CN115184744B - GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN - Google Patents
GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN Download PDFInfo
- Publication number
- CN115184744B CN115184744B CN202210733546.2A CN202210733546A CN115184744B CN 115184744 B CN115184744 B CN 115184744B CN 202210733546 A CN202210733546 A CN 202210733546A CN 115184744 B CN115184744 B CN 115184744B
- Authority
- CN
- China
- Prior art keywords
- module
- gis
- discharge signal
- detection
- discharge
- 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
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
- G01R31/1227—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 of components, parts or materials
- G01R31/1254—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 of components, parts or materials of gas-insulated power appliances or vacuum gaps
-
- 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
- G01R31/1209—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 using acoustic measurements
-
- 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
- G01R31/1218—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 using optical methods; using charged particle, e.g. electron, beams or X-rays
-
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
-
- 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
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Databases & Information Systems (AREA)
- Medical Informatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Acoustics & Sound (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a device and a method for detecting a GIS ultrahigh frequency discharge signal based on a fast-RCNN, wherein the device comprises the following steps: the system comprises a data acquisition module, a preprocessing module, a transmission module, a cloud server module and an output module; the data acquisition module acquires a GIS partial discharge signal and generates a PRPD map of the partial discharge signal, the preprocessing module receives the PRPD map, performs preprocessing to obtain a gray image, constructs a target data set, the transmission module receives the target data set, performs preprocessing and sends processed data to the cloud server, the cloud server module deploys a trained detection model, detects data to be detected and stores a visual result, and the output module receives the data and accesses the cloud detection service, observes and records the detected result. The invention can accurately detect the discharge types contained in the PRPD patterns of the mixed discharge signals and the single discharge signals in real time, also realizes more effective maintenance and real-time monitoring of the GIS, and improves the accuracy and reliability of GIS partial discharge detection.
Description
Technical Field
The invention relates to the field of GIS ultrahigh frequency partial discharge signal detection, in particular to a device and a method for detecting a GIS ultrahigh frequency discharge signal based on fast-RCNN.
Background
GIS (Gas Insulated Switchgear, gas insulated switch) has advantages such as area is little, anti-pollution ability is strong, has obtained extensive application in electric power system. However, during the manufacturing and assembly process of the GIS, small defects, such as metal particles, insulating air gaps, etc., are left in the GIS due to the technical and design problems, and these defects easily cause partial discharge of the GIS. Partial discharge is a pulse discharge form of discharge or breakdown of an insulator in an electric field caused by electric field deformity and field intensity concentration in a local range, and is accompanied by phenomena such as electromagnetic radiation, electric pulse, ultrasonic wave, light and the like, if the phenomenon cannot be timely detected and effectively controlled and processed, insulation faults or power system faults can be further caused.
Most of the existing methods for identifying GIS partial discharge signals can only identify single discharge signals, and can not detect mixed discharge signals of multiple discharge signals respectively, and the traditional detection mode is low in efficiency, so that a detector is required to spend a great deal of effort to observe and count.
Disclosure of Invention
The invention aims to provide a device and a method for detecting GIS ultrahigh frequency discharge signals based on fast-RCNN, which are used for solving the problems that mixed discharge signals of various discharge signals cannot be detected respectively and the efficiency of the traditional detection mode is very low.
The invention solves the technical problems by adopting the following technical scheme:
a GIS ultrahigh frequency discharge signal detection device based on Faster-RCNN comprises: the system comprises a data acquisition module, a preprocessing module, a transmission module, a cloud server module and an output module; the data acquisition module acquires a partial discharge signal of the GIS and generates a PRPD map of the partial discharge signal, the preprocessing module preprocesses the PRPD map, outputs a gray level image, constructs a target data set, the transmission module receives the target data set and sends the target data set to the cloud server, the cloud server module receives the data of the transmission module, deploys a trained detection model, detects data to be detected and stores a visual result, the cloud server module transmits the visual result to the output module, and the output module transmits the visual result to an upper computer or intelligent terminal device.
Preferably, a back-end service of the cloud server module is connected with a front-end page through a Websocket, wherein the back-end service is built through a flash frame, and the front-end page is built through a Vue frame.
Preferably, the backend service includes: and the module is used for receiving the image module to be detected, detecting the image module to be detected and outputting and storing the detection result.
Preferably, the front-end service includes: the visual detection result module, the visual history result module and the control and management module.
A GIS ultrahigh frequency discharge signal detection method based on Faster-RCNN comprises the following specific steps:
step S1: acquiring a GIS ultrahigh frequency partial discharge signal and generating a PRPD map;
step S2: preprocessing the PRPD map, constructing a target data set, and obtaining a corresponding gray level image;
step S3: setting a target detection model as a network based on a fast-RCNN, and training a target data set;
step S4: and acquiring a trained target detection model, detecting data to be detected, identifying the types of all discharge signals in the PRPD map, and further judging according to the characteristics of the discharge signals.
Preferably, the target data set includes: noise signals, tip discharge signals, floating discharge signals, solid insulation discharge signals, and particle discharge signals.
Preferably, in the step S3, the target detection model based on the fast-RCNN includes: the device comprises a gray image feature extraction module, a region candidate network module, a region of interest pooling module and a classification model construction module; the gray image feature extraction module extracts and trains gray image features through a convolutional neural network, the region candidate network module screens reference frames and predicts labels and coordinate offsets of candidate frames, the region-of-interest pooling module receives feature images and candidate frames, extracts feature candidate frames and transmits the feature candidate frames to the classification model component module, and the classification model component module acquires categories of the candidate frames and acquires final positions of detection frames.
Preferably, the region of interest pooling module extracts a feature candidate frame to obtain a feature map with a size of 7×7×512.
Preferably, the classification model component module obtains the candidate frame category through joint training of the detection classification probability and the detection frame regression.
Preferably, in step S4, a data acquisition module is started, after the acquired PRPD map passes through the preprocessing module, the transmission module is used to send the gray level map to be detected to the cloud server for detection and judgment, and finally, the upper computer or the intelligent terminal device is used to access the Web service of the cloud server, so that the detection result is observed in real time.
The invention not only can accurately and real-timely detect the corresponding area and type of each signal source in the PRPD map of the mixed discharge signal, but also can detect and identify the discharge type contained in the PRPD map of the single discharge signal, thereby realizing more effective partial discharge monitoring on the GIS and improving the accuracy and reliability of the partial discharge detection of the GIS.
Drawings
FIG. 1 is a schematic flow chart of the detection method of the present invention;
FIG. 2 is a schematic diagram of the structure of the detecting device of the present invention;
FIG. 3 is a PRPD pattern normalization schematic of the present invention;
in the figure: (a) - (l) is the gray scale graphic intent of the PRPD pattern normalized of the discharge signal and the noise signal, respectively;
FIG. 4 is a schematic diagram of labeling a target dataset and generating a json file containing labeling information;
in the figure: (a) - (f) is the discharge signal and noise signal data annotation schematic drawings respectively, and (g) - (h) are annotation information schematic drawings;
FIG. 5 is a schematic diagram of a target detection network based on the Faster-RCNN of the present invention;
FIG. 6 is a schematic diagram of a convolutional neural network based on VGG-16 of the present invention.
Detailed Description
The technical scheme of the invention is further described below with reference to the embodiment and the attached drawings.
Example 1
The present embodiment provides a device for detecting a GIS ultrahigh frequency discharge signal based on a fast-RCNN (fast region candidate method and a region-based convolutional neural network), where the fast RCNN is a typical representation in a two-step target detection model, and includes: the system comprises a data acquisition module, a preprocessing module, a transmission module, a cloud server module and an output module; wherein: the data acquisition module acquires a GIS partial discharge signal and generates a PRPD map of the partial discharge signal, the preprocessing module preprocesses the PRPD map to obtain a gray image and constructs a target data set, the transmission module receives the target data set and sends the target data set to the cloud server, the cloud server module receives the data of the transmission module, deploys a trained detection model, detects data to be detected and stores a visual result, the cloud server module transmits the visual result to the output module, and the output module transmits the visual result to an upper computer or intelligent terminal equipment for a user to observe and record the detected result.
In a further implementation manner of this embodiment, the connection between the back-end service and the front-end page of the cloud server module through Websocket is a protocol for performing full duplex communication on a single TCP connection, websocket allows the server to actively push data to the client, so that data exchange between the client and the server becomes simpler, in Websocket API, the browser and the server only need to complete one handshake, and can directly create persistent connection between the two, and perform bidirectional data transmission, where the back-end service is built through a flash framework (a lightweight Web application framework written by Python, with a core structure that is simple and has very strong expandability and compatibility), and the front-end page is built through a Vue framework (a progressive framework for building a user interface, with advantages of lightweight, bidirectional data binding and faster running speed).
Further implementation manner of the embodiment, the back-end service includes: and the module is used for receiving the image module to be detected, detecting the image module to be detected and outputting and storing the detection result.
Further implementation manner of the embodiment, the front-end service includes: the visual detection result module, the visual history result module and the control and management module.
Example 2
The embodiment provides a GIS ultrahigh frequency discharge signal detection method based on fast-RCNN, which comprises the following specific steps:
step S1: acquiring a GIS ultrahigh frequency partial discharge signal and generating a PRPD map;
step S2: preprocessing the PRPD map, constructing a target data set, and obtaining a corresponding gray level image;
step S3: setting a target detection model as a network based on a fast-RCNN, and training a target data set;
step S4: and acquiring a trained target detection model, detecting data to be detected, identifying the types of all discharge signals in the PRPD map, and further judging according to the characteristics of the discharge signals.
Further implementations of the present embodiments include the target data set including: noise signals, tip discharge signals, floating discharge signals, solid insulation discharge signals, and particle discharge signals.
The PRPD pattern of the partial discharge signal generated in step S1 has the format: the abscissa represents the phase 0-360 degrees, the ordinate represents the magnitude of the signal amplitude, and the magnitude of each point is the number of pulses accumulated at that phase.
In a further embodiment of the present example, in step S2, the gray scale is quantized to 600 parts on the abscissa, 800 parts on the ordinate, and the pixel gray scale value range of the gray scale is [0,255].
In a further implementation manner of the present embodiment, in step S3, the target detection model based on the fast-RCNN includes: the device comprises a gray image feature extraction module, a region candidate network module, a region of interest pooling module and a classification model construction module;
as shown in fig. 6, the gray image feature extraction module extracts features of an input gray image by using a convolutional neural network based on VGG-16, wherein the convolutional neural network convolutional layers comprise 13 convolutional layers+13 relu (Rectified Linear Unit, modified linear units) activation function layers+4 pooling layers, and training is performed through the convolutional neural network to obtain a feature map;
all convolution (conv) layers are: kernel_size=3, coding=1, stride=1
All pooling (pooling) layers are: kernel_size=2, coding=0, stride=2
The convolution is performed with the processing of pad=1 (filling 1 layer of edge pixels), and then the convolution is performed with 3x3 convolution and then output, so that the convolution layer in convolution layers does not change the size of the input matrix and the output matrix. Only the feature extraction layer changes the output length and width into 1/2 of the input, and a matrix with the size of 600 x 800 x 1 is changed into 38 x 50 x 512 through convolution layers fixation;
the regional candidate network layer (RPN layer) uses a 3x3 sliding window (slide window) on the feature map extracted by the convolution layer to traverse the whole feature map, a prediction frame cluster and a scale are obtained through a K-means clustering algorithm according to a data set, three aspect ratios are obtained, 9 prior frames (anchors) are generated by the aspect ratio of the three prior frames at the center of each window in the traversing process, 256 anchors are selected from 50 x 38 x 9=17100 candidate anchors, one branch is used for predicting whether a feature map (proposal) label is a foreground or a background, the other branch is used for predicting coordinate bias of the feature map, the output channel number 9*2 for classification is carried out, and the convolution kernel channel number for regression is 9 x 4;
the interest region pooling module (ROI layer) collects the input feature images and feature frames, extracts feature image candidate frames after integrating the information to obtain 7 x 512 feature images with fixed sizes, and then sends the feature images and feature frames to a subsequent full-connection layer to judge target types;
the classification model construction module performs classification and regression model construction, and combines training on classification probability and frame regression (Bounding box regression) by using Softmax Loss (detection classification probability) and Smooth L1 Loss (detection frame regression), so that the classification of the feature frame is obtained, a Loss function is calculated, counter-propagation is performed, and the weight is modified, so that the final accurate position of the detection frame is obtained;
the final loss function formula is:
p i :Anchor[i]is a predictive classification probability of (2);
t i :Anchor[i]the parameterized coordinates of the predicted binding Box (parameterized coordinates);
Anchor[i]parameterized coordinates of a binding Box of the group Truth;
N cls :mini-batch size;
N reg : number of Anchor locations;
r is a Smooth L1 function;
indicating that the binding Box is only returned when the sample is positive;
in a further implementation manner of the embodiment, in step S3, a labelme image labeling tool is used to label the target data set; for each gray level image in a sample, marking a [ phase, amplitude ] area of each partial discharge signal and noise signal in the map and the type of the signal respectively in the form of a rectangular frame; for the case that two clusters of 180-degree phase deviation clusters exist in a spectrum of single discharge signals such as suspension discharge, solid insulation discharge, particle discharge and the like, the two clusters are respectively marked by 2 rectangular frames.
In a further implementation manner of this embodiment, in step S3, the noted data set includes: training set, validation set and test set, wherein: the training set, the verification set and the test set are deployed to the cloud server according to the proportion of 6:2:2.
In a further implementation of this example, the Faster RCNN adopts an alternate optimization training, which is divided into 4 steps:
1. training the RPN network on a pre-trained model;
2. training a Fast-RCNN network for the first time by using the trained RPN;
3. initializing an RPN network by using the model trained in the second step, and training the RPN network for the second time;
4. step 3, collecting proposals by the trained RPN network, and training the Fast-RCNN network for the second time;
in a further implementation manner of this embodiment, in step S4, a data acquisition module is started, after the acquired PRPD map passes through a preprocessing module, a transmission module is used to send a gray level map to be detected to a cloud server for detection and judgment, and finally, an upper computer or an intelligent terminal device is used to access Web services of the cloud server, so that the detection result is observed in real time.
The judging method comprises the following steps: and analyzing and confirming the targets detected in the gray level diagram corresponding to the map to be detected one by one. For 3 kinds of targets of suspension discharge, solid insulation discharge and particle discharge, in a region (corresponding to 300 pixels horizontally shifted and 100 pixels longitudinally shifted on a gray level diagram) which is 180 degrees away from the phase position of the target and has an amplitude within 10dB, if the similar targets exist, the targets are paired and then used as the same detected partial discharge signal; if the similar object does not exist, discarding the object.
The sequence of the above embodiments is only for convenience of description, and does not represent the advantages and disadvantages of the embodiments.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and are not limiting; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for detecting a GIS ultrahigh frequency discharge signal based on Faster-RCNN is characterized by comprising the following specific steps:
step S1: acquiring a GIS ultrahigh frequency partial discharge signal and generating a PRPD map;
step S2: preprocessing the PRPD map, constructing a target data set, and obtaining a corresponding gray level image;
step S3: setting a target detection model as a network based on a fast-RCNN, and training a target data set;
step S4: acquiring a trained target detection model, detecting data to be detected, identifying the types of all discharge signals in the PRPD map, and further judging according to the characteristics of the discharge signals;
in the step S2, the abscissa of the gray scale map is quantized to 600 parts, the ordinate is quantized to 800 parts, and the pixel gray scale value range of the gray scale map is [0,255];
in the step S3, labeling the target data set by using a labelme image labeling tool; for each gray level image in a sample, marking a [ phase, amplitude ] area of each partial discharge signal and noise signal in the map and the type of the signal respectively in the form of a rectangular frame; for the situation that two clusters of single discharge signals of suspension discharge, solid insulation discharge and particle discharge have 180-degree phase deviation in a map, marking the two clusters with 2 rectangular frames respectively;
in the step S3, the noted data set includes: training set, validation set and test set, wherein: the training set, the verification set and the test set are deployed into the cloud server according to the proportion of 6:2:2;
the Faster RCNN adopts alternate optimization training, and comprises 4 steps:
step 1, training an RPN network on a pre-trained model;
step 2, training the Fast-RCNN network for the first time by utilizing the trained RPN;
step 3, initializing the RPN network by using the model trained in the second step, and training the RPN network for the second time;
and step 4, collecting proposals by using the trained RPN network again in the step 3, and training the Fast-RCNN network for the second time.
2. The GIS uhf discharge signal detection method of claim 1, wherein the target data set comprises: noise signals, tip discharge signals, floating discharge signals, solid insulation discharge signals, and particle discharge signals.
3. The method for detecting a GIS uhf discharge signal according to claim 1, wherein in the step S3, the fast-RCNN-based target detection model includes: the device comprises a gray image feature extraction module, a region candidate network module, a region of interest pooling module and a classification model construction module; the gray image feature extraction module extracts and trains gray image features through a convolutional neural network, the region candidate network module screens reference frames and predicts labels and coordinate offsets of candidate frames, the region-of-interest pooling module receives feature images and candidate frames, extracts feature candidate frames and transmits the feature candidate frames to the classification model construction module, and the classification model construction module acquires categories of the candidate frames and acquires final positions of detection frames.
4. The method for detecting a GIS uhf discharge signal according to claim 3, wherein the region of interest pooling module extracts feature candidate frames to obtain a feature map with a size of 7×7×512.
5. The method for detecting the ultrahigh frequency discharge signal of the GIS according to claim 3, wherein the classification model construction module obtains the candidate frame class through joint training of the detection classification probability and the detection frame regression.
6. The method for detecting the ultrahigh frequency discharge signal of the GIS according to claim 1, wherein in the step S4, a data acquisition module is started, the acquired PRPD pattern is subjected to a preprocessing module, a transmission module is used for transmitting a gray level diagram to be detected to a cloud server for detection and judgment, and finally an upper computer or intelligent terminal device is used for accessing Web service of the cloud server to observe the detection result in real time.
7. The GIS ultrahigh frequency discharge signal detection device is used for realizing the GIS ultrahigh frequency discharge signal detection method based on fast-RCNN as claimed in claim 1, and is characterized by comprising the following steps: the system comprises a data acquisition module, a preprocessing module, a transmission module, a cloud server module and an output module; the data acquisition module acquires a partial discharge signal of the GIS and generates a PRPD map of the partial discharge signal, the preprocessing module preprocesses the PRPD map, outputs a gray level image, constructs a target data set, the transmission module receives the target data set and sends the target data set to the cloud server, the cloud server module receives the data of the transmission module, deploys a trained detection model, detects data to be detected and stores a visual result, the cloud server module transmits the visual result to the output module, and the output module transmits the visual result to an upper computer or intelligent terminal device.
8. The GIS ultrahigh frequency discharge signal detection device according to claim 7, wherein a back-end service of the cloud server module is connected with a front-end page through Websocket, wherein the back-end service is built through a flash frame, and the front-end page is built through a Vue frame.
9. The GIS uhf discharge signal detection apparatus of claim 8, wherein the back-end service comprises: and the module is used for receiving the image module to be detected, detecting the image module to be detected and outputting and storing the detection result.
10. The GIS uhf discharge signal detection apparatus of claim 8, wherein the front-end service comprises: the visual detection result module, the visual history result module and the control and management module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210733546.2A CN115184744B (en) | 2022-06-27 | 2022-06-27 | GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202210733546.2A CN115184744B (en) | 2022-06-27 | 2022-06-27 | GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN |
Publications (2)
Publication Number | Publication Date |
---|---|
CN115184744A CN115184744A (en) | 2022-10-14 |
CN115184744B true CN115184744B (en) | 2023-09-05 |
Family
ID=83516068
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202210733546.2A Active CN115184744B (en) | 2022-06-27 | 2022-06-27 | GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN115184744B (en) |
Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109635661A (en) * | 2018-11-13 | 2019-04-16 | 同济大学 | A kind of far field wireless charging reception object detection method based on convolutional neural networks |
CN110569717A (en) * | 2019-07-26 | 2019-12-13 | 深圳供电局有限公司 | partial discharge detection method, device, system, equipment and readable storage medium |
CN111431287A (en) * | 2020-05-14 | 2020-07-17 | 国网上海市电力公司 | Integrated perception terminal system of power distribution room |
CN111862119A (en) * | 2020-07-21 | 2020-10-30 | 武汉科技大学 | Semantic information extraction method based on Mask-RCNN |
CN112034310A (en) * | 2020-07-31 | 2020-12-04 | 国网山东省电力公司东营供电公司 | Partial discharge defect diagnosis method and system for combined electrical appliance |
CN114462521A (en) * | 2022-01-26 | 2022-05-10 | 浙江天铂云科光电股份有限公司 | Efficient intelligent classification and detection method for power equipment |
-
2022
- 2022-06-27 CN CN202210733546.2A patent/CN115184744B/en active Active
Patent Citations (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109635661A (en) * | 2018-11-13 | 2019-04-16 | 同济大学 | A kind of far field wireless charging reception object detection method based on convolutional neural networks |
CN110569717A (en) * | 2019-07-26 | 2019-12-13 | 深圳供电局有限公司 | partial discharge detection method, device, system, equipment and readable storage medium |
CN111431287A (en) * | 2020-05-14 | 2020-07-17 | 国网上海市电力公司 | Integrated perception terminal system of power distribution room |
CN111862119A (en) * | 2020-07-21 | 2020-10-30 | 武汉科技大学 | Semantic information extraction method based on Mask-RCNN |
CN112034310A (en) * | 2020-07-31 | 2020-12-04 | 国网山东省电力公司东营供电公司 | Partial discharge defect diagnosis method and system for combined electrical appliance |
CN114462521A (en) * | 2022-01-26 | 2022-05-10 | 浙江天铂云科光电股份有限公司 | Efficient intelligent classification and detection method for power equipment |
Non-Patent Citations (1)
Title |
---|
林刚 ; 王波 ; 彭辉 ; 王晓阳 ; 陈思远 ; 张黎明 ; .基于改进Faster-RCNN的输电线巡检图像多目标检测及定位.电力自动化设备.2019,(第05期),第220-225页. * |
Also Published As
Publication number | Publication date |
---|---|
CN115184744A (en) | 2022-10-14 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN112380952B (en) | Power equipment infrared image real-time detection and identification method based on artificial intelligence | |
CN109118479B (en) | Capsule network-based insulator defect identification and positioning device and method | |
Liu et al. | High precision detection algorithm based on improved RetinaNet for defect recognition of transmission lines | |
CN112034310A (en) | Partial discharge defect diagnosis method and system for combined electrical appliance | |
Modava et al. | Hierarchical coastline detection in SAR images based on spectral‐textural features and global–local information | |
CN115187527B (en) | Separation and identification method for multi-source mixed ultrahigh frequency partial discharge spectrum | |
CN111121797B (en) | Road screening method, device, server and storage medium | |
CN112365475A (en) | Power transmission line icing monitoring method and system based on image processing mode | |
CN112232371A (en) | American license plate recognition method based on YOLOv3 and text recognition | |
CN113205039A (en) | Power equipment fault image identification and disaster investigation system and method based on multiple DCNNs | |
CN113255580A (en) | Method and device for identifying sprinkled objects and vehicle sprinkling and leaking | |
CN113252701A (en) | Cloud edge cooperation-based power transmission line insulator self-explosion defect detection system and method | |
CN114612741A (en) | Defect recognition model training method and device, electronic equipment and storage medium | |
CN113962973A (en) | Power transmission line unmanned aerial vehicle intelligent inspection system and method based on satellite technology | |
CN111553500B (en) | Railway traffic contact net inspection method based on attention mechanism full convolution network | |
CN115184744B (en) | GIS ultrahigh frequency discharge signal detection device and method based on fast-RCNN | |
Liu et al. | Research on surface defect detection based on semantic segmentation | |
CN113781436B (en) | High-voltage switch state monitoring method based on camera and angle measurement | |
CN111354191B (en) | Lane driving condition determining method, device and equipment and storage medium | |
CN113962955A (en) | Method and device for identifying target object from image and electronic equipment | |
CN113643234A (en) | Composite insulator damage detection method, terminal equipment and readable storage medium | |
Wu et al. | Detection method based on improved faster R-CNN for pin defect in transmission lines | |
Yang et al. | Intelligent overheating fault diagnosis for overhead transmission line using semantic segmentation | |
Ma et al. | Unsupervised semantic segmentation of high-resolution UAV imagery for road scene parsing | |
Murthy et al. | Video surveillance-based insulator condition monitoring analysis for substation monitoring system (SMS) |
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 |