A method and system for detecting the vacuum degree of a vacuum interrupter

Abstract

The present application relates to the technical field of electrical engineering, more particularly to a kind of vacuum interrupter vacuum degree detection method and system, comprising: the voltage signal of vacuum interrupter under different vacuum degree when breaking is collected, vacuum degree identification model is constructed, vacuum degree identification model is trained using dataset, and the vacuum degree identification model after training is obtained;Vacuum degree of vacuum interrupter is detected using the vacuum degree identification model after training.The detection method of the present application uses the dataset formed by the features extracted from the data collected in the experiment to train the vacuum degree identification model, significantly improves the accuracy of the vacuum degree classification of the vacuum degree identification model.The detection method of the present application does not need to contact the vacuum interrupter, does not need to change the structure of the vacuum interrupter, and will not threaten the insulation of each part, to realize real-time, accurate and reliable online detection of the vacuum interrupter.

Description

Technical Field

[0001] This invention relates to the field of electrical engineering technology, and more specifically to a method and system for detecting the vacuum level of a vacuum interrupter. Background Technology

[0002] Vacuum circuit breakers utilize vacuum as the arc-extinguishing and insulating medium, offering advantages such as light weight, small size, long lifespan, excellent breaking performance, high reliability, and environmental friendliness. Currently, they hold a dominant position in medium-voltage power systems and are developing towards high-voltage, high-current applications. The vacuum interrupter, as the core component of a vacuum circuit breaker, plays a crucial role in the successful interruption of the circuit.

[0003] Vacuum interrupters are an important component of vacuum circuit breakers, and their vacuum level testing is crucial for normal operation. National standards require a minimum vacuum level of 6.6 × 10⁻⁶ for interrupters. -2 Currently, existing detection methods are broadly categorized into online and offline detection. Online detection mainly includes methods such as coupling capacitance, photoelectric conversion, and rotating electric field probe detection. The coupling capacitance method is relatively mature, simple in structure, and low in cost, and is currently used in many applications; however, its accuracy needs further improvement. The photoelectric conversion method requires a highly accurate probe; however, due to the harsh operating environment of vacuum circuit breakers, using this method for online monitoring would significantly impact the probe's accuracy, and the optical components themselves inherently have stability and measurement error issues. The rotating electric field probe detection method involves cumbersome subsequent data processing and complex circuitry, which can affect detection accuracy to some extent, thus limiting its practicality. Therefore, a vacuum degree detection method is needed that does not require direct contact with the vacuum interrupter, does not alter the interrupter's structure, and does not threaten the insulation of any part of the circuit breaker. Summary of the Invention

[0004] (I) Purpose of the Invention

[0005] The purpose of this invention is to provide a method and system for detecting the vacuum level of a vacuum interrupter that does not require direct contact with the vacuum interrupter, does not require changes to the structure of the vacuum interrupter, and does not threaten the insulation of various parts; and can improve the detection accuracy of the vacuum level of the vacuum interrupter.

[0006] (II) Technical Solution

[0007] To address the above problems, this invention provides a method for detecting the vacuum level of a vacuum interrupter, comprising:

[0008] Collect voltage signals when the vacuum interrupter is switched on under different vacuum levels;

[0009] Perform feature analysis on the voltage signal to extract its features;

[0010] Based on the characteristics of the voltage signal, a dataset is constructed;

[0011] Construct a vacuum degree recognition model, and train the vacuum degree recognition model using the dataset to obtain the trained vacuum degree recognition model;

[0012] The vacuum level of the vacuum interrupter is detected using a trained vacuum level recognition model.

[0013] In another aspect of the present invention, preferably, acquiring the voltage signal when the vacuum interrupter is switched on at different vacuum levels includes:

[0014] Set up the data acquisition channel;

[0015] The acquisition channel includes a capacitor, a switch, a vacuum interrupter, and an exhaust component;

[0016] The capacitor, switch, and vacuum interrupter are connected to form a circuit;

[0017] The capacitor is used to supply power to the circuit, and the switch is used to control the opening and closing of the circuit;

[0018] The exhaust component is connected to the vacuum interrupter, and the vacuum interrupter achieves a vacuum state through the exhaust component.

[0019] In another aspect of the present invention, preferably, the step of collecting the voltage signal when the vacuum interrupter is switched on under different vacuum levels further includes:

[0020] Seal the vacuum interrupter and evacuate it to a vacuum state;

[0021] Charge the capacitor, disconnect the switch, and obtain the voltage signal under vacuum conditions;

[0022] The vacuum state includes abnormal vacuum state and normal vacuum state;

[0023] The vacuum state is cyclically set to abnormal vacuum state and normal vacuum state for voltage signal acquisition.

[0024] In another aspect of the invention, preferably, the abnormal vacuum state is a pressure higher than 6.6 × 10⁻⁶. -2 Pa;

[0025] The normal vacuum state is defined as a pressure below 6.6 × 10⁻⁶. -2 Pa.

[0026] In another aspect of the present invention, preferably, feature analysis is performed on the voltage signal to extract features of the voltage signal, including:

[0027] Perform time-domain waveform analysis on the voltage signal to obtain its time-domain characteristics;

[0028] The time-domain features are converted into frequency-domain features using the Fast Fourier Transform algorithm, and the spectral information is extracted. The spectral information is a feature of the voltage signal.

[0029] In another aspect of the present invention, preferably, the vacuum degree recognition model is based on a Bayesian neural network algorithm;

[0030] The Bayesian neural network algorithm includes a feedforward neural network structure and Bayesian inference;

[0031] The feedforward neural network structure consists of an input layer, a hidden layer, and an output layer. The input data of the input layer includes spectral information, and the output data of the output layer indicates whether the vacuum degree is abnormal.

[0032] The Bayesian inference uses Bayes' theorem to estimate the posterior distribution of the parameters:

[0033]

[0034] Where p(θ|D) is the posterior distribution of weight θ given data D; p(D|θ) is the likelihood function, representing the probability of training data D occurring given weight θ; p(θ) is the prior distribution of the weights, which is a Gaussian distribution; and p(D) is the evidence, i.e. the marginal probability of the data, which is set as a constant.

[0035] In another aspect of the present invention, preferably, training the vacuum degree recognition model using the dataset includes:

[0036] The dataset is divided into a training set and a test set;

[0037] Using the characteristics of the voltage signal as input and whether the vacuum degree is abnormal as output, the vacuum degree identification model is trained using the training set to obtain a first vacuum degree identification model.

[0038] The accuracy of the first vacuum degree recognition model is verified using the test set. If the accuracy meets a preset threshold, the trained vacuum degree recognition model is obtained.

[0039] If the accuracy does not meet the preset threshold, the parameters of the first vacuum degree recognition model are adjusted until the accuracy meets the preset threshold, and the trained vacuum degree recognition model is obtained.

[0040] In another aspect of the present invention, preferably, the vacuum degree identification model is trained using the training set, with the characteristics of the voltage signal as input and whether the vacuum degree is abnormal as output, to obtain a first vacuum degree identification model, including:

[0041] During training, the model parameters are adjusted using gradient descent; the gradient descent method is calculated using the following formula:

[0042] θ 1 =θ 0 -α▽J(θ)

[0043] Where θ is the model parameter, α is the learning rate, and J(θ) is the loss function.

[0044] In another aspect of the present invention, preferably, the method further includes:

[0045] The test results of the vacuum degree of the vacuum interrupter are displayed;

[0046] The test results include abnormal and normal;

[0047] If the detection result is abnormal, an alarm will be triggered;

[0048] If the test result is normal, continue the test.

[0049] In another aspect, preferably, a system for detecting the vacuum level of a vacuum interrupter includes:

[0050] Acquisition module: Acquires voltage signals when the vacuum interrupter is switched on at different vacuum levels;

[0051] Extraction module: Performs feature analysis on the voltage signal and extracts the features of the voltage signal;

[0052] Construction module: Constructs a dataset based on the characteristics of the voltage signal;

[0053] Training module: Construct a vacuum degree recognition model, train the vacuum degree recognition model using the dataset, and obtain the trained vacuum degree recognition model;

[0054] Detection module: The vacuum level of the vacuum interrupter is detected using a trained vacuum level recognition model.

[0055] (III) Beneficial Effects

[0056] The above-described technical solution of the present invention has the following beneficial technical effects:

[0057] The detection method of this invention utilizes a dataset formed by extracting features from experimentally collected data to train a vacuum degree identification model, significantly improving the accuracy of vacuum degree classification. This method does not require contact with the vacuum interrupter, does not alter its structure, and poses no threat to the insulation of any parts, enabling real-time, accurate, and reliable online detection of the vacuum interrupter. Attached Figure Description

[0058] Figure 1 This is an overall flowchart of one embodiment of the present invention;

[0059] Figure 2 This is a schematic diagram of the acquisition channel according to an embodiment of the present invention;

[0060] Figure 3 This is a schematic diagram of the detection system structure according to an embodiment of the present invention;

[0061] Figure label:

[0062] 1: Capacitor, 2: Inductor, 3: Switch, 4: Vacuum interrupter, 5: Oscilloscope, 6: High-voltage coil. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments and the accompanying drawings. It should be understood that these descriptions are merely exemplary and not intended to limit the scope of the invention. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concept of the invention.

[0064] The accompanying drawings show structural schematic diagrams according to embodiments of the present invention. These drawings are not drawn to scale, and some details have been enlarged for clarity, and some details may have been omitted. The shapes of the various regions and layers shown in the drawings, as well as their relative sizes and positional relationships, are merely exemplary and may deviate from reality due to manufacturing tolerances or technical limitations. Furthermore, those skilled in the art can design regions / layers with different shapes, sizes, and relative positions as needed.

[0065] Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0066] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0067] The invention will now be described in more detail with reference to the accompanying drawings. In the various drawings, the same elements are indicated by similar reference numerals. For clarity, the various parts in the drawings are not drawn to scale.

[0068] Example 1

[0069] A method for detecting the vacuum level of a vacuum interrupter. Figure 1 An overall flowchart of one embodiment of the present invention is shown, as follows: Figure 1As shown, it includes:

[0070] The step involves collecting voltage signals from the vacuum interrupter at different vacuum levels during the interruption process. This step aims to obtain voltage response data of the vacuum interrupter under different vacuum conditions. Different vacuum levels can be simulated using experimental equipment, and the vacuum interrupter can be interrupted at each vacuum level. Simultaneously, a high-precision voltage sensor or oscilloscope is used to record the voltage signal at the moment of interruption. During the experiment, it is essential to ensure the consistency of experimental conditions, such as temperature and pressure, to reduce the influence of external factors on the experimental results.

[0071] Feature analysis is performed on the voltage signal to extract its features; key information reflecting changes in vacuum level is extracted from the acquired voltage signal. Signal processing techniques, such as Fourier transform and wavelet transform, can be used to perform frequency domain or time-frequency domain analysis on the voltage signal to extract characteristic parameters such as amplitude, frequency, and phase.

[0072] Based on the characteristics of the voltage signal, a dataset is constructed; the extracted feature parameters are combined with the corresponding vacuum degree values ​​to form a dataset for subsequent model training. Each feature parameter under a vacuum degree is treated as a sample, and whether the vacuum degree is abnormal is used as a label, thus constructing the dataset required for supervised learning.

[0073] A vacuum degree recognition model is constructed and trained using the dataset to obtain the trained vacuum degree recognition model. Overtraining enables the model to accurately identify voltage signal characteristics under different vacuum degrees. During training, the model continuously adjusts its parameters to minimize prediction error.

[0074] The vacuum level of the vacuum interrupter is detected using a trained vacuum level recognition model. The trained model is then applied to actual testing to quickly and accurately determine the vacuum level of the vacuum interrupter. In practical applications, the voltage signal of the vacuum interrupter to be tested is first acquired, and feature parameters are extracted. These feature parameters are then input into the trained vacuum level recognition model, which outputs a predicted vacuum level indicating whether it is abnormal.

[0075] Furthermore, in this embodiment, the acquisition of voltage signals when the vacuum interrupter is switched on at different vacuum levels includes:

[0076] Set up the data acquisition channel; Figure 2 A schematic diagram of the acquisition channel according to an embodiment of the present invention is shown, as follows: Figure 2 As shown,

[0077] The acquisition channel includes capacitor 1, switch 3, vacuum interrupter 4, and exhaust component;

[0078] The capacitor 1, switch 3, and vacuum interrupter 4 are connected to form a circuit;

[0079] The capacitor 1 is used to supply power to the circuit, and the switch 3 is used to control the opening and closing of the circuit;

[0080] The exhaust component is connected to the vacuum interrupter 4, and the vacuum interrupter achieves a vacuum state through the exhaust component.

[0081] The acquisition channel also includes an inductor 2, an oscilloscope 5, and a high-voltage coil 6. The inductor 2 and the high-voltage coil 6 are arranged in the circuit, and the oscilloscope 5 is connected to the high-voltage coil 6 for acquiring voltage signals.

[0082] The method for collecting voltage signals when the vacuum interrupter is switched on under different vacuum levels also includes:

[0083] Seal the vacuum interrupter and evacuate it to a vacuum state; by opening the mechanical pump and molecular valve of the exhaust component, evacuate the cavity in the detachable interrupter to an abnormal vacuum state;

[0084] Charge the capacitor, disconnect the switch, and open the circuit breaker using the operating mechanism; acquire the voltage signal under vacuum conditions;

[0085] The vacuum state includes abnormal vacuum state and normal vacuum state;

[0086] The vacuum state is cyclically set to abnormal vacuum state and normal vacuum state for voltage signal acquisition.

[0087] The abnormal vacuum state is defined as a pressure higher than 6.6 × 10⁻⁶. -2 Pa;

[0088] The normal vacuum state is defined as a pressure below 6.6 × 10⁻⁶. -2 Pa.

[0089] Furthermore, in this embodiment, feature analysis is performed on the voltage signal, and the features of the voltage signal are extracted, including:

[0090] Time-domain waveform analysis is performed on the voltage signal to obtain its time-domain characteristics, including the extraction of parameters such as peak value, mean value, RMS value, and waveform factor. These parameters can reflect the transient characteristics of the cathode spot under different vacuum conditions.

[0091] The time-domain features are converted into frequency-domain features using a Fast Fourier Transform (FFT) algorithm to extract spectral information, such as the main frequency components, frequency distribution, and energy spectral density. These frequency-domain features help identify the correlation between the frequency characteristics of the arc voltage and the vacuum level.

[0092] Wavelet packet transform is used to perform multi-scale analysis on voltage signals, thereby obtaining the frequency characteristics of the signals at different time scales.

[0093] Furthermore, in this embodiment, the vacuum degree recognition model is based on a Bayesian neural network algorithm.

[0094] The Bayesian neural network algorithm consists of a feedforward neural network and Bayesian inference:

[0095] The feedforward neural network includes an input layer, a hidden layer, and an output layer. The input layer represents the characteristics of the voltage signal, and the output layer represents the vacuum state, including normal and abnormal states.

[0096] Bayesian inference uses Bayes' theorem to estimate the posterior distribution of parameters:

[0097]

[0098] Where p(θ|D) is the posterior distribution of weight θ given data D; p(D|θ) is the likelihood function, representing the probability of training data D occurring given weight θ; p(θ) is the prior distribution of the weights, which is a Gaussian distribution; and p(D) is the evidence, i.e. the marginal probability of the data, which is set as a constant.

[0099] Training the vacuum degree recognition model using the dataset includes:

[0100] The dataset is divided into a training set and a test set; the ratio of the division can be 8:2.

[0101] Using the characteristics of the voltage signal as input and whether the vacuum degree is abnormal as output, the vacuum degree identification model is trained using the training set to obtain a first vacuum degree identification model.

[0102] The accuracy of the first vacuum degree recognition model is verified using the test set. If the accuracy meets a preset threshold, the trained vacuum degree recognition model is obtained. The preset threshold can be set to 90%.

[0103] If the accuracy does not meet the preset threshold, the parameters of the first vacuum degree recognition model are adjusted until the accuracy meets the preset threshold, and the trained vacuum degree recognition model is obtained.

[0104] Using the characteristics of the voltage signal as input and whether the vacuum level is abnormal as output, the vacuum level identification model is trained using the training set to obtain the first vacuum level identification model, which includes:

[0105] During training, gradient descent is used to adjust model parameters in order to minimize prediction error.

[0106] The gradient descent method is based on the following algorithm:

[0107] Where θ is the model parameter, α is the learning rate, and J(θ) is the loss function.

[0108] Furthermore, in this embodiment, the method further includes:

[0109] The interface displays the detection results of the vacuum level of the vacuum interrupter; it can show real-time monitoring data, alarm information, and historical records. This interface can be used on a PC or mobile device, allowing maintenance personnel to easily check the equipment status at any time. It supports data export, allowing users to save monitoring data and analysis results as reports for subsequent analysis and recording.

[0110] The test results include abnormal and normal;

[0111] If the test result is abnormal, an alarm will be triggered; the alarm will be automatically triggered to notify relevant maintenance personnel to ensure timely maintenance and inspection of the equipment.

[0112] If the test result is normal, continue the test.

[0113] The detection method of this invention utilizes a dataset formed by extracting features from experimentally collected data to train a vacuum degree identification model, significantly improving the accuracy of vacuum degree classification. This method does not require contact with the vacuum interrupter, does not alter its structure, and poses no threat to the insulation of any parts, enabling real-time, accurate, and reliable online detection of the vacuum interrupter.

[0114] Example 2

[0115] A system for detecting the vacuum level of a vacuum interrupter. Figure 3 A schematic diagram of the detection system structure according to an embodiment of the present invention is shown, as follows: Figure 3 As shown, it includes:

[0116] Acquisition module: Acquires voltage signals when the vacuum interrupter is switched on at different vacuum levels;

[0117] Extraction module: Performs feature analysis on the voltage signal and extracts the features of the voltage signal;

[0118] Construction module: Constructs a dataset based on the characteristics of the voltage signal;

[0119] Training module: Construct a vacuum degree recognition model, train the vacuum degree recognition model using the dataset, and obtain the trained vacuum degree recognition model;

[0120] Detection module: The vacuum level of the vacuum interrupter is detected using a trained vacuum level recognition model.

[0121] It should be understood that the specific embodiments described above are merely illustrative or explanatory of the principles of the invention and do not constitute a limitation thereof. Therefore, any modifications, equivalent substitutions, improvements, etc., made without departing from the spirit and scope of the invention should be included within the protection scope of the invention. Furthermore, the appended claims are intended to cover all variations and modifications falling within the scope and boundaries of the appended claims, or equivalent forms of such scope and boundaries.

[0122] The above description does not provide detailed explanations of the technical aspects of each layer's patterning and etching. However, those skilled in the art should understand that various methods existing in the prior art can be used to form layers and regions of the desired shape. Furthermore, to form the same structure, those skilled in the art can also design methods that are not entirely identical to those described above.

[0123] The present invention has been described above with reference to embodiments thereof. However, these embodiments are merely illustrative and not intended to limit the scope of the invention. The scope of the invention is defined by the appended claims and their equivalents. Various substitutions and modifications can be made by those skilled in the art without departing from the scope of the invention, and all such substitutions and modifications should fall within the scope of the invention.

[0124] Although embodiments of the present invention have been described in detail, it should be understood that various changes, substitutions, and modifications can be made to the embodiments of the present invention without departing from the spirit and scope of the invention.

[0125] Obviously, the above embodiments are merely illustrative examples for clear explanation and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations here. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for detecting the vacuum degree of a vacuum interrupter, characterized in that, include: Collect voltage signals when the vacuum interrupter is switched on under different vacuum levels; Perform feature analysis on the voltage signal to extract its features; Based on the characteristics of the voltage signal, a dataset is constructed; Construct a vacuum degree recognition model, and train the vacuum degree recognition model using the dataset to obtain the trained vacuum degree recognition model; The vacuum level of the vacuum interrupter is detected using a trained vacuum level recognition model. The voltage signals collected when the vacuum interrupter is switched on at different vacuum levels include: Set up the data acquisition channel; The acquisition channel includes a capacitor (1), a switch (3), a vacuum interrupter (4), and an exhaust component; The capacitor (1), switch (3) and vacuum interrupter are connected to form a circuit; The capacitor (1) is used to supply power to the circuit, and the switch (3) is used to control the opening and closing of the circuit; The exhaust component is connected to the vacuum interrupter (4), and the vacuum interrupter (4) achieves a vacuum state through the exhaust component; Seal the vacuum interrupter (4) and evacuate the vacuum interrupter (4) to a vacuum state; Charge the capacitor (1), disconnect the switch (3), and obtain the voltage signal under the vacuum state; The vacuum state includes abnormal vacuum state and normal vacuum state; The vacuum state is cyclically set to abnormal vacuum state and normal vacuum state for voltage signal acquisition; The acquisition channel also includes an inductor (2), an oscilloscope (5), and a high-voltage coil (6). The inductor (2) and the high-voltage coil (6) are arranged in the circuit. The oscilloscope (5) is connected to the high-voltage coil (6) for acquiring voltage signals. The vacuum degree identification model is based on a Bayesian neural network algorithm; The Bayesian neural network algorithm includes a feedforward neural network structure and Bayesian inference; The feedforward neural network structure consists of an input layer, a hidden layer, and an output layer. The input data of the input layer includes spectral information, and the output data of the output layer indicates whether the vacuum degree is abnormal. The Bayesian inference uses Bayes' theorem to estimate the posterior distribution of the parameters: in, Given data D, the weights The posterior distribution of; It is the likelihood function, representing the likelihood given weights. The probability of training data D occurring at that time; The prior distribution of the weights is a Gaussian distribution; This is evidence, i.e., the marginal probability of the data, which is set as a constant.

2. The method for detecting the vacuum degree of a vacuum interrupter according to claim 1, characterized in that, The abnormal vacuum state is defined as a pressure higher than 6.6 × 10⁻² Pa; The normal vacuum state is when the air pressure is below 6.6 × 10⁻² Pa.

3. The method for detecting the vacuum degree of a vacuum interrupter according to claim 1, characterized in that, Feature analysis is performed on the voltage signal, and the features of the voltage signal are extracted, including: Perform time-domain waveform analysis on the voltage signal to obtain its time-domain characteristics; The time-domain features are converted into frequency-domain features using the Fast Fourier Transform algorithm, and the spectral information is extracted. The spectral information is a feature of the voltage signal.

4. The method for detecting the vacuum degree of a vacuum interrupter according to claim 1, characterized in that, Training the vacuum degree recognition model using the dataset includes: The dataset is divided into a training set and a test set; Using the characteristics of the voltage signal as input and whether the vacuum degree is abnormal as output, the vacuum degree identification model is trained using the training set to obtain a first vacuum degree identification model. The accuracy of the first vacuum degree recognition model is verified using the test set. If the accuracy meets a preset threshold, the trained vacuum degree recognition model is obtained. If the accuracy does not meet the preset threshold, the parameters of the first vacuum degree recognition model are adjusted until the accuracy meets the preset threshold, and the trained vacuum degree recognition model is obtained.

5. The method for detecting the vacuum degree of a vacuum interrupter according to claim 4, characterized in that, Using the characteristics of the voltage signal as input and whether the vacuum level is abnormal as output, the vacuum level identification model is trained using the training set to obtain the first vacuum level identification model, which includes: During training, the model parameters are adjusted using gradient descent, which is calculated using the following formula: in, These are model parameters. It's the learning rate. It is the loss function.

6. The method for detecting the vacuum degree of a vacuum interrupter according to claim 1, characterized in that, The method further includes: The test results of the vacuum degree of the vacuum interrupter are displayed; The test results include abnormal and normal; If the detection result is abnormal, an alarm will be triggered; If the test result is normal, continue the test.

7. A system for detecting the vacuum level of a vacuum interrupter, characterized in that, include: Acquisition module: Acquires voltage signals when the vacuum interrupter is switched on at different vacuum levels; Extraction module: Performs feature analysis on the voltage signal and extracts the features of the voltage signal; Construction module: Constructs a dataset based on the characteristics of the voltage signal; Training module: Construct a vacuum degree recognition model, train the vacuum degree recognition model using the dataset, and obtain the trained vacuum degree recognition model; Detection module: Utilizes a trained vacuum degree recognition model to detect the vacuum degree of the vacuum interrupter. The voltage signals collected when the vacuum interrupter is switched on at different vacuum levels include: Set up the data acquisition channel; The acquisition channel includes a capacitor (1), a switch (3), a vacuum interrupter (4), and an exhaust component; The capacitor (1), switch (3) and vacuum interrupter are connected to form a circuit; The capacitor (1) is used to supply power to the circuit, and the switch (3) is used to control the opening and closing of the circuit; The exhaust component is connected to the vacuum interrupter (4), and the vacuum interrupter (4) achieves a vacuum state through the exhaust component; Seal the vacuum interrupter (4) and evacuate the vacuum interrupter (4) to a vacuum state; Charge the capacitor (1), disconnect the switch (3), and obtain the voltage signal under the vacuum state; The vacuum state includes abnormal vacuum state and normal vacuum state; The vacuum state is cyclically set to abnormal vacuum state and normal vacuum state for voltage signal acquisition; The acquisition channel also includes an inductor (2), an oscilloscope (5), and a high-voltage coil (6). The inductor (2) and the high-voltage coil (6) are arranged in the circuit. The oscilloscope (5) is connected to the high-voltage coil (6) for acquiring voltage signals. The vacuum degree identification model is based on a Bayesian neural network algorithm; The Bayesian neural network algorithm includes a feedforward neural network structure and Bayesian inference; The feedforward neural network structure consists of an input layer, a hidden layer, and an output layer. The input data of the input layer includes spectral information, and the output data of the output layer indicates whether the vacuum degree is abnormal. The Bayesian inference uses Bayes' theorem to estimate the posterior distribution of the parameters: in, Given data D, the weights The posterior distribution of; It is the likelihood function, representing the likelihood given weights. The probability of training data D occurring at that time; The prior distribution of the weights is a Gaussian distribution; This is evidence, i.e., the marginal probability of the data, which is set as a constant.

Images