CN111579939A - Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning - Google Patents
Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning Download PDFInfo
- Publication number
- CN111579939A CN111579939A CN202010330069.6A CN202010330069A CN111579939A CN 111579939 A CN111579939 A CN 111579939A CN 202010330069 A CN202010330069 A CN 202010330069A CN 111579939 A CN111579939 A CN 111579939A
- Authority
- CN
- China
- Prior art keywords
- voltage power
- partial discharge
- voltage
- time sequence
- power cable
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
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
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/04—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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a method for detecting a partial discharge phenomenon of a high-voltage power cable based on deep learning, which comprises the following steps: preprocessing the voltage time sequence signal data of the original high-voltage power cable to acquire the characteristics of the high-voltage power line signal, and skipping the step if the data preprocessing is not needed; converting the voltage time sequence signal of the preprocessed high-voltage power cable or the unprocessed signal into time sequence data input by a partial discharge phenomenon detection model; and training the training set to obtain a final partial discharge phenomenon detection model, judging the performance of the detection model according to the evaluation index, and then improving the detection model until a satisfactory-effect model is obtained. The invention can accurately judge whether the partial discharge phenomenon occurs in the voltage time sequence signal of a certain section of the high-voltage power line.
Description
Technical Field
The invention relates to the field of high-voltage power cables, in particular to a method for detecting partial discharge phenomenon of a high-voltage power cable based on deep learning.
Background
As an important medium for urban power transmission, high-voltage transmission lines play an important role in various fields such as national development and people's life. Under the new situation of rapid development of Chinese economy, the total amount of electricity used by people, enterprises, factories and the like is continuously increased, and the scale of an electric power system is more complicated and enlarged. In the huge circuit network, the high-voltage power cable distributes as if the whole skeleton of a human body, and bears the needs of daily life and production work of people.
In the early stage of power failure, a phenomenon called Partial Discharge (PD) often occurs in a high-voltage power line, which is equivalent to a discharge phenomenon generated after an insulation medium is accidentally punctured, and the partial discharge phenomenon can slowly damage an insulation layer of a high-voltage power cable distribution line, and if the insulation layer is left alone, the power failure or fire can be finally caused.
The partial discharge phenomenon is a concomitant product of a power cable system and is also a problem to be faced and solved by the power industry. With the development of partial discharge detection technology, there are some mature methods for detecting partial discharge information. However, these conventional methods often require close proximity to the high voltage power cable, which makes it complicated and time consuming to troubleshoot, particularly in mountains, remote areas, and in areas with somewhat complex terrain. These conventional partial discharge detection algorithms are based on various phenomena accompanying partial discharge, and represent the state and characteristics of partial discharge by physical quantities that can express the phenomena. Electrical pulses, electromagnetic radiation, ultrasound, light and some new products are generated during partial discharges and cause local overheating. Accordingly, various detection methods such as a pulse current detection method, a UHF method, an ultrasonic detection method, a photometric method, a chemical detection method, and an infrared detection method have appeared.
1) Pulse current detection method: the pulse current method is the most widely applied detection method at home and abroad, and forms corresponding international and domestic standards. The method is mainly characterized in that a transformer is equivalent to a capacitor, when partial discharge occurs, instantaneous voltage changes can be generated at two ends of the capacitor, the capacitor is led to a detection impedance through a coupling capacitor, pulse current can be obtained, the pulse current corresponds to the partial discharge, and parameters of the transformer during the partial discharge can be obtained after processing. But the method has low measuring frequency, cannot avoid air corona interference, is not suitable for on-line monitoring, and is the only standard detection method at present.
2) Ultrasonic detection method: the partial discharge often releases sound waves, and in the ultrasonic detection method, a sensor is used for receiving locally generated ultrasonic waves, so that the size and the position of the partial discharge are determined. It is difficult to quantify and it is not easy to distinguish between in-service plant interference signals.
3) A photometric method: the property of the optical fiber is changed mainly by using the sound wave compression, and the characteristics of the output information of the optical fiber are changed, so that the partial discharge signal is measured, but the technology is not mature.
4) Infrared detection method: the main principle of the method is that when partial discharge occurs, the corresponding part is overheated, if the temperature of the part to be detected is higher than the absolute temperature, radiation energy converted from heat energy is generated, and the radiation energy is also infrared rays, so that the detection purpose is achieved by utilizing the thermal imaging principle of the infrared detector. The method is suitable for detecting the overheating phenomenon of the external wiring end of the equipment and the like, and the internal condition of the equipment in operation is not easy to monitor.
5) Ultra high frequency detection technology (UHF): the ultrahigh frequency signal of the transformer during partial discharge is acquired through the sensor, so that the partial discharge is detected and positioned. The method has high detection frequency band, and can avoid corona interference; the intensity of discharge can be reflected, the sudden failure can be responded in time, and the online monitoring is suitable.
In recent years, as machine learning has been developed, deep learning methods have been advanced and developed. Due to the strong flexibility and self-adaptive capacity of the deep learning method, more and more applications are put into use in various fields. The deep learning method is also applied to time sequence classification, and as the availability of time sequence data is improved, a plurality of TSC (time sequence classification) algorithms are available.
However, for the task of detecting the partial discharge phenomenon, a deep learning method has not been applied to the field. In this field, there are problems of less samples, non-uniformity of positive and negative samples, and noise in data, which have a great influence on the task of classifying the voltage sequence. How to obtain a higher accuracy of the final classification under the influence of these problems remains a challenging problem.
Disclosure of Invention
The invention provides a method for detecting partial discharge phenomena of a high-voltage power cable based on deep learning, which extracts the characteristics, the characteristics and the partial discharge mode of the partial discharge phenomena hidden in a voltage time sequence signal through the deep learning method, thereby being capable of more accurately judging whether the partial discharge phenomena occur in the voltage time sequence signal of a certain section of a high-voltage power line, and the method is described in detail as follows:
a method for detecting partial discharge phenomena of a high-voltage power cable based on deep learning comprises the following steps:
preprocessing the voltage time sequence signal data of the original high-voltage power cable to acquire the characteristics of the high-voltage power line signal, and skipping the step if the data preprocessing is not needed;
converting the voltage time sequence signal of the preprocessed high-voltage power cable or the unprocessed signal into time sequence data input by a partial discharge phenomenon detection model;
and training the training set to obtain a final partial discharge phenomenon detection model, judging the performance of the detection model according to the evaluation index, and then improving the detection model until a satisfactory-effect model is obtained.
The preprocessing of the voltage time sequence signal data of the original high-voltage power cable specifically comprises the following steps:
dividing the time sequence voltage signal data of each phase limit into a plurality of sections, and solving the mean value mean, the standard deviation std, the mean plus-minus standard deviation mean +/-std and the amplitude max _ range of the time sequence voltage signal data;
respectively forming a plurality of groups of percent _ calc and a plurality of groups of relative _ percent after the average value is subtracted from the signal data of 0 percent, 1 percent, 25 percent, 50 percent, 75 percent, 99 percent and 100 percent, splicing the data into a one-dimensional array to be used as a preprocessing characteristic value of the segment, and then stacking a plurality of obtained one-dimensional arrays into a characteristic matrix;
and splicing the processed data of the three phases together as the characteristics of the corresponding high-voltage power line.
Furthermore, the partial discharge phenomenon detection model is formed by sequentially overlapping an input layer, two bidirectional LSTM layers, an attention layer and a full-connection layer.
The technical scheme provided by the invention has the beneficial effects that:
1. according to the invention, the characteristics and the partial discharge mode of the partial discharge phenomenon hidden in the voltage time sequence signal are extracted through a deep learning method, so that whether the partial discharge phenomenon occurs in the monitored voltage time sequence signal of a certain section of high-voltage power line can be accurately judged.
2. The method provided by the invention can simplify the fault detection of the high-voltage power line, so that the fault detection becomes more efficient, the section where the fault occurs can be determined, the safety of related workers is ensured, the partial discharge phenomenon can be timely found and maintained, the frequency of the fault occurrence of the high-voltage power line can be greatly reduced, and further unnecessary loss is avoided.
Drawings
FIG. 1 is a schematic diagram of voltage timing signal data without partial discharge;
FIG. 2 is a schematic diagram of voltage timing signal data with partial discharge;
FIG. 3 is a schematic diagram of a sample data set distribution;
FIG. 4 is a schematic diagram of a loss function variation during training;
FIG. 5 is a schematic view of a model structure;
FIG. 6 is a schematic diagram of a PR curve of a model obtained after training;
fig. 7 is a flowchart of a method for detecting partial discharge of a high-voltage power cable based on deep learning.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, embodiments of the present invention are described in further detail below.
Time series (Time series) is a set of data point sequences arranged according to Time occurrence sequence, and the Time interval of a set of Time series is a constant value. The time sequence classification task is to judge the classification of the time sequence by learning and extracting the characteristics of the time sequence. The method is to judge whether the local discharge phenomenon occurs in the time period by extracting the characteristics of the voltage time sequence signal.
In the existing industrial field, the data volume of the voltage time sequence signal of the standard high-voltage power cable is not large, and the problem of unbalance of the positive and negative samples exists; in addition, certain noise exists in a voltage signal of the high-voltage power circuit monitored by the equipment; in addition, the manually extracted features do not extract the time sequence information in the time sequence well. These reasons can make the classification effect on the voltage sequence less than satisfactory, thereby affecting the detection effect of the partial discharge phenomenon.
The invention provides a method for detecting partial discharge phenomenon in a high-voltage power line by using a deep learning method. The raw data set is first subjected to appropriate preprocessing operations depending on the size of the data set. In the process of power transmission, a high-voltage power line often uses three-phase alternating current, so that three-phase voltage time sequence signal data can be simultaneously monitored in the process of monitoring voltage signals of the same power transmission line. In the data preprocessing process, the most important is to integrate the voltage timing signal data of the three quadrants together, and the whole is used as the input of the deep learning method. The preprocessed data can achieve a final better effect after five times of cross validation iterative training and parameter fine tuning retraining. The deep learning method can effectively extract the characteristic information of the time dimension of the voltage time sequence signal data of the high-voltage power line, capture the relation among three phase limits and acquire the signal mode when the partial discharge phenomenon occurs, so that the method can save labor, improve the detection efficiency of the partial discharge phenomenon and eliminate potential hidden dangers of the voltage power line as soon as possible.
Example 1
A method for detecting partial discharge phenomena in a high-voltage power line by using a deep learning method comprises the following steps:
101: determining whether manual processing of original high-voltage power cable voltage time sequence signal data is needed;
for example: processing of normalization, data compression, data enhancement, and the like is performed.
The pretreatment method in step 101 specifically comprises the following steps:
for three-phase alternating current, firstly, processing signals of each phase: dividing the time sequence voltage signal data of each phase limit into a plurality of sections, then calculating the mean value mean, the standard deviation std, the mean plus minus standard deviation mean +/-std (representing that the dispersion degree of the single measurement standard deviation and the random error normal distribution curve are used as the standard to describe), the amplitude max _ range of the section, the array percent _ calc formed by signal data of 0%, 1%, 25%, 50%, 75%, 99% and 100% in the section and the array relative _ percent after the mean value is subtracted from the array percent _ calc, and splicing the array into a one-dimensional array as the preprocessing characteristic value of the section. And then stacking the obtained multiple one-dimensional arrays into a feature matrix.
And finally, splicing the processed data of the three phases together as the characteristics of the corresponding high-voltage power line, and also as the input of the model.
102: preprocessing the original high-voltage power cable voltage time sequence signal data by using the method determined in the step 101 to obtain the characteristics of the high-voltage power line signal, and skipping the step if the data preprocessing is not needed;
103: converting the voltage time sequence signal of the preprocessed high-voltage power cable or the unprocessed signal into time sequence data input by a partial discharge phenomenon detection model;
the partial discharge phenomenon detection model is mainly formed by sequentially overlapping an input layer, two-way LSTM (Long Short-Term Memory) layers, an attention layer and a full connection layer.
104: and training the training set to obtain a final partial discharge phenomenon detection model, judging the performance of the detection model according to the evaluation index, and then improving the detection model until a satisfactory-effect model is obtained.
The effect satisfaction degree of the model can be measured according to a set threshold, if the threshold is larger than the threshold, the effect is satisfactory, and if the threshold is smaller than the threshold, the effect is unsatisfactory.
Example 2
The scheme of embodiment 1 is further described below with reference to specific examples and fig. 1 to 4, and mainly includes four parts:
1) preprocessing the voltage time sequence signal data of the original high-voltage power cable; 2) the overall framework of the adopted model; 3) fine adjustment of parameters; 4) and analyzing a model prediction result.
The data set used in this example was collected from the high voltage power line by detecting some of the partial discharge related voltage timing signals from a new meter designed by ENET Centre of Ottotra technology university (VSB-STUDIO). 2904 high-voltage wire samples are contained in the data set, and as shown in fig. 1, the distribution of positive and negative samples in the data set is shown. Each sample contains 3 phases of data because the line to be measured is a high-voltage ac line, and the data is collected every 0.02s, and each sample contains 800000 time points of data for a single phase, where fig. 1 shows voltage time sequence signal data without partial discharge phenomenon, and fig. 2 shows voltage time sequence signal data with partial discharge phenomenon. Since in a practical application scenario, this data set represents the data that is most easily collected over the high voltage power lines, the current data set is used as the main data for the entire algorithm training.
Firstly, preprocessing the voltage time sequence signal data of the original high-voltage power cable
Because the original high-voltage power cable voltage time sequence signal data adopted in the embodiment is too large, the training set is about 10g, if all the data are loaded into the memory, certain pressure is caused on equipment, and the original high-voltage power cable voltage time sequence signal data are directly used as input, so that the training time of the model can be increased. In order to reduce training time and pressure on equipment memory, and to combine voltage timing signal data of three phase limits in a high-voltage power line, the original high-voltage power cable voltage timing signal data is preprocessed before model training.
The signal of each phase is processed first: the 800000 data of each phase limit is divided into 160 segments, each segment contains 5000 time point data, namely data within the time period of 100 seconds is taken, then the average mean, standard deviation std, average plus-minus standard deviation mean + -std, amplitude max _ range of the time period, a group of percent _ calc composed of 0%, 1%, 25%, 50%, 75%, 99%, 100% of the original high voltage power cable voltage timing signal data in the time period, and a group of percent _ percent after the average value is subtracted from the percent _ calc are obtained, and the (mean, std, mean + -std, max _ range, percent _ calc, relative _ percent) are spliced into a one-dimensional array to serve as the preprocessing characteristic value of the time period. The resulting 160 one-dimensional arrays were then stacked into a 160 x 19 matrix.
And finally, splicing the processed data of the three phases together as the characteristics of the corresponding high-voltage power line as the input of the model.
Second, the overall framework of the model adopted
As shown in fig. 5, the model is mainly composed of an input layer, a bidirectional LSTM layer, an attention layer and a full link layer. In this model, the bidirectional LSTM layer is used to better extract information of the power cable voltage timing signal data in the time dimension, and two bidirectional LSTM layers are used to extract higher-level timing information. The attention layer is added to make the model focus more on the information related to the partial discharge. In the model, the activation function of the dense _1 layer is Relu, the activation function of the dense _2 layer is Sigmoid, the loss function adopted by the model is a binary cross entropy (binary _ cross entropy) loss function, and the optimization algorithm is an Adam optimization algorithm. During the training process, the batch _ size is set to 128 and the epochs to 50, and five cross-validation methods are used for training. The predicted value corresponding to the input high-voltage power line voltage timing signal is finally output through the model, and the predicted value is the probability of the partial discharge phenomenon.
Third, model result prediction analysis
On the time series classification task, generally, accuracy, recall rate, AP, and mapp are taken as evaluation indexes, and the calculation method is described as follows:
true Positive (TP): predicting positive class as a positive class number
True Negative, TN: predicting negative classes as negative class numbers
False Positive (FP): predicting negative class as a false positive class number (Type I error)
False Negative (FN): predict positive class as negative class number → missing report (Type II error)
The accuracy is as follows:
the recall ratio is as follows:
the AP and the like divide each class of training samples into a plurality of blocks, average precision of samples in all the blocks, and measure the quality of the trained model on each class; and the mAP is measured, the model obtained through training is good or bad in all categories, and the average value of all the APs is taken for the obtained AP, so that the mAP can be obtained.
As in fig. 4, the variation of the loss function during five cross-validation training sessions is shown. FIG. 6 shows PR graphs of the model of the proposed method of the present invention, showing recall, precision corresponding to the classification result. The average Accuracy (AP) of the method can be known to reach 0.72091, and the partial discharge phenomenon can be effectively detected through the voltage time sequence signal data of the high-voltage power cable.
Experiments show that the model can effectively acquire the time sequence information hidden in the power signal, and excavate a signal mode during normal power transmission and a signal mode during partial discharge, so that the future detected signal data can be judged. Meanwhile, the detection method provided by the invention has better detection capability under the condition of polar imbalance of the positive sample and the negative sample.
In the embodiment of the present invention, except for the specific description of the model of each device, the model of other devices is not limited, as long as the device can perform the above functions.
Those skilled in the art will appreciate that the drawings are only schematic illustrations of preferred embodiments, and the above-described embodiments of the present invention are merely provided for description and do not represent the merits of the embodiments.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents, improvements and the like that fall within the spirit and principle of the present invention are intended to be included therein.
Claims (3)
1. A method for detecting partial discharge phenomena of a high-voltage power cable based on deep learning is characterized by comprising the following steps:
preprocessing the voltage time sequence signal data of the original high-voltage power cable to acquire the characteristics of the high-voltage power line signal, and skipping the step if the data preprocessing is not needed;
converting the voltage time sequence signal of the preprocessed high-voltage power cable or the unprocessed signal into time sequence data input by a partial discharge phenomenon detection model;
and training the training set to obtain a final partial discharge phenomenon detection model, judging the performance of the detection model according to the evaluation index, and then improving the detection model until a satisfactory-effect model is obtained.
2. The method for detecting the partial discharge phenomenon of the high-voltage power cable based on the deep learning as claimed in claim 1, wherein the preprocessing of the voltage time sequence signal data of the original high-voltage power cable is specifically as follows:
dividing the time sequence voltage signal data of each phase limit into a plurality of sections, and solving the mean value mean, the standard deviation std, the mean plus-minus standard deviation mean +/-std and the amplitude max _ range of the time sequence voltage signal data;
respectively forming a plurality of groups of percent _ calc and a plurality of groups of relative _ percent after the average value is subtracted from the signal data of 0 percent, 1 percent, 25 percent, 50 percent, 75 percent, 99 percent and 100 percent, splicing the data into a one-dimensional array to be used as a preprocessing characteristic value of the segment, and then stacking a plurality of obtained one-dimensional arrays into a characteristic matrix;
and splicing the processed data of the three phases together as the characteristics of the corresponding high-voltage power line.
3. The method for detecting the partial discharge phenomenon of the high-voltage power cable based on the deep learning of claim 1, wherein the partial discharge phenomenon detection model is formed by sequentially overlapping an input layer, two bidirectional LSTM layers, an attention layer and a full connection layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010330069.6A CN111579939A (en) | 2020-04-23 | 2020-04-23 | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010330069.6A CN111579939A (en) | 2020-04-23 | 2020-04-23 | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN111579939A true CN111579939A (en) | 2020-08-25 |
Family
ID=72125443
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010330069.6A Pending CN111579939A (en) | 2020-04-23 | 2020-04-23 | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111579939A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112067960A (en) * | 2020-09-30 | 2020-12-11 | 广东电网有限责任公司中山供电局 | High-voltage cable partial discharge mode identification method based on PSO and bidirectional LSTM |
| CN112183745A (en) * | 2020-09-25 | 2021-01-05 | 广东电网有限责任公司中山供电局 | High-voltage cable partial discharge mode identification method based on particle swarm algorithm and DBN |
| CN112287789A (en) * | 2020-10-20 | 2021-01-29 | 国网四川省电力公司电力科学研究院 | GIS ultrahigh frequency method partial discharge map identification method based on deep learning |
| CN115128410A (en) * | 2022-06-27 | 2022-09-30 | 国网上海市电力公司 | TPA-LSTM-based direct current cable partial discharge fault mode identification method |
| CN116482501A (en) * | 2023-06-14 | 2023-07-25 | 西安邮电大学 | Ultrasonic imaging method and system for partial discharge detection of high-voltage cable connector |
| CN117890740A (en) * | 2024-03-14 | 2024-04-16 | 云南电投绿能科技有限公司 | Partial discharge positioning method, device and equipment for power station cable and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104777410A (en) * | 2015-04-22 | 2015-07-15 | 东北电力大学 | Partial discharge pattern identification method for crosslinked polyethylene cable |
| CN109766930A (en) * | 2018-12-24 | 2019-05-17 | 太原理工大学 | A kind of method for predicting residual useful life of the mine machinery equipment based on DCNN model |
| CN109784473A (en) * | 2018-12-13 | 2019-05-21 | 天津大学 | A kind of short-term wind power prediction method based on Dual Clocking feature learning |
| CN110208660A (en) * | 2019-06-05 | 2019-09-06 | 国网江苏省电力有限公司电力科学研究院 | A kind of training method and device for power equipment shelf depreciation defect diagonsis |
| CN110543557A (en) * | 2019-09-06 | 2019-12-06 | 北京工业大学 | construction method of medical intelligent question-answering system based on attention mechanism |
| CN110826527A (en) * | 2019-11-20 | 2020-02-21 | 南京邮电大学 | Electroencephalogram negative emotion recognition method and system based on aggressive behavior prediction |
| CN110837028A (en) * | 2019-09-27 | 2020-02-25 | 中国船舶重工集团公司第七一九研究所 | Method for rapidly identifying partial discharge mode |
| CN110927535A (en) * | 2019-11-12 | 2020-03-27 | 国网天津市电力公司电力科学研究院 | Power equipment partial discharge severity evaluation method based on extreme learning machine |
-
2020
- 2020-04-23 CN CN202010330069.6A patent/CN111579939A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104777410A (en) * | 2015-04-22 | 2015-07-15 | 东北电力大学 | Partial discharge pattern identification method for crosslinked polyethylene cable |
| CN109784473A (en) * | 2018-12-13 | 2019-05-21 | 天津大学 | A kind of short-term wind power prediction method based on Dual Clocking feature learning |
| CN109766930A (en) * | 2018-12-24 | 2019-05-17 | 太原理工大学 | A kind of method for predicting residual useful life of the mine machinery equipment based on DCNN model |
| CN110208660A (en) * | 2019-06-05 | 2019-09-06 | 国网江苏省电力有限公司电力科学研究院 | A kind of training method and device for power equipment shelf depreciation defect diagonsis |
| CN110543557A (en) * | 2019-09-06 | 2019-12-06 | 北京工业大学 | construction method of medical intelligent question-answering system based on attention mechanism |
| CN110837028A (en) * | 2019-09-27 | 2020-02-25 | 中国船舶重工集团公司第七一九研究所 | Method for rapidly identifying partial discharge mode |
| CN110927535A (en) * | 2019-11-12 | 2020-03-27 | 国网天津市电力公司电力科学研究院 | Power equipment partial discharge severity evaluation method based on extreme learning machine |
| CN110826527A (en) * | 2019-11-20 | 2020-02-21 | 南京邮电大学 | Electroencephalogram negative emotion recognition method and system based on aggressive behavior prediction |
Non-Patent Citations (5)
| Title |
|---|
| MING DONG ETC.: "Partial discharge detection on aerial covered conductors using time-series decomposition and long short-term memory network", 《ELECTRIC POWER SYSTEMS RESEARCH》 * |
| 吕松南: "面向绝缘架空导线的局部放电故障智能检测方法研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 * |
| 周念成等: "深度学习在智能电网中的应用现状分析与展望", 《电力系统自动化》 * |
| 周秀等: "基于深度学习的变压器局部放电模式识别研究", 《高压电器》 * |
| 宋辉等: "运行条件下GIS局部放电严重程度评估方法", 《中国电机工程学报》 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112183745A (en) * | 2020-09-25 | 2021-01-05 | 广东电网有限责任公司中山供电局 | High-voltage cable partial discharge mode identification method based on particle swarm algorithm and DBN |
| CN112067960A (en) * | 2020-09-30 | 2020-12-11 | 广东电网有限责任公司中山供电局 | High-voltage cable partial discharge mode identification method based on PSO and bidirectional LSTM |
| CN112287789A (en) * | 2020-10-20 | 2021-01-29 | 国网四川省电力公司电力科学研究院 | GIS ultrahigh frequency method partial discharge map identification method based on deep learning |
| CN115128410A (en) * | 2022-06-27 | 2022-09-30 | 国网上海市电力公司 | TPA-LSTM-based direct current cable partial discharge fault mode identification method |
| CN115128410B (en) * | 2022-06-27 | 2024-05-14 | 国网上海市电力公司 | Direct-current cable partial discharge fault mode identification method based on TPA-LSTM |
| CN116482501A (en) * | 2023-06-14 | 2023-07-25 | 西安邮电大学 | Ultrasonic imaging method and system for partial discharge detection of high-voltage cable connector |
| CN116482501B (en) * | 2023-06-14 | 2023-08-29 | 西安邮电大学 | Ultrasonic imaging method and system for partial discharge detection of high-voltage cable connector |
| CN117890740A (en) * | 2024-03-14 | 2024-04-16 | 云南电投绿能科技有限公司 | Partial discharge positioning method, device and equipment for power station cable and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111579939A (en) | Method for detecting partial discharge phenomenon of high-voltage power cable based on deep learning | |
| CN103645425B (en) | High-voltage cable insulation defect partial discharge on-line monitoring diagnosis method | |
| CN112147444B (en) | Power transformer working state monitoring method and system | |
| CN110261746B (en) | Cable defect detection method based on periodic attenuation characteristics of oscillating wave voltage | |
| CN103576059A (en) | Integrated fault diagnosis method and system for turn-to-turn discharging of transformer | |
| CN111650451A (en) | Converter transformer fault reason identification method and system | |
| CN112880750A (en) | Transformer multidimensional comprehensive online monitoring intelligent diagnosis system | |
| CN115542099B (en) | Online GIS partial discharge detection method and device | |
| CN113780401A (en) | Composite insulator fault detection method and system based on principal component analysis method | |
| CN114169237A (en) | Power cable joint temperature abnormity early warning method combining EEMD-LSTM and isolated forest algorithm | |
| Ding et al. | Gear fault diagnosis based on VMD sample entropy and discrete hopfield neural network | |
| CN109459662B (en) | High-voltage cable defect state evaluation system | |
| Mishra et al. | Diagnostics analysis of partial discharge events of the power cables at various voltage levels using ramping behavior analysis method | |
| Liu et al. | Partial Discharge Detection Method for Distribution Network Based on Feature Engineering | |
| Oh et al. | A study on improvement of resource efficiency for IoT-based pipe leak detection | |
| US20230341477A1 (en) | Series arc detection firmware | |
| CN112949497B (en) | GIS partial discharge pattern recognition method based on improved generalized regression neural network | |
| CN113722659B (en) | Fault mode-based substation main equipment diagnosis method and system | |
| Li et al. | Distribution grid fault diagnosis based on fault indicator and random matrix | |
| Wang et al. | A motor fault diagnosis method based on industrial wireless sensor networks | |
| Das et al. | Time-frequency representation aided deep transfer learning approach for localization and identification of single and multiple partial discharge events | |
| CN113551157B (en) | Oil-gas data acquisition system and method based on sensor network | |
| Zhao et al. | Research on Intelligent Diagnosis of Switch Partial Discharge Based on Power Internet of Things | |
| Lu et al. | Design of Power Transformer Fault Diagnosis System Based on Wavelet Transform | |
| CN117233551A (en) | Ring main unit partial discharge fault early warning and identifying method based on multi-element combined characteristics |
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 | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20200825 |
|
| WD01 | Invention patent application deemed withdrawn after publication |