GIS (Geographic Information System) submillimeter particle defect accumulation and development stage identification method and device

Abstract

The invention provides a GIS submillimeter particle defect accumulation and development stage identification method and device, and belongs to the technical field of power test equipment. Comprising the following steps: acquiring ultrasonic signals of submillimeter particles moving in a GIS (Geographic Information System) and extracting ultrasonic signal feature data; collecting pulse current signals and extracting pulse features; carrying out reliability verification on the ultrasonic signal characteristic data and the pulse characteristic data, and removing abnormal data; establishing a feature matrix based on the ultrasonic signal feature data and the pulse feature data, and inputting the feature matrix into a multilayer convolutional neural network CNN model for analysis to obtain an accumulation form and a development stage of particles in the GIS; the accumulation form and the development stage comprise a small-amplitude vertical jumping form in the early stage, a large-amplitude vertical jumping form in the middle stage, a horizontal diffusion form in the last stage and a fuzzy state. According to the detection method, the accumulation form and the development stage of the particles can be accurately judged, and the detection reliability is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power test equipment, and in particular to a method and a device for identifying the accumulation and development stages of GIS sub-millimeter particle defects. Background Art

[0002] Gas Insulated Switchgear (GIS) plays an important role in the field of power transmission and distribution, and is widely used in high-voltage and ultra-high-voltage transmission systems. It has good sealing, small footprint and reliable operation, and has gradually become a key equipment in the power system. However, as the operation time of GIS increases, submillimeter metal particles may appear inside. These particles will move under the influence of electric field forces and airflow, and under certain conditions may cause problems such as partial discharge, resulting in insulation degradation, and even insulation breakdown and equipment failure in severe cases. Therefore, accurately detecting and identifying the accumulation morphology and development stage of submillimeter particles inside GIS is crucial for the safe operation of the equipment.

[0003] At present, the commonly used methods for detecting particles inside GIS mainly include pulse current detection method, ultrasonic detection method and ultra-high frequency detection method. The pulse current method determines the particle motion state and discharge characteristics by measuring the instantaneous current signal generated by the discharge; the ultrasonic detection method detects the ultrasonic signal generated by the particle motion to achieve defect detection; the ultra-high frequency detection method is mainly used for the detection of high-frequency discharge phenomena. These methods can detect the existence of particles and preliminarily analyze their motion state to a certain extent. However, in practical applications, it is difficult for a single detection method to comprehensively and accurately identify the accumulation morphology and development stage of particles. Due to the randomness and complexity of particle motion, the existing technology has limitations in judging the motion characteristics, discharge behavior and development stage, which is prone to misjudgment or omission, affecting the reliability of the detection results.

[0004] With the increasing requirements for particle detection accuracy and safe operation inside GIS, there is an urgent need for a detection method that can effectively combine multiple detection methods and comprehensively analyze the particle movement characteristics and discharge characteristics to more accurately determine the accumulation form and development stage of particles. Summary of the invention

[0005] In view of this, the present invention proposes a method and device for identifying the accumulation and development stage of submillimeter particle defects in GIS, aiming to combine pulse current signals with ultrasonic signals to improve the accuracy of particle development stage judgment. The detection method of the present invention can accurately determine the particle accumulation morphology and development stage, improve the reliability of detection, and provide technical guarantee for the safe and stable operation of GIS equipment.

[0006] The technical solution adopted by the embodiment of the present invention to solve the technical problem is:

[0007] A method for identifying accumulation and development stages of GIS sub-millimeter particle defects, comprising:

[0008] Step S1, collecting ultrasonic signals of sub-millimeter particles moving inside the GIS and extracting ultrasonic signal characteristic data, wherein the ultrasonic signal characteristic data includes ultrasonic signal amplitude, ultrasonic signal repetition rate, and ultrasonic signal spectrum distribution;

[0009] Step S2, collecting the pulse current signal and extracting the pulse characteristics, wherein the pulse characteristic data includes the pulse signal amplitude, the pulse signal frequency, and the PRPD spectrum of the pulse signal;

[0010] Step S3, verifying the reliability of the ultrasonic signal characteristic data and the pulse characteristic data, and removing abnormal data;

[0011] Step S4, establishing a feature matrix based on the ultrasonic signal feature data and the pulse feature data, inputting the feature matrix into a multi-layer convolutional neural network (CNN) model for analysis, and obtaining the accumulation morphology and development stage of the particles in the GIS, wherein the accumulation morphology and development stage include a small vertical jump morphology in the early stage, a large vertical jump morphology in the middle stage, a horizontal diffusion morphology in the late stage, and a fuzzy state.

[0012] Preferably, the step S3 comprises:

[0013] First, multiple sampling window averaging is performed: within a certain period of time, the ultrasonic signal and the pulse current signal are averaged by multiple sampling windows to smooth the signal fluctuations. Assuming that N samplings are performed in each sampling window, the average value of the data in the window is calculated.

[0014]

[0015] Among them, x i represents the i-th sampling value, N is the number of sampling times;

[0016] Then anomaly detection: normal data patterns are constructed from the training set, and metrics such as cosine similarity or Euclidean distance are used to evaluate the deviation of real-time data from normal patterns:

[0017]

[0018] Where x and y represent the feature vectors of real-time data and normal mode respectively, and D(x,y) is the Euclidean distance between them;

[0019] If D exceeds the set threshold, it is determined to be abnormal data and the abnormal data is removed.

[0020] Preferably, the step S4 comprises:

[0021] Establishing the characteristic matrix based on the ultrasonic signal characteristic data and the pulse characteristic data;

[0022] A multi-layer convolutional neural network (CNN) model is used to perform deep learning training on the feature matrix, and the multi-layer convolutional neural network (CNN) model is composed of an input layer, a convolution layer, a pooling layer, and a fully connected layer in sequence; the feature matrix is ​​sent to the input layer, the convolution layer performs convolution operations on the feature matrices in the time domain and the frequency domain to extract motion morphological features, the pooling layer downsamples the motion morphological features and then sends them to the fully connected layer, the fully connected layer performs morphological recognition, and outputs classification results of the accumulated morphology and development stage.

[0023] Preferably, the small vertical jump pattern in the early stage has three signal characteristics at the same time: the ultrasonic signal amplitude is less than 20dB, the ultrasonic signal repetition rate is greater than 15 times / s, and the pulse signal amplitude is less than 10pC;

[0024] The large vertical jump pattern in the middle stage has three signal characteristics at the same time: the ultrasonic signal amplitude is in the range of [20,30]dB, the ultrasonic signal repetition rate is in the range of [5,15] times / s, and the pulse signal amplitude is greater than 10pC;

[0025] The horizontal diffusion morphology at the end stage has three signal characteristics at the same time: the ultrasonic signal amplitude is greater than 30dB, the ultrasonic signal repetition rate is less than 5 times / s, and the pulse signal amplitude is less than 10pC;

[0026] The fuzzy state refers to a state that does not conform to the early small vertical jump pattern, the middle large vertical jump pattern, and the late horizontal diffusion pattern.

[0027] Preferably, after step S4, the method further includes:

[0028] Step S5, further performing risk assessment based on the classification results of the accumulation morphology and development stage of the particles.

[0029] Preferably, the step S4 comprises:

[0030] Step S51: according to the average amplitude value of the ultrasonic signal The particle size d is obtained and scored to obtain the size score; d is calculated as:

[0031]

[0032] In the formula, k is the coefficient and n is the exponent;

[0033] Particle size satisfies d>d high For large particles, the size score is 3;high and d medium is the particle size threshold;

[0034] The particle size meets d medium ≤d≤d high It is a medium particle with a size score of 2;

[0035] The particle size satisfies d<d medium For small particles, the size score is 1;

[0036] Step S52, obtaining a beating amplitude score according to the classification results of the accumulation form and development stage:

[0037] The ultrasonic signal amplitude in the range of (25dB, 30dB] corresponds to a beating amplitude score of 3;

[0038] The corresponding beating amplitude score for the ultrasonic signal amplitude in the interval of (15dB, 25dB] is 2;

[0039] The corresponding beating amplitude score when the ultrasonic signal amplitude is in the interval of [0dB, 15dB] is 1;

[0040] The horizontal diffusion pattern at the end stage corresponds to a beating amplitude score of 0.5;

[0041] Step S53, obtaining a diffusion intensity score according to the classification results of the accumulation morphology and development stage:

[0042] The diffusion intensity score corresponding to the ultrasound signal repetition rate in the interval of [1,5) times / s is 3;

[0043] The diffusion intensity score corresponding to the ultrasonic signal repetition rate in the interval [0,1) times / s is 2;

[0044] The diffusion intensity score corresponding to the small vertical jump pattern in the early stage and the large vertical jump pattern in the middle stage is 1;

[0045] Step S54, calculate the risk score and output the risk assessment result; wherein, risk score = a×size score+b×jump amplitude score+c×diffusion intensity score, a+b+c=1; when the risk score is greater than 2.5, the risk assessment result is high risk, and prevention and control measures should be taken immediately; when the risk score is between 1.5 and 2.5, the risk assessment result is medium risk, and monitoring and detection should be strengthened; when the risk score is less than 1.5, the risk assessment result is low risk, and the regular detection frequency should be maintained; in addition, the risk assessment result of the fuzzy form is an unknown result.

[0046] A GIS sub-millimeter particle defect accumulation and development stage identification device, comprising:

[0047] Ultrasonic signal acquisition module, used to collect ultrasonic signals of sub-millimeter particles moving inside the GIS. The acquisition position is located on the outer surface of the bottom of the GIS device shell, facing the gathering position of the sub-millimeter particles.

[0048] Pulse current signal module, used for collecting pulse current signals;

[0049] The data analysis and monitoring module is used to implement the aforementioned method.

[0050] It can be seen from the above technical scheme that the method and device for identifying the accumulation morphology and development stage of GIS submillimeter particle defects provided by the embodiment of the present invention first collects the ultrasonic signal of the submillimeter particles when they move inside the GIS and extracts the ultrasonic signal feature data; collects the pulse current signal and extracts the pulse feature; verifies the reliability of the ultrasonic signal feature data and the pulse feature data, and removes the abnormal data; establishes a feature matrix based on the ultrasonic signal feature data and the pulse feature data, and inputs it into the multi-layer convolutional neural network CNN model for analysis, and obtains the accumulation morphology and development stage of the particles in the GIS, which includes a small vertical jump morphology in the early stage, a large vertical jump morphology in the middle stage, and a horizontal diffusion morphology in the late stage. The detection method of the present invention can realize the accurate judgment of the accumulation morphology and development stage of particles, improve the reliability of detection, and provide technical guarantee for the safe and stable operation of GIS equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 The flowchart of the method for identifying the accumulation morphology and development stage of GIS sub-millimeter particle defects of the present invention. DETAILED DESCRIPTION

[0052] The technical scheme and technical effects of the present invention are further elaborated in detail below in conjunction with the accompanying drawings of the present invention.

[0053] The GIS sub-millimeter particle defect accumulation and development stage identification method provided by the present invention can simultaneously consider the pulse current signal and ultrasonic signal caused by partial discharge, and realize the identification of the accumulation form and development stage of sub-millimeter particle defects in GIS. The present invention monitors the movement process of sub-millimeter metal particles inside GIS in real time, and combines the characteristic changes of pulse current signals and ultrasonic signals to judge the accumulation form and development stage of particles. The particle movement can be divided into three accumulation forms: small vertical jump, large vertical jump and horizontal diffusion, which correspond to the early, middle and late stages of the development stage respectively.

[0054] refer to Figure 1As shown, the method of the present invention is based on the acoustic-electrical joint identification method, which integrates and analyzes the time-frequency characteristics of the ultrasonic signal with the statistical characteristics of the pulse current signal, so as to more accurately identify the different accumulation forms and development stages of the particles. The specific implementation steps include:

[0055] Step S1, collecting ultrasonic signals of sub-millimeter particles moving inside the GIS, and performing time-frequency domain analysis on the ultrasonic signals to extract ultrasonic signal feature data, wherein the ultrasonic signal feature data includes ultrasonic signal amplitude, ultrasonic signal repetition rate, and ultrasonic signal spectrum distribution;

[0056] (1) Amplitude (A): During the collision and jumping of particles, the amplitude gradually increases with the movement stage;

[0057] (2) Repetition rate (f): The repetition rate refers to the frequency of the particles touching the GIS shell, which is usually higher during small-amplitude pulsation and decreases when entering the horizontal diffusion stage;

[0058] (3) Spectral distribution (PSD): The ultrasonic signal is Fourier transformed and the spectral power density (PSD) distribution is analyzed to obtain the motion frequency characteristics of the particles. This characteristic is closely related to the size, speed and motion shape of the particles.

[0059] Step S2, during the particle movement and discharge process, the current signal is monitored by the pulse current method (the current signal may be from the ground wire of the GIS equipment), the pulse current signal is collected and the pulse characteristics are extracted, and the pulse characteristic data includes the pulse signal amplitude, the pulse signal frequency, and the PRPD spectrum of the pulse signal;

[0060] (1) Pulse amplitude (P): reflects the discharge intensity of particles in different motion forms. The amplitude decreases significantly in the horizontal diffusion stage.

[0061] (2) Pulse frequency (Fp): the repetition frequency of the pulse current signal, which is affected by the frequency of the particles touching the shell;

[0062] (3) PRPD spectrum: The PRPD spectrum of the statistical pulse signal is used to characterize the changing characteristics of the pulse signal at different development stages.

[0063] Step S3, verifying the reliability of the ultrasonic signal characteristic data and the pulse characteristic data, and removing abnormal data;

[0064] Step S4, establishing a feature matrix based on the ultrasonic signal feature data and the pulse feature data, integrating the amplitude, repetition rate, spectral distribution of the ultrasonic signal with the amplitude, frequency, and shape factor of the pulse current signal into a feature matrix. By constructing the feature matrix, the motion and discharge characteristics of the particles in different accumulation stages can be more comprehensively described; the feature matrix is ​​input into a multi-layer convolutional neural network CNN model for analysis, and the convolutional neural network is used to perform deep learning training on the feature matrix. The CNN model automatically identifies the time-frequency pattern in the feature matrix, and then determines the motion state of the particles, and obtains the accumulation morphology and development stage of the particles in the GIS. The accumulation morphology and development stage include a small vertical jump morphology in the early stage, a large vertical jump morphology in the middle stage, a horizontal diffusion morphology in the late stage, and a fuzzy state.

[0065] Step S3 includes:

[0066] First, multiple sampling window averaging is performed: within a certain period of time, the ultrasonic signal and the pulse current signal are averaged by multiple sampling windows to smooth the signal fluctuations. Assuming that N samplings are performed in each sampling window, the average value of the data in the window is calculated.

[0067]

[0068] Among them, x i represents the i-th sampling value, N is the number of sampling times;

[0069] Then anomaly detection: normal data patterns are constructed from the training set, and metrics such as cosine similarity or Euclidean distance are used to evaluate the deviation of real-time data from normal patterns:

[0070]

[0071] Where x and y represent the feature vectors of real-time data and normal mode respectively, and D(x,y) is the Euclidean distance between them;

[0072] If D exceeds the set threshold, it is determined as abnormal data and is removed.

[0073] Step S4 includes:

[0074] Establishing a feature matrix based on ultrasonic signal feature data and pulse feature data;

[0075] The multi-layer convolutional neural network (CNN) model is used to perform deep learning training on the feature matrix. The multi-layer convolutional neural network (CNN) model consists of an input layer, a convolutional layer, a pooling layer, and a fully connected layer in sequence:

[0076] Input layer: accepts input of feature matrix;

[0077] Convolution layer: Perform convolution operations on the features in the time domain and frequency domain to extract motion morphology features;

[0078] Pooling layer: downsample the features to reduce the impact of noise; then send them to the fully connected layer

[0079] Fully connected layer: performs morphological recognition through feature combination and outputs the classification results of the accumulated morphology.

[0080] Accumulation pattern classification and stage judgment:

[0081] The small vertical jump pattern in the early stage has three signal characteristics at the same time: the ultrasonic signal amplitude is less than 20dB, the ultrasonic signal repetition rate is greater than 15 times / s, and the pulse signal amplitude is less than 10pC; under this condition, the preliminary judgment accuracy of the CNN model can reach more than 90%.

[0082] The large vertical jump pattern in the middle stage has three signal characteristics at the same time: the ultrasonic signal amplitude is in the range of [20,30]dB, the ultrasonic signal repetition rate is in the range of [5,15] times / s, and the pulse signal amplitude is greater than 10pC; when these signal characteristics are met, the recognition accuracy of the CNN model is further improved to 95%.

[0083] The horizontal diffusion form in the final stage has three signal characteristics at the same time: the ultrasonic signal amplitude is greater than 30dB, the ultrasonic signal repetition rate is less than 5 times / s, and the pulse signal amplitude is less than 10pC; in the final stage, the particle movement gradually shifts from vertical jumping to horizontal diffusion. The signal changes in this stage are more significant, the ultrasonic signal amplitude increases to more than 30dB, and the repetition rate decreases to less than 5 times per second; the amplitude of the pulse current signal also gradually decreases to less than 10pC again, at which time the system determines that the particles have entered the horizontal diffusion stage. In the horizontal diffusion stage, if only the ultrasonic signal repetition rate decreases or the pulse current signal amplitude decreases, the system will assign a 50% possibility of the horizontal diffusion stage; when both the ultrasonic signal and the pulse current signal decrease, the system confidence reaches 99%.

[0084] The fuzzy state refers to a state that does not conform to the early small vertical jump pattern, the middle large vertical jump pattern, and the late horizontal diffusion pattern.

[0085] Preferably, after step S4, the method further includes:

[0086] Step S5, further conduct risk assessment based on the classification results of the accumulation form and development stage of the particles. The system generates the discrimination results of three accumulation forms through the output and reliability verification of the CNN model, and further conducts risk assessment based on the size and movement characteristics of the particles. By outputting the judgment results of the movement stage, the equipment can provide early warning and preventive maintenance for the accumulation risk of sub-millimeter particles in GIS. The specific implementation includes:

[0087] Step S51: according to the average amplitude value of the ultrasonic signal The particle size d is obtained and scored to obtain the size score; d is calculated as:

[0088]

[0089] In the formula, k and n are coefficients, and n is an exponent;

[0090] Particle size satisfies d>d high For large particles, the size score is 3; high and d medium is the particle size threshold;

[0091] The particle size meets d medium ≤d≤d high It is a medium particle with a size score of 2;

[0092] The particle size satisfies d<d medium For small particles, the size score is 1;

[0093] Step S52, deriving a beat amplitude score based on the classification results of the accumulation form and development stage:

[0094] The ultrasonic signal amplitude in the range of (25dB, 30dB] corresponds to a beating amplitude score of 3;

[0095] The corresponding beating amplitude score for the ultrasonic signal amplitude in the interval of (15dB, 25dB] is 2;

[0096] The corresponding beating amplitude score when the ultrasonic signal amplitude is in the interval of [0dB, 15dB] is 1;

[0097] The horizontal diffusion pattern at the end stage corresponds to a beating amplitude score of 0.5;

[0098] Step S53, obtaining a diffusion intensity score according to the classification results of the accumulation morphology and development stage:

[0099] The diffusion intensity score corresponding to the ultrasound signal repetition rate in the interval of [1,5) times / s is 3;

[0100] The diffusion intensity score corresponding to the ultrasonic signal repetition rate in the interval [0,1) times / s is 2;

[0101] The diffusion intensity score corresponding to the small vertical jump pattern in the early stage and the large vertical jump pattern in the middle stage is 1;

[0102] Step S54, calculate the risk score and output the risk assessment result; wherein, risk score = a×size score+b×jump amplitude score+c×diffusion intensity score, a+b+c=1; when the risk score is greater than 2.5, the risk assessment result is high risk, and prevention and control measures should be taken immediately; when the risk score is between 1.5 and 2.5, the risk assessment result is medium risk, and monitoring and detection should be strengthened; when the risk score is less than 1.5, the risk assessment result is low risk, and the regular detection frequency should be maintained; in addition, the risk assessment result of the fuzzy form is an unknown result.

[0103] A GIS sub-millimeter particle defect accumulation and development stage identification device, comprising:

[0104] Ultrasonic signal acquisition module, used to collect ultrasonic signals of sub-millimeter particles moving inside the GIS. The acquisition position is located on the outer surface of the bottom of the GIS device shell, facing the gathering position of the sub-millimeter particles.

[0105] The pulse current signal module is used to collect pulse current signals; the collection location may be the grounding location of the GIS equipment, such as the pulse current of the grounding wire;

[0106] The data analysis and monitoring module is used to implement the aforementioned method.

[0107] This device is designed to combine pulse current signals with ultrasonic signals to improve the accuracy of particle development stage judgment, achieve accurate determination of particle accumulation morphology and development stage, improve detection reliability, and provide technical guarantee for the safe and stable operation of GIS equipment.

[0108] According to the embodiments disclosed in the present invention, the present invention also discloses an electronic device, a readable storage medium and a computer program product. The electronic device is intended to represent various forms of digital computers, including a computing unit, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) or a computer program loaded from a storage unit into a random access memory (RAM). In the RAM, various programs and data required for the operation of the device can also be stored. The computing unit, ROM and RAM are connected to each other via a bus. The input / output (I / O) interface is also connected to the bus.

[0109] Multiple components in the device are connected to the I / O interface, including: input unit, output unit, storage unit, and communication unit. The communication unit allows the device to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0110] The computing unit may be a variety of general and / or special processing components with processing and computing capabilities to execute the various methods and processes described above, such as the GIS sub-millimeter particle defect accumulation morphology and development stage identification method.

[0111] What is disclosed above is only a preferred embodiment of the present invention, which certainly cannot be used to limit the scope of rights of the present invention. A person skilled in the art can understand that all or part of the processes of the above embodiments and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A method for identifying accumulation and development stages of GIS sub-millimeter particle defects, characterized in that: include: Step S1, collecting ultrasonic signals of sub-millimeter particles moving inside the GIS and extracting ultrasonic signal characteristic data, wherein the ultrasonic signal characteristic data includes ultrasonic signal amplitude, ultrasonic signal repetition rate, and ultrasonic signal spectrum distribution; Step S2, collecting the pulse current signal and extracting the pulse characteristics, wherein the pulse characteristic data includes the pulse signal amplitude, the pulse signal frequency, and the PRPD spectrum of the pulse signal; Step S3, verifying the reliability of the ultrasonic signal characteristic data and the pulse characteristic data, and removing abnormal data; Step S4, establishing a feature matrix based on the ultrasonic signal feature data and the pulse feature data, inputting the feature matrix into a multi-layer convolutional neural network (CNN) model for analysis, and obtaining the accumulation morphology and development stage of the particles in the GIS, wherein the accumulation morphology and development stage include a small vertical jump morphology in the early stage, a large vertical jump morphology in the middle stage, a horizontal diffusion morphology in the late stage, and a fuzzy state.

2. The GIS sub-millimeter particle defect accumulation and development stage identification method according to claim 1, characterized in that: The step S3 comprises: First, multiple sampling window averaging is performed: within a certain period of time, the ultrasonic signal and the pulse current signal are averaged by multiple sampling windows to smooth the signal fluctuations. Assuming that N samplings are performed in each sampling window, the average value of the data in the window is calculated. Among them, x i represents the i-th sampling value, N is the number of sampling times; Then anomaly detection: normal data patterns are constructed from the training set, and metrics such as cosine similarity or Euclidean distance are used to evaluate the deviation of real-time data from normal patterns: Where x and y represent the feature vectors of real-time data and normal mode respectively, and D(x,y) is the Euclidean distance between them; If D exceeds the set threshold, it is determined to be abnormal data and the abnormal data is removed.

3. The GIS sub-millimeter particle defect accumulation and development stage identification method according to claim 2, characterized in that: The step S4 comprises: Establishing the characteristic matrix based on the ultrasonic signal characteristic data and the pulse characteristic data; A multi-layer convolutional neural network (CNN) model is used to perform deep learning training on the feature matrix, and the multi-layer convolutional neural network (CNN) model is composed of an input layer, a convolution layer, a pooling layer, and a fully connected layer in sequence; the feature matrix is ​​sent to the input layer, the convolution layer performs convolution operations on the feature matrices in the time domain and the frequency domain to extract motion morphological features, the pooling layer downsamples the motion morphological features and then sends them to the fully connected layer, the fully connected layer performs morphological recognition, and outputs classification results of the accumulated morphology and development stage.

4. The GIS sub-millimeter particle defect accumulation and development stage identification method according to claim 3, characterized in that: The small vertical jump pattern in the early stage has three signal characteristics at the same time: the ultrasonic signal amplitude is less than 20dB, the ultrasonic signal repetition rate is greater than 15 times / s, and the pulse signal amplitude is less than 10pC; The large vertical jump pattern in the middle stage has three signal characteristics at the same time: the ultrasonic signal amplitude is in the range of [20,30]dB, the ultrasonic signal repetition rate is in the range of [5,15] times / s, and the pulse signal amplitude is greater than 10pC; The horizontal diffusion morphology at the end stage has three signal characteristics at the same time: the ultrasonic signal amplitude is greater than 30dB, the ultrasonic signal repetition rate is less than 5 times / s, and the pulse signal amplitude is less than 10pC; The fuzzy state refers to a state that does not conform to the early small vertical jump pattern, the middle large vertical jump pattern, and the late horizontal diffusion pattern.

5. The GIS sub-millimeter particle defect accumulation and development stage identification method according to claim 1, characterized in that: After step S4, the method further includes: Step S5, further performing risk assessment based on the classification results of the accumulation morphology and development stage of the particles.

6. The GIS sub-millimeter particle defect accumulation and development stage identification method according to claim 5, characterized in that: The step S5 comprises: Step S51: according to the average amplitude value of the ultrasonic signal The particle size d is obtained and scored to obtain the size score; d is calculated as: In the formula, k and n are coefficients, and n is an exponent; Particle size satisfies d>d high For large particles, the size score is 3; high and d medium is the amplitude threshold; The particle size meets d medium ≤d≤d high It is a medium particle with a size score of 2; The particle size satisfies d<d medium For small particles, the size score is 1; Step S52, obtaining a beating amplitude score according to the classification results of the accumulation form and development stage: The ultrasonic signal amplitude in the range of (25dB, 30dB] corresponds to a beating amplitude score of 3; The corresponding beating amplitude score for the ultrasonic signal amplitude in the interval of (15dB, 25dB] is 2; The corresponding beating amplitude score when the ultrasonic signal amplitude is in the interval of [0dB, 15dB] is 1; The horizontal diffusion pattern at the end stage corresponds to a beating amplitude score of 0.5; Step S53, obtaining a diffusion intensity score according to the classification results of the accumulation morphology and development stage: The diffusion intensity score corresponding to the ultrasound signal repetition rate in the interval of [1,5) times / s is 3; The diffusion intensity score corresponding to the ultrasonic signal repetition rate in the interval [0,1) times / s is 2; The diffusion intensity score corresponding to the small vertical jump pattern in the early stage and the large vertical jump pattern in the middle stage is 1; Step S54, calculate the risk score and output the risk assessment result; wherein, risk score = a×size score+b×jump amplitude score+c×diffusion intensity score, a+b+c=1; when the risk score is greater than 2.5, the risk assessment result is high risk, and prevention and control measures should be taken immediately; when the risk score is between 1.5 and 2.5, the risk assessment result is medium risk, and monitoring and detection should be strengthened; when the risk score is less than 1.5, the risk assessment result is low risk, and the regular detection frequency should be maintained; in addition, the risk assessment result of the fuzzy form is an unknown result.

7. A GIS sub-millimeter particle defect accumulation and development stage identification device, characterized in that: include: Ultrasonic signal acquisition module, used to collect ultrasonic signals of sub-millimeter particles moving inside the GIS. The acquisition position is located on the outer surface of the bottom of the GIS device shell, facing the gathering position of the sub-millimeter particles. Pulse current signal module, used for collecting pulse current signals; A data analysis and monitoring module, used to implement the method described in any one of claims 1-6.

8. An electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 6.

9. A non-transitory computer-readable storage medium storing computer instructions, wherein: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-6.

10. A computer program product, comprising a computer program, which, when executed by a processor, implements the method according to any one of claims 1 to 6.

Images