Multi-parameter edge type portable equipment for intelligent electric power inspection

Abstract

The invention provides a multi-parameter edge type portable device for electric power intelligent inspection, and belongs to the technical field of electric power inspection. The equipment comprises a partial discharge acquisition module, an infrared and visible light MSX double-vision fusion module and an intelligent hand-held terminal, and the partial discharge acquisition module comprises an HFCT detection unit, a UHF detection unit, an AE detection unit and a transient earth electric wave partial discharge detection unit; the infrared and visible light MSX double-vision fusion module adds image information shot by the visible light camera into the whole infrared image in real time, and the infrared image and the visible light image are matched, superposed and fused; and the partial discharge acquisition module and the infrared and visible light MSX double-vision fusion module are arranged on the intelligent handheld terminal. According to the invention, intelligent inspection and in-situ diagnosis auxiliary decision making of typical power equipment are realized, and the labor intensity of operation and maintenance inspection personnel is effectively reduced; the intelligent inspection level of an electric power field can be comprehensively improved, typical equipment abnormity can be found as early as possible, and the service life and the utilization rate of electric power equipment are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of electric power inspection, and specifically is a multi-parameter edge-type portable equipment for electric power intelligent inspection. Background Art

[0002] Power equipment status inspection is an important measure to ensure the safe operation of the power system. In terms of partial discharge status inspection, partial discharge detectors are indispensable detection tools in the power industry, providing a basis for insulation defect detection and diagnosis of power equipment. As a detection technology, its application includes three aspects: technical factors, detection data utilization rate and equipment investment. Only when the above three aspects are closely integrated with production practice can it have sustainable application value. There are three main types of partial discharge sensors for power equipment: UHF, AE and HFCT. Among them, UHF and AE sensors are the most commonly used. In terms of temperature measurement, there are many domestic and foreign equipment, including infrared imagers and various temperature measurement terminal applications. A series of temperature measurement devices such as infrared imagers, monochrome and dual-color thermometers have achieved certain practical application effects. However, the current infrared imagers are generally expensive and lack the function of on-site image analysis and processing. They usually need to use the background system for offline analysis or build a remote communication network for online analysis. However, the reliable communication transmission of massive images occupies a lot of network resources on the one hand, and the reliability needs to be further improved. In addition, in terms of equipment insulation monitoring and defect auxiliary identification and evaluation, the evaluation indicators are relatively rough, the reliability of the diagnostic results is poor, and the reference value is not high.

[0003] In the prior art, there are no reports on the application of mobile inspection intelligent detection terminals based on temperature maps and partial discharge multi-feature comprehensive detection. For the manual inspection of equipment appearance along the line, traditional inspections require the configuration of different types of partial discharge, temperature measurement and other detection equipment, which are inconvenient to carry, cumbersome to operate, and have high manpower, material and time costs. In addition, it is difficult to achieve standardized inspection records of power equipment inspections, on-site comprehensive diagnosis and automatic file management on site, and it is impossible to achieve reliable instrumented reconnaissance for intensive collection and joint identification of multiple parameters, and it is impossible to accurately identify abnormal conditions in equipment operation. Summary of the invention

[0004] The present invention aims to address the technical defects of the prior art and provide a multi-parameter edge-type portable equipment for smart power inspection, so as to solve the technical problems that conventional detection equipment is inconvenient to carry, cumbersome to operate, and has functions that need to be improved.

[0005] Another technical problem to be solved by the present invention is how to develop a mobile inspection intelligent detection terminal based on temperature spectrum and partial discharge multi-feature comprehensive detection.

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

[0007] The multi-parameter edge-type portable equipment for smart power inspection comprises a partial discharge acquisition module, an infrared and visible light MSX dual-view fusion module, and an intelligent handheld terminal, wherein the partial discharge acquisition module comprises an HFCT detection unit, a UHF detection unit, an AE detection unit, and a transient ground wave partial discharge detection unit; the infrared and visible light MSX dual-view fusion module adds image information captured by a visible light camera to the entire infrared image in real time, and matches, superimposes and fuses the infrared image with the visible light image; the partial discharge acquisition module and the infrared and visible light MSX dual-view fusion module are arranged on the intelligent handheld terminal.

[0008] Preferably, the partial discharge acquisition module also includes a signal conditioning unit, an MCU, a communication unit, and a power supply phase synchronization unit. The partial discharge acquisition module detects, conditions, acquires and sends partial discharge signals, receives data, and displays, stores and uploads the data.

[0009] Preferably, the partial discharge acquisition module uses FastICA and frequency slicing wavelet transform as narrowband interference suppression algorithms for partial discharge detection, and suppresses white noise interference based on frequency slicing wavelet decomposition of EEMD.

[0010] Preferably, the partial discharge acquisition module establishes a typical power equipment partial discharge database, and the typical power equipment partial discharge database includes a self-verification module, a self-reduction module and a self-growth module; the self-verification module searches for signals similar to the signal to be judged through the sample library, and when a similar signal exists, the category of the similar signal is used as the category of the signal, and when no similar signal exists, it is judged to be a partial discharge signal or an interference signal through an automatic spectrum recognition method; the self-reduction module calculates the similarity between the signal and the historical sample library signal, and deletes the signal whose similarity is higher than a certain threshold; the self-growth module calculates the similarity between the signal and the signal in the sample library, reduces the signal whose similarity is greater than a certain threshold, and inputs the reduced signal into the sample library to update the sample library data.

[0011] Preferably, the equipment compares the image information collected on-site with the power equipment database through an artificial intelligence deep neural network model, and identifies, segments and filters the pre-processed images.

[0012] Preferably, the equipment diagnoses the monitoring data through an algorithm model, and the algorithm model includes a surface temperature judgment method, a relative temperature difference judgment method, a similar comparison method, and an archive analysis method.

[0013] Preferably, the equipment is equipped with a typical power equipment health diagnosis system, and the method for establishing the typical power equipment health diagnosis system includes: extracting and constructing multi-dimensional feature parameters of typical power equipment; reducing the dimension to select effective feature parameters; dividing the database; selecting a neural network model; and training and optimizing the neural network model.

[0014] Preferably, in the process of establishing a typical power equipment health diagnosis system, an image feature extraction method based on a grayscale co-occurrence matrix is ​​used to obtain multidimensional feature parameters of partial discharge and infrared temperature measurement that characterize the health status of typical power equipment, and the constructed multidimensional feature parameters are ranked by importance using a random forest algorithm; feature parameters that distinguish partial discharge and infrared temperature measurement signals of different defect types are screened out, and the data sets in the database are divided into a training set and a test set, with the training set and the test set accounting for 75% and 25% respectively; a convolutional neural network model is selected and the network parameters are initialized, the training set is used as input, and the algorithm is used to train and optimize the neural network.

[0015] Preferably, the process of establishing a typical power equipment health diagnosis system also includes the following steps: determining the parameters of the neural network, retraining the optimized neural network model using the training set, evaluating the performance of the typical power equipment defect diagnosis system based on deep learning through the test set, and proposing a typical power equipment defect diagnosis method based on the recognition accuracy and efficiency of the test system.

[0016] Preferably, the intelligent handheld terminal is integrated with a wireless communication module, a GPS positioning module or a Beidou positioning module, and a geographic information system.

[0017] The present invention proposes a multi-mode partial discharge sensor integrated technology and an infrared temperature measurement technology based on artificial intelligence, as well as a typical power equipment intelligent inspection technology based on a mobile operation mode, to achieve the partial discharge data collection, processing, and feature extraction of typical power inspection equipment, and provide basic support for state diagnosis and evaluation. Taking into account the influence of external environmental interference, a new algorithm model for simultaneous collection, processing, and analysis of various types of monitoring data diagnosis is proposed, the signal processing filter circuit is improved, the temperature compensation mode is optimized, and the MSX dual-view fusion technology is used to match, superimpose, and fuse infrared images with visible light images, highlight the identification target, and improve the accurate capture of abnormal targets in the image and the temperature measurement accuracy. At the same time, a trend analysis decision-making model is established based on the typical intelligent inspection technology of power equipment in mobile operation mode, and the dynamic operating conditions and trends related to the status of typical power equipment based on convolutional neural networks are evaluated. On this basis, multi-parameter edge-type portable equipment such as integrated intelligent handheld terminal devices, wireless communication technology, GPS / Beidou global positioning, and geographic information system (GIS) are developed for smart power inspection, and a unified power equipment operation status assessment, diagnosis, learning and improvement management platform is established to realize the electronic, information, standardization, visualization and intelligentization of power field inspection operations, effectively improving the quality and efficiency of field operations.

[0018] The present invention realizes intelligent inspection and on-site diagnosis and decision-making assistance for typical power equipment such as transformers, switch cabinets, circuit breakers, and cables, effectively reduces the labor intensity of operation and maintenance inspection personnel, reduces the heavy work under the rapid expansion of the power grid scale and the safety hazards of staff inspection under complex on-site conditions, prevents personal injury losses, and improves the safety of power grid operation and power supply reliability. The present invention can comprehensively improve the level of intelligent inspection on-site, innovate the live operation and inspection mode, improve the efficiency of operation and inspection, optimize the operation and inspection effect, facilitate early detection of typical equipment abnormalities, and improve the service life and utilization rate of power equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 It is a structural schematic diagram of the small integrated partial discharge detection part in the present invention.

[0020] Figure 2 It is a flow chart of interference signal suppression in the partial discharge detection link in the present invention.

[0021] Figure 3 It is a flow chart of establishing a typical partial discharge database of power equipment in the present invention.

[0022] Figure 4 It is a schematic diagram of training and optimizing a neural network in the present invention.

[0023] Figure 5 It is a schematic diagram for evaluating dynamic operating conditions and trends related to typical power equipment in the present invention. DETAILED DESCRIPTION

[0024] The specific embodiments of the present invention will be described in detail below. In order to avoid too much unnecessary details, the well-known structures or functions will not be described in detail in the following examples. The approximate language used in the following examples can be used for quantitative expression, indicating that the quantity can be allowed to have a certain change without changing the basic function. Unless otherwise defined, the technical and scientific terms used in the following examples have the same meaning as those generally understood by those skilled in the art to which the present invention belongs.

[0025] A multi-parameter edge-type portable equipment for intelligent power inspection.

[0026] 1. Multi-mode partial discharge sensing integrated technology

[0027] (1) A small integrated detection section for partial discharge that integrates high frequency, ultra-high frequency, ultrasonic, ground wave, and ground current sensing technologies

[0028] Small-scale integrated partial discharge detection involves the collection and detection of different types and characteristic information of partial discharge. There are many partial discharge sensor modules with rich functions. It has high requirements on the size and endurance of the equipment. At the same time, it also has the most core sensitivity and accuracy. It is necessary to ensure the goal of optimal coordination and realize the integrated fusion collection and detection of multiple types of partial discharge characteristics.

[0029] An integrated design method for partial discharge that integrates high frequency, ultra-high frequency, ultrasonic, and ground current sensing technologies is proposed. Figure 1 As shown, accurate identification and extraction of multiple types of partial discharge characteristic signals and wireless and reliable transmission are achieved to meet the needs of partial discharge status detection of power electrical equipment.

[0030] 1) HFCT detection: When partial discharge occurs inside the power equipment, the high-frequency discharge pulse will propagate to the earth along the grounding wire. The high-frequency (HFCT) method detects high-frequency current signals to achieve partial discharge detection. It is a non-invasive detection method. The detected equipment does not need to be shut down. It is currently the main means of live detection. In this method, the signal coupled out by the high-frequency sensor is not a pure discharge signal, but a mixture of electromagnetic interference noise. At the same time, it is affected by factors such as sensor performance, and there are deficiencies in sensitivity verification and detection reliability.

[0031] 2) UHF detection: detects ultra-high frequency electromagnetic wave signals, has a large detection range, is not affected by corona interference in the air, is sensitive to various defects, and the detection sensitivity can reach several pc.

[0032] 3)AE detection: detects ultrasonic signals, is resistant to electrical interference, and is sensitive to free particle defects and vibrations of metal parts. When close to the defective part, the detection sensitivity can reach several pc.

[0033] 4) Transient Earth Electromagnetic Discharge Detection (TEV): When partial discharge occurs, the electromagnetic waves generated spread from the discharge point to the surrounding areas. The transient low voltage signal is detected using the TEV sensor, and the detection frequency band is within the range of 300kHZ-100MHz. TEV has a high sensitivity to internal discharge, and can effectively determine the partial discharge fault inside the switch cabinet. It can also locate the partial discharge source based on the time difference or amplitude comparison of the transient low voltage signal reaching different sensors.

[0034] (2) Small integrated intelligent wireless partial discharge sensing effect improvement technology

[0035] Partial discharge can accurately reflect the deterioration state of power equipment, and the accumulation of partial discharge will cause further deterioration of its insulation performance and affect the safe operation of the equipment. However, environmental interference in the partial discharge detection process cannot be ignored, and interference signal suppression is one of the key issues. The interference in partial discharge testing can be summarized into three types: periodic narrowband interference, white noise interference, and random pulse interference. Periodic narrowband interference basically comes from signals such as radio broadcasts inside the power network and some environmental spaces. In particular, narrowband signals that coincide with the frequency used by the detection platform will have a greater impact on the accuracy of partial discharge signal measurement; white noise interference is mainly a stable random signal caused by the thermal noise of electrical equipment. Its power spectrum density is uniformly distributed in the frequency domain, and its mathematical expectation and variance are independent of time; random pulse interference is generally a random interference signal generated when switches, relays, and thyristor devices are closed or opened; these interference signals will seriously affect the accuracy of the obtained partial discharge signal.

[0036] According to the characteristics of various interference signals, random pulse interference is usually strong, low frequency and easy to identify, periodic narrowband interference has a high degree of recognition in the frequency domain, and white noise interference is a broadband interference random signal, which has similar characteristics to partial discharge in the frequency domain. Therefore, the following process can be used step by step in actual interference suppression. Figure 2 shown.

[0037] The present invention adopts FastICA and frequency slicing wavelet transform as narrowband interference suppression algorithms for partial discharge detection, and frequency slicing wavelet suppression white noise interference method based on EEMD decomposition. Since random pulse interference has a relatively large amplitude and a relatively dispersed time domain distribution.

[0038] (3) Constructing partial discharge characteristics and database of typical power equipment defects

[0039] Automatic preprocessing of partial discharge data, based on partial discharge data to summarize the influence of different defects on the health status of power equipment, and establish a typical partial discharge database of power equipment. Figure 3 shown.

[0040] The health diagnosis method of typical power equipment requires certain training data. Based on the evaluation learning of the historical sample library, the original pattern recognition model is first used to determine whether the signal belongs to partial discharge or interference. If the pattern recognition model cannot make a judgment, a noise reduction system based on K-Means clustering analysis is used to make further judgments. Finally, accurate partial discharge signals and interference signals are obtained and classified into the partial discharge library or interference library. With the partial discharge library and interference library, more accurate pattern recognition models can be built through them, reducing the number of manual participation times, and finally realizing a fully automatic or nearly fully automatic partial discharge automatic preprocessing system.

[0041] The data processing strategy of the typical power equipment partial discharge characteristic information database includes three modules: self-verification, self-reduction and self-increase.

[0042] Self-verification refers to automatically judging whether a signal is a partial discharge or interference. First, the sample library is used to find a signal similar to the signal to be judged. If there is a similar signal, the category of the similar signal is used as the category of the signal. If there is no similar signal, it is judged by automatic spectrum recognition technology. By combining the sample library and automatic spectrum recognition method, the module can automatically judge whether the signal is a partial discharge or interference, realizing the function of self-verification.

[0043] The self-reduction module refers to automatically reducing the number of signals to reduce the overhead of system memory and other resources. This module calculates the similarity between the signal and the historical sample library signal and deletes the signal with high similarity to achieve the purpose of reducing the number of signals.

[0044] The self-growth module means that the signal judged by the system will be automatically added to the sample database to update the historical sample library data. The updated sample library data contains more and richer samples, which enhances the recognition ability of the self-verification link. The similarity between the signal and the signal in the sample library is calculated, and the signal with greater similarity is reduced. The reduced signal is input into the sample library to update the sample library data.

[0045] 2. Infrared temperature measurement technology based on artificial intelligence

[0046] (1) Visual temperature measurement in dual spectrum mode

[0047] The present invention adopts high-resolution digital infrared temperature measurement technology, comprehensively considers environmental interference factors, establishes an algorithm model, improves the signal processing filter circuit, optimizes the temperature compensation mode, and realizes non-contact high-precision temperature measurement and digital display.

[0048] By introducing image registration and fusion technology, the infrared image and the visible light image are matched, superimposed and fused. The conventional process of image fusion is generally to perform registration first and then further fusion. The main purpose of image fusion is to improve the reliability of the image by processing redundant data between multiple images, and to improve the clarity of the image by processing complementary information between multiple images. According to the different levels of information representation, the processing methods of image fusion can generally be divided into three levels: pixel-level fusion, feature-level fusion and decision-level fusion. The present invention mainly marks the temperature information obtained from the infrared image on the visible light image, so only the fusion at the pixel level needs to be considered. Consider using a dedicated MSX dual-view fusion technology to add the key detail information captured by the visible light camera to the entire infrared image in real time. By matching, superimposing and fusing the infrared image with the visible light image, the picture quality and contour clarity can be greatly improved.

[0049] (2) Temperature detection and diagnosis of typical power equipment based on artificial intelligence

[0050] Through the artificial intelligence deep neural network model, the image information collected on-site is intelligently compared with the power equipment database, and the pre-processed image is identified, the target is segmented and filtered to highlight the identified target, ultimately improving the accurate capture of abnormal targets in the image, and realizing intelligent analysis, diagnosis and early warning of equipment abnormalities and defects in infrared images and visible light images.

[0051] In the present invention, the influence of external environmental interference is comprehensively considered, and a new algorithm model for simultaneously collecting, processing and analyzing various types of monitoring data diagnosis is proposed. The algorithm model includes surface temperature judgment method, relative temperature difference judgment method, similar comparison method, archive analysis method, etc. At the same time, the signal processing filter circuit is improved, and the temperature compensation mode is optimized, so as to achieve accurate judgment of the attributes, location and severity of potential faults in the equipment.

[0052] 3. Typical intelligent inspection technology of power equipment based on mobile operation mode

[0053] (1) Typical power equipment condition assessment and diagnosis based on convolutional neural network, deep learning and artificial neural network analysis and early warning

[0054] The main contents include: extracting and constructing multi-dimensional feature parameters of typical power equipment → reducing the dimension to select effective feature parameters → dividing the database → selecting the neural network model → training and optimizing the neural network model → establishing a typical power equipment health diagnosis system.

[0055] The image feature extraction method based on gray level co-occurrence matrix (GLCM) is used to obtain multidimensional feature parameters of partial discharge and infrared temperature measurement that can characterize the health status of typical power equipment. The importance of the constructed multidimensional feature parameters is ranked using the random forest algorithm.

[0056] Effective characteristic parameters that are helpful in distinguishing partial discharge and infrared temperature measurement signals of different defect types are screened out, and the data sets in the database are divided into training sets and test sets (the proportions of the two are 75% and 25% respectively).

[0057] Select a suitable convolutional neural network model and initialize the network parameters, take the training set as input, and use different algorithms to train and optimize the neural network. Figure 4 shown.

[0058] Determine the parameters of the neural network, use the training set to retrain the optimized neural network model, and use the test set to evaluate the performance of the typical power equipment defect diagnosis system based on deep learning. Combined with the recognition accuracy and efficiency of the test system, a typical power equipment defect diagnosis method is proposed, and corresponding solutions to the fault problems are proposed to improve on-site work efficiency and reduce on-site working time.

[0059] The multi-parameter early warning system for typical power equipment based on learning and reasoning artificial neural networks, based on monitoring data, equipment parameters, abnormal fault feature library, etc., establishes a trend analysis decision model through a multi-parameter decision tree early warning system. The calculation results can evaluate the dynamic operating conditions and trends related to typical power equipment. At the same time, it supports hierarchical early warning and assists in carrying out value-added services such as equipment and operation and maintenance management. Figure 5 shown.

[0060] (2) Typical power equipment inspection business assistance and evidence collection identification and positioning implementation

[0061] The power inspection business has professional and rigorous requirements, including management elements and technical elements. The specific processes and requirements vary depending on the situation of each unit and the actual site. The present invention first clarifies the power inspection business process, risk points, defects and countermeasures requirements and plans, builds a power inspection business auxiliary system, and combines the core inspection objectives to design and implement the evidence collection, identification and positioning functions of typical power equipment anomalies, hidden dangers or failures.

[0062] By integrating intelligent handheld terminal devices, wireless communication technology, GPS / Beidou global positioning, geographic information system (GIS), etc., based on the concepts of "mobility", "standardization" and "normalization", a unified digital management system for power inspection management is established to realize the electronic, information-based, standardized, visual and intelligent power on-site inspection operations, effectively improve the quality and efficiency of on-site operations, and effectively prevent hidden dangers by tracking and analyzing the data accumulated in the system, provide a strong data basis for defect tracking management, achieve long-term efficient and stable operation, and play a leading and guiding role in the enterprise informatization construction of power supply companies.

[0063] 4. Multi-parameter edge portable equipment for smart power inspection

[0064] The software part mainly includes the terminal embedded software system and the power equipment operation status assessment diagnosis learning and improvement management platform software. The power equipment operation status assessment diagnosis learning and improvement management platform software not only realizes the basic functions of terminal equipment overall management, but also has the core functions of optimizing the characteristics of typical power inspection equipment and updating the characteristic library, as well as iterative training and lightweight upgrade of the evaluation algorithm for typical power equipment status.

[0065] The hardware part mainly includes a small integrated partial discharge acquisition module that integrates high frequency, ultra-high frequency, ultrasonic, geomagnetic wave, and ground current sensing technologies, an infrared and visible light MSX dual-view fusion module, and multi-parameter edge-type portable equipment for smart power inspection.

[0066] The software part mainly includes the terminal embedded software system and the power equipment operation status assessment diagnosis learning and improvement management platform software. Among them, the power equipment operation status assessment diagnosis learning and improvement management platform software not only realizes the basic functions of terminal equipment overall management, but also has the core functions of optimizing the characteristics of typical power equipment and updating the characteristic library, as well as iterative training and lightweight upgrade of the status assessment algorithm of typical power equipment. The intelligent assistance of the inspection business of typical power equipment mainly includes plan management, scheduling tasks, information collection, diagnosis and evaluation, event reporting, statistical analysis, hidden danger tracing, data management, resource layers, as well as personnel positioning, map navigation, electronic fences, track playback and one-key alarm.

[0067] The hardware part mainly includes a small integrated acquisition module for partial discharge that integrates high frequency, ultra-high frequency, ultrasonic, ground wave, and ground current sensing technologies, an infrared and visible light MSX dual-view fusion module, and an intelligent handheld terminal device. High frequency current detection band: 0.5-50MHz, test range 0-80dB; ultra-high frequency detection band: 300-1500MHz, test range 0-70dB; transient ground wave detection band: 3-100MHz, test range 0-70dB; ultrasonic detection band: 20kHz-300kHz, test range -10-70dB; ground current detection range: 0-6000A, error ±1A, resolution 0.1A, to meet the needs of partial discharge inspection of different types of electrical equipment; infrared spectrum: intelligent recognition and early warning, measurement range -20-300℃, wavelength 750-1400nm, accuracy ±2%.

[0068] The embodiments of the present invention are described in detail above, but the contents are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the scope of the present invention shall be included in the protection scope of the present invention.

Claims

1. Multi-parameter edge portable equipment for power intelligent inspection, characterized by: It includes a partial discharge acquisition module, an infrared and visible light MSX dual-view fusion module, and an intelligent handheld terminal, wherein the partial discharge acquisition module includes an HFCT detection unit, a UHF detection unit, an AE detection unit, and a transient ground wave partial discharge detection unit; the infrared and visible light MSX dual-view fusion module adds the image information taken by the visible light camera to the entire infrared image in real time, and matches, superimposes and fuses the infrared image with the visible light image; the partial discharge acquisition module and the infrared and visible light MSX dual-view fusion module are arranged on the intelligent handheld terminal.

2. The multi-parameter edge-type portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The partial discharge acquisition module also includes a signal conditioning unit, an MCU, a communication unit, and a power supply phase synchronization unit. The partial discharge acquisition module detects, conditions, acquires and sends partial discharge signals, receives data, and displays, stores and uploads the data.

3. The multi-parameter edge-type portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The partial discharge acquisition module uses FastICA and frequency slicing wavelet transform as narrowband interference suppression algorithms for partial discharge detection, and suppresses white noise interference based on frequency slicing wavelet decomposition of EEMD.

4. The multi-parameter edge portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The partial discharge acquisition module establishes a typical power equipment partial discharge database, and the typical power equipment partial discharge database includes a self-verification module, a self-reduction module and a self-growth module; the self-verification module searches for signals similar to the signal to be judged through the sample library, and when a similar signal exists, the category of the similar signal is used as the category of the signal, and when no similar signal exists, it is judged that it belongs to a partial discharge signal or an interference signal through an automatic spectrum recognition method; the self-reduction module calculates the similarity between the signal and the historical sample library signal, and deletes the signal whose similarity is higher than a certain threshold; the self-growth module calculates the similarity between the signal and the signal in the sample library, reduces the signal whose similarity is greater than a certain threshold, and inputs the reduced signal into the sample library to update the sample library data.

5. The multi-parameter edge portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The equipment uses an artificial intelligence deep neural network model to compare the image information collected on-site with the power equipment database, and recognizes, segments and filters the pre-processed images.

6. The multi-parameter edge-type portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The equipment diagnoses the monitoring data through an algorithm model, which includes a surface temperature judgment method, a relative temperature difference judgment method, a similar comparison method, and an archive analysis method.

7. The multi-parameter edge portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: The equipment is equipped with a typical power equipment health diagnosis system, and the method for establishing the typical power equipment health diagnosis system includes: extracting and constructing multi-dimensional characteristic parameters of typical power equipment; reducing the dimension to select effective characteristic parameters; dividing the database; selecting a neural network model; and training and optimizing the neural network model.

8. The multi-parameter edge portable equipment for electric power intelligent inspection according to claim 1 is characterized in that: In the process of establishing a typical power equipment health diagnosis system, the image feature extraction method based on the gray level co-occurrence matrix is ​​used to obtain the multi-dimensional feature parameters of partial discharge and infrared temperature measurement that characterize the health status of typical power equipment, and the constructed multi-dimensional feature parameters are ranked by importance using the random forest algorithm; the feature parameters that distinguish the partial discharge and infrared temperature measurement signals of different defect types are screened out, and the data sets in the database are divided into a training set and a test set, and the proportions of the training set and the test set are 75% and 25% respectively; the convolutional neural network model is selected and the network parameters are initialized, the training set is used as input, and the algorithm is used to train and optimize the neural network.

9. The multi-parameter edge-type portable equipment for electric power intelligent inspection according to claim 8 is characterized in that: The process of establishing a typical power equipment health diagnosis system also includes the following steps: determining the parameters of the neural network, retraining the optimized neural network model using the training set, evaluating the performance of the typical power equipment defect diagnosis system based on deep learning through the test set, and proposing a typical power equipment defect diagnosis method based on the recognition accuracy and efficiency of the test system.

10. The multi-parameter edge portable equipment for electric power intelligent inspection according to claim 1, characterized in that: The intelligent handheld terminal is integrated with a wireless communication module, a GPS positioning module or a Beidou positioning module, and a geographic information system.

Images