CN113687176B - Deep neural network-based power consumption abnormity detection method and system - Google Patents
Deep neural network-based power consumption abnormity detection method and system Download PDFInfo
- Publication number
- CN113687176B CN113687176B CN202111237405.3A CN202111237405A CN113687176B CN 113687176 B CN113687176 B CN 113687176B CN 202111237405 A CN202111237405 A CN 202111237405A CN 113687176 B CN113687176 B CN 113687176B
- Authority
- CN
- China
- Prior art keywords
- electricity
- power
- data
- user side
- electricity utilization
- 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.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Abstract
The invention discloses a method and a system for detecting abnormal electricity consumption based on a deep neural network, wherein the method considers the influence of air temperature and electricity consumption peak on the electricity consumption power, marking the real-time electricity data acquired by a meter at the user side, constructing an electricity data characteristic matrix, meanwhile, the line loss value of the line on which the user side is located is also calculated, difference calculation is carried out through the sum of the output power of the transformer on which the user side belongs and the electric power consumption of the meter to obtain a power consumption difference value, the number of days that the change rate of the power difference value continuously rises is counted, if the number of days that the change rate of the power difference value continuously rises is larger than a preset second threshold value, and judging that the electricity stealing or electric leakage suspicion exists at the user side, and carrying out abnormity detection on the real-time electricity utilization data of the user side through a pre-trained deep neural network model so as to output a detection result to determine that the user side is electricity stealing or electric leakage. Therefore, the efficiency of detecting the electricity utilization abnormity is improved, and the accuracy of detecting the electricity utilization abnormity is also improved.
Description
Technical Field
The invention relates to the technical field of power data detection, in particular to a method and a system for detecting power utilization abnormity based on a deep neural network.
Background
In the process of power transmission, abnormal behaviors such as power leakage or power stealing and the like generated at a user side often cause serious loss of power supply enterprises. When detecting the behavior of electric leakage or electricity stealing on the user side, the existing electricity utilization abnormality detection method generally needs to utilize a large amount of accumulated data of a power grid to perform a large amount of complex data analysis so as to predict the possible abnormal behaviors of electric leakage or electricity stealing and the like on the user side, but the analysis method has overlarge calculated amount and low prediction precision, thereby reducing the efficiency and the accuracy of electricity utilization abnormality detection.
Disclosure of Invention
The invention provides a deep neural network-based power consumption abnormity detection method and system, which are used for solving the technical problem that the efficiency and the accuracy of power consumption abnormity detection are reduced.
In view of this, the first aspect of the present invention provides a method for detecting abnormal power consumption based on a deep neural network, including the following steps:
the method comprises the steps that real-time electricity utilization data of a user side are obtained based on a meter of the user side, wherein the real-time electricity utilization data comprise electricity utilization power;
marking the real-time electricity utilization data, wherein the marking content comprises the temperature of the day corresponding to the real-time electricity utilization data and whether the current temperature is the peak electricity utilization period;
constructing a power utilization data characteristic matrix based on the real-time power utilization data, wherein each element in the power utilization data characteristic matrix is the real-time power utilization data of the user side and the corresponding labeled content;
normalizing the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
calculating a line loss value of a line where the user side is located, and calculating the sum of the electric power consumed by the meter according to the line loss value and the electric power consumed;
performing difference calculation based on the output power of the transformer belonging to the user side and the sum of the meter power consumption to obtain a power consumption difference;
judging whether the power consumption difference value is larger than a preset first threshold value or not, if so, calculating a power difference value change rate corresponding to the power consumption difference value, and counting the number of days for which the power difference value change rate continuously rises;
judging whether the number of days that the power difference value change rate continuously rises is larger than a preset second threshold, and if the number of days that the power difference value change rate continuously rises is larger than the preset second threshold, judging that electricity stealing or electric leakage suspicion exists on the user side;
and inputting the normalized power consumption data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking historical power consumption data of the user side and corresponding marked contents thereof as a training set based on a deep learning algorithm.
Preferably, the method further comprises:
and preprocessing the real-time power utilization data, wherein the preprocessing mode comprises data cleaning and interpolation missing completion.
Preferably, the step of calculating the line loss value of the line on which the user side is located specifically includes:
calculating the line loss power of the line where the user side is located according to the following formula:
in the formula，
Line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance.
Preferably, the electricity data feature matrix is normalized by a maximum and minimum value method.
Preferably, the step of inputting the normalized electricity data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, where the detection result is an electricity stealing flag or an electricity leakage flag, and the pre-trained deep neural network model is obtained by training, based on a deep learning algorithm, with the historical electricity data of the user side and the corresponding labeled content thereof as a training set, includes:
extracting historical electricity utilization data of a preset time length according to a current electricity utilization time period based on a meter at a user side, wherein the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in a power utilization peak period;
preprocessing the historical electricity utilization data, and marking the preprocessed historical electricity utilization data based on an expert database, wherein the marking content comprises an electricity stealing mark and an electricity leakage mark;
constructing a historical electricity utilization data matrix based on the marked historical electricity utilization data, wherein each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and substituting the test data set into the deep neural network model for testing to obtain a test result, judging whether the test result is in an acceptable range, if not, adjusting network parameters to perform iterative training again, and if so, obtaining a corresponding deep neural network model.
In a second aspect, the present invention further provides a deep neural network-based power consumption anomaly detection system, including:
the real-time electricity consumption data acquisition module is used for acquiring real-time electricity consumption data of a user side based on a meter of the user side, and the real-time electricity consumption data comprises electricity consumption power;
the data marking module is used for marking the real-time electricity utilization data, and the marking content of the data marking module comprises the temperature of the current day corresponding to the real-time electricity utilization data and whether the current day is the peak electricity utilization period;
the electricity utilization matrix construction module is used for constructing an electricity utilization data characteristic matrix based on the real-time electricity utilization data, and each element in the electricity utilization data characteristic matrix is the real-time electricity utilization data of the user side and the corresponding labeled content;
the normalization module is used for performing normalization processing on the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
the line loss calculation module is used for calculating a line loss value of a line where the user side is located, and calculating the sum of the electric power consumed by the meter according to the line loss value and the electric power consumed;
the difference calculation module is used for performing difference calculation based on the output power of the transformer to which the user side belongs and the sum of the meter power consumption to obtain a power consumption difference;
the difference value change judging module is used for judging whether the power consumption difference value is larger than a preset first threshold value or not, if the power consumption difference value is larger than the preset first threshold value, calculating a power difference value change rate corresponding to the power consumption difference value, and counting the number of days for which the power difference value change rate continuously rises;
the difference change day judging module is used for judging whether the number of days that the power difference change rate continuously rises is larger than a preset second threshold or not, and judging that electricity stealing or electricity leakage suspicion exists at the user side if the number of days that the power difference change rate continuously rises is larger than the preset second threshold;
and the anomaly detection module is used for inputting the normalized electricity consumption data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking the historical electricity consumption data of the user side and the corresponding marked content thereof as a training set on the basis of a deep learning algorithm.
Preferably, the system further comprises a preprocessing module, wherein the preprocessing module is used for preprocessing the real-time electricity utilization data, and the preprocessing mode comprises data cleaning and interpolation missing completion.
Preferably, the line loss calculating module is specifically configured to calculate the line loss power of the line where the user side is located according to the following formula:
in the formula (I), the compound is shown in the specification,
line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance.
Preferably, the normalization module is specifically configured to perform normalization processing on the electricity consumption data feature matrix by using a maximum and minimum value method.
Preferably, the system further comprises:
the historical data acquisition module is used for extracting historical electricity utilization data of a preset time length according to the current electricity utilization time period based on a meter at the user side, and the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in an electricity utilization peak period;
the marking module is used for preprocessing the historical electricity utilization data, marking the preprocessed historical electricity utilization data based on an expert database, wherein the marking content comprises an electricity stealing mark and an electricity leakage mark;
the historical data matrix building module is used for building a historical electricity utilization data matrix based on the marked historical electricity utilization data, and each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
the dividing module is used for dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
the training module is used for substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and the test module is used for substituting the test data set into the deep neural network model to test to obtain a test result, judging whether the test result is in an acceptable range, adjusting network parameters to carry out iterative training again if the test result is not in the acceptable range, and obtaining a corresponding deep neural network model if the test result is in the acceptable range.
According to the technical scheme, the invention has the following advantages:
the invention obtains real-time electricity data of a user side by a meter at the user side by considering the influence of air temperature and electricity consumption peak on electricity consumption power, marks the real-time electricity data of the user side, constructs an electricity consumption data characteristic matrix, normalizes the electricity consumption data characteristic matrix, calculates the line loss value of a line where the user side is positioned, calculates the sum of electricity consumption power of the meter through the line loss value and the electricity consumption power, calculates the difference value through the output power of a transformer which the user side belongs to and the sum of the electricity consumption power of the meter to obtain an electricity consumption power difference value, counts the number of days when the electricity consumption power difference value is larger than a preset first threshold value, determines that electricity stealing or electricity leakage is suspected at the user side if the number of days when the electricity consumption power difference value continuously rises is larger than a preset second threshold value, and performs abnormity detection on the real-time electricity data of the user side through a pre-trained deep neural network model, thereby outputting the detection result to determine that the user side is electricity stealing or electricity leakage. Therefore, the efficiency of detecting the electricity utilization abnormity is improved, and the accuracy of detecting the electricity utilization abnormity is also improved.
Drawings
Fig. 1 is a flowchart of a power consumption abnormality detection method based on a deep neural network according to an embodiment of the present invention;
fig. 2 is a block diagram of a power consumption abnormality detection system based on a deep neural network according to an embodiment of the present invention.
Detailed Description
In the process of power transmission, abnormal behaviors such as power leakage or power stealing and the like generated at a user side often cause serious loss of power supply enterprises. The existing method for predicting the possible actions of electricity stealing due to electric leakage and fire at the user side generally needs to utilize a large amount of accumulated data of a power grid to perform a large amount of complex data analysis so as to predict the possible abnormal actions of electric leakage or electricity stealing and the like at the user side, but the analysis method has overlarge calculated amount and low prediction precision, so that the efficiency and the accuracy of electricity abnormal detection are reduced.
In order to make the technical solutions of the present invention better understood, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
For convenience of understanding, please refer to fig. 1, the method for detecting power consumption abnormality based on a deep neural network provided by the present invention includes the following steps:
s1, acquiring real-time electricity utilization data of the user side based on the meter of the user side, wherein the real-time electricity utilization data comprise electricity utilization power;
it is understood that the user side meter is the electricity data displayed by the electric energy meter.
S2, marking the real-time electricity consumption data, wherein the marking content comprises the temperature of the current day corresponding to the real-time electricity consumption data and whether the current day is the electricity consumption peak period;
it should be noted that, the temperature may affect the resistance of the line, which may further cause the power consumption to change, and meanwhile, the power consumption peak time of the user side may also affect the power consumption, and the power consumption peak time of different types of user sides may also be different, for example, an individual user may consume more power in holidays, for example, an enterprise user may consume less power in holidays.
S3, constructing a power utilization data characteristic matrix based on the real-time power utilization data, wherein each element in the power utilization data characteristic matrix is the real-time power utilization data of the user side and the corresponding labeled content;
in the present embodiment, the electricity consumption data feature matrix is a one-row and multi-column matrix.
S4, carrying out normalization processing on the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
s5, calculating a line loss value of a line where the user side is located, and calculating the sum of the electricity consumption power of the meter through the line loss value and the electricity consumption power;
it should be noted that the line loss value of the line on which the user side is located is fixed and unchanged for a certain time, and therefore, the line loss value of the line can be calculated according to a certain period.
S6, calculating a difference value based on the output power of the transformer to which the user side belongs and the sum of the electric power used by the meter to obtain a difference value of the used electric power;
it should be noted that the output power of the transformer belonging to the user side is set by the station area, therefore, the output power of the transformer is generally fixed, the line loss value of the line where the user side is located is fixed within a certain time, the sum of the power consumption for meter measurement changes with the power consumption, when the power consumption increases, the sum of the power consumption for meter measurement increases, and the difference of the power consumption decreases with the increase of the power consumption, wherein the difference of the power consumption may be the power consumption which is not counted by the meter at the user side.
S7, judging whether the power consumption difference is larger than a preset first threshold, if so, calculating the power difference change rate corresponding to the power consumption difference, and counting the number of days for which the power difference change rate continuously rises;
it can be understood that, if the power consumption difference is greater than the preset first threshold, it indicates that the numerical error counted by the meter at the user side is large, and in order to improve the accuracy of the subsequent suspected judgment of electricity stealing or electricity leakage, it counts the number of days that the power difference change rate continuously rises, where the power difference change rate is calculated as the power difference per day, and the number of days that the power difference change rate continuously rises can indicate that the power difference gradually increases.
S8, judging whether the number of days of continuous rising of the power difference value change rate is larger than a preset second threshold, and if the number of days of continuous rising of the power difference value change rate is larger than the preset second threshold, judging that electricity stealing or electric leakage suspicion exists on the user side;
the number of days that the power difference change rate continuously rises may indicate that the power difference gradually increases, and if the number of days that the power difference change rate continuously rises is greater than a preset second threshold, it indicates that the electricity consumption power displayed by the meter is gradually decreased, and the electricity consumption power not displayed by the meter is gradually increased, which indicates that there is a suspicion of electricity stealing or electricity leakage at the user side. If the number of days in which the power difference change rate continuously rises is not greater than the preset second threshold, the flow returns to step S7 to repeat the monitoring and count the number of days in which the power difference change rate continuously rises.
And S9, inputting the normalized electricity utilization data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking the historical electricity utilization data of the user side and the corresponding marked content thereof as a training set based on a deep learning algorithm.
It should be noted that, in the method for detecting power consumption abnormality based on the deep neural network provided in this embodiment, by considering the influence of the air temperature and the power consumption peak on the power consumption, the meter at the user side is labeled with the real-time power consumption data obtained by the meter at the user side, a power consumption data feature matrix is constructed, the power consumption data feature matrix is normalized, and meanwhile, the line loss value of the line where the user side is located is also calculated, the power consumption power sum of the meter is calculated through the line loss value and the power consumption power, a difference value calculation is performed through the output power of the transformer belonging to the user side and the power consumption power sum of the meter, so as to obtain the power consumption difference value, when the power consumption difference value is greater than a preset first threshold, the number of days when the power difference value continuously increases is counted, it is determined that there is electricity stealing or suspected electricity leakage at the user side if the number of days when the power difference value continuously increases is greater than a preset second threshold, and carrying out anomaly detection on the real-time power consumption data of the user side through a pre-trained deep neural network model, thereby outputting a detection result to determine whether the user side is power stealing or power leakage. Therefore, the efficiency of detecting the electricity utilization abnormity is improved, and the accuracy of detecting the electricity utilization abnormity is also improved.
The following is a detailed description of an embodiment of the power consumption abnormality detection method based on the deep neural network provided by the invention.
The invention provides a deep neural network-based power utilization abnormity detection method, which comprises the following steps:
s100, acquiring real-time electricity utilization data of a user side based on a meter of the user side, wherein the real-time electricity utilization data comprise electricity utilization power;
it is understood that the user side meter is the electricity data displayed by the electric energy meter.
And S200, preprocessing the real-time power utilization data, wherein the preprocessing mode comprises data cleaning and interpolation missing completion.
In the interpolation missing completion processing process, an interpolation algorithm is adopted to complete missing data, and the interpolation can be specifically realized by Lagrange interpolation, Newton interpolation, Hermite interpolation, segmented interpolation, spline interpolation and the like.
S300, marking the real-time electricity utilization data, wherein the marking content comprises the temperature of the day corresponding to the real-time electricity utilization data and whether the current temperature is in the peak electricity utilization period;
it should be noted that, the temperature may affect the resistance of the line, which may further cause the power consumption to change, and meanwhile, the power consumption peak time of the user side may also affect the power consumption, and the power consumption peak time of different types of user sides may also be different, for example, an individual user may consume more power in holidays, for example, an enterprise user may consume less power in holidays.
S400, constructing a power utilization data characteristic matrix based on the real-time power utilization data, wherein each element in the power utilization data characteristic matrix is the real-time power utilization data of the user side and the corresponding labeled content of the real-time power utilization data;
in the present embodiment, the electricity consumption data feature matrix is a one-row and multi-column matrix.
S500, normalizing the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
in this embodiment, the electricity consumption data feature matrix is normalized by the maximum and minimum value method.
S600, calculating a line loss value of a line where a user side is located, and calculating the sum of electric power consumption of a meter through the line loss value and the electric power consumption;
it should be noted that the line loss value of the line on which the user side is located is fixed and unchanged for a certain time, and therefore, the line loss value of the line can be calculated according to a certain period.
In this embodiment, the line loss power of the line on which the user side is located is calculated by the following formula:
in the formula (I), the compound is shown in the specification,
line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance.
Wherein, the material of the circuit is generally copper or aluminum, and a takes the value of 0.004.
S700, calculating a difference value based on the output power of the transformer belonging to the user side and the sum of the electric power used by the meter to obtain a difference value of the used electric power;
it should be noted that the output power of the transformer belonging to the user side is set by the station area, therefore, the output power of the transformer is generally fixed, the line loss value of the line where the user side is located is fixed within a certain time, the sum of the power consumption for meter measurement changes with the power consumption, when the power consumption increases, the sum of the power consumption for meter measurement increases, and the difference of the power consumption decreases with the increase of the power consumption, wherein the difference of the power consumption may be the power consumption which is not counted by the meter at the user side.
S800, judging whether the power consumption difference is larger than a preset first threshold, if so, calculating the power difference change rate corresponding to the power consumption difference, and counting the number of days for which the power difference change rate continuously rises;
it can be understood that, if the power consumption difference is greater than the preset first threshold, it indicates that the numerical error counted by the meter at the user side is large, and in order to improve the accuracy of the subsequent suspected judgment of electricity stealing or electricity leakage, it counts the number of days that the power difference change rate continuously rises, where the power difference change rate is calculated as the power difference per day, and the number of days that the power difference change rate continuously rises can indicate that the power difference gradually increases.
S900, judging whether the number of days of continuous rising of the power difference change rate is larger than a preset second threshold, and if the number of days of continuous rising of the power difference change rate is larger than the preset second threshold, judging that electricity stealing or electric leakage suspicion exists on a user side;
the number of days that the power difference change rate continuously rises may indicate that the power difference gradually increases, and if the number of days that the power difference change rate continuously rises is greater than a preset second threshold, it indicates that the electricity consumption power displayed by the meter is gradually decreased, and the electricity consumption power not displayed by the meter is gradually increased, which indicates that there is a suspicion of electricity stealing or electricity leakage at the user side. If the number of days in which the power difference value change rate continuously rises is not greater than the preset second threshold, the process returns to step S800 to perform repeated monitoring and count the number of days in which the power difference value change rate continuously rises.
S1000, inputting the normalized electricity utilization data matrix corresponding to the user side with the suspicion of electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking historical electricity utilization data of the user side and corresponding marked contents thereof as a training set based on a deep learning algorithm.
In this embodiment, step S1000 includes a process of training the deep neural network model, which specifically includes:
s1001, extracting historical electricity utilization data of a preset time length according to a current electricity utilization time period based on a meter at a user side, wherein the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in a peak electricity utilization period;
wherein the time length is generally one year, half year or one month.
S1002, preprocessing historical electricity utilization data, and marking the preprocessed historical electricity utilization data based on an expert database, wherein the marking contents comprise an electricity stealing mark and an electricity leakage mark;
the preprocessing mode is consistent with the preprocessing mode of the real-time electricity utilization data. Meanwhile, the expert database comprises the standard for judging whether electricity stealing or electricity leakage is carried out on the historical electricity utilization data, so that the historical electricity utilization data can be marked through the expert database.
S1003, constructing a historical electricity utilization data matrix based on the marked historical electricity utilization data, wherein each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
s1004, dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
in a general example, the training data set and the test data set are divided in a ratio of 9: 1.
S1005, substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and S1006, substituting the test data set into the deep neural network model for testing to obtain a test result, judging whether the test result is in an acceptable range, if not, adjusting network parameters to perform iterative training again, and if so, obtaining a corresponding deep neural network model.
In this embodiment, the network parameters are adjusted using a back-propagation algorithm, and then retraining is performed until convergence or the number of training times is reached.
Meanwhile, the deep neural network model is a converged deep neural network model; in the training process, firstly, network node compression processing needs to be performed on the deep neural network model before training, in this embodiment, the network nodes of all layers are compressed in a manner of dividing 2, node compression can achieve model training convergence by using less sample data, and can reduce the probability of overfitting in the training process; after the node compression, selecting a relevant loss function, in this embodiment, adopting a deep neural network classical loss function; in this embodiment, during updating, regularization rules are used to regularize network nodes of all layers of the neural network model before training to form regularization terms, and the regularization terms are added to the original loss function to form an updated loss function.
The above is a detailed description of an embodiment of the power consumption abnormality detection method based on the deep neural network provided by the invention, and the following is a detailed description of an embodiment of the power consumption abnormality detection system based on the deep neural network provided by the invention.
For convenience of understanding, referring to fig. 2, the present invention provides a deep neural network-based power consumption anomaly detection system, including:
the real-time electricity consumption data acquisition module 100 is configured to acquire real-time electricity consumption data of a user side based on a meter of the user side, where the real-time electricity consumption data includes electricity consumption power;
the data marking module 200 is configured to mark the real-time electricity consumption data, where the marking content includes a current temperature corresponding to the real-time electricity consumption data and whether the current temperature is a peak electricity consumption time;
the electricity utilization matrix building module 300 is configured to build an electricity utilization data feature matrix based on the real-time electricity utilization data, and each element in the electricity utilization data feature matrix is the real-time electricity utilization data of the user side and the corresponding labeled content thereof;
the normalization module 400 is used for performing normalization processing on the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
the line loss calculation module 500 is configured to calculate a line loss value of a line where the user side is located, and calculate a power consumption sum by using the line loss value and the power consumption calculation table;
a difference calculation module 600, configured to perform difference calculation based on the sum of the output power of the transformer to which the user side belongs and the meter power consumption to obtain a power consumption difference;
the difference change judging module 700 is configured to judge whether the power consumption difference is greater than a preset first threshold, and if the power consumption difference is greater than the preset first threshold, calculate a power difference change rate corresponding to the power consumption difference, and count the number of days in which the power difference change rate continuously rises;
the difference change day number judging module 800 is configured to judge whether the number of days that the power difference change rate continuously rises is greater than a preset second threshold, and if the number of days that the power difference change rate continuously rises is greater than the preset second threshold, determine that there is a suspicion of electricity stealing or electricity leakage at the user side;
and the anomaly detection module 900 is configured to input the normalized electricity consumption data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and output a detection result, where the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training based on a deep learning algorithm and using the historical electricity consumption data of the user side and the corresponding labeled content thereof as a training set.
Furthermore, the system also comprises a preprocessing module, wherein the preprocessing module is used for preprocessing the real-time power utilization data, and the preprocessing mode comprises data cleaning and interpolation missing completion.
Further, the line loss calculating module is specifically configured to calculate the line loss power of the line where the user side is located according to the following formula:
in the formula (I), the compound is shown in the specification,
line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance.
And further, the normalization module is specifically used for performing normalization processing on the electricity utilization data feature matrix by adopting a maximum and minimum value method.
Further, the system also includes:
the historical data acquisition module is used for extracting historical electricity utilization data of a preset time length according to the current electricity utilization time period based on a meter at the user side, and the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in a power utilization peak period;
the marking module is used for preprocessing the historical electricity utilization data, marking the preprocessed historical electricity utilization data based on the expert database, wherein the marking content comprises an electricity stealing mark and an electricity leakage mark;
the historical data matrix construction module is used for constructing a historical electricity utilization data matrix based on the marked historical electricity utilization data, and each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
the dividing module is used for dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
the training module is used for substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and the test module is used for substituting the test data set into the deep neural network model to test to obtain a test result, judging whether the test result is in an acceptable range, adjusting network parameters to carry out iterative training again if the test result is not in the acceptable range, and obtaining a corresponding deep neural network model if the test result is in the acceptable range.
It should be noted that the working process of the power consumption abnormality detection system based on the deep neural network provided in this embodiment is consistent with the flow of the power consumption abnormality detection method based on the deep neural network provided above, and details are not repeated here.
In the power consumption anomaly detection system based on the deep neural network, by considering the influence of the air temperature and the power consumption peak on the power consumption, the meter at the user side is marked with the real-time power consumption data at the user side, a power consumption data feature matrix is constructed, the power consumption data feature matrix is normalized, meanwhile, the line loss value of the line where the user side is located is calculated, the power consumption sum of the meter is calculated through the line loss value and the power consumption power, the difference value calculation is performed through the output power of the transformer belonging to the user side and the power consumption sum of the meter to obtain the power consumption difference value, when the power consumption difference value is larger than a preset first threshold value, the number of days that the power difference value change rate continuously rises is counted, when the number of days that the power difference value change rate continuously rises is larger than a preset second threshold value, it is determined that electricity stealing or electricity leakage suspicion exists at the user side, and the real-time power consumption data at the user side is subjected to anomaly detection through a pre-trained deep neural network model And detecting, thereby outputting a detection result to determine that the user side is electricity stealing or electricity leakage. Therefore, the efficiency of detecting the electricity utilization abnormity is improved, and the accuracy of detecting the electricity utilization abnormity is also improved.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative, and for example, the division of the units is only one logical division, and other divisions may be realized in practice, for example, a plurality of units or components may be combined or integrated into another system, or some features may be omitted, or not executed. In addition, the shown or discussed mutual coupling or direct coupling or communication connection may be an indirect coupling or communication connection through some interfaces, devices or units, and may be in an electrical, mechanical or other form.
The above examples are only intended to illustrate the technical solution of the present invention, but not to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (6)
1. A power utilization abnormity detection method based on a deep neural network is characterized by comprising the following steps:
the method comprises the steps that real-time electricity utilization data of a user side are obtained based on a meter of the user side, wherein the real-time electricity utilization data comprise electricity utilization power;
preprocessing the real-time power utilization data, wherein the preprocessing mode comprises data cleaning and interpolation missing completion;
marking the real-time electricity utilization data, wherein the marking content comprises the temperature of the day corresponding to the real-time electricity utilization data and whether the current temperature is the peak electricity utilization period;
constructing a power utilization data characteristic matrix based on the real-time power utilization data, wherein each element in the power utilization data characteristic matrix is the real-time power utilization data of the user side and the corresponding labeled content;
normalizing the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
calculating a line loss value of a line where the user side is located, and calculating the sum of the electric power consumed by the meter according to the line loss value and the electric power consumed;
the step of calculating the line loss value of the line on which the user side is located specifically includes:
calculating the line loss power of the line where the user side is located according to the following formula:
in the formula (I), the compound is shown in the specification,
line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance;
performing difference calculation based on the output power of the transformer belonging to the user side and the sum of the meter power consumption to obtain a power consumption difference;
judging whether the power consumption difference value is larger than a preset first threshold value or not, if so, calculating a power difference value change rate corresponding to the power consumption difference value, and counting the number of days for which the power difference value change rate continuously rises;
judging whether the number of days that the power difference value change rate continuously rises is larger than a preset second threshold, and if the number of days that the power difference value change rate continuously rises is larger than the preset second threshold, judging that electricity stealing or electric leakage suspicion exists on the user side;
and inputting the normalized power consumption data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking historical power consumption data of the user side and corresponding marked contents thereof as a training set based on a deep learning algorithm.
2. The deep neural network-based power consumption anomaly detection method according to claim 1, wherein the power consumption data feature matrix is subjected to normalization processing by adopting a maximum and minimum value method.
3. The method for detecting the abnormal electricity consumption based on the deep neural network as claimed in claim 1, wherein the step of inputting the normalized electricity consumption data matrix corresponding to the user side with the suspected electricity stealing or electricity leakage into a pre-trained deep neural network model, and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, the pre-trained deep neural network model is based on a deep learning algorithm, and the step of training the historical electricity consumption data and the corresponding marked content of the historical electricity consumption data on the user side as a training set comprises the steps of:
extracting historical electricity utilization data of a preset time length according to a current electricity utilization time period based on a meter at a user side, wherein the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in a power utilization peak period;
preprocessing the historical electricity utilization data, and marking the preprocessed historical electricity utilization data based on an expert database, wherein the marking content comprises an electricity stealing mark and an electricity leakage mark;
constructing a historical electricity utilization data matrix based on the marked historical electricity utilization data, wherein each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and substituting the test data set into the deep neural network model for testing to obtain a test result, judging whether the test result is in an acceptable range, if not, adjusting network parameters to perform iterative training again, and if so, obtaining a corresponding deep neural network model.
4. An electricity consumption abnormality detection system based on a deep neural network, characterized by comprising:
the real-time electricity consumption data acquisition module is used for acquiring real-time electricity consumption data of a user side based on a meter of the user side, and the real-time electricity consumption data comprises electricity consumption power;
the data marking module is used for marking the real-time electricity utilization data, and the marking content of the data marking module comprises the temperature of the current day corresponding to the real-time electricity utilization data and whether the current day is the peak electricity utilization period;
the electricity utilization matrix construction module is used for constructing an electricity utilization data characteristic matrix based on the real-time electricity utilization data, and each element in the electricity utilization data characteristic matrix is the real-time electricity utilization data of the user side and the corresponding labeled content;
the normalization module is used for performing normalization processing on the electricity utilization data characteristic matrix to obtain a normalized electricity utilization data matrix;
the line loss calculation module is used for calculating a line loss value of a line where the user side is located, and calculating the sum of the electric power consumed by the meter according to the line loss value and the electric power consumed;
the line loss calculation module is specifically configured to calculate the line loss power of the line where the user side is located according to the following formula:
in the formula (I), the compound is shown in the specification,
line loss power, I load current,
the resistance value of the wiring is represented, wherein,
wherein R represents a line resistance value at a reference temperature, the reference temperature is 20 degrees, R1 represents a temperature added resistance value,
a represents a lead temperature coefficient, T is a reference temperature, and T1 is the current environment temperature; r2 represents a load current additional resistance;
the difference calculation module is used for performing difference calculation based on the output power of the transformer to which the user side belongs and the sum of the meter power consumption to obtain a power consumption difference;
the difference value change judging module is used for judging whether the power consumption difference value is larger than a preset first threshold value or not, if the power consumption difference value is larger than the preset first threshold value, calculating a power difference value change rate corresponding to the power consumption difference value, and counting the number of days for which the power difference value change rate continuously rises;
the difference change day judging module is used for judging whether the number of days that the power difference change rate continuously rises is larger than a preset second threshold or not, and judging that electricity stealing or electricity leakage suspicion exists at the user side if the number of days that the power difference change rate continuously rises is larger than the preset second threshold;
the abnormal detection module is used for inputting the normalized electricity data matrix corresponding to the user side with the suspicion of electricity stealing or electricity leakage into a pre-trained deep neural network model and outputting a detection result, wherein the detection result is an electricity stealing mark or an electricity leakage mark, and the pre-trained deep neural network model is obtained by training by taking historical electricity data of the user side and corresponding marked contents thereof as a training set on the basis of a deep learning algorithm;
the preprocessing module is used for preprocessing the real-time power utilization data, and the preprocessing mode comprises data cleaning and interpolation missing completion.
5. The deep neural network-based power consumption anomaly detection system according to claim 4, wherein the normalization module is specifically configured to perform normalization processing on the power consumption data feature matrix by using a maximum and minimum value method.
6. The deep neural network-based power consumption abnormality detection system according to claim 4, further comprising:
the historical data acquisition module is used for extracting historical electricity utilization data of a preset time length according to the current electricity utilization time period based on a meter at the user side, and the historical electricity utilization data comprise historical electricity utilization power, historical temperature corresponding to the historical electricity utilization power and whether the historical electricity utilization power is in an electricity utilization peak period;
the marking module is used for preprocessing the historical electricity utilization data, marking the preprocessed historical electricity utilization data based on an expert database, wherein the marking content comprises an electricity stealing mark and an electricity leakage mark;
the historical data matrix building module is used for building a historical electricity utilization data matrix based on the marked historical electricity utilization data, and each element in the historical electricity utilization data matrix is historical electricity utilization power and corresponding mark content;
the dividing module is used for dividing all elements in the historical electricity utilization data matrix into a training data set and a testing data set;
the training module is used for substituting the training data set into a deep learning algorithm for training, and taking the marked content as output to obtain a deep neural network model;
and the test module is used for substituting the test data set into the deep neural network model to test to obtain a test result, judging whether the test result is in an acceptable range, adjusting network parameters to carry out iterative training again if the test result is not in the acceptable range, and obtaining a corresponding deep neural network model if the test result is in the acceptable range.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111237405.3A CN113687176B (en) | 2021-10-25 | 2021-10-25 | Deep neural network-based power consumption abnormity detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111237405.3A CN113687176B (en) | 2021-10-25 | 2021-10-25 | Deep neural network-based power consumption abnormity detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113687176A CN113687176A (en) | 2021-11-23 |
| CN113687176B true CN113687176B (en) | 2022-02-15 |
Family
ID=78587847
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111237405.3A Active CN113687176B (en) | 2021-10-25 | 2021-10-25 | Deep neural network-based power consumption abnormity detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113687176B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11899516B1 (en) | 2023-07-13 | 2024-02-13 | T-Mobile Usa, Inc. | Creation of a digital twin for auto-discovery of hierarchy in power monitoring |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115018403B (en) * | 2022-08-10 | 2023-04-07 | 国网天津市电力公司营销服务中心 | Checking method for illegal electricity utilization, storage medium and electronic equipment |
| CN115575754B (en) * | 2022-11-21 | 2023-05-02 | 浙江万胜智能科技股份有限公司 | Intelligent industrial park electricity consumption anomaly identification method and system |
| CN117491792B (en) * | 2023-12-27 | 2024-03-22 | 四川中威能电力科技有限公司 | Power consumption abnormality detection method, system and storage medium based on intelligent ammeter |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104345192A (en) * | 2014-11-25 | 2015-02-11 | 国家电网公司 | Photovoltaic power generation subsidy cheating prevention device based on neural network algorithm |
| CN105205531A (en) * | 2014-06-30 | 2015-12-30 | 国家电网公司 | Anti-electric-larceny prediction method based on machine learning and apparatus thereof |
| CN105573997A (en) * | 2014-10-09 | 2016-05-11 | 普华讯光（北京）科技有限公司 | Method and device for determining electric larceny suspect user |
| CN106373025A (en) * | 2016-08-22 | 2017-02-01 | 重庆邮电大学 | Outlier detection-based real-time anti-power-theft monitoring method for power utilization information acquisition system |
| CN107024641A (en) * | 2017-03-29 | 2017-08-08 | 国网上海市电力公司 | A kind of platform area line loss per unit exception monitoring system |
| CN107966600A (en) * | 2017-09-30 | 2018-04-27 | 国家电网公司 | A kind of electricity anti-theft system and its electricity anti-theft method based on deep learning algorithm |
| CN108416693A (en) * | 2018-02-01 | 2018-08-17 | 东华大学 | A kind of multiplexing electric abnormality Activity recognition method based on fuzzy neural network |
| CN109614997A (en) * | 2018-11-29 | 2019-04-12 | 武汉大学 | A kind of stealing Risk Forecast Method and device based on deep learning |
| CN110824270A (en) * | 2019-10-09 | 2020-02-21 | 中国电力科学研究院有限公司 | Electricity stealing user identification method and device combining transformer area line loss and abnormal events |
| CN113011997A (en) * | 2021-02-20 | 2021-06-22 | 上海电机学院 | Power grid user electricity utilization abnormal behavior detection method |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101263467B1 (en) * | 2012-11-15 | 2013-05-10 | 엔텍월드(주) | The noise monitoring device for power plant control system |
| CN106680557A (en) * | 2017-01-05 | 2017-05-17 | 云南电网有限责任公司电力科学研究院 | Electricity theft prevention method based on user behavior analysis |
| US10530151B2 (en) * | 2018-01-09 | 2020-01-07 | Timothy A Carty | System and method for suppressing electromagnetic pulse-induced electrical system surges |
| CN109507502A (en) * | 2018-11-13 | 2019-03-22 | 国家电网有限公司 | A kind of system of quick check low-voltage customer multiplexing electric abnormality situation |
| CN110108914B (en) * | 2019-05-21 | 2021-06-25 | 国网湖南省电力有限公司 | Intelligent decision-making method, system, equipment and medium for preventing electricity stealing |
-
2021
- 2021-10-25 CN CN202111237405.3A patent/CN113687176B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105205531A (en) * | 2014-06-30 | 2015-12-30 | 国家电网公司 | Anti-electric-larceny prediction method based on machine learning and apparatus thereof |
| CN105573997A (en) * | 2014-10-09 | 2016-05-11 | 普华讯光（北京）科技有限公司 | Method and device for determining electric larceny suspect user |
| CN104345192A (en) * | 2014-11-25 | 2015-02-11 | 国家电网公司 | Photovoltaic power generation subsidy cheating prevention device based on neural network algorithm |
| CN106373025A (en) * | 2016-08-22 | 2017-02-01 | 重庆邮电大学 | Outlier detection-based real-time anti-power-theft monitoring method for power utilization information acquisition system |
| CN107024641A (en) * | 2017-03-29 | 2017-08-08 | 国网上海市电力公司 | A kind of platform area line loss per unit exception monitoring system |
| CN107966600A (en) * | 2017-09-30 | 2018-04-27 | 国家电网公司 | A kind of electricity anti-theft system and its electricity anti-theft method based on deep learning algorithm |
| CN108416693A (en) * | 2018-02-01 | 2018-08-17 | 东华大学 | A kind of multiplexing electric abnormality Activity recognition method based on fuzzy neural network |
| CN109614997A (en) * | 2018-11-29 | 2019-04-12 | 武汉大学 | A kind of stealing Risk Forecast Method and device based on deep learning |
| CN110824270A (en) * | 2019-10-09 | 2020-02-21 | 中国电力科学研究院有限公司 | Electricity stealing user identification method and device combining transformer area line loss and abnormal events |
| CN113011997A (en) * | 2021-02-20 | 2021-06-22 | 上海电机学院 | Power grid user electricity utilization abnormal behavior detection method |
Non-Patent Citations (1)
| Title |
|---|
| 基于时空关联矩阵的配电台区反窃电预警方法;唐冬来 等;《电力系统自动化》;20201010;第44卷(第19期);第168-176页 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11899516B1 (en) | 2023-07-13 | 2024-02-13 | T-Mobile Usa, Inc. | Creation of a digital twin for auto-discovery of hierarchy in power monitoring |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113687176A (en) | 2021-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN113687176B (en) | Deep neural network-based power consumption abnormity detection method and system | |
| CN114298863B (en) | Data acquisition method and system of intelligent meter reading terminal | |
| CN112734128B (en) | 7-day power load peak prediction method based on optimized RBF | |
| CN110222991B (en) | Metering device fault diagnosis method based on RF-GBDT | |
| CN112149873B (en) | Low-voltage station line loss reasonable interval prediction method based on deep learning | |
| CN113111596B (en) | Carbon peak-reaching real-time visual measuring and calculating method based on electric brain center | |
| CN109033513A (en) | Method for diagnosing fault of power transformer and diagnosing fault of power transformer device | |
| CN110865260A (en) | Method for monitoring and evaluating MOV actual state based on outlier detection | |
| CN111062620B (en) | Intelligent electric power charging fairness analysis system and method based on hybrid charging data | |
| CN104281779A (en) | Abnormal data judging and processing method and device | |
| CN115686124B (en) | Energy storage battery output power self-adjusting system and method based on safety protection | |
| CN116031888A (en) | Dynamic load prediction-based power flow optimization method, system and storage medium | |
| CN115081795A (en) | Enterprise energy consumption abnormity cause analysis method and system under multidimensional scene | |
| CN111881961A (en) | Power distribution network fault risk grade prediction method based on data mining | |
| CN116227637A (en) | Active power distribution network oriented refined load prediction method and system | |
| CN110880044B (en) | Markov chain-based load prediction method | |
| CN116148753A (en) | Intelligent electric energy meter operation error monitoring system | |
| CN115293257A (en) | Detection method and system for abnormal electricity utilization user | |
| CN117235664A (en) | Fault diagnosis method and system for power distribution communication equipment and computer equipment | |
| CN116151799A (en) | BP neural network-based distribution line multi-working-condition fault rate rapid assessment method | |
| CN110852906A (en) | Method and system for identifying electricity stealing suspicion based on high-dimensional random matrix | |
| CN116384223A (en) | Nuclear equipment reliability assessment method and system based on intelligent degradation state identification | |
| CN112926801B (en) | Load curve combined prediction method and device based on quantile regression | |
| CN114692729A (en) | New energy station bad data identification and correction method based on deep learning | |
| CN114385403A (en) | Distributed cooperative fault diagnosis method based on double-layer knowledge graph framework |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |