changed into 'one-area one-index' new management mode construction, so that the loss reduction is scientifically guided. The accurate theoretical line loss calculation provides important basis for the management mode of one area and one index and the establishment of loss reduction measures.

The transformer area line loss is affected by load rate, power supply radius, three-phase unbalance degree, number of users, load characteristics, terminal electric quantity duty ratio, head-end voltage, grid structure, power factor and the like, and because the transformer area power supply mode is various and the structure is complex, accurate topological structure, line parameters and operation data cannot be obtained in actual conditions, and the electric energy loss of a low-voltage power network is difficult to calculate in detail and accurately.

Disclosure of Invention

The invention provides a method, a system and computer equipment for measuring and calculating reasonable line loss values of a low-voltage transformer area, which are used for extracting relevant original data of the line loss of the transformer area, constructing a characteristic index of the line loss of the transformer area, classifying the transformer area and respectively constructing a corresponding reasonable line loss calculation model of the transformer area, so that theoretical line loss of the transformer area is accurately calculated, and the problems in the existing service mode are solved.

The technical scheme adopted by the method is as follows: the method for measuring and calculating the reasonable line loss value of the low-voltage station area comprises the following steps:

s1, acquiring original data, wherein the original data sources comprise a marketing business application system, an electricity consumption information acquisition system, a power grid PMS2.0 system and a marketing GIS system;

s2, preprocessing the obtained original data;

s3, extracting multidimensional characteristic indexes including capacity type basic indexes, operation type indexes, platform type indexes and electric quantity type indexes for each platform area according to the preprocessed data, and constructing a platform area line loss potential characteristic index system;

s4, calculating mutual information between each dimension characteristic index and the line loss rate of the transformer area according to the constructed potential characteristic index system of the line loss of the transformer area, and realizing importance sorting of the characteristic indexes of the transformer area; screening important characteristic indexes to form a station area line loss characteristic index by combining the correlation among the dimension characteristic indexes;

s5, constructing a low-voltage platform region classification model, and classifying the platform region by adopting a clustering algorithm based on four indexes of the network power occupation ratio, the grid structure, the power supply radius and the power supply quantity of the platform region;

s6, analyzing data samples of various areas, and screening typical samples of various areas for model training;

s7, aiming at each type of area, building an area reasonable line loss calculation model by using an MLP neural network algorithm.

The technical scheme adopted by the system of the invention is as follows: a low-voltage station reasonable line loss value measuring and calculating system comprises:

the data acquisition module is used for acquiring original data, and the original data sources comprise a marketing business application system, an electricity consumption information acquisition system, a power grid PMS2.0 system and a marketing GIS system;

the preprocessing module is used for preprocessing the acquired original data;

the index system construction module is used for extracting multidimensional characteristic indexes including capacity type basic indexes, operation type indexes, platform type indexes and electric quantity type indexes for each platform area according to the preprocessed data, and constructing a platform area line loss potential characteristic index system;

the characteristic index screening module is used for calculating mutual information between each dimension characteristic index and the line loss rate of the transformer area according to the constructed potential characteristic index system of the line loss of the transformer area, and realizing importance sorting of the characteristic indexes of the transformer area; screening important characteristic indexes to form a station area line loss characteristic index by combining the correlation among the dimension characteristic indexes;

the system comprises a platform area classification module, a clustering module and a power supply module, wherein the platform area classification module is used for constructing a low-voltage platform area classification model and classifying the platform areas by adopting the clustering algorithm based on four indexes of the network power occupation ratio, the grid structure, the power supply radius and the power supply quantity of the platform areas;

the sample screening module is used for analyzing data samples of various areas and screening typical samples of various areas for model training;

the line loss calculation module is used for respectively constructing a reasonable line loss calculation model of each type of the area by using an MLP neural network algorithm.

The computer equipment comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, and when the processor executes the computer program, the method for measuring and calculating the reasonable line loss value of the low-voltage station area is realized.

Compared with the prior art, the invention has the following advantages and beneficial effects:

the invention extracts the relevant original data of the line loss of the area based on a big data algorithm, constructs the line loss characteristic index of the area, realizes the classification of the area through an EM clustering algorithm, constructs corresponding reasonable line loss calculation models of the area respectively for different types of areas based on an MLP neural network algorithm, thereby accurately calculating the theoretical line loss of the area and providing important basis for the management mode of one area and one index and the establishment of loss reduction measures.

Drawings

FIG. 1 is a flow chart of a method for measuring and calculating reasonable line loss values of a low-voltage transformer area in an embodiment of the invention;

FIG. 2 is a schematic diagram illustrating importance ranking of a region feature index according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a result of classifying a region by using an unsupervised EM clustering algorithm in an embodiment of the present invention;

FIG. 4 is a schematic diagram of an MLP neural network constructed in an embodiment of the invention;

fig. 5 shows the prediction result of the MLP neural network in the embodiment of the present invention.

Detailed Description

The technical scheme of the present invention will be described in further detail below with reference to examples and drawings, but the embodiments of the present invention are not limited thereto.

Example 1

As shown in fig. 1, the process of the method for measuring and calculating reasonable line loss value of low-voltage area includes the steps of data acquisition, data analysis and processing, construction of area line loss characteristic indexes, construction of low-voltage area classification models, screening of training samples of various areas, construction of area reasonable line loss calculation models and the like, and specifically includes the following steps:

step 1, data acquisition: raw data are acquired from a marketing business application system, an electricity consumption information acquisition system, a power grid PMS2.0 system, a marketing GIS system and the like. The method comprises the steps of mainly extracting archive data such as a platform area, an electric energy meter, an electric meter box, a metering point and the like from a marketing business application system; the method mainly comprises the steps of mainly extracting electric quantity, current, voltage and other collected data from an electricity consumption information collecting system; information data such as a circuit, a transformer and the like are mainly extracted from a power grid PMS2.0 system; and mainly extracting the coordinate information of the meter box and the transformer from the marketing GIS system.

Step 1.1, data acquisition of a marketing business application system: obtaining the information of the area file, wherein the information comprises fields such as a statistic date, a unit number, an area ID, an area number, an area name, an area capacity, an installation address and the like; acquiring electric energy meter information, wherein the electric energy meter information comprises electric energy meter identification, asset number, electric energy meter bar code, metering point identification, power supply unit number, power supply area code, user number, comprehensive multiplying power and other fields; obtaining table box information, wherein the table box information comprises fields such as table box ID, table box asset number and the like; and acquiring metering point information, wherein the metering point information comprises fields such as metering point identification, user number, station area identification, line identification, meter reading section number, metering point classification, wiring mode and the like.

Step 1.2, data acquisition of an electricity consumption information acquisition system: acquiring user daily electric quantity detail data, wherein the user daily electric quantity detail data comprises a unit number, a user number, an electric quantity type, an indication type, a last indication, a current indication, an acquired electric quantity, an estimated electric quantity and other fields; acquiring user solar voltage information, wherein the user solar voltage information comprises a metering point identifier, a power supply unit, a user number, an electric energy meter identifier, an A-phase voltage, a B-phase voltage, a C-phase voltage and other fields; the method comprises the steps of obtaining platform area assessment table information, wherein the platform area assessment table information comprises platform area identification, CT multiplying power, PT multiplying power, self multiplying power, A-phase voltage, B-phase voltage, C-phase voltage, A-phase current, B-phase current, C-phase current, active total power, reactive total power and other fields.

Step 1.3, data acquisition of a PMS2.0 system: extracting line information, including fields such as low-voltage line ID, line name, electric nameplate operation library ID, equipment code, operation number, transformer type, belonging transformer, belonging city, operation maintenance unit, maintenance team, voltage class, operation state, line type, line total length, overhead line length, cable line length, etc.; and extracting the circuit information of the transformer, wherein the circuit information comprises a column transformer ID, an electric nameplate operation library ID, the belonging circuit, the belonging tower, the equipment name, the operation number, the voltage class, the operation and maintenance unit, the belonging city and other fields.

Step 1.4, data acquisition of a marketing GIS system: the method mainly extracts the longitude and latitude space coordinate information of the meter box and the longitude and latitude space coordinate information of the transformer.

Step 2, data analysis and processing: the obtained original data is subjected to preprocessing such as cleaning and the like, and the integrity, consistency, uniqueness, null value and the like of the data are mainly checked.

Step 2.1, integrity and uniqueness check: for archival data, firstly checking the integrity of the data, for example checking the integrity of all users, meters and meter box archival information under a platform area, and if the record is missing, completing the archival data; and secondly, carrying out uniqueness check on the platform area, the user, the meter and the meter box, and taking only one record if repeated records appear.

Step 2.2, consistency check: and checking whether the data is satisfactory or not according to the reasonable value range and the interrelation of each variable, and eliminating the data which is beyond the normal range, is unreasonable in logic or contradicts each other. And (3) for the numerical variable, counting the variable distribution condition, judging the normal value range according to service experience, automatically identifying each variable value out of range, and screening and checking out-of-range example data. For example, for metrology acquired data, there are few data jumps, even negative values, for which outliers are directly replaced with null values.

Step 2.3, checking null value: for measurement and collection data, due to the fact that the data collection frequency is high, partial collection point loss phenomenon exists, due to the fact that time continuity exists in collection type data, the missing data and a plurality of front and rear sampling data have strong correlation, so that fewer records are deleted in the data, null filling can be carried out according to the average value of the front and rear 10 sampling points of the missing data, the records with more missing data are checked whether collection and transmission communication is faulty, if so, a large amount of missing data is caused due to communication reasons, and the data can be retransmitted after the communication is normal.

Step 3, constructing a potential characteristic index system of the line loss of the transformer area: and 2, extracting multidimensional characteristic indexes such as capacity type basic indexes, operation type indexes, platform type indexes, electric quantity type indexes and the like from each platform area according to the data preprocessed in the step 2, and constructing a platform area line loss potential characteristic index system, wherein the characteristic indexes and the index descriptions of the dimensions are shown in the following table 1.

TABLE 1 potential characteristic index of line loss for regions

Let epsilon _i For the three-phase unbalance, in this embodiment, a percent form is adopted, and the calculation formula of the three-phase unbalance in the operation index is as follows:

wherein I is _max Is the maximum one-phase load current, i.e. the maximum of the three-phase load currents; i _av The average value of the load currents of the three phases is the average value of the load currents of the phase A, the phase B and the phase C.

Let the total surface root mean square current I of the platform area _rms Average value I of three-phase load current _av And taking the equivalent relation k as a load shape factor in the operation class index, the calculation formula of the load shape factor is as follows:

wherein T is the number of current measurement points determined by a calculation period, and i is the current sequence number of the measurement points; i _i Is the current at the i-th measurement point.

Defining the power supply radius in the area category index as the space coordinate distance from the farthest meter box to the transformer, and calculating the power supply radius by the following formula: assume that the longitude and latitude (i.e. space coordinates) of the two points of the farthest meter box and the transformer are respectively (lon) ₁ ,lat ₁ )、(lon ₂ ,lat ₂ ) The power supply radius (namely the distance between the farthest meter box and the two points of the transformer) v _d The calculation formula is as follows:

in the formula, v _h For the maximum circular section radian passing through the center of the earth, v _d Is the space coordinate distance between the farthest meter box and the transformer.

Step 4, screening the characteristic index of the line loss of the transformer area

Calculating mutual information between each dimension characteristic index and the line loss rate of the transformer area according to the potential characteristic index system of the transformer area line loss constructed in the step 3, and realizing importance sequencing of the characteristic indexes of the transformer area, as shown in fig. 2; and screening important characteristic indexes to form a line loss characteristic index of the platform area by combining the correlation among the dimension characteristic indexes, and taking the important characteristic indexes as the input quantity of a later model.

The mutual information is an information measure in the information theory, is information shared between two random variables X, Y, and the larger the mutual information is, the higher the correlation between the variable X and the variable Y is. Mutual information can also be expressed as an amount by which uncertainty of the variable Y is reduced due to the introduction of the variable X, the larger the amount of reduction, the more advantageous the variable X is for determining the variable Y. Mutual information I (X; Y) is the relative entropy of the joint distribution p (X, Y) and the edge distribution p (X) p (Y), namely:

the mutual information between the potential characteristic index of the line loss of the area and the statistical line loss rate of the area in table 1 is calculated by using the sklearn. Feature_selection library in the python tool, as shown in fig. 2. The mutual information calculation is one-time calculation, and the embodiment is mainly used for determining the input index of the model; the statistical line loss rate of the station area is directly calculated by the total table of the station area according to the data directly acquired by the electricity consumption information acquisition system.

1) Important characteristic index

The influence of the loss electric quantity, voltage drop, the capacity of the station area, the voltage of the head end, the power supply quantity, the sales electric quantity and the voltage of the tail end on the line loss rate of the station area is large, and the method belongs to important characteristic indexes; the power supply quantity, the sales power quantity and the loss power quantity are all power quantity information representing the area, linear relations exist among the power supply quantity, the loss power quantity is calculated based on the power supply quantity, and therefore the three characteristic indexes only need the power supply quantity as the characteristic index; the head end voltage, the tail end voltage and the voltage drop are used for representing the voltage information of the platform region, and considering that the two indexes of the tail end voltage and the voltage drop are difficult to cover the full-quantity platform region, the three indexes only need the head end voltage as characteristic indexes.

2) Secondary characteristic index

The power supply radius, the three-phase unbalance, the terminal electric quantity duty ratio, the load shape factor, the single-phase meter counting, the number of users, the network electric quantity duty ratio, the three-phase meter counting, the grid structure and the power factor have general influence on the line loss rate, and belong to secondary characteristic indexes; the number of users, the table count, the single-phase table count and the three-phase table count are in strong linear correlation, belong to the same kind of characteristics, and only the number of users is needed to be used as a characteristic index.

3) General features

The user average capacity, the photovoltaic user number, the photovoltaic electric quantity and the urban and rural categories have small influence on the line loss rate, belong to general characteristics, and are temporarily not considered when a model is built in the later period.

In summary, finally, 11 characteristic indexes of head-end voltage, power supply quantity, power supply radius, three-phase unbalance degree, terminal electric quantity duty ratio, load shape factor, load rate, number of users, network-surfing electric quantity duty ratio, grid structure and power factor are selected as input quantity of a later model.

Step 5, constructing a low-voltage station classification model: based on four indexes of the power ratio of the network access of the transformer area, the grid structure, the power supply radius and the power supply quantity of the transformer area, the transformer area is classified by adopting an unsupervised EM clustering algorithm.

Because the line loss of the transformer area is affected by various factors, only partial characteristics can be grasped through service experience, and the overall rule is difficult to grasp, the embodiment utilizes the power ratio of the network access of the transformer area, the grid structure, the power supply radius and the index data of the power supply quantity of the transformer area, and adopts an unsupervised EM clustering algorithm to divide the low-voltage transformer area into a plurality of groups so as to support the work of constructing a reasonable line loss calculation model of the transformer area in the subsequent stage. In this embodiment, the number of clusters is set to 1 to 25, the BIC value for each cluster number is calculated, and the optimal cluster number is determined.

The reason for adopting EM clustering is mainly that the method is based on Gaussian distribution, and clustering results are elliptical geometrically, so that the method is wider in applicability. The EM clustering algorithm is mainly to calculate probabilities, and estimate the probability that a data point belongs to a cluster by using a given multi-element gaussian probability distribution model, namely, each cluster is regarded as a gaussian model, and is mainly performed by two steps alternately, and the total log likelihood is increased by the continuous alternation of E-step and M-step until the model converges.

Taking a platform area (i.e. an active platform area) with the network electric quantity duty ratio of the extraction platform area being larger than zero as an example, extracting the network electric quantity duty ratio, a grid structure, a power supply radius and platform area power supply quantity index data of the platform area, adopting an EM clustering algorithm, setting the clustering number to be 1 to 25, automatically calculating a BIC value under each clustering data by a system, outputting the optimal clustering number to be 8, wherein the clustering result of the platform area is shown in figure 3, and the main characteristics of various platform areas are shown in the following table 2:

table 2 main characteristic table of various areas

Category(s)	Internet power duty cycle	Grid structure	Radius of power supply (Rice)	Table area power supply (kWh)
					1	Small (less than 5%)	Overhead, hybrid line	Middle (400 to 800)	Larger (5000 to 10000)
2	Small (less than 5%)	Overhead, hybrid line	Big (greater than 800)	Middle (2000 to 5000)
					3	Small (less than 5%)	Cable with improved cable characteristics	Small (less than 400)	Larger (5000 to 10000)
4	Small (less than 5%)	Cable with improved cable characteristicsOverhead, hybrid line	Small (less than 400)	Super (greater than 10000)
					5	Small (less than 5%)	Cable, overhead, hybrid line	Middle (400 to 800)	Super (greater than 10000)
6	Big (greater than 5%)	Cable, overhead, hybrid line	Middle (400 to 800)	Small (less than 2000)
					7	Big (greater than 5%)	Hybrid circuit	Small (less than 400)	Small (less than 2000)
8	Small (less than 5%)	Cable, overhead, hybrid line	Big (greater than 800)	Small (less than 2000)

Step 6, screening various training samples of the platform area: and analyzing data samples of various areas, and screening typical samples of various areas for model training.

Extracting data of solar line loss and lunar line loss of platform area and screening platformThe area with stable area solar line loss and moon line loss is a training sample, and the specific screening rules are as follows: 1) Removing areas with the monthly line loss rate exceeding the range of 0% -10%; 2) Removing areas with day loss rate exceeding 0% -10% and day being greater than 15 days; 3) Removing days with the daily line loss exceeding 0% -10%, and taking the mean square error of the daily line loss rate in the rest days as the daily line loss fluctuation rate sigma of the station area; 4) According to the cluster result of the areas, the fluctuation threshold value when the number of the areas in each category is 70% is obtained and is used as the range of the allowable fluctuation of the line loss rate of the areas5) Screening the range of the fluctuation of the solar line loss rate sigma < the allowable fluctuation of the line loss rate of the area>And marking the stable area as a training sample after screening.

Step 7, constructing a reasonable line loss calculation model of the platform region: and for each type of area, respectively constructing an area reasonable line loss calculation model by using an MLP neural network algorithm.

Step 7.1, model selection: because the line loss of the platform area is comprehensively influenced by various line loss factors, the line loss factors are nonlinear, and the MLP neural network has strong self-learning and self-adapting capability, generalization capability and fault tolerance capability, the MLP neural network model is selected when the reasonable line loss calculation model of the platform area is built, and according to experience and multiple experiments, the platform area line loss prediction model is finally built by adopting the MLP network with four layers of multiple inputs and single outputs and 2 hidden layers, as shown in figure 4.

Step 7.2 model training:

based on sample data, firstly carrying out normalization processing, then training model parameters for each type of area through BP algorithm, and in the training process, carrying out parameter optimization based on cross-validation grid search, wherein the parameter settings are shown in the following table 3:

table 3 network parameter search scope

Parameter name	Parameter range
		Number of Layer1 neurons	range(6，100，2)
Number of Layer2 neurons	range(6100，2)
		activation function	[’relu’，’tanh’，’logistic’，’identity’]

By performing grid search on the parameters, calculating possible combinations of all parameters (Layer 1, layer2, activation) to obtain optimal parameters of a reasonable line loss calculation model of each type of area, for example, performing parameter optimization on a certain type of area, and performing an iterative process as follows:

0.000597 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，6)，′solver′：′lbfgs′}

0.000389 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，6)，′solver′：′sgd′}

0.000289 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，6)，′solver′：′adam′}

0.000420 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，8)，′solver′：′lbfgs′}

0.000843 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，8)，′solver′：′sgd′}

0.900445 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，8)，′solver′：′adam′}

0.000580 with：{′activation′：′relu′，′hidden_layer_sizes′：(6，10)，′solver′：′lbfgs′}

the optimal model parameters of the type of platform area are obtained through grid search:

{activation：’relu’，hidden_layer1_sizes’：10，hidden_layer2_sizes’：8}

the optimal MLP four-layer network model after training is obtained, 100 stable sample areas are predicted, and the predicted value and the actual value are shown in figure 5.

Example 2

Based on the same inventive concept as embodiment 1, this embodiment provides a low-voltage station reasonable line loss value measuring and calculating system, which includes the following modules:

the preprocessing module is used for preprocessing the acquired original data;

The operation indexes comprise three-phase unbalance degree and load shape coefficients in the index system constructed by the index system construction module; the three-phase unbalance is calculated as follows:

wherein ε _i For three-phase imbalance degree, I _max Is the maximum value of the three-phase load current; i _av Is the average value of three-phase load current;

let the total surface root mean square current I of the platform area _rms Average value I of three-phase load current _av And taking the equivalent relation k as a load shape factor, the calculation formula of the load shape factor is as follows:

In addition, the area category index comprises a power supply radius, wherein the power supply radius is defined as the space coordinate distance from the farthest meter box to the transformer, and the power supply radius v _d The calculation formula of (2) is as follows:

in the formula, v _h For the maximum circular section radian passing through the center of the earth, v _d The space coordinate distance between the farthest meter box and the transformer; space coordinates of the furthest meter boxIs (lon) ₁ ,lat ₁ ) The spatial coordinates of the transformer are (lon ₂ ,lat ₂ )。

Example 3

Based on the same inventive concept as that of embodiment 1, this embodiment provides a corresponding computer device, including a memory, a processor, and a computer program stored in the memory and capable of running on the processor, where the processor implements the steps of the method for measuring and calculating a reasonable line loss value of a low-voltage area in embodiment 1 when executing the computer program.

The above examples are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above examples, and any other changes, modifications, substitutions, combinations, and simplifications that do not depart from the spirit and principle of the present invention should be made in the equivalent manner, and the embodiments are included in the protection scope of the present invention.

Claims

1. A reasonable line loss value measuring and calculating method for a low-voltage station area is characterized by comprising the following steps:

s2, preprocessing the obtained original data;

s7, respectively constructing a reasonable line loss calculation model of each type of area by using an MLP neural network algorithm;

the operation class index comprises a load shape factor; let the total surface root mean square current I of the platform area _rms Average value I of three-phase load current _av And taking the equivalent relation k as a load shape factor, the calculation formula of the load shape factor is as follows:

wherein T is the number of current measurement points determined by a calculation period, and i is the current sequence number of the measurement points; i _i The current of the ith measuring point;

the region category index comprises a power supply radius, wherein the power supply radius is defined as the space coordinate distance from the farthest meter box to the transformer, and the power supply radius v _d The calculation formula of (2) is as follows:

in the formula, v _h For the maximum circular section radian passing through the center of the earth, v _d The space coordinate distance between the farthest meter box and the transformer; the spatial coordinates of the furthest bin are (lon ₁ ,lat ₁ ) The spatial coordinates of the transformer are (lon ₂ ,lat ₂ )；

In the step S5, an unsupervised EM clustering algorithm is adopted to divide a low-voltage area into a plurality of groups, the area with the network power ratio larger than zero is extracted, the number of clusters is set, the BIC value under each cluster data is calculated, and the optimal number of clusters is output;

step S6, extracting data of solar line loss and lunar line loss of the station area, and screening the station area with the solar line loss and lunar line loss tending to be stable as a training sample; the screening rule for the training samples of various areas is as follows:

1) Removing areas with the monthly line loss rate exceeding the range of 0% -10%;

2) Removing areas with day loss rate exceeding 0% -10% and day being greater than 15 days;

3) Removing days with the daily line loss exceeding 0% -10%, and taking the mean square error of the daily line loss rate in the rest days as the daily line loss fluctuation rate sigma of the station area;

4) According to the area division result, calculating the fluctuation threshold value when the area number in each category is 70%, and taking the fluctuation threshold value as the range of the allowable fluctuation of the area line loss rate

5) Screening station area solar line loss fluctuation rate sigma<Range of allowed fluctuation of line loss rate of such a stationAnd marking the stable area as a training sample after screening.

2. The method for measuring and calculating reasonable line loss values of a low-voltage transformer area according to claim 1, wherein the operation type index comprises three-phase unbalance, and the calculation formula is as follows:

wherein ε _i For three-phase imbalance degree, I _max Is the maximum value of the three-phase load current; i _av Is the average value of the three-phase load current.

3. A low-voltage station reasonable line loss value measuring and calculating system is characterized by comprising:

the preprocessing module is used for preprocessing the acquired original data;

the line loss calculation module is used for respectively constructing a reasonable line loss calculation model of each type of the transformer area by using an MLP neural network algorithm;

The station area classification module adopts an unsupervised EM clustering algorithm to divide a low-voltage station area into a plurality of groups, extracts station areas with the network power occupation ratio larger than zero from the station areas, sets the clustering number, calculates BIC values under each clustering data, and outputs the optimal clustering number;

the sample screening module extracts data of solar line loss and lunar line loss of the platform area, and screens the platform area with the solar line loss and lunar line loss tending to be stable as a training sample; the screening rule for the training samples of various areas is as follows:

4. The low-voltage transformer area reasonable line loss value measuring and calculating system according to claim 3, wherein the operation type indexes comprise three-phase unbalance degree and load shape factor; the three-phase unbalance is calculated as follows:

5. Computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the method for measuring and calculating reasonable line loss values of a low voltage area according to claim 1 or 2 when executing the computer program.