CN115828156A - Power grid metadata monitoring-based electricity stealing and leakage identification method and system - Google Patents
Power grid metadata monitoring-based electricity stealing and leakage identification method and system Download PDFInfo
- Publication number
- CN115828156A CN115828156A CN202211509049.0A CN202211509049A CN115828156A CN 115828156 A CN115828156 A CN 115828156A CN 202211509049 A CN202211509049 A CN 202211509049A CN 115828156 A CN115828156 A CN 115828156A
- Authority
- CN
- China
- Prior art keywords
- layer
- electricity stealing
- power grid
- leaking
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The disclosure belongs to the technical field of power monitoring, and particularly relates to a power stealing and leakage identification method and system based on power grid metadata monitoring, which comprises the steps of obtaining power grid metadata; constructing an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data; and obtaining the user type according to the obtained multidimensional time sequence data and the user classification neural network model, and finishing the recognition of the electricity stealing and leaking. The method has the advantages that the efficiency of recognizing the electricity stealing and leaking behaviors is required to be improved, the positioning method is redundant, powerful data support is provided for power distribution network optimization and rectification of power supply enterprises, and the income increase of economic benefits can be realized.
Description
Technical Field
The disclosure belongs to the technical field of power monitoring, and particularly relates to an electricity stealing and leaking identification method and system based on power grid metadata monitoring.
Background
The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.
The increasing prominence of the problem of electricity stealing and leakage causes power supply enterprises to suffer huge economic loss, easily jeopardizes the safe operation of a power grid and disturbs the normal electricity utilization order of users; the method is imperative for standardizing the management of power utilization users, promoting the benign development of the power industry and identifying the stealing and leaking behavior in time; therefore, power supply enterprises need to perfect the technical means of detecting and positioning the electricity stealing and leaking so as to improve the attack on the electricity stealing and leaking behaviors.
However, the existing electricity stealing and leaking identification technology is complex, the electricity stealing behavior cannot be accurately identified, and the reliability of identification and positioning of the electricity stealing and leaking behavior is low.
Disclosure of Invention
In order to solve the problems, the invention provides an electricity stealing and leaking identification method and system based on power grid metadata monitoring, the identification efficiency of electricity stealing and leaking behaviors needs to be improved, the positioning method is redundant, powerful data support is provided for power distribution network optimization and rectification of power supply enterprises, and income increase of economic benefits can be realized.
According to some embodiments, a first aspect of the present disclosure provides a power stealing and leakage identification method based on power grid metadata monitoring, which adopts the following technical solutions:
a power stealing and leakage identification method based on power grid metadata monitoring comprises the following steps:
acquiring power grid metadata;
constructing an electric network element data analysis model based on the acquired electric network element data to obtain multi-dimensional time sequence data;
and obtaining the user type according to the obtained multidimensional time sequence data and the user classification neural network model, and finishing the recognition of the electricity stealing and leaking.
The obtained power grid metadata are information describing data items related to power transmission quantity and electricity stealing and leaking, simple and efficient management of the power grid information is achieved through power grid metadata definition, and electricity stealing and leaking conditions are effectively detected.
As one or more implementation modes, extracting power utilization acceleration, instantaneous frequency, power spectrum and entropy rate from power grid metadata of normal users and electricity stealing and leaking users to form multidimensional time sequence data; and coding the user types of the normal user and the electricity stealing and leaking user, and obtaining a training sample according to the multi-dimensional time sequence data and the user type codes corresponding to the multi-dimensional time sequence data, namely completing the construction of the electric network element data analysis model.
Further, the specific calculation method of the multidimensional time series data is as follows:
where Ea is the electrical acceleration, ec is the periodic electrical power usage, T is the period length,
representing the variation value of the electrical speed in the period, ps representing the power spectrum, x (t) representing the data signal of the user electrical network element, t being time, g (tau-t) representing the window function of Fourier transform, tau being the step factor, f representing the frequency, f ins Representing instantaneous frequency, P representing entropy rate;
calculating a spectrum entropy signal according to the power spectrum and the entropy rate, wherein the spectrum entropy signal describes the relation between the power spectrum and the entropy rate, is a measure of the distribution of the power spectrum, and can show the uncertainty and the disorder of the user metadata signal through the spectrum entropy signal;
the spectral entropy calculation method comprises the following steps: se = -Sigma Plog 2 P; wherein Se is spectral entropy.
In one or more embodiments, the user classification neural network model includes an input layer, a PCA space layer, a memory layer, a convolutional layer group, a forgetting layer, and an output layer; the input layer transmits the multi-dimensional time sequence data X to the PCA space layer, the PCA space layer reduces the dimension of the multi-dimensional time sequence data, and main characteristic information of original data is reserved; the PCA spatial layer transmits the data after dimensionality reduction to a memory layer, and the memory layer performs mapping after multiplying the hidden state of the previous layer, the input of the current layer and the coefficient matrix and keeps the neuron state of the previous layer; the memory layer transmits the output to the convolution layer group, the convolution layer transmits to the pooling layer after calculation, the pooling layer transmits the calculated data to the next convolution layer by adopting a maximum pooling mode, and iteration is repeated; the last pooling layer transmits the calculated data to the forgetting layer, and the forgetting layer reads, transmits and screens information by using the principle of a forgetting gate to identify the data; the forgetting layer adopts a dual-channel mode.
Further, the update process of the neuron state is as follows: c (C) = (C-1) < '> a (C) <' > d (C); wherein, an indicates a hadamard product; the output of the forgetting layer is as follows:
wherein OP 5 In order to output the forgetting layer,
the coefficient matrix of the hidden state of the previous layer,
is a matrix of the coefficient of the layer, b o A bias for a forgetting layer; the forgetting layer transmits the calculation result to the output layer, and the output layer outputs y; the loss function of the user classification neural network model is as follows: l = -ylogY- (1-) log (1-Y); where L is the loss function, Y is the neural network output, and Y is the desired output.
As one or more implementation modes, inputting multidimensional time series data into a user classification neural network, and outputting a corresponding user type; and if the user type is the electricity stealing and leaking user, positioning the user position according to the data node corresponding to the power grid metadata acquisition source of the current user, and sending an electricity stealing and leaking alarm.
According to some embodiments, a second aspect of the present disclosure provides an electricity stealing and leaking identification system based on power grid metadata monitoring, which adopts the following technical solutions:
a power grid metadata monitoring-based electricity stealing and leakage identification system comprises:
an acquisition module configured to acquire grid metadata;
the modeling module is configured to construct an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data;
and the identification module is configured to obtain a user type according to the obtained multi-dimensional time sequence data and the user classification neural network model, and finish the identification of the electricity stealing and leaking.
According to some embodiments, a third aspect of the present disclosure provides a computer-readable storage medium, which adopts the following technical solutions:
a computer readable storage medium, on which a program is stored, which when executed by a processor, performs the steps in the method for grid metadata monitoring based electrical theft and leakage identification according to the first aspect of the present disclosure.
According to some embodiments, a fourth aspect of the present disclosure provides an electronic device, which adopts the following technical solutions:
an electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the method for identifying electricity stealing and leakage based on power grid metadata monitoring according to the first aspect of the present disclosure.
Compared with the prior art, the beneficial effect of this disclosure is:
according to the method, the power utilization acceleration, the instantaneous frequency, the power spectrum and the entropy rate are extracted from the power grid metadata to form multi-dimensional time sequence data, so that obvious difference information between normal users and electricity stealing and leaking users can be obtained, and the user classification accuracy is improved; coding the user type corresponding to each group of data, constructing a sample set consisting of multi-dimensional time sequence data and the corresponding user type codes, and providing a data basis for the subsequent detection of the electricity stealing and leakage behavior to reach the expected precision;
constructing a user classification neural network, converting multidimensional time sequence data into a matrix X as a training sample to be input into the user classification neural network, taking a user type corresponding to the multidimensional time sequence data as the output of the user classification neural network, training the neural network, identifying the electricity stealing and leaking behaviors through the output user type codes, and positioning the user position, thereby improving the efficiency of identifying the electricity stealing and leaking behaviors;
the method can effectively solve the problems that the prior art lacks a simple, feasible, accurate and reliable electricity stealing behavior identification method, the efficiency of identifying the electricity stealing and leaking behaviors needs to be improved, the positioning method is redundant, and the system or the method is subjected to a series of effect researches and verifications, can finally provide powerful data support for power distribution network optimization and rectification of power supply enterprises, can realize income increase of economic benefits, and has very high practical value.
Drawings
The accompanying drawings, which are included to provide a further understanding of the disclosure, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure and are not to limit the disclosure.
Fig. 1 is a structural diagram of an electricity stealing and leaking identification system based on power grid metadata monitoring in a first embodiment of the present disclosure;
fig. 2 is a flowchart of an electricity stealing and leaking identification method based on power grid metadata monitoring in a second embodiment of the disclosure;
fig. 3 is a block diagram of a structure of an electricity stealing and leaking identification system based on power grid metadata monitoring in an electricity stealing and leaking identification method based on data monitoring of an electrical network element in the third embodiment of the present disclosure.
Detailed Description
The present disclosure is further described with reference to the following drawings and examples.
It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the disclosure. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.
The embodiments and features of the embodiments in the present disclosure may be combined with each other without conflict.
Example one
The embodiment of the disclosure introduces an electricity stealing and leakage identification system based on power grid metadata monitoring.
The system for identifying electricity stealing and leakage based on power grid metadata monitoring as shown in fig. 1 comprises a data acquisition module 10, an online generation module 20, a metadata storage module 30, an analysis module 40 and an alarm module 50;
specifically, the data acquisition module 10 is configured to acquire service data generated by power grid operation in real time, and send the service data to the online generation module 20 in a data transmission manner;
the online generation module 20 is configured to extract power grid metadata and source information describing the service data from each service data, and the online generation module 20 sends the extracted information to the metadata storage module 30 in a data transmission manner;
the metadata storage module 30 is used for storing the electric network element data and the source information, receiving the query request from the analysis module 40, and sending the request data to the analysis module 40 in a data transmission manner;
the analysis module 40 comprises a model construction unit 401, a feature extraction unit 402, a user type coding unit 403 and a user classification unit 404, wherein the model construction unit 401 is used for constructing a power grid metadata analysis model and sending information used by the model to the feature extraction unit 402, the user type coding unit 403 and the user classification unit 404; the feature extraction unit 402 is configured to extract power utilization acceleration, instantaneous frequency, power spectrum and entropy rate from the power grid metadata of the normal user and the electricity stealing and leakage users to form multidimensional time series data, and send the multidimensional time series data to the user classification unit 404; the user type encoding unit 403 is used for encoding user types of normal users and electricity stealing and leaking users and sending the user types to the user classification unit 404; the user classification unit 404 is configured to construct a user classification neural network, transform the multidimensional timing sequence data into a matrix X, input the matrix X as a training sample into the user classification neural network, and train the neural network by using a user type corresponding to the multidimensional timing sequence data as an output of the user classification neural network; the user classification neural network comprises an input layer, a PCA spatial layer, a memory layer, a convolution layer group, a forgetting layer and an output layer; the analysis module 40 sends the calculation result to the alarm module 50 in a data transmission manner;
and the alarm module 50 is used for sending an electricity stealing and leaking alarm to the position of the electricity stealing and leaking node.
Example two
The second embodiment of the disclosure introduces an electricity stealing and leaking identification method based on power grid metadata monitoring, and adopts the electricity stealing and leaking identification system based on power grid metadata monitoring introduced in the first embodiment.
As shown in fig. 2, a power theft and leakage identification method based on power grid metadata monitoring includes:
acquiring power grid metadata;
constructing an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data;
and obtaining the user type according to the obtained multidimensional time sequence data and the user classification neural network model, and finishing the recognition of the electricity stealing and leaking.
The power grid system can generate a large amount of service data in the operation process, information linkage monitoring equipment is installed on the power grid system, and the information linkage monitoring equipment is provided with a monitoring system which monitors and collects the service data in real time aiming at electricity stealing and leakage detection items. The monitoring system comprises a data acquisition module 10, an online generation module 20, a metadata storage module 30 and an analysis module 40, wherein the data acquisition module 10 acquires service data generated by the operation of a power grid in real time, and the online generation module 20 extracts power grid metadata and source information describing the service data from the service data and stores the power grid metadata and the source information into the metadata storage module 30.
Wherein extracting grid metadata from the business data comprises:
and extracting the structural information and the vocabulary information in the service data under the requirement condition according to the electricity stealing and leakage detection requirement.
The structure information in the service data comprises data field attribute labels, data characteristics, instance quantity, meta attribute quantity, missing value quantity and unknown value quantity; the vocabulary information in the business data includes the range of data field attribute values and the data type.
The power grid metadata are information describing data items related to power transmission quantity and the like and electricity stealing and leaking, simple and efficient management of the power grid information is achieved through power grid metadata definition, and electricity stealing and leaking conditions are effectively detected.
For convenience of description and distinction, in the embodiment of the present application, different grid metadata collection sources are used as one data node, and at least one data node is provided in a grid system.
The metadata storage module 30 receives the query request from the analysis module 40, and the analysis module 40 processes and analyzes the metadata to identify the metadata characteristics of each metadata, where the metadata analysis steps are specifically as follows:
the model construction unit 401 constructs an electric network element data analysis model, and collects the electric network element data of normal users and electricity stealing and leaking users;
specifically, the number of samples of the power grid metadata of the normal user is equal to that of the power grid metadata of the electricity stealing and leaking user;
furthermore, the obvious difference information between normal users and electricity stealing and leaking users can be obtained by extracting the electricity acceleration, the instantaneous frequency, the power spectrum and the entropy rate, and the classification accuracy of the users is improved; the specific calculation method of the multidimensional time series data is as follows:
wherein Ea is the electrical acceleration, ec is the periodic electrical consumption, T is the period length,
representing the variation value of the electrical speed in the period, ps representing the power spectrum, x (t) representing the data signal of the user electrical network element, t being time, g (tau-t) representing the window function of Fourier transform, tau being the step factor, f representing the frequency, f ins Representing instantaneous frequency and P representing entropy rate.
And calculating a spectral entropy signal according to the power spectrum and the entropy rate, wherein the spectral entropy signal describes the relation between the power spectrum and the entropy rate, is a measure of the distribution of the power spectrum, and can show the uncertainty and the disorder of the user metadata signal. The spectral entropy calculation method comprises the following steps:
Se＝-∑P log 2 P
wherein Se is spectral entropy.
The power utilization acceleration, instantaneous frequency, power spectrum and entropy rate are extracted from the power grid metadata to form multi-dimensional time sequence data, so that obvious difference information between normal users and electricity stealing and leaking users can be obtained, and the user classification accuracy is improved; and coding the user type corresponding to each group of data, constructing a sample set consisting of the multidimensional time sequence data and the corresponding user type codes, and providing a data basis for the subsequent detection of the electricity stealing and leakage behavior to reach the expected precision.
Constructing a user classification neural network, inputting multidimensional time sequence data serving as a training sample into the user classification neural network, and outputting the user type corresponding to the multidimensional time sequence data serving as the output of the user classification neural network to realize real-time detection and positioning of electricity stealing and leaking; the user classifying unit 404 constructs a user classification neural network, transforms the multidimensional timing sequence data into a matrix X as a training sample and inputs the matrix X into the user classification neural network, and trains the neural network by using the user type corresponding to the multidimensional timing sequence data as the output of the user classification neural network. The user classification neural network comprises an input layer, a PCA spatial layer, a memory layer, a convolution layer group, a forgetting layer and an output layer.
The input layer transmits the multi-dimensional time sequence data X to the PCA space layer, the PCA space layer reduces the dimension of the multi-dimensional time sequence data, main characteristic information of original data is reserved, and the specific calculation is as follows:
IP 2 ＝ω 12 X+b 1
OP 2 ＝ω 2 mf
wherein, IP 2 Representing the input, ω, of a PCA spatial layer 12 Representing the connection weights of the input layer, PCA spatial layer, b 1 To be offset, mf i The data of the main characteristics are represented,
representing the load vector, T is the bias, i ∈ [1, N]N denotes the total number of data, OP 2 Representing the output of the PCA spatial layer, ω 2 Representing the weights of the neurons in the PCA spatial layer.
The PCA space layer transmits the data after dimensionality reduction to a memory layer, the memory layer maps the hidden state of the previous layer after multiplying the input of the current layer by a coefficient matrix, and finally the neuron state of the previous layer is reserved, and the specific calculation process is as follows:
r(c)＝σ[δ 1 z(c-1)+δ 2 OP 2 +b 2 ]
OP 3 ＝r(c)×C(c-1)
where r (c) represents the mapping result, σ represents the activation function, δ 1 A coefficient matrix representing the hidden state of the previous layer, z (c-1) representing the hidden state of the previous layer, δ 2 Representing the principal layer coefficient matrix, b 2 To be biased, OP 3 For the output of the memory layer, C (C-1) is the neuron state of the previous layer, and C is the number of layers.
The memory layer transmits output to a convolution layer group, the convolution layer group consists of a convolution layer and a pooling layer, the number of convolution kernels is Q, the size of the convolution layer is qxq, the convolution layer is transmitted to the pooling layer after calculation, a maximum pooling mode is adopted, then the pooling layer transmits calculated data to the next convolution layer, iteration is repeated, and the specific calculation process is as follows:
OP 4 ＝h(ω 4 OP 3 + 3 )
OP 4 ′ ＝max(OP 4 )
wherein OP 4 For the output of the convolutional layer, OP 4 ′ Is the output of the pooling layer, h is the activation function, ω 4 As weights of convolution kernels, b 3 Is an offset.
And the last pooling layer transmits the calculated data to the forgetting layer, and the forgetting layer reads, transmits and screens information by using the principle of a forgetting gate and identifies the data. The forgetting layer adopts a dual-channel mode, and the specific calculation process is as follows:
wherein a (c) is the first channel output and d (c) is the second channel output,
a coefficient matrix of hidden states of an upper layer in the first channel,
is a matrix of the system coefficients in the first channel, b 41 Is an offset of the first channel and,
a coefficient matrix of hidden states of a previous layer in the second channel, tanh is an activation function,
is a matrix of the system coefficients in the second channel, b 42 Is an offset of the second channel.
The update process of the neuron state is as follows:
C(c)＝(c-1)⊙a(c)⊙d(c)
wherein, l represents a hadamard product. The output of the forgetting layer is:
wherein OP 5 In order to output the forgetting layer,
the coefficient matrix of the hidden state of the previous layer,
is a matrix of the coefficient of the layer, b o Is the bias of the forgetting layer. The forgetting layer transmits the calculation result to the output layer, and the output layer outputs y.
Constructing a loss function:
L＝-ylogY-(1-)log(1-Y)
where L is the loss function, Y is the neural network output, and Y is the desired output. According to the prior art, the parameters of the neural network are corrected until the accuracy of the neural network reaches the expectation.
And inputting the multidimensional time sequence data into a user classification neural network, and outputting a corresponding user type. If the user type is the electricity stealing and leaking user, the position of the user is located according to the data node corresponding to the power grid metadata acquisition source of the current user, and the position of the electricity stealing and leaking node is sent to the alarm module 50 to send out an electricity stealing and leaking alarm.
The method comprises the steps of constructing a user classification neural network, converting multidimensional time sequence data into a matrix X as a training sample to be input into the user classification neural network, taking a user type corresponding to the multidimensional time sequence data as the output of the user classification neural network, training the neural network, identifying the electricity stealing and leaking behaviors through the output user type codes, positioning the position of a user, and improving the efficiency and the accuracy of identifying the electricity stealing and leaking behaviors.
In the embodiment, the power utilization acceleration, the instantaneous frequency, the power spectrum and the entropy rate are extracted from the power grid metadata to form multi-dimensional time sequence data, so that the obvious difference information between normal users and electricity stealing and leaking users can be obtained, and the user classification accuracy is improved; coding the user type corresponding to each group of data, constructing a sample set consisting of multi-dimensional time sequence data and the corresponding user type codes, and providing a data basis for the subsequent detection of the electricity stealing and leakage behavior to achieve the expected precision; according to the invention, the user classification neural network is constructed, multidimensional time sequence data are converted into the matrix X which is used as a training sample and input into the user classification neural network, the user type corresponding to the multidimensional time sequence data is used as the output of the user classification neural network, the neural network is trained, the electricity stealing and leaking behaviors are identified through the output user type codes, the user position is positioned, and the efficiency of identifying the electricity stealing and leaking behaviors is improved.
EXAMPLE III
The third embodiment of the disclosure introduces an electricity stealing and leakage identification system based on power grid metadata monitoring.
Fig. 3 shows a power theft and leakage identification system based on power grid metadata monitoring, which includes:
an acquisition module configured to acquire grid metadata;
the modeling module is configured to construct an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data;
and the identification module is configured to obtain the user type according to the obtained multi-dimensional time sequence data and the user classification neural network model, and finish the identification of the electricity stealing and leaking.
The detailed steps are the same as those of the power grid metadata monitoring-based electricity stealing and leaking identification method provided in the second embodiment, and are not described herein again.
Example four
The fourth embodiment of the disclosure provides a computer-readable storage medium.
A computer-readable storage medium, on which a program is stored, which when executed by a processor implements the steps in the method for identifying electricity stealing and leakage based on grid metadata monitoring according to the second embodiment of the present disclosure.
The detailed steps are the same as those of the power grid metadata monitoring-based electricity stealing and leaking identification method provided in the second embodiment, and are not described herein again.
EXAMPLE five
The fifth embodiment of the disclosure provides electronic equipment.
An electronic device includes a memory, a processor, and a program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the power grid metadata monitoring-based electricity stealing and leakage identification method according to the second embodiment of the present disclosure.
The detailed steps are the same as those of the power grid metadata monitoring-based electricity stealing and leaking identification method provided in the second embodiment, and are not described herein again.
The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure, and various modifications and changes may be made to the present disclosure by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present disclosure should be included in the protection scope of the present disclosure.
Claims (10)
1. A power stealing and leakage identification method based on power grid metadata monitoring is characterized by comprising the following steps:
acquiring power grid metadata;
constructing an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data;
and obtaining the user type according to the obtained multidimensional time sequence data and the user classification neural network model, and finishing the recognition of the electricity stealing and leaking.
2. The power grid metadata monitoring-based electricity stealing and leaking identification method as claimed in claim 1, wherein the obtained power grid metadata are information describing data items related to power transmission amount and electricity stealing and leaking, simple and efficient management of power grid information is realized through power grid metadata definition, and electricity stealing and leaking conditions are effectively detected.
3. The power grid metadata monitoring-based electricity stealing and leaking identification method as claimed in claim 1, wherein electricity utilization acceleration, instantaneous frequency, power spectrum and entropy rate are extracted from the power grid metadata of normal users and electricity stealing and leaking users to form multidimensional time sequence data; and coding the user types of the normal user and the electricity stealing and leaking user, and obtaining a training sample according to the multi-dimensional time sequence data and the user type codes corresponding to the multi-dimensional time sequence data, namely completing the construction of the electric network element data analysis model.
4. The power grid metadata monitoring-based electricity stealing and leakage identification method as claimed in claim 3, wherein the specific calculation method of the multidimensional time series data is as follows:
wherein Ea is the electrical acceleration, ec is the periodic electrical consumption, T is the period length,
representing the variation value of the electrical speed in the period, ps representing the power spectrum, x (t) representing the data signal of the user electrical network element, t being time, g (tau-t) representing the window function of Fourier transform, tau being the step factor, f representing the frequency, f ins Representing instantaneous frequency, P representing entropy rate;
calculating a spectrum entropy signal according to the power spectrum and the entropy rate, wherein the spectrum entropy signal describes the relation between the power spectrum and the entropy rate, is a measure of the distribution of the power spectrum, and can show the uncertainty and the disorder of the user metadata signal through the spectrum entropy signal;
the spectral entropy calculation method comprises the following steps: se = -Sigma Plog 2 P; wherein Se is spectral entropy.
5. The power grid metadata monitoring-based electricity stealing and leaking identification method as claimed in claim 1, wherein the user classification neural network model comprises an input layer, a PCA spatial layer, a memory layer, a convolutional layer group, a forgetting layer and an output layer; the input layer transmits the multi-dimensional time sequence data X to the PCA space layer, the PCA space layer reduces the dimension of the multi-dimensional time sequence data, and main characteristic information of original data is reserved; the PCA space layer transmits the data after dimensionality reduction to a memory layer, and the memory layer maps the hidden state of the previous layer after multiplying the input of the current layer by a coefficient matrix and reserves the neuron state of the previous layer; the memory layer transmits the output to the convolution layer group, the convolution layer transmits to the pooling layer after calculation, the pooling layer transmits the calculated data to the next convolution layer by adopting a maximum pooling mode, and iteration is repeated; the last pooling layer transmits the calculated data to the forgetting layer, and the forgetting layer reads, transmits and screens information by using the principle of a forgetting gate to identify the data; the forgetting layer adopts a dual-channel mode.
6. The power grid metadata monitoring-based electricity stealing and leaking identification method as claimed in claim 5, wherein the update process of the neuron state is as follows: c (C) = C (C-1) </al a (C) </al > d (C); wherein, an indicates a hadamard product; the output of the forgetting layer is as follows:
wherein OP 5 In order to output the forgetting layer,
the coefficient matrix of the hidden state of the previous layer,
is a matrix of the coefficient of the layer, b o A bias for a forgetting layer; the forgetting layer transmits the calculation result to the output layer, and the output layer outputs y; the loss function of the user classification neural network model is as follows: l = -ylogY- (1-) log (1-Y); where L is the loss function, Y is the neural network output, and Y is the desired output.
7. The power grid metadata monitoring-based electricity stealing and leaking identification method as claimed in claim 1, wherein multidimensional time series data are input into a user classification neural network, and corresponding user types are output; and if the user type is an electricity stealing and leaking user, positioning the user position according to the data node corresponding to the power grid metadata acquisition source of the current user, and sending an electricity stealing and leaking alarm.
8. An electricity stealing and leakage identification system based on power grid metadata monitoring is characterized by comprising:
an acquisition module configured to acquire grid metadata;
the modeling module is configured to construct an electric network element data analysis model based on the acquired electric network metadata to obtain multi-dimensional time sequence data;
and the identification module is configured to obtain the user type according to the obtained multi-dimensional time sequence data and the user classification neural network model, and finish the identification of the electricity stealing and leaking.
9. A computer-readable storage medium, on which a program is stored, which program, when being executed by a processor, carries out the steps of the method for grid metadata monitoring based electrical theft and leakage identification according to any one of claims 1 to 7.
10. An electronic device comprising a memory, a processor and a program stored on the memory and executable on the processor, wherein the processor when executing the program implements the steps in the method for power grid metadata monitoring based electric theft and leakage identification according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211509049.0A CN115828156A (en) | 2022-11-29 | 2022-11-29 | Power grid metadata monitoring-based electricity stealing and leakage identification method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211509049.0A CN115828156A (en) | 2022-11-29 | 2022-11-29 | Power grid metadata monitoring-based electricity stealing and leakage identification method and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN115828156A true CN115828156A (en) | 2023-03-21 |
Family
ID=85532458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211509049.0A Pending CN115828156A (en) | 2022-11-29 | 2022-11-29 | Power grid metadata monitoring-based electricity stealing and leakage identification method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115828156A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117194588A (en) * | 2023-11-07 | 2023-12-08 | 江苏龙虎网信息科技股份有限公司 | Business data integrated supervision system and method based on big data |
-
2022
- 2022-11-29 CN CN202211509049.0A patent/CN115828156A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117194588A (en) * | 2023-11-07 | 2023-12-08 | 江苏龙虎网信息科技股份有限公司 | Business data integrated supervision system and method based on big data |
| CN117194588B (en) * | 2023-11-07 | 2024-01-19 | 江苏龙虎网信息科技股份有限公司 | Business data integrated supervision system and method based on big data |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wang et al. | Load profiling and its application to demand response: A review | |
| CN103728551B (en) | A kind of analog-circuit fault diagnosis method based on cascade integrated classifier | |
| Yuan et al. | Learning-based real-time event identification using rich real PMU data | |
| El Kababji et al. | A data-driven approach for generating synthetic load patterns and usage habits | |
| CN113962259A (en) | Multi-mode double-layer fault diagnosis method for fuel cell system | |
| CN111080105A (en) | Transformer area user-to-user relationship identification method and system based on voltage time sequence data | |
| He et al. | Feature extraction of analogue circuit fault signals via cross‐wavelet transform and variational Bayesian matrix factorisation | |
| CN112598303A (en) | Non-invasive load decomposition method based on combination of 1D convolutional neural network and LSTM | |
| Himeur et al. | On the applicability of 2d local binary patterns for identifying electrical appliances in non-intrusive load monitoring | |
| Gu et al. | An improved sensor fault diagnosis scheme based on TA-LSSVM and ECOC-SVM | |
| CN115828156A (en) | Power grid metadata monitoring-based electricity stealing and leakage identification method and system | |
| Zu et al. | A simple gated recurrent network for detection of power quality disturbances | |
| CN115204536A (en) | Building equipment fault prediction method, device, equipment and storage medium | |
| Liao et al. | Recognition of partial discharge patterns | |
| Egri et al. | Cross-correlation based clustering and dimension reduction of multivariate time series | |
| CN117034123B (en) | Fault monitoring system and method for fitness equipment | |
| CN117219124A (en) | Switch cabinet voiceprint fault detection method based on deep neural network | |
| Wan et al. | Software Defect Prediction Using Dictionary Learning. | |
| CN114462707B (en) | Web service multidimensional QoS joint prediction method based on feature depth fusion | |
| Ren et al. | A clustering-based method for health conditions evaluation of aero-engines | |
| CN112463964B (en) | Text classification and model training method, device, equipment and storage medium | |
| CN115598459A (en) | Power failure prediction method for 10kV feeder line fault of power distribution network | |
| CN114397521A (en) | Fault diagnosis method and system for electronic equipment | |
| Livani et al. | A hybrid machine learning method and its application in municipal waste prediction | |
| Karthik et al. | Detection of anomalies in time series data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |