A method and system for cable defect identification based on online sequence limit learning machine

Abstract

This invention discloses a cable defect identification method and system based on an online sequence extreme learning machine, relating to the field of cable inspection technology. The method includes the following steps: when other defects exist in the cable under test, the processor acquires the location information of the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, activates the nearest image detection unit to collect an image of the cable under test based on the location information, and sends it to the processor. The processor converts the image of the cable under test into a corresponding partial discharge spectrum, constructs and trains an online sequence extreme learning machine recognition model, and processes the partial discharge spectrum using the online sequence extreme learning machine recognition model to obtain the defect identification result. This invention can simultaneously identify possible moisture defects, appearance defects, or other physical defects in the cable, improving the identification efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cable detection, in particular to a cable defect identification method and system based on an online sequence extreme learning machine. BACKGROUND

[0002] At present, with the continuous advancement of urbanization, cross-Linked Polyethylene (XLPE) power cables are widely used in urban power grids due to their good performance. However, during the installation and operation of the cables, the cables may be affected by production processes, construction quality, operating environment, external damage and other factors, and may produce various types of defects, such as appearance defects, moisture defects, insulation defects, etc. If the local defects of the cable body are not timely located and the defect types are not identified, hidden dangers will be left for the operation of the cable line, which may easily cause insulation breakdown accidents.

[0003] However, the existing cable defect identification method only identifies a single type of defect. Since there are many different types of cable defects, different positioning and identification methods need to be used, and only a single identification method has the disadvantage of low identification efficiency. At the same time, the continuously input data cannot be processed in the process of cable defect identification, and the data processing accuracy is low.

[0004] Therefore, how to provide a cable defect identification method that can solve the above problems is a problem that those skilled in the art need to solve. SUMMARY

[0005] Therefore, the present application provides a cable defect identification method and system based on an online sequence extreme learning machine, which can simultaneously identify possible moisture defects, appearance defects or other physical defects of the cable, and improves the identification efficiency.

[0006] In order to achieve the above purpose, the present application adopts the following technical solutions:

[0007] A cable defect identification method based on an online sequence extreme learning machine, comprising the following steps:

[0008] A plurality of wireless radio frequency signal transmitting units, wireless radio frequency signal receiving units, moisture detection units, wireless communication units and image detection units are arranged at equal intervals on the cable to be tested;

[0009] The moisture detection unit detects the moisture detection parameters of the cable to be tested in real time, the wireless radio frequency signal transmitting unit transmits radio frequency signals to the cable to be tested, the wireless radio frequency signal receiving unit receives radio frequency signals, and the moisture detection parameters and radio frequency signal reflection results are sent to the processor through the wireless communication unit for analysis to determine whether there is a moisture defect or other defect.

[0010] When the to-be-tested cable has other defects, the processor acquires position information corresponding to the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the to-be-tested cable according to the position information to collect a to-be-tested cable image and sends the to-be-tested cable image to the processor;

[0011] The processor converts the to-be-tested cable image into a corresponding partial discharge atlas, constructs and trains an online sequence extreme learning machine identification model, processes the partial discharge atlas by using the online sequence extreme learning machine identification model, and obtains a defect identification result.

[0012] Preferably, when the to-be-tested cable has a damp defect, position information corresponding to the damp detection unit that sends a damp detection parameter is acquired, and a position where the damp defect occurs is determined according to the position information.

[0013] Preferably, the method further comprises:

[0014] According to the defect identification result or the damp defect, the processor processes by calling a corresponding defect processing method through big data.

[0015] Preferably, the specific processing process of transmitting a radio frequency signal to the to-be-tested cable by the wireless radio frequency signal transmitting unit, receiving the radio frequency signal by the wireless radio frequency signal receiving unit, and sending a radio frequency signal reflection result to the processor for analysis by the wireless communication unit comprises:

[0016] Acquiring a reflection coefficient of the to-be-tested cable, a radio frequency signal strength transmitted by the wireless radio frequency signal transmitting unit, an attenuation coefficient, and a distance between the wireless radio frequency signal receiving unit and the wireless radio frequency signal transmitting unit;

[0017] Determining a preset arrival time according to the reflection coefficient, the radio frequency signal strength, the attenuation coefficient, and the distance;

[0018] Acquiring an actual arrival time of a signal received by the wireless radio frequency signal receiving unit, if an absolute value of a difference between the preset arrival time and the actual arrival time is within a preset threshold range, no defect is generated, if the absolute value exceeds the preset threshold range, a defect exists, and the image detection unit closest to the to-be-tested cable is started to collect a to-be-tested cable image.

[0019] Preferably, the specific processing process of constructing and training an online sequence extreme learning machine identification model, processing the partial discharge atlas by using the online sequence extreme learning machine identification model, and obtaining a defect identification result comprises:

[0020] Pretreating the partial discharge atlas;

[0021] An online sequence extreme learning machine identification model is constructed, and a historical data set is obtained to train and test the online sequence extreme learning machine identification model.

[0022] The preprocessed partial discharge atlas is identified by using the trained and tested online sequence extreme learning machine identification model to obtain a defect identification result.

[0023] Preferably, the online sequence extreme learning machine identification model comprises an input layer, a hidden layer and a classifier connected in sequence.

[0024] Preferably, the specific process of constructing an online sequence extreme learning machine identification model and obtaining a historical data set to train and test the online sequence extreme learning machine identification model comprises:

[0025] A historical cable defect image data set containing multiple defect types is obtained, and is classified according to each defect type to obtain n classification data sets n0, n1, n2,..., n i , wherein i represents the number of classification data sets.

[0026] For i = 2, 3, 4,..., i-1, the model m0, m1, m2,..., m j , wherein j represents the number of cached models, and the following operations are performed multiple times:

[0027] The classification data set n i-1 is selected. The online sequence extreme learning machine identification model is trained, and after the training is completed, the training model at this time is taken as the cached model m j-1 . The n i is used to train the cached model m j-1 to obtain the cached model m j .

[0028] The cached model m j at this time is verified by using a verification set, and when the loss is minimum, it is taken as the final online sequence extreme learning machine identification model.

[0029] Preferably, the moisture detection unit adopts a capacitive sensor.

[0030] The application also provides a system using the cable defect identification method based on the online sequence extreme learning machine.

[0031] The detection module comprises a plurality of wireless radio frequency signal transmitting units, a wireless radio frequency signal receiving unit, a moisture detection unit, a wireless communication unit and an image detection unit, which are used to detect the moisture detection parameters of the cable to be detected in real time, transmit radio frequency signals to the cable to be detected through the wireless radio frequency signal transmitting unit, and receive radio frequency signals through the wireless radio frequency signal receiving unit;

[0032] The processor is connected with the wireless radio frequency signal transmitting unit, the wireless radio frequency signal receiving unit, the moisture detection unit, the wireless communication unit and the image detection unit, is used to analyze the moisture detection parameters and the radio frequency signal reflection results through the moisture detection unit and the wireless communication unit, judge whether there is a moisture defect or other defects, and when the cable to be detected has other defects, the processor acquires the position information corresponding to the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the cable to be detected to collect the image of the cable to be detected according to the position information, and sends the image to the processor, and the processor converts the image of the cable to be detected into a corresponding partial discharge atlas;

[0033] The recognition module is used to construct and train an online sequence extreme learning machine recognition model, process the partial discharge atlas by using the online sequence extreme learning machine recognition model, and obtain a defect recognition result.

[0034] According to the above technical solution, compared with the prior art, the cable defect recognition method and system based on the online sequence extreme learning machine have the following beneficial effects:

[0035] (1) By setting a plurality of types of detection units, the moisture defect or other defects of the cable can be detected, the types of recognition are rich, and the recognition efficiency is improved;

[0036] (2) When the cable to be detected has other defects, the processor acquires the position information corresponding to the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the cable to be detected to collect the image of the cable to be detected according to the position information, and sends the image to the processor; an online sequence extreme learning machine recognition model is constructed and trained, the image of the cable to be detected is processed by using the online sequence extreme learning machine recognition model, and a defect recognition result is obtained; the training process of the online sequence extreme learning machine recognition model selects incremental learning, the model can process new image data while processing existing data, and the efficiency of the model in processing data is improved. BRIEF DESCRIPTION OF DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description only constitute a part of the embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of the provided drawings.

[0038] Figure 1 A flowchart of a cable defect identification method based on an online sequence extreme learning machine provided by the present application is provided.

[0039] Figure 2 A structural schematic diagram of a cable defect identification system based on an online sequence extreme learning machine provided by the present application is provided. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments only constitute a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0041] Referring to the drawings shown in the embodiments of the present application, a cable defect identification method based on an online sequence extreme learning machine is disclosed, comprising the following steps: Figure 1

[0042] The multiple wireless radio frequency signal transmitting units, the wireless radio frequency signal receiving units, the moisture detection units, the wireless communication units and the image detection units are arranged at equal intervals on the cable to be tested;

[0043] The moisture detection parameters of the cable to be tested are detected in real time by the moisture detection units, the radio frequency signals are transmitted to the cable to be tested by the wireless radio frequency signal transmitting units, the radio frequency signals are received by the wireless radio frequency signal receiving units, and the moisture detection parameters and the radio frequency signal reflection results are sent to the processor for analysis through the wireless communication units to determine whether there is a moisture defect or other defects;

[0044] When the cable to be tested has other defects, the processor acquires the position information corresponding to the wireless radio frequency signal transmitting units and the wireless radio frequency signal receiving units, starts the image detection unit closest to the cable to be tested according to the position information to collect the image of the cable to be tested, and sends the image to the processor;

[0045] The processor converts the image of the cable to be tested into a corresponding partial discharge atlas, constructs and trains an online sequence extreme learning machine identification model, processes the partial discharge atlas by using the online sequence extreme learning machine identification model, and obtains a defect identification result.

[0046] ​In a specific embodiment, when the to-be-tested cable has a moisture defect, the position information corresponding to the moisture detection unit that sends the moisture detection parameter is acquired, and the position where the moisture defect occurs is determined according to the position information.

[0047] In a specific embodiment, the processor further processes the defect according to the defect identification result or the moisture defect by calling a corresponding defect processing method through big data.

[0048] In a specific embodiment, the specific processing process in which the wireless radio frequency signal emitting unit emits a radio frequency signal to the to-be-tested cable, the wireless radio frequency signal receiving unit receives the radio frequency signal, and the wireless communication unit sends the radio frequency signal reflection result to the processor for analysis includes:

[0049] The reflection coefficient of the to-be-tested cable, the radio frequency signal strength emitted by the wireless radio frequency signal emitting unit, the attenuation coefficient, and the distance between the wireless radio frequency signal receiving unit and the wireless radio frequency signal emitting unit are acquired.

[0050] The preset arrival time is determined according to the reflection coefficient, the radio frequency signal strength, the attenuation coefficient, and the distance.

[0051] The actual arrival time of the signal received by the wireless radio frequency signal receiving unit is acquired, if the absolute value of the difference between the preset arrival time and the actual arrival time is within a preset threshold range, then no defect is generated, if the preset threshold range is exceeded, then a defect exists, and the image detection unit closest to the to-be-tested cable is started to collect the image of the to-be-tested cable.

[0052] In a specific embodiment, the specific processing process in which the online sequence extreme learning machine identification model is constructed and trained, and the online sequence extreme learning machine identification model is used to process the partial discharge atlas to obtain the defect identification result includes:

[0053] The partial discharge atlas is preprocessed.

[0054] The online sequence extreme learning machine identification model is constructed, and historical data sets are acquired to train and test the online sequence extreme learning machine identification model.

[0055] The preprocessed partial discharge atlas is identified by using the trained and tested online sequence extreme learning machine identification model to obtain the defect identification result.

[0056] In a specific embodiment, the online sequence extreme learning machine identification model includes an input layer, a hidden layer, and a classifier connected in sequence.

[0057] In a specific embodiment, the specific process in which the online sequence extreme learning machine identification model is constructed, and historical data sets are acquired to train and test the online sequence extreme learning machine identification model includes:

[0058] A historical cable defect image dataset containing multiple defect types is obtained, and classified according to each defect type, to obtain n classified datasets n0, n1, n2,..., n i where i represents the number of classified datasets;

[0059] For i = 2, 3, 4,..., i-1, cache the model m0, m1, m2,..., m j where j represents the number of cache models, and the following operations are performed multiple times:

[0060] Select the classified dataset n i-1 Train the online sequence extreme learning machine recognition model, and after the training is completed, the trained model at this time is taken as the cache model m j-1 , and n i is used to train the cache model m j-1 to obtain the cache model m j ;

[0061] The cache model m j at this time is verified using the validation set, and when the loss is minimum, it is taken as the final online sequence extreme learning machine recognition model.

[0062] Specifically, the above training process first divides the original dataset into multiple datasets according to the type, and iterates each time to train the model obtained in the last round of training combined with the new dataset. The model can process new image data while processing existing data, improving the efficiency of the model in processing data.

[0063] In one specific embodiment, the moisture detection unit adopts any one or any two of a capacitive sensor and a dielectric sensor.

[0064] Referring to FIG. 1, Figure 2 The embodiment of the application also provides a system using the cable defect recognition method based on the online sequence extreme learning machine of any one of the above embodiments, comprising:

[0065] The detection module includes a plurality of wireless radio frequency signal transmitting units, wireless radio frequency signal receiving units, moisture detection units, wireless communication units, and image detection units, for real-time detection of moisture detection parameters of the cable to be detected, simultaneously transmitting radio frequency signals to the cable to be detected through the wireless radio frequency signal transmitting units, and receiving the radio frequency signals through the wireless radio frequency signal receiving units;

[0066] The processor is connected with the wireless radio frequency signal transmitting unit, the wireless radio frequency signal receiving unit, the moisture detection unit, the wireless communication unit and the image detection unit, is used for analyzing the moisture detection parameter and the radio frequency signal reflection result sent to the processor by the moisture detection unit through the wireless communication unit, judging whether there is moisture defect or other defect; and when the cable to be measured has other defects, the processor obtains the position information corresponding to the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the cable to be measured according to the position information, collects the image of the cable to be measured and sends it to the processor, and the processor converts the image of the cable to be measured into the corresponding partial discharge atlas;

[0067] The recognition module is used for constructing and training an online sequence extreme learning machine recognition model, processing the partial discharge atlas by using the online sequence extreme learning machine recognition model, and obtaining a defect recognition result.

[0068] The various embodiments in the specification are described in a progressive manner, and each embodiment focuses on the difference from other embodiments. The same or similar parts of each embodiment can be referred to each other. For the device disclosed by the embodiments, since it corresponds to the method disclosed by the embodiments, the description is relatively simple, and the related parts can be referred to the method part.

[0069] The above description of the disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications of the embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A cable defect recognition method based on an online sequence extreme learning machine, characterized in that, Comprise the following steps: A plurality of wireless radio frequency signal transmitting units, wireless radio frequency signal receiving units, moisture detection units, wireless communication units and image detection units are arranged at equal intervals on the cable to be tested; The moisture detection parameter of the cable to be tested is detected in real time by the moisture detection unit, and the radio frequency signal is transmitted to the cable to be tested by the wireless radio frequency signal transmitting unit, and the radio frequency signal is received by the wireless radio frequency signal receiving unit, and the moisture detection parameter and the radio frequency signal reflection result are sent to the processor for analysis through the wireless communication unit to determine whether there is a moisture defect or other defect, and the specific processing process comprises: The reflection coefficient of the cable to be tested, the radio frequency signal strength transmitted by the wireless radio frequency signal transmitting unit, the attenuation coefficient and the distance between the wireless radio frequency signal receiving unit and the wireless radio frequency signal transmitting unit are obtained; The preset arrival time is determined according to the reflection coefficient, the radio frequency signal strength, the attenuation coefficient and the distance; The actual arrival time of the signal received by the wireless radio frequency signal receiving unit is obtained, if the absolute value of the difference between the preset arrival time and the actual arrival time is within the preset threshold range, no defect is generated, if it exceeds the preset threshold range, there is a defect; When the cable to be tested has other defects, the processor obtains the position information corresponding to the wireless radio frequency signal transmitting unit and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the cable to be tested according to the position information, and sends the image to the processor; The processor converts the image of the cable to be tested into a corresponding partial discharge atlas, constructs and trains an online sequence extreme learning machine recognition model, processes the partial discharge atlas by using the online sequence extreme learning machine recognition model, and obtains a defect recognition result. An online sequence extreme learning machine recognition model is constructed, and historical data sets are obtained to train and test the online sequence extreme learning machine recognition model, and the specific process comprises: A historical cable defect image dataset containing a plurality of defect types is acquired and classified by each defect type to obtain n classified datasets n0, n1, n2,..., n i where i represents the number of classified datasets; For i = 2, 3, 4,..., i - 1, cache model m0, m1, m2,... m j where j denotes the number of cache models, the following operations are performed multiple times: Selecting a classification data set n i-1 Training the online sequence extreme learning machine identification model, and taking the trained model as a cache model m j-1 i Training the cache model m j-1 j ;​​ The cache model m at this time is obtained j The training process first divides the historical cable defect image dataset into multiple datasets according to the defect type, uses a loop to iteratively train the model obtained in the last round of training combined with a new dataset each time, and the model realizes processing of the existing data volume while processing new image data.

2. The cable defect recognition method based on online sequential extreme learning machine according to claim 1, characterized in that, When the cable to be tested has a moisture defect, the position information corresponding to the moisture detection unit sending the moisture detection parameter is obtained, and the position of the moisture defect is determined according to the position information.

3. The cable defect recognition method based on online sequential extreme learning machine according to claim 2, characterized in that, Also includes: According to the defect recognition result or the moisture defect, the processor processes by calling the corresponding defect processing method through big data.

4. The cable defect recognition method based on online sequential extreme learning machine according to claim 1, characterized in that, The specific processing process of constructing and training an online sequence extreme learning machine recognition model, processing the partial discharge atlas by using the online sequence extreme learning machine recognition model, and obtaining a defect recognition result further comprises: The partial discharge atlas is preprocessed; The online sequence extreme learning machine recognition model is trained and tested, and the partial discharge atlas after preprocessing is recognized to obtain a defect recognition result.

5. The cable defect recognition method based on online sequential extreme learning machine according to claim 4, characterized in that, The online sequence extreme learning machine recognition model comprises an input layer, a hidden layer and a classifier connected in turn.

6. The cable defect recognition method based on online sequential extreme learning machine according to claim 2, characterized in that, The moisture detection unit adopts a capacitive sensor.

7. A system for cable defect recognition using the online sequence extreme learning machine based method according to any one of claims 1-6, characterized in that, Comprise: The detection module comprises a plurality of wireless radio frequency signal transmitting units, a wireless radio frequency signal receiving unit, a moisture detection unit, a wireless communication unit and an image detection unit, which are used to detect moisture detection parameters of the cable to be detected in real time, transmit radio frequency signals to the cable to be detected through the wireless radio frequency signal transmitting units, and receive radio frequency signals through the wireless radio frequency signal receiving unit; The processor is connected with the wireless radio frequency signal transmitting units, the wireless radio frequency signal receiving unit, the moisture detection unit, the wireless communication unit and the image detection unit, is used to analyze whether there is a moisture defect or other defects by sending the moisture detection parameters and the radio frequency signal reflection results to the processor through the wireless communication unit through the moisture detection unit; and when the cable to be detected has other defects, the processor acquires position information corresponding to the wireless radio frequency signal transmitting units and the wireless radio frequency signal receiving unit, starts the image detection unit closest to the cable to be detected to collect an image of the cable to be detected according to the position information, and sends the image to the processor, and the processor converts the image of the cable to be detected into a corresponding partial discharge atlas; The recognition module is used to construct and train an online sequence extreme learning machine recognition model, process the partial discharge atlas by using the online sequence extreme learning machine recognition model, and obtain a defect recognition result.

Images