CN110703006B - Three-phase power quality disturbance detection method based on convolutional neural network - Google Patents
Three-phase power quality disturbance detection method based on convolutional neural network Download PDFInfo
- Publication number
- CN110703006B CN110703006B CN201910832724.5A CN201910832724A CN110703006B CN 110703006 B CN110703006 B CN 110703006B CN 201910832724 A CN201910832724 A CN 201910832724A CN 110703006 B CN110703006 B CN 110703006B
- Authority
- CN
- China
- Prior art keywords
- phase
- electric energy
- disturbance
- neural network
- convolutional neural
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2415—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on parametric or probabilistic models, e.g. based on likelihood ratio or false acceptance rate versus a false rejection rate
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Systems or methods specially adapted for specific business sectors, e.g. utilities or tourism
- G06Q50/06—Electricity, gas or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/02—Preprocessing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Abstract
The invention relates to the field of electrical automation, and discloses a three-phase power quality disturbance detection method based on a convolutional neural network, which comprises the following steps: A) collecting three-phase electric energy disturbance signal data; B) carrying out data preprocessing on the three-phase electric energy disturbance signal to obtain a three-phase electric energy RGB picture; C) making a training set and a testing set; D) constructing a convolutional neural network model; E) and obtaining a three-phase power quality disturbance detection result. The method has high efficiency, reduces the influence of artificial subjective factors, converts the collected three-phase electric energy signals into RGB pictures, ensures that the signal characteristics are more compact, has less calculation amount, intelligently extracts the characteristics of the disturbance signals by establishing a convolutional neural network model, realizes the accurate classification of the disturbance signals, can simultaneously detect a plurality of disturbance signals, and can detect the condition of faults between two phases or three phases.
Description
Technical Field
The invention relates to the technical field of electrical automation, in particular to a three-phase power quality disturbance detection method based on a convolutional neural network.
Background
In recent years, with the dual development of national economy and science and technology, the demand of people for daily electricity utilization is continuously increased, and the requirement on the quality of electric power is more strict. However, since various new electrical devices and loads with different performances are connected to the power system, the voltage in the power grid is more likely to be disturbed, causing various problems related to the power quality, which may shorten the service life of the electrical devices, increase the line loss rate, and even cause abnormal or damaged operation of the electrical devices, causing a large-scale power failure event, and bringing huge economic loss and consequences. Nowadays, disturbance detection on the quality of electric energy becomes an important means for evaluating and purifying the quality of electric power, and has important significance on safe power supply and distribution optimization of electric power resources. At present, for power quality disturbance detection, a common method is to extract relevant features from an original voltage signal by an empirical formula, and then send the features into a manufactured classifier to realize classification. The common method for feature extraction comprises the following steps: fourier transform, wavelet transform, S transform, etc.; the classification is usually performed by: support vector machines, decision trees, and the like. The Fourier transform transducer maps the time domain signal to the frequency domain, the use is simple, and the detection precision is high, but when the noise of the signal is large and the period is unstable, the method is easy to cause the problem of frequency spectrum leakage, so that the result has large deviation, and the detection precision is obviously reduced. The wavelet transform is extended to a certain extent on the theoretical basis of Fourier transform, so that the size of a window can change along with the frequency, unbalanced signals can be better processed, but the accuracy of a final detection result is low because the wavelet transform is insensitive to a high-frequency band part, frequency band aliasing often occurs when the method decomposes a disturbance signal, and the instantaneity and the stability are poor. The support vector machine is a typical two-classification method, but in reality, many problems are multi-classification problems, so that the use of the SVM is limited, and suitable parameters are not easy to configure when the optimal hyperplane is solved. The decision tree has a tree structure which increases gradually from top to bottom layer by layer, is greatly influenced by data quality, has an inflexible algorithm, and is difficult to process when the data volume is large enough.
For example, a method for identifying a power quality disturbing signal disclosed in chinese patent document, whose publication No. CN 109324250a, includes the following steps: s1, obtaining a known power quality disturbance signal; s2, performing wavelet packet transformation on the power quality disturbance signal in the step S1, and extracting a feature vector; s3, classifying the power quality disturbance signals based on the characteristic vectors of the power quality disturbance signals in the step S2 and based on fuzzy clustering, and obtaining central vectors of clusters; and S4, extracting characteristic vectors of the power quality disturbance signals to be identified according to the step S2, respectively calculating the distance between the characteristic vectors and the clustering center vectors obtained in the step S3, and further judging the type of the power quality disturbance signals to be identified. The electric energy quality disturbance signal is subjected to wavelet packet transformation, the calculated amount is large, when the data amount is enough, the processing is troublesome, and the classification accuracy of the electric energy quality disturbance signal is low.
Disclosure of Invention
The invention provides a three-phase power quality disturbance detection method based on a convolutional neural network, which aims to solve the problems of large calculated amount and low accuracy of detection results of the traditional power quality disturbance detection method.
In order to achieve the purpose, the invention adopts the following technical scheme:
a three-phase power quality disturbance detection method based on a convolutional neural network comprises the following steps:
A) collecting three-phase electric energy disturbance signal data;
B) carrying out data preprocessing on the three-phase electric energy disturbance signal to obtain a three-phase electric energy RGB picture;
C) making a training set and a testing set;
D) constructing a convolutional neural network model, taking a three-phase electric energy RGB picture as the input of the convolutional neural network model, training the convolutional neural network model by using a training set, adjusting parameters, and testing the convolutional neural network model by using a test set;
E) and acquiring a three-phase power disturbance signal to be detected, inputting the three-phase power disturbance signal into the trained convolutional neural network model, and taking the output of the convolutional neural network model as a three-phase power quality disturbance detection result.
The method comprises the steps of performing two-dimensional mapping on a one-dimensional signal, converting a one-dimensional three-phase electric energy disturbance signal into a two-dimensional matrix through data preprocessing, obtaining a three-phase electric energy RGB picture, effectively reducing the scale of the disturbance signal, enabling the feature distribution of the disturbance signal to be more concentrated, being beneficial to adopting a convolutional neural network to perform disturbance feature learning and disturbance identification, inputting a manufactured training set into the convolutional neural network for training through constructing a convolutional neural network model, performing parameter adjustment, and using a test set to detect the accuracy. After the test is passed, in practical application, only newly acquired signals are required to be made into a data set and sent into a trained convolutional neural network, so that a corresponding classification result can be obtained, and the intelligent detection of the power quality disturbance is realized.
Further, step B) comprises:
B1) setting sampling frequency, sampling the three-phase electric energy disturbance signal in the period for h times to obtain 3 multiplied by h sampling data, and recording the data as a matrix
H-th sampling data representing the i-th perturbation signal;
B2) uniformly dividing the sampling data of each phase disturbance signal for n times, wherein each division comprises m sampling data to obtain an m multiplied by n two-dimensional matrix of the first phase disturbance signal
The m x n two-dimensional matrix of the second-phase disturbance signal is H2The m × n two-dimensional matrix of the third phase disturbance signal is H3；
B3) Setting a compression amplitude value interval [ -X, X ], carrying out compression truncation processing on each element in the two-dimensional matrix, and obtaining a sampling value f (u) corresponding to the element after processing;
B4) performing linear transformation on the processed sampling values f (u) to obtain transformed values g (u) and obtain 3 transformed two-dimensional matrixes;
B5) and respectively taking the 3 transformed two-dimensional matrixes as three single-channel matrixes of the RGB image to obtain the three-phase electric energy RGB image.
The method comprises the steps of taking each m sampling points of a one-dimensional three-phase electric energy disturbing signal as a column to be cut off, obtaining n columns in total, obtaining one sampling data for each phase when the three-phase electric energy disturbing signal is sampled once, obtaining 3 sampling data for three phases, obtaining 3 m multiplied by n two-dimensional matrixes in total, compressing the electric energy disturbing signal into a compression amplitude range of (-X, X) through compression processing, enabling the electric energy disturbing signal to be in a range of 0 to 255 through linear transformation because an RGB color single channel has 256 levels of brightness and is represented by numbers of 0, 1, 2, 1.
By calculation of
To f*(u) rounding to obtain a processed sampling value f (u), wherein u0The average value of the maximum values of the electric energy signals is u, and the amplitude of the disturbance electric energy signals is u; by calculating g (u) ═ 255(f (u) + X)/2X]The transformed values g (u) are obtained.
Further, setting time periods T, monitoring the three-phase electric energy signals through the power grid in each time period T, and recording the maximum value of the three-phase electric energy signals in the ith time period
Obtaining the average value u of the maximum value of the electric energy signal0，
The average value of the maximum value of the electric energy signal is updated in real time by using historical data of a power grid, so that the three-phase disturbance electric energy signal can have elasticity when being processed, the electric energy disturbance signal can be better distinguished according to actual conditions when linear transformation is carried out, and serious voltage sag events and voltage interruption events can be better distinguished.
The three-phase electric energy disturbance signals comprise three-phase voltage temporary rise, three-phase voltage temporary fall, single-phase grounding short-circuit fault, two-phase grounding short-circuit fault, interphase short-circuit fault, three-phase grounding short-circuit fault, three-phase interruption fault, three-phase voltage flicker, three-phase harmonic wave and three-phase transient oscillation.
Further, step C) comprises: and B), generating w random samples for each three-phase electric energy disturbance signal, repeating the step B) for each random sample to obtain w three-phase electric energy RGB pictures, and dividing the w three-phase electric energy RGB pictures into a training set and a test set according to a proportion.
Further, the convolutional neural network model in step D) includes: two convolutional layers, two pooling layers, two full-link layers and an output layer, wherein the convolutional layers use ReLu activation functions,adopting 64 convolution kernels with the step length of 1 and the size of 5 multiplied by 5, adopting kernels with the step length of 2 and the size of 3 multiplied by 3 in the pooling layer, carrying out one-dimensional treatment on the output matrix of the second pooling layer, rearranging the output matrix into a column of vectors according to the sequence of rows, and carrying out L treatment on the weights by the full-connection layer2And (3) regularization, wherein a ReLu activation function is used in the first full-connection layer, 10 labels are set as output in the output layer, and a softmax function is used as a classifier.
The convolution layer of the convolution neural network performs convolution on all parts of an input three-phase electric energy disturbance signal to extract local features to obtain a feature map, and each feature map is subjected to independent operation such as average pooling or maximum pooling through the pooling layer. The expression of the ReLu activation function is f (x) ═ max (0, x), in the case that the input is a negative value, the output is 0 after passing through the ReLu activation function, and the neurons are not activated, which means that only part of the neurons are activated, so that the network is sparse, and after the sparseness is realized through the ReLu activation function, the calculation amount is reduced, and the model can better mine relevant features. L is2The regularization has the effect of limiting the parameters to be too high or too large, and avoiding the model to be more complex. The softmax function maps the outputs of a plurality of neurons into the (0, 1) interval, thereby performing multi-classification.
Therefore, the invention has the following beneficial effects: the method has the advantages that the efficiency is high, the influence of artificial subjective factors is reduced, the collected three-phase electric energy signals are converted into RGB pictures, the signal characteristics are more compact after the three-phase electric energy signals are converted into the RGB pictures, the calculated amount is small, three-phase electric energy quality disturbance data are quickly and effectively processed, the disturbance signal characteristics are intelligently extracted by establishing a convolutional neural network model, the disturbance signals are accurately classified, a plurality of disturbance signals can be simultaneously detected, and the condition that faults occur between two phases or three phases can be detected.
Drawings
FIG. 1 is a flow chart of a three-phase power quality disturbance detection method based on a convolutional neural network.
Fig. 2 is a schematic diagram of the structure of the convolutional neural network of the present invention.
Detailed Description
The invention is further described with reference to the following detailed description and accompanying drawings.
A three-phase power quality disturbance detection method based on a convolutional neural network is disclosed, as shown in FIG. 1, and comprises the following steps: A) collecting three-phase electric energy disturbance signal data, wherein the three-phase electric energy disturbance signal data comprise three-phase voltage temporary rise, three-phase voltage temporary fall, single-phase grounding short circuit fault, two-phase grounding short circuit fault, interphase short circuit fault, three-phase grounding short circuit fault, three-phase interruption fault, three-phase voltage flicker, three-phase harmonic wave and three-phase transient oscillation.
B) Carrying out data preprocessing on the three-phase electric energy disturbance signal to obtain a three-phase electric energy RGB picture, comprising the following steps of: B1) setting sampling frequency, sampling three-phase electric energy disturbance signals in a period for h times to obtain 3 multiplied by h sampling data, and recording the data as a matrix
H-th sampling data representing the i-th perturbation signal;
B2) uniformly dividing the sampling data of each phase disturbance signal for n times, wherein each division comprises m sampling data to obtain an m multiplied by n two-dimensional matrix of the first phase disturbance signal
The m x n two-dimensional matrix of the second-phase disturbance signal is H2The m x n two-dimensional matrix of the third phase disturbance signal is H3；
B3) Setting time periods T, monitoring three-phase electric energy signals through the power grid in each time period T, and recording the maximum value of the three-phase electric energy signals in the ith time period
Obtaining the average value u of the maximum value of the electric energy signal0，
Setting a compression amplitude interval [ -2,2 [ -2]Performing compression truncation processing on each element in the two-dimensional matrix through calculation
To f*(u) rounding to obtain a processed sampling value f (u), wherein u0The average value of the maximum values of the electric energy signals is u, and the amplitude of the disturbance electric energy signals is u;
B4) performing linear transformation on the processed sampling values f (u), and calculating g (u) ═ 255(f (u) +2)/4 to obtain transformed values g (u) so as to obtain 3 transformed two-dimensional matrixes;
B5) and (3) respectively taking the three transformed two-dimensional matrixes as three single-channel matrixes of the RGB image to form the three-phase electric energy RGB image.
And B) generating w random samples for each three-phase electric energy disturbance signal, repeating the step B) for each random sample to obtain w three-phase electric energy RGB pictures, and dividing the w three-phase electric energy RGB pictures into a training set and a test set according to a proportion.
And constructing a convolutional neural network model, taking a three-phase electric energy RGB picture as the input of the convolutional neural network model, training the convolutional neural network model by using a training set, adjusting parameters, and testing the convolutional neural network model by using a test set.
As shown in fig. 2, the convolutional neural network model includes: the convolution layer comprises two convolution layers, two pooling layers, two full-connection layers and an output layer, wherein the convolution layers all use ReLu activation functions, 64 convolution kernels with the step length of 1 and the size of 5 multiplied by 5 are used in the first convolution layer, due to the fact that the size of the picture is not large, two circles of elements with the value of 0 are arranged on the periphery of a three-channel image matrix, the output result after convolution is the same as the size of the original input, the input data are added with offset after convolution, and the convolution layers are output as the input of the pooling layers through the ReLu activation functions. The pooling layers adopt kernels with the step size of 2 and the size of 3 x 3, and the step size is slightly smaller than the size of the kernels, so that the feature extraction is richer. For the second layer poolThe output matrix of the layer is unidimensionalized, rearranged into a column of vectors according to the sequence of rows, the first layer of full-connection layer uses a ReLu activation function, and then the weight is L-shaped2Regularization, the sixth full-connection layer carries out L on the weight2And (4) regularization, wherein the output layer sets 10 labels as output and adopts a softmax function as a classifier.
And acquiring a three-phase power disturbance signal to be detected, inputting the three-phase power disturbance signal into the trained convolutional neural network model, and taking the output of the convolutional neural network model as a three-phase power quality disturbance detection result.
The method has high efficiency, reduces the influence of artificial subjective factors, converts the acquired three-phase electric energy signal into RGB pictures, ensures that the converted signal characteristics are tighter, has less calculation amount, quickly and effectively processes three-phase electric energy quality disturbance data, intelligently extracts the disturbance signal characteristics by establishing a convolutional neural network model, realizes the accurate classification of the disturbance signals, can simultaneously detect a plurality of disturbance signals, and can detect the condition of faults between two phases or three phases.
Although the embodiments of the present invention have been described with reference to the accompanying drawings, it is not intended to limit the scope of the present invention, and it should be understood by those skilled in the art that various modifications and variations can be made without inventive efforts by those skilled in the art based on the technical solution of the present invention.
Claims (4)
1. A three-phase power quality disturbance detection method based on a convolutional neural network is characterized by comprising the following steps:
A) collecting three-phase electric energy disturbance signal data;
B) carrying out data preprocessing on the three-phase electric energy disturbance signal to obtain a three-phase electric energy RGB picture;
C) making a training set and a testing set;
D) constructing a convolutional neural network model, taking a three-phase electric energy RGB picture as the input of the convolutional neural network model, training the convolutional neural network model by using a training set, adjusting parameters, and testing the convolutional neural network model by using a test set;
E) collecting a three-phase power disturbance signal to be detected, inputting the three-phase power disturbance signal into a trained convolutional neural network model, and taking the output of the convolutional neural network model as a three-phase power quality disturbance detection result;
the step B) comprises the following steps:
B1) setting sampling frequency, sampling the three-phase electric energy disturbance signal in the period for h times to obtain 3 multiplied by h sampling data, and recording the data as a matrix
H-th sampling data representing the i-th perturbation signal;
B2) uniformly dividing the sampling data of each phase disturbance signal for n times, wherein each division comprises m sampling data to obtain an m multiplied by n two-dimensional matrix of the first phase disturbance signal
The m x n two-dimensional matrix of the second-phase disturbance signal is H2The m × n two-dimensional matrix of the third phase disturbance signal is H3；
B3) Setting a compression amplitude interval [ -X, X ], performing compression truncation processing on each element in the two-dimensional matrix, and obtaining a sampling value f (u) corresponding to the element after processing;
B4) performing linear transformation on the processed sampling values f (u) to obtain transformed values g (u) and obtain 3 transformed two-dimensional matrixes;
B5) respectively taking the 3 transformed two-dimensional matrixes as three single-channel matrixes of the RGB image to obtain a three-phase electric energy RGB image; the step C) comprises the following steps: generating w random samples from each three-phase electric energy disturbance signal, repeating the step B) on each random sample to obtain w three-phase electric energy RGB pictures, and dividing the w three-phase electric energy RGB pictures into a training set and a test set according to a proportion;
the convolutional neural network model in the step D) comprises the following steps: two layers of convolution layer, two layers of pooling layer, two layers of full connectionConnecting layers and output layers, wherein the convolutional layers use ReLu activation functions, 64 convolutional kernels with the step length of 1 and the size of 5 x 5 are adopted, kernels with the step length of 2 and the size of 3 x 3 are adopted in the pooling layers, the output matrix of the second pooling layer is subjected to one-dimensional treatment, the output matrixes are rearranged into a column of vectors according to the sequence of rows, and the weights of all the connecting layers are subjected to L treatment2And (4) regularization, wherein a ReLu activation function is used in the first full-connection layer, 10 labels are set as output in the output layer, and a softmax function is used as a classifier.
2. The method as claimed in claim 1, wherein the method comprises calculating the disturbance of the three-phase power quality
To f*(u) rounding to obtain a processed sampling value f (u), wherein u0The average value of the maximum values of the electric energy signals is u, and the amplitude of the disturbance electric energy signals is u; by calculating g (u) ═ 255(f (u) + X)/2X]The transformed values g (u) are obtained.
3. The method as claimed in claim 2, wherein a time period T is set, the three-phase power signal is monitored through the power grid in each time period T, and the maximum value of the three-phase power signal in the ith time period is recorded
Obtaining the average value u of the maximum value of the electric energy signal0，
4. The convolutional neural network-based three-phase power quality disturbance detection method as recited in claim 1, wherein the three-phase power disturbance signal comprises a three-phase voltage sag, a single-phase ground short fault, a two-phase ground short fault, an interphase short fault, a three-phase ground short fault, a three-phase interruption fault, a three-phase voltage flicker, a three-phase harmonic wave and a three-phase transient oscillation.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910832724.5A CN110703006B (en) | 2019-09-04 | 2019-09-04 | Three-phase power quality disturbance detection method based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910832724.5A CN110703006B (en) | 2019-09-04 | 2019-09-04 | Three-phase power quality disturbance detection method based on convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110703006A CN110703006A (en) | 2020-01-17 |
| CN110703006B true CN110703006B (en) | 2022-06-21 |
Family
ID=69194315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910832724.5A Active CN110703006B (en) | 2019-09-04 | 2019-09-04 | Three-phase power quality disturbance detection method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110703006B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112069883B (en) * | 2020-07-28 | 2024-03-29 | 浙江工业大学 | Deep learning signal classification method integrating one-dimensional two-dimensional convolutional neural network |
| CN115430815A (en) * | 2022-08-09 | 2022-12-06 | 衡阳镭目科技有限责任公司 | Crystallizer liquid level control method and device, electronic equipment and storage medium |
| CN117169657B (en) * | 2023-11-03 | 2024-01-12 | 山东和兑智能科技有限公司 | High-voltage cable state monitoring method and system based on artificial intelligence |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102445620A (en) * | 2011-11-21 | 2012-05-09 | 东北大学 | Transient power quality detection device and method for the same |
| CN108664950A (en) * | 2018-05-22 | 2018-10-16 | 天津大学 | A kind of electrical energy power quality disturbance identification and sorting technique based on deep learning |
| CN108830127A (en) * | 2018-03-22 | 2018-11-16 | 南京航空航天大学 | A kind of rotating machinery fault feature intelligent diagnostic method based on depth convolutional neural networks structure |
| CN108896296A (en) * | 2018-04-18 | 2018-11-27 | 北京信息科技大学 | A kind of wind turbine gearbox method for diagnosing faults based on convolutional neural networks |
| CN109359693A (en) * | 2018-10-24 | 2019-02-19 | 国网上海市电力公司 | A kind of Power Quality Disturbance Classification Method |
| CN109612708A (en) * | 2018-12-28 | 2019-04-12 | 东北大学 | Based on the power transformer on-line detecting system and method for improving convolutional neural networks |
| CN109635928A (en) * | 2018-12-06 | 2019-04-16 | 华北电力大学 | A kind of voltage sag reason recognition methods based on deep learning Model Fusion |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR2981751B1 (en) * | 2011-10-19 | 2020-04-17 | Schneider Electric Industries Sas | METHOD AND DEVICE FOR ANALYZING THE QUALITY OF ELECTRICAL ENERGY IN A THREE-PHASE ELECTRICAL NETWORK |
| US11734704B2 (en) * | 2018-02-17 | 2023-08-22 | Ei Electronics Llc | Devices, systems and methods for the collection of meter data in a common, globally accessible, group of servers, to provide simpler configuration, collection, viewing, and analysis of the meter data |
| CN109765333A (en) * | 2018-12-05 | 2019-05-17 | 国网辽宁省电力有限公司信息通信分公司 | A kind of Diagnosis Method of Transformer Faults based on GoogleNet model |
-
2019
- 2019-09-04 CN CN201910832724.5A patent/CN110703006B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102445620A (en) * | 2011-11-21 | 2012-05-09 | 东北大学 | Transient power quality detection device and method for the same |
| CN108830127A (en) * | 2018-03-22 | 2018-11-16 | 南京航空航天大学 | A kind of rotating machinery fault feature intelligent diagnostic method based on depth convolutional neural networks structure |
| CN108896296A (en) * | 2018-04-18 | 2018-11-27 | 北京信息科技大学 | A kind of wind turbine gearbox method for diagnosing faults based on convolutional neural networks |
| CN108664950A (en) * | 2018-05-22 | 2018-10-16 | 天津大学 | A kind of electrical energy power quality disturbance identification and sorting technique based on deep learning |
| CN109359693A (en) * | 2018-10-24 | 2019-02-19 | 国网上海市电力公司 | A kind of Power Quality Disturbance Classification Method |
| CN109635928A (en) * | 2018-12-06 | 2019-04-16 | 华北电力大学 | A kind of voltage sag reason recognition methods based on deep learning Model Fusion |
| CN109612708A (en) * | 2018-12-28 | 2019-04-12 | 东北大学 | Based on the power transformer on-line detecting system and method for improving convolutional neural networks |
Non-Patent Citations (7)
| Title |
|---|
| Classification_of_power_quality_events_using_deep_learning_on_event_images;Ebrahim Balouji等;《3rd International Conference on Pattern Recognition and Image Analysis》;20170420;全文 * |
| Learning_and_Transferring_Convolutional_Neural_Network_Knowledge_to_Ocean_Front_Recognition;Estanislau Lima等;《IEEE GEOSCIENCE AND REMOTE SENSING LETTERS》;20170330;全文 * |
| 基于CS_CNN的电能质量扰动数据分析研究;齐敏;《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》;20180915;全文 * |
| 基于卷积神经网络的电能质量扰动分类;王知芳等;《第九届电能质量研讨会论文集》;20180628;全文 * |
| 基于卷积神经网络的电能质量扰动分类;瞿合祚等;《武汉大学学报（工学版）》;20180628;第1-5节 * |
| 基于数字图像处理技术的暂态电能质量扰动分类;江辉等;《电力系统保护与控制》;20150701;全文 * |
| 基于相空间重构和支持向量机的电能扰动分类方法;李智勇等;《电力系统自动化》;20070310;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110703006A (en) | 2020-01-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Guo et al. | Deep-learning-based fault classification using Hilbert–Huang transform and convolutional neural network in power distribution systems | |
| CN110703006B (en) | Three-phase power quality disturbance detection method based on convolutional neural network | |
| CN107340456B (en) | Power distribution network operating condition intelligent identification Method based on multiple features analysis | |
| CN110956220B (en) | Non-invasive household appliance load identification method | |
| Reddy et al. | A DOST based approach for the condition monitoring of 11 kV distribution line insulators | |
| CN110441654B (en) | Power quality disturbance detection method for power distribution network based on improved EWT and CMPE | |
| WO2019056753A1 (en) | Dynamic equivalent modeling method for distributed photovoltaic power station cluster | |
| CN106548013B (en) | Utilize the voltage sag source identification method for improving incomplete S-transformation | |
| CN111080105A (en) | Transformer area user-to-user relationship identification method and system based on voltage time sequence data | |
| CN106250904B (en) | Electric energy disturbance analyzer based on improved S transformation and classification method | |
| CN115510913B (en) | Fault diagnosis method of H-bridge cascade inverter based on data driving | |
| CN111308260B (en) | Electric energy quality monitoring and electric appliance fault analysis system based on wavelet neural network and working method thereof | |
| CN105572501A (en) | Power quality disturbance identification method based on SST conversion and LS-SVM | |
| CN109946543B (en) | V-I pattern sample data set screening method for non-invasive load monitoring | |
| CN113158781B (en) | Lightning trip type identification method | |
| CN112444758B (en) | Intelligent power distribution network line fault diagnosis and classification evaluation method | |
| Gao et al. | Series arc fault diagnosis method of photovoltaic arrays based on GASF and improved DCGAN | |
| Dag et al. | Fault classification for power distribution systems via a combined wavelet-neural approach | |
| CN111277221A (en) | Photovoltaic fault diagnosis method and device | |
| CN110245617B (en) | Artificial intelligence analysis method based on transient recording waveform | |
| CN113238110B (en) | Power quality disturbance diagnosis method | |
| CN111539275B (en) | Electrical load classification method and system based on load characteristic visualization | |
| CN114545147A (en) | Voltage sag source positioning method based on deep learning in consideration of time-varying topology | |
| CN114355110A (en) | Fault current mode identification method based on convolutional neural network | |
| CN114252725A (en) | HHT and ResNet 18-based single-phase earth fault type comprehensive identification method |
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 |