CN112016684B - Electric power terminal fingerprint identification method of deep parallel flexible transmission network - Google Patents

Electric power terminal fingerprint identification method of deep parallel flexible transmission network Download PDF

Info

Publication number
CN112016684B
CN112016684B CN202010780676.2A CN202010780676A CN112016684B CN 112016684 B CN112016684 B CN 112016684B CN 202010780676 A CN202010780676 A CN 202010780676A CN 112016684 B CN112016684 B CN 112016684B
Authority
CN
China
Prior art keywords
power terminal
deep
parallel
model
transmitting network
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
Application number
CN202010780676.2A
Other languages
Chinese (zh)
Other versions
CN112016684A (en
Inventor
殷林飞
马晨骁
罗仕逵
高放
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangxi University
Original Assignee
Guangxi University
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangxi University filed Critical Guangxi University
Priority to CN202010780676.2A priority Critical patent/CN112016684B/en
Publication of CN112016684A publication Critical patent/CN112016684A/en
Application granted granted Critical
Publication of CN112016684B publication Critical patent/CN112016684B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • 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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R19/00—Arrangements for measuring currents or voltages or for indicating presence or sign thereof
    • G01R19/04—Measuring peak values or amplitude or envelope of ac or of pulses
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
    • G01R23/16—Spectrum analysis; Fourier analysis
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R25/00—Arrangements for measuring phase angle between a voltage and a current or between voltages or currents
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
    • Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
    • Y04S10/00—Systems supporting electrical power generation, transmission or distribution
    • Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications

Abstract

The invention provides a power terminal fingerprint identification method of a deep parallel flexible transmitting network. The method is based on a deep flexible transmitting network model combining a flexible transmitting network and deep learning; and in training, the method uses different deep flexible transmitting network parameters to carry out parallel learning. When a power terminal is newly connected into the power system, the method completes the identification of the power terminal according to the fingerprint of the power terminal. The method can identify the power terminal accessed to the power system, provides reference for safe operation and control and load prediction of the power system, and effectively improves the intelligent level of the power system.

Description

Electric power terminal fingerprint identification method of deep parallel flexible transmission network
Technical Field
The invention belongs to the field of intellectualization and control of an electric power system, and relates to a method for learning and identifying fingerprints of an electric power terminal by using a deep learning model of a flexible transmitting network in parallel, which is suitable for load identification and intellectualization level improvement of the electric power system.
Background
With the rapid development of computer science, the intelligence level of the power system is also continuously improved. Today's smart grid technology has penetrated into various links of power generation, power transmission, power distribution, power transformation and utilization. Based on the smart grid technology, the operating efficiency and stability of the power system are improved. Under the background of rapid development in the field of artificial intelligence nowadays, one direction of improvement of the power grid intelligence level is to combine with artificial intelligence. However, training of the artificial intelligent model requires large-scale data to support, and the data source of the current power grid is often power grid equipment, so that the abundance of power grid operation data is effectively improved by adding the power terminal equipment into the data source, and a foundation is laid for improving the intelligence level of the power system. One important data in the power terminal data is the type, name of the power terminal accessing the power grid. With the large-scale application of terminal monitoring equipment such as an intelligent ammeter and the like, the problem of obtaining power consumption data of a power grid terminal does not exist. Therefore, the research on the effective power terminal fingerprint identification method for identifying the power consumption data has important guiding significance for knowing the power consumption characteristics of users, improving the intelligent level of a power grid and adjusting a power generation plan.
Disclosure of Invention
The invention provides a power terminal fingerprint identification method of a deep parallel flexible transmitting network, which uses a deep learning model of the parallel flexible transmitting network to learn and identify the fingerprint of a power terminal, and comprises the following steps:
acquiring power utilization data: collecting power consumption data of voltage and current of each power terminal in different working states when the power terminal operates independently;
a data preprocessing step: carrying out Fourier transform on the voltage and current waveforms obtained in the electricity data acquisition step to form a 4n matrix as a fingerprint characteristic of the power terminal;
training: the method aims to establish a deep parallel flexible transmitting network model, establish a voltage and current Fourier transform matrix model and a data set corresponding to the name of a power terminal and train the model in parallel; when the fingerprint data of the power terminal is trained, a deep flexible transmitting network model based on the combination of a flexible transmitting network and deep learning is used, and the number of the used flexible transmitting network neurons is not less than three. And in training, parallel learning is carried out by using different deep flexible transmitting network parameters, and the number of parallel models is not less than three. The flexible transmitting network neuron framework comprises two feedforward channel parameters (w, v) and an iterative memory intensity parameter M, and each parallel deep flexible transmitting network training step specifically comprises the following steps:
(1) a feed-forward process: taking the t-th neuron iteration for the first time as an example, the feedforward channel has two parameters (w, v), the memory strength of the neuron itself has an iteration parameter M, and after the feedforward function is described in a complex number, the following feedforward process function is obtained as
Figure BDA0002620070540000021
Wherein
Figure BDA0002620070540000022
For the output of the current iteration,
Figure BDA0002620070540000023
for the memory strength of the next iteration, sigma is a sigmoid excitation function and comprises a real part function and an imaginary part function, and both alpha and beta are adjustable parameters of the parallel model;
(2) and (3) a back propagation process: after one iteration is completed, the parameters of the flexible transmitting network are corrected according to the iteration result.
First, the gradient of the model is calculated,
Figure BDA0002620070540000024
wherein
Figure BDA0002620070540000025
For transmission in the reverse directionDifference of broadcast correction
Figure BDA0002620070540000026
In order to perform the dot-product operation,
Figure BDA0002620070540000027
for the activated point-state derivative vector,
Figure BDA0002620070540000028
and
Figure BDA0002620070540000029
the method for calculating the core for back propagation comprises two methods respectively connected with
Figure BDA00026200705400000210
And
Figure BDA00026200705400000211
an associated counter-propagating conduit;
correcting W, V by a given step length eta
Figure BDA00026200705400000212
The loss function E should be calculated after the back propagation process is finished. The loss function E is an integral loss function E (W, V) of the flexible transmit network neuron model, which fits the loss by the square of the difference of the real and predicted values, specifically:
Figure BDA00026200705400000213
wherein Y is t In order to be the output signal of the final layer,
Figure BDA00026200705400000214
is a real signal;
Figure BDA00026200705400000215
is a direction composed of 0 and 1Quantity, if the device number is k, then
Figure BDA00026200705400000216
Is 1, otherwise is 0.
And judging whether the continuous iteration or the training reaches the expected index according to the loss function calculation result. The last iteration is carried out on M after the expected index is reached,
Figure BDA00026200705400000217
an identification step: and inputting a Fourier transform matrix of the voltage and current waveforms of the power terminal into the deep parallel flexible transmitting network classifier model after the training step is completed, judging and calculating parallel calculation results obtained by parallel models with different parameters according to an accumulation sum mode to obtain the maximum possible power terminal number, and taking the maximum possible power terminal number as an identification result.
The fingerprint identification method for the power terminal can learn voltage and current waveforms of a large number of power terminal devices, namely the fingerprints of the power terminal, and accurately identify the power terminal devices when the power terminal devices are accessed. Compared with a Macarocke-Pettes model, the flexible transmission network model has the advantages that the operation speed and the prediction accuracy are remarkably improved by introducing the flexible transmission network model, and the requirements of high precision and low calculation power are met under the condition that a large number of power terminals are connected. Through the identification of the fingerprint of the power terminal, the smart grid can sense the information of the power terminal equipment accessed to the power grid, so that more references are provided for the operation of the smart grid.
Drawings
FIG. 1 is a general design flow diagram of the method of the present invention.
Fig. 2 is a schematic diagram of a flexible transmission network of the method of the present invention.
Detailed Description
The invention provides a power terminal fingerprint identification method of a deep parallel flexible transmission network, which is explained in detail by combining the attached drawings as follows:
FIG. 1 is a general design flow diagram of the method of the present invention. Firstly, voltage and current waveforms under different operating conditions are obtained from a large number of electric power terminal devices, waveform characteristics are learned through a deep parallel flexible transmitting network, electric power terminal fingerprint identification after intervention of new devices to generate new waveforms is achieved, and the method is specifically achieved through the following steps.
Step S1, collecting working conditions of the power terminal equipment needing to be learned, connecting the power terminal equipment to the voltage and current monitoring equipment, and applying different working conditions to the power terminal equipment to obtain voltage and current waveforms of the power terminal equipment under different working conditions. And carrying out high-frequency acquisition on the waveform, and storing to form test data.
And step S2, processing the voltage and current data acquired at high frequency in the step S1, performing discrete Fourier decomposition operation on the voltage and current data to obtain amplitude and phase information of each harmonic of a voltage and current waveform, and integrating the amplitude and phase information to form a 4n matrix which is regarded as the fingerprint of the power terminal equipment. Because the load in the power grid always presents the characteristic of a near-sine waveform and the content of higher harmonics is gradually reduced, the fingerprint matrix of the power terminal can be reasonably discarded, a decomposition value with higher frequency is ignored, and the operation speed of subsequent steps is increased.
And step S3, training a deep parallel flexible transmitting network by using the fingerprint data of the power terminal processed in the step S2, wherein the flexible transmitting network is improved on the basis of a Macarock-Pettes model, a new bionics principle is introduced, the research result of the neurons is improved according to biology, the parameters of feedforward channels are improved to 2, the 3 rd parameter is introduced to simulate the memory intensity of the neurons, and the improved model has higher target fitting speed and higher fitting precision. Before training, the depth and iteration parameters of the deep flexible transmitting network are reasonably set, the number of deep learning neurons is more than 3, and a deep flexible transmitting network model with different parameters or depths is trained by using parallel calculation.
The flexible transmitting network neuron framework comprises two feedforward channel parameters (w, v) and an iterative memory intensity parameter M, and each parallel deep flexible transmitting network training step specifically comprises the following steps:
(1) a feed-forward process: taking the t-th neuron as an example, the feedforward channel has two parameters (w, v), the memory strength of the neuron itself has an iterative parameter M, and after complex description of the feedforward function, the following feedforward process function is obtained as
Figure BDA0002620070540000041
Wherein
Figure BDA0002620070540000042
Is output for the current iteration and is used as a starting point,
Figure BDA0002620070540000043
for the memory strength of the next iteration, sigma is a sigmoid excitation function and comprises a real part function and an imaginary part function, and both alpha and beta are adjustable parameters of the parallel model;
(2) and (3) a back propagation process: after one iteration is finished, the flexible transmitting network parameters are corrected according to the iteration result, firstly, the model gradient is calculated,
Figure BDA0002620070540000044
wherein
Figure BDA0002620070540000045
In order to correct for the difference in the back-propagation,
Figure BDA0002620070540000046
in order to perform the dot-product operation,
Figure BDA0002620070540000047
point state derivative vector for activation
Figure BDA0002620070540000048
Figure BDA00026200705400000415
And
Figure BDA0002620070540000049
a method is computed for the back propagation core. Comprises two of
Figure BDA00026200705400000410
And
Figure BDA00026200705400000411
an associated counter-propagating conduit;
wherein:
Figure BDA00026200705400000412
Figure BDA00026200705400000413
wherein:
Figure BDA00026200705400000414
wherein:
Figure BDA0002620070540000051
after the calculation is finished, correcting W and V by a given step length eta
Figure BDA0002620070540000052
The loss function E should be calculated after the back propagation process is finished. The loss function E is an integral loss function E (W, V) of the flexible transmit network neuron model, which fits the loss by squaring the difference between the true and predicted values, specifically:
Figure BDA0002620070540000053
wherein Y is t In order to be the output signal of the final layer,
Figure BDA0002620070540000054
is a true signal;
Figure BDA0002620070540000055
is a vector composed of 0 and 1, if the device number is k, then
Figure BDA0002620070540000056
Is 1, otherwise is 0.
And judging whether the continuous iteration or the training reaches the expected index according to the loss function calculation result. The last iteration is carried out on M after the expected index is reached,
Figure BDA0002620070540000057
and step S4, after the loss function of each parallel model meets the requirement, the fingerprint of the power terminal can be identified. Collecting voltage and current waveforms of the power terminal equipment, performing discrete Fourier transform on the voltage and current waveforms, inputting a transform result into a depth parallel flexible transmitting network model which completes learning, calculating in each parallel model, and adding each parallel result to obtain a maximum possible result, namely, a recognition result.
Fig. 2 is a schematic diagram of a flexible transmission network of the method of the present invention. According to the power terminal fingerprint identification method, the flexible transmission network model is introduced, and compared with a Macarocke-Pettes model, the operation speed and the prediction accuracy are remarkably improved, so that the requirements of high precision and low calculation power are met in actual engineering under the condition that a large number of power terminals are accessed. Through the identification of the fingerprint of the power terminal, the smart grid can sense the information of the power terminal equipment accessed to the power grid, so that more references are provided for the operation of the smart grid.

Claims (5)

1. A power terminal fingerprint identification method of a deep parallel flexible transmission network is characterized in that the power terminal fingerprint can be learned by using the deep parallel flexible transmission network, and the power terminal is identified when being accessed into a power system; the method comprises the following steps in the using process:
(1) acquiring power utilization data: collecting power consumption data of voltage and current of each power terminal in different working states when the power terminal operates independently;
(2) a data preprocessing step: performing discrete Fourier transform on the voltage and current waveforms obtained in the step (1) to form a 4n matrix as a fingerprint feature of the power terminal;
(3) training: establishing a deep flexible transmitting network based on the combination of the flexible transmitting network and deep learning; the flexible transmitting network is improved on the basis of a Macarok-Pittss model, the parameters of feed-forward channels are improved to 2, and the 3 rd parameter is introduced to simulate the memory intensity of neurons, and the flexible transmitting network neuron framework comprises two feed-forward channel parameters (w, v), an iterative memory intensity parameter M and a loss function E; the loss function E is an integral loss function E (W, V) of the flexible transmitting network neuron model, and the loss is fitted by the square of the difference between a real value and a predicted value:
Figure FDA0003730889290000011
wherein Y is t Is the output signal of the final layer; y is t true Is a true signal; y is t true Is a vector consisting of 0 and 1, if the device number is k, then Y t true | k Is 1, otherwise is 0;
the improved model has higher target fitting speed and higher fitting precision; before training is carried out, firstly, the depth and iteration parameters of the deep flexible transmitting network are set, the number of deep learning neurons is more than 3, and a deep flexible transmitting network model with different parameters or depths is trained by using parallel calculation; the number of parallel models is not less than three; establishing a voltage and current Fourier transform matrix model and a data set corresponding to the name of the power terminal to carry out parallel training on the model;
(4) an identification step: and inputting a discrete Fourier transform matrix of the voltage and current waveforms of the power terminal in the deep parallel flexible transmitting network classifier model after the training step is completed, and outputting a fingerprint matching result of the power terminal after the judgment of a parallel system.
2. The power terminal fingerprint identification method of the deep parallel flexible transmission network according to claim 1, wherein the power terminal fingerprint characteristics in the step (2) are voltage and current waveforms of the power terminal in different operating states, and a 4n matrix obtained by performing discrete fourier transform on the power terminal fingerprint characteristics is a mathematical model of the characteristics.
3. The power terminal fingerprint identification method of the deep parallel flexible transmission network according to claim 1, wherein in the step (4) of identifying the power terminal fingerprint data, the deep flexible transmission network is used for identification, a plurality of predicted values are output through a parallel model, and a judgment is made and an identification result is determined in an accumulation and sum manner.
4. The power terminal fingerprint identification method of the deep parallel flexible transmission network according to claim 1, characterized in that each parallel deep flexible transmission network training step comprises the following steps,
(1) a feed-forward process: taking the t-th neuron as an example in the first iteration, the feedforward channel has two parameters (w, v), the neuron itself; the memory intensity has an iterative parameter M, and after the feedforward function is described in a complex number, the following feedforward process function is obtained as
Figure FDA0003730889290000021
Wherein
Figure FDA0003730889290000022
For the output of the current iteration,
Figure FDA0003730889290000023
for the memory strength of the next iteration, sigma is a sigmoid excitation function and has a real part function and an imaginary part function, and alpha and beta are both adjustable parameters of the parallel model;
(2) and (3) a back propagation process: after one iteration is finished, the flexible transmitting network parameters are corrected according to the iteration result, firstly, the model gradient is calculated,
Figure FDA0003730889290000024
wherein
Figure FDA0003730889290000025
In order to correct for the difference in the back-propagation,
Figure FDA0003730889290000026
in order to perform the dot-product operation,
Figure FDA0003730889290000027
for the activated point-state derivative vector,
Figure FDA0003730889290000028
and
Figure FDA0003730889290000029
the method for calculating the core for back propagation comprises two methods which are respectively connected with W t l And V t l An associated counter-propagating conduit;
correcting W and V with a given step length eta,
Figure FDA00037308892900000210
after the back propagation process is finished, judging whether the iteration is continued or the training reaches the expected index according to the calculation result of the loss function, carrying out the last iteration on the M after the expected index is reached,
Figure FDA00037308892900000211
5. the power terminal fingerprint identification method of the deep parallel flexible transmission network according to claim 1, characterized in that, the real-time voltage and current waveforms of the power terminal operation are subjected to discrete fourier transform to form a 4n matrix, and then the 4n matrix is sent to the deep parallel flexible transmission network which has been trained for calculation, and the obtained parallel calculation result is judged and calculated according to the accumulation sum mode, so as to obtain the maximum possible power terminal number, which is regarded as the identification result.
CN202010780676.2A 2020-08-06 2020-08-06 Electric power terminal fingerprint identification method of deep parallel flexible transmission network Active CN112016684B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010780676.2A CN112016684B (en) 2020-08-06 2020-08-06 Electric power terminal fingerprint identification method of deep parallel flexible transmission network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010780676.2A CN112016684B (en) 2020-08-06 2020-08-06 Electric power terminal fingerprint identification method of deep parallel flexible transmission network

Publications (2)

Publication Number Publication Date
CN112016684A CN112016684A (en) 2020-12-01
CN112016684B true CN112016684B (en) 2022-09-02

Family

ID=73499055

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010780676.2A Active CN112016684B (en) 2020-08-06 2020-08-06 Electric power terminal fingerprint identification method of deep parallel flexible transmission network

Country Status (1)

Country Link
CN (1) CN112016684B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113205177B (en) * 2021-04-25 2022-03-25 广西大学 Electric power terminal identification method based on incremental collaborative attention mobile convolution

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101900761A (en) * 2009-11-05 2010-12-01 中国航天科技集团公司第五研究院第五一四研究所 High-accuracy non-integer-period sampled harmonic analysis and measurement method
CN102129013A (en) * 2011-01-21 2011-07-20 昆明理工大学 Distribution network fault location method utilizing natural frequency and artificial neural network
CN107045649A (en) * 2014-07-21 2017-08-15 徐志强 Simulative neural network with short-term memory and long-term memory
CN109727678A (en) * 2017-10-28 2019-05-07 徐志强 Variable neuron, cynapse simulator and simulated substrate neural circuitry
CN110503188A (en) * 2019-08-01 2019-11-26 三星（中国）半导体有限公司 The method and apparatus for realizing the deep neural network model based on isomery illustraton of model
CN110705456A (en) * 2019-09-29 2020-01-17 华南理工大学 Micro motor abnormity detection method based on transfer learning
CN110738253A (en) * 2019-10-14 2020-01-31 南京信息工程大学 short-term wind power prediction method based on FCM and AFSA-Elman

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101900761A (en) * 2009-11-05 2010-12-01 中国航天科技集团公司第五研究院第五一四研究所 High-accuracy non-integer-period sampled harmonic analysis and measurement method
CN102129013A (en) * 2011-01-21 2011-07-20 昆明理工大学 Distribution network fault location method utilizing natural frequency and artificial neural network
CN107045649A (en) * 2014-07-21 2017-08-15 徐志强 Simulative neural network with short-term memory and long-term memory
CN109727678A (en) * 2017-10-28 2019-05-07 徐志强 Variable neuron, cynapse simulator and simulated substrate neural circuitry
CN110503188A (en) * 2019-08-01 2019-11-26 三星（中国）半导体有限公司 The method and apparatus for realizing the deep neural network model based on isomery illustraton of model
CN110705456A (en) * 2019-09-29 2020-01-17 华南理工大学 Micro motor abnormity detection method based on transfer learning
CN110738253A (en) * 2019-10-14 2020-01-31 南京信息工程大学 short-term wind power prediction method based on FCM and AFSA-Elman

Non-Patent Citations (5)

* Cited by examiner, † Cited by third party
Title
Deep Flexible Transmitter Networks for Non-Intrusive Load Monitoring of Power Distribution Networks;Ma XiaoChen等;《IEEE Access》;20210730;第9卷;第107424-107436页 *
Flexible Transmitter Network;Shao-Qun Zhang等;《arXiv:2004.03839v2 [cs.NE]》;20200410;第1-27页 *
Waveform contenders for 5G: Description, analysis and comparison;Jean-BaptisteDoré等;《Physical communication》;20171230;第24卷;第46-61页 *
基于多神经网络融合的短期负荷预测方法;庞昊等;《电力自动化设备》;20200630;第40卷(第6期);第37-42页 *
基于小波和神经网络算法的电力系统短期负荷预测的研究;黄丹;《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》;20091215(第12期);第C042-159页 *

Also Published As

Publication number Publication date
CN112016684A (en) 2020-12-01

Similar Documents

Publication Publication Date Title
CN109214575B (en) Ultrashort-term wind power prediction method based on small-wavelength short-term memory network
CN109685314B (en) Non-intrusive load decomposition method and system based on long-term and short-term memory network
CN111880044B (en) Online fault positioning method for distribution network containing distributed power supply
CN110726898B (en) Power distribution network fault type identification method
CN110070228B (en) BP neural network wind speed prediction method for neuron branch evolution
CN113988449B (en) Wind power prediction method based on transducer model
CN114157539B (en) Data-aware dual-drive modulation intelligent identification method
CN111105321A (en) Power load prediction method based on EEMD secondary decomposition
CN112016684B (en) Electric power terminal fingerprint identification method of deep parallel flexible transmission network
CN113804997B (en) Voltage sag source positioning method based on bidirectional WaveNet deep learning
CN115146538A (en) Power system state estimation method based on message passing graph neural network
CN113313198B (en) Cutter wear prediction method based on multi-scale convolution neural network
CN109459609B (en) Distributed power supply frequency detection method based on artificial neural network
CN112070625A (en) User electricity consumption behavior pattern recognition method and system
CN113158134B (en) Method, device and storage medium for constructing non-invasive load identification model
CN113780661B (en) Wind power ultra-short-term power prediction method for abrupt wind speed
CN110610203A (en) Electric energy quality disturbance classification method based on DWT and extreme learning machine
CN115469184A (en) New energy transmission line fault identification method based on convolutional network
CN111797991A (en) Deep network model compression system, method and device
CN111539275B (en) Electrical load classification method and system based on load characteristic visualization
CN113887794A (en) Method and device for reactive power optimization of power distribution network
CN113761777A (en) Ultra-short-term photovoltaic power prediction method based on HP-OVMD
CN112836876A (en) Power distribution network line load prediction method based on deep learning
CN111882284A (en) Operation monitoring method and system for distributed power supply grid connection
CN113205177B (en) Electric power terminal identification method based on incremental collaborative attention mobile convolution

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