Vacuum switch vacuum degree detection method and device based on multiple pre-detection modules, and computer equipment

Abstract

The invention relates to a vacuum switch vacuum degree detection method and device based on multiple pre-detection modules and computer equipment. The method comprises the following steps: acquiring a one-dimensional spectral signal of a to-be-detected vacuum switch generated based on laser-induced plasma, and inputting the one-dimensional spectral signal to a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal; respectively inputting the spectral feature information into a plurality of vacuum degree pre-detection modules to obtain a plurality of vacuum degree predicted values; and inputting the plurality of vacuum degree predicted values into a comprehensive detection module to obtain a target vacuum degree detection result. By adopting the method, the vacuum degree detection accuracy can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technologies, and particularly to a method, device, computer equipment, computer-readable storage medium, and computer program product for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules. Background Art

[0002] Currently, the on-line detection of the vacuum degree of a vacuum switch faces the dual challenges of detection accuracy and real-time performance. The traditional sensor method is limited by physical principles, with the detection range limited to the order of 0.1 - 10 Pa, and it requires invasive installation, making it impossible to achieve high-vacuum live detection at the level of 10-4 Pa. Compared with traditional methods, the vacuum degree detection method based on laser-induced plasma technology has the advantages of non-contact, high sensitivity, and no need for externally added calibration gas, and has gradually become a research hotspot in the field of vacuum degree detection. However, the detection deviation caused by spectral fluctuations based on laser-induced plasma technology lacks dynamic adaptability, and the prediction error in the low-pressure range is high, resulting in low accuracy in detecting the vacuum degree of vacuum switches. Summary of the Invention

[0003] Based on this, in view of the above technical problems, it is necessary to provide a method, device, computer equipment, computer-readable storage medium, and computer program product for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules, which can improve the accuracy of vacuum degree detection.

[0004] In a first aspect, the present application provides a method for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules, including:

[0005] Obtain a one-dimensional spectral signal of the to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0006] Respectively input the spectral feature information into multiple vacuum degree pre-detection modules to obtain multiple vacuum degree prediction values;

[0007] Input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0008] In one embodiment, the step of inputting the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal includes:

[0009] Perform smoothing and noise reduction processing on the one-dimensional spectral signal to obtain a noise-reduced spectral signal;

[0010] Determine spectral peaks in the noise-reduced spectral signal according to the first derivative and the second derivative of the noise-reduced spectral signal;

[0011] Perform validity screening on the spectral peaks according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks;

[0012] Extract the spectral peak features of the target spectral peaks to obtain the spectral feature information.

[0013] In one embodiment, the extracting the spectral peak features of the target spectral peaks to obtain the spectral feature information includes:

[0014] Extract the peak position, peak height, spectral peak width, spectral peak area, and peak contrast of each target spectral peak;

[0015] Encode the peak position, peak height, spectral peak width, spectral peak area, and peak contrast corresponding to each target spectral peak into a feature vector, and combine the feature vectors corresponding to each target spectral peak to form the spectral feature information.

[0016] In one embodiment, after performing validity screening on the spectral peaks according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks, it further includes:

[0017] Obtain the theoretical spectral peaks corresponding to the known elements contained in the vacuum switch to be detected as theoretical reference spectral peaks;

[0018] Identify the reference target spectral peaks corresponding to the theoretical reference spectral peaks from the target spectral peaks;

[0019] Perform offset correction processing on the wavelength position of the target spectral peaks based on the deviation between the theoretical reference spectral peaks and the reference target spectral peaks, and perform normalization processing on the intensity of the target spectral peaks based on the intensity of the reference target spectral peaks to obtain processed target spectral peaks;

[0020] The extracting the spectral peak features of the target spectral peaks to obtain the spectral feature information includes:

[0021] Extract the spectral peak features of the processed target spectral peaks to obtain the spectral feature information.

[0022] In one embodiment, the obtaining the one-dimensional spectral signal of the vacuum switch to be detected generated based on laser-induced plasma includes:

[0023] Perform multiple laser excitation samplings on the same detection position of the vacuum switch to be detected to obtain multiple candidate one-dimensional spectral signals;

[0024] Perform average fusion processing on the multiple candidate one-dimensional spectral signals to obtain the one-dimensional spectral signal.

[0025] In one embodiment, after obtaining the target vacuum degree detection result, it further includes:

[0026] Compare the target vacuum degree detection result of the current detection cycle with the target vacuum degree detection results of several previous detection cycles of the current detection cycle to determine whether the vacuum degree change rate exceeds a preset change threshold;

[0027] Generate a trend warning message when the vacuum degree change rate exceeds the preset change threshold.

[0028] In a second aspect, the present application also provides a vacuum switch vacuum degree detection device based on multiple pre-detection modules, including:

[0029] A spectral feature acquisition module, configured to acquire a one-dimensional spectral signal of a to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0030] A vacuum degree pre-detection module, configured to input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values;

[0031] A vacuum degree determination module, configured to input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0032] In a third aspect, the present application also provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0033] Acquire a one-dimensional spectral signal of a to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0034] Input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values;

[0035] Input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0036] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0037] Acquire a one-dimensional spectral signal of a to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0038] Input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values;

[0039] Input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0040] In a fifth aspect, the present application also provides a computer program product, including a computer program, which when executed by a processor implements the following steps:

[0041] Obtain a one-dimensional spectral signal of a vacuum switch to be detected generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0042] Input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values;

[0043] Input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0044] The above vacuum degree detection method, device, computer equipment, computer-readable storage medium and computer program product based on multiple pre-detection modules. First, obtain the one-dimensional spectral signal of the vacuum switch to be detected generated based on laser-induced plasma, and input the one-dimensional spectral signal into the feature extraction module to obtain the spectral feature information of the one-dimensional spectral signal, which can extract representative spectral feature parameters from the original spectral data, help reduce the noise interference and fluctuation influence in the original spectrum, provide a more stable and controllable data input for subsequent prediction, and improve the processing adaptability in the spectral fluctuation environment; then, input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values, which can utilize the complementary advantages of different models in aspects such as data learning ability, non-linear fitting ability, and noise tolerance, enhance the modeling accuracy of the vacuum degree under different spectral feature patterns, effectively improve the prediction accuracy in the low-pressure range or weak signal environment, and alleviate the problem of inaccurate prediction of a single model under specific working conditions; finally, input the multiple vacuum degree prediction values into the comprehensive detection module to obtain the target vacuum degree detection result, and combine and optimize the prediction results through a fusion strategy, thereby reducing the risk of error amplification or failure that may exist in a single model, enhancing the robustness and credibility of the final output through the consistency and complementarity of the results between models, and making the vacuum degree detection result more stable and engineering practical. In the above method, through the processing flow of sequentially performing spectral feature extraction, multi-module prediction and fusion output, an integrated vacuum degree detection mechanism for spectral signals is realized. First, the one-dimensional spectral signal is converted into structured feature parameters through the feature extraction module, making the original spectral information have stronger expression stability and computational availability; secondly, multiple vacuum degree pre-detection modules process the same feature information in parallel, enabling the models to form a complementarity when responding to different spectral distributions and feature changes, and enhancing the fitting coverage ability for the input data; finally, the comprehensive detection module uniformly fuses the output results of each module, reduces the influence of the prediction errors of individual models, and enhances the robustness and credibility of the final detection result. Through the coordinated action of each module in the data expression, modeling and fusion links, it can stably output vacuum degree detection results with good accuracy and consistency, and has the advantages of engineering practicality and algorithm adaptability. Description of the Drawings

[0045] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the following will briefly introduce the drawings required to be used in the description of the embodiments of the present application or related technologies. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.

[0046] Figure 1Schematic diagram of the process of the vacuum degree detection method of a vacuum switch based on multiple pre-detection modules in an embodiment;

[0047] Figure 2 Schematic diagram of the process of the spectral feature information acquisition step in an embodiment;

[0048] Figure 3 Schematic diagram of the process of the vacuum degree detection method of a vacuum switch based on multiple pre-detection modules in another embodiment;

[0049] Figure 4 Schematic diagram of the model framework of the vacuum degree detection method of a vacuum switch based on multiple pre-detection modules in an embodiment;

[0050] Figure 5 Block diagram of the structure of the vacuum degree detection device of a vacuum switch based on multiple pre-detection modules in an embodiment;

[0051] Figure 6 Internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0052] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0053] In one embodiment, as Figure 1 shown, a vacuum degree detection method of a vacuum switch based on multiple pre-detection modules is provided. In this embodiment, this method is illustrated by taking its application to a terminal as an example. It can be understood that this method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is implemented through the interaction between the terminal and the server. Among them, the terminal can be, but is not limited to, various personal computers, laptop computers, smart phones, tablet computers, etc. The server can be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. In this embodiment, the method includes the following steps:

[0054] Step S101, obtain the one-dimensional spectral signal of the to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into the feature extraction module to obtain the spectral feature information of the one-dimensional spectral signal.

[0055] Among them, the one-dimensional spectral signal refers to the wavelength-intensity sequence data obtained by the terminal through the spectral acquisition device under the action of laser excitation, and usually shows the intensity change curve of the laser plasma emission spectrum within a certain wavelength range, and is used to reflect the emission characteristics of the measured substance and the plasma state.

[0056] Exemplarily, based on the Laser-Induced Breakdown Spectroscopy (LIBS) technology, the terminal applies a high-energy laser pulse to the target area of the vacuum switch to be detected to excite plasma emission; the emission signal is acquired by a spectral acquisition device and converted into digital spectral data to form an original spectrum; the terminal performs preprocessing operations on the original spectral data, including background removal, noise smoothing, and intensity normalization, etc., to enhance signal stability and peak shape clarity; subsequently, the processed one-dimensional spectral signal is input into a feature extraction module, and spectral peaks are identified through an automatic peak-seeking algorithm, and spectral feature information including peak position, peak height, full width at half maximum, peak area, etc. is extracted to provide reliable input features for subsequent vacuum degree prediction.

[0057] Step S102: Input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values.

[0058] Among them, the spectral feature information usually includes numerical values in multiple dimensions such as the central wavelength (peak position) of the spectral peak, the peak top intensity (peak height), the full width at half maximum, the spectral peak area, the peak contrast, etc., which are used to characterize the physical properties and distribution characteristics of each spectral peak in the one-dimensional spectrum.

[0059] Among them, the vacuum degree pre-detection module refers to multiple regression models preset inside the terminal, which are used to predict the corresponding vacuum degree value based on the input spectral feature information. Different modules can adopt different types of regression algorithms to improve the adaptability and robustness of the overall prediction.

[0060] Exemplarily, the terminal inputs the spectral feature information into multiple regression models with different structures for independent inference and prediction. The regression models can include but are not limited to Support Vector Regression (SVR), Random Forest Regression (RF), Multilayer Perceptron (MLP), and Gradient Boosted Regression (GBR), etc. Each model calculates the input features according to its internal training parameters and outputs a corresponding vacuum degree prediction value. The terminal sequentially obtains the prediction results corresponding to multiple models for subsequent comprehensive detection and processing.

[0061] Step S103: Input the multiple vacuum degree prediction values into a comprehensive detection module to obtain the target vacuum degree detection result.

[0062] Among them, the comprehensive detection module refers to the processing unit inside the terminal for fusing the prediction results of multiple models to output the final vacuum degree detection value. Its function is to perform weighted or fitting integration on multiple predicted values through a fixed fusion strategy to improve the overall prediction accuracy and stability.

[0063] Exemplarily, the terminal forms a prediction vector with each vacuum degree predicted value and inputs it into the comprehensive detection module for processing; the comprehensive detection module can perform weighted fusion on the input prediction vector based on the linear regression algorithm, where the weighted coefficients of each model are fixed values preset or determined during the model training stage. Through this fusion process, the terminal can eliminate the influence brought by the prediction deviation of individual models, thereby outputting a target vacuum degree detection result with higher confidence and robustness for subsequent state judgment or control response.

[0064] In the above vacuum degree detection method of a vacuum switch based on multiple pre-detection modules, first, a one-dimensional spectral signal of the vacuum switch to be detected generated based on laser-induced plasma is obtained, and the one-dimensional spectral signal is input into the feature extraction module to obtain the spectral feature information of the one-dimensional spectral signal, which can extract representative spectral feature parameters from the original spectral data, helping to reduce the noise interference and fluctuation influence in the original spectrum, providing a more stable and controllable data input for subsequent prediction, and improving the processing adaptability in a spectral fluctuation environment; then, the spectral feature information is respectively input into multiple vacuum degree pre-detection modules to obtain multiple vacuum degree prediction values, which can utilize the complementary advantages of different models in aspects such as data learning ability, non-linear fitting ability, and noise tolerance, enhance the modeling accuracy of the vacuum degree under different spectral feature patterns, effectively improve the prediction accuracy in a low-pressure range or a weak signal environment, and alleviate the problem of inaccurate prediction of a single model under specific working conditions; finally, the multiple vacuum degree prediction values are input into the comprehensive detection module to obtain the target vacuum degree detection result, and the prediction results are combined and optimized through a fusion strategy, thereby reducing the risk of error amplification or failure that may exist in a single model, enhancing the robustness and credibility of the final output through the consistency and complementarity of the results between models, and making the vacuum degree detection result more stable and engineering practical. In the above method, through the processing flow of sequentially performing spectral feature extraction, multi-module prediction, and fusion output, an integrated vacuum degree detection mechanism for spectral signals is realized. First, the one-dimensional spectral signal is converted into structured feature parameters through the feature extraction module, making the original spectral information have stronger expression stability and computational availability; second, multiple vacuum degree pre-detection modules process the same feature information in parallel, enabling the models to form a complementarity when responding to different spectral distributions and feature changes, and enhancing the fitting coverage ability for the input data; finally, the comprehensive detection module uniformly fuses the output results of each module, reduces the influence of the prediction error of individual models, and enhances the robustness and credibility of the final detection result. Through the synergistic effect of each module in the data expression, modeling, and fusion links, a vacuum degree detection result with good accuracy and consistency can be stably output, having the advantages of engineering practicality and algorithm adaptability.

[0065] In an exemplary embodiment, as Figure 2 shown, the above step S101 of inputting the one-dimensional spectral signal into the feature extraction module to obtain the spectral feature information of the one-dimensional spectral signal can also be realized through the following steps:

[0066] Step S201, perform smoothing and noise reduction processing on the one-dimensional spectral signal to obtain a noise-reduced spectral signal;

[0067] Step S202, determine the spectral peaks in the noise-reduced spectral signal according to the first derivative and the second derivative of the noise-reduced spectral signal;

[0068] Step S203: Screen the spectral peaks for validity according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks;

[0069] Step S204: Extract the spectral peak features of the target spectral peaks to obtain spectral feature information.

[0070] Exemplarily, the terminal first performs smoothing and noise reduction processing on the acquired original spectral signal to suppress the interference of high-frequency noise and background fluctuations on spectral peak recognition and improve the spectral line continuity and curve differentiability. Subsequently, the first derivative and the second derivative of the smoothed spectral signal are calculated. Among them, the first derivative reflects the slope change of the signal and is used to identify the extreme points of intensity change, while the second derivative represents the curvature direction. The maximum value of the spectral peak corresponds to the point where the first derivative is zero and the second derivative is less than zero. Therefore, the terminal locates the possible spectral peak positions based on this criterion.

[0071] To avoid false peak recognition caused by noise, overlapping spectral lines or background fluctuations, the terminal further sets a spectral intensity threshold and a minimum peak spacing parameter to screen the validity of the preliminarily located spectral peaks. Among them, the intensity threshold is used to eliminate weak signal points with insufficient intensity and low signal-to-noise ratio; while the minimum peak spacing prevents repeated counting of adjacent pseudo-peaks due to high data sampling density, thereby obtaining a physically more reliable set of target spectral peaks.

[0072] After the terminal identifies the target spectral peaks, it further extracts parameters such as the central wavelength, peak height, full width at half maximum, spectral peak area, and contrast of each spectral peak. These parameters can effectively characterize the intensity distribution, morphology, and signal characteristics of the spectral peaks and are the basis for constructing spectral feature information. Finally, the terminal encodes the feature parameters of multiple spectral peaks into a feature vector structure as the input basis for subsequent vacuum degree modeling and prediction.

[0073] In this embodiment, by performing smoothing and noise reduction, first and second derivative analysis, spectral peak validity screening, and spectral peak feature extraction on the one-dimensional spectral signal, it is possible to extract feature parameters with high stability, strong anti-noise ability, and physical significance from the original complex spectral data, and construct spectral feature information with good identifiability and modelability. This processing process not only improves the accuracy of spectral peak recognition, reduces the risk of interference from pseudo-peaks and weak peaks, but also provides more discriminative and expressive input data for the subsequent vacuum degree prediction model, thereby enhancing the reliability and accuracy of the vacuum degree detection results as a whole, and is particularly suitable for vacuum detection scenarios with large spectral signal fluctuations or weak spectral line features.

[0074] In an exemplary embodiment, the above step S204 extracts the spectral peak features of the target spectral peaks to obtain spectral feature information, and further includes: extracting the peak position, peak height, spectral peak width, spectral peak area, and peak contrast of each target spectral peak; encoding the peak position, peak height, spectral peak width, spectral peak area, and peak contrast corresponding to each target spectral peak into a feature vector, and combining the feature vectors corresponding to each target spectral peak to form spectral feature information.

[0075] Among them, the peak position refers to the central wavelength position corresponding to the target spectral peak, which is used to characterize the emission energy level or element type to which the spectral line belongs; the peak height refers to the intensity value at the peak vertex of the spectral peak, which reflects the relative energy output of the spectral line under the current excitation conditions; the spectral peak width (usually the full width at half maximum) is used to describe the broadening degree of the spectral peak, indirectly reflecting the influence of factors such as collisions, pressure, or laser stability in the plasma; the spectral peak area represents the total energy integral of the spectral line within this interval, which can be regarded as a comprehensive expression of the peak intensity; the peak contrast is the intensity ratio between the spectral peak and its neighboring background, which is used to measure the significance and anti-interference ability of the spectral peak in the spectrum.

[0076] Exemplarily, after the terminal obtains the target spectral peak set, it sequentially performs feature extraction operations on each spectral peak:

[0077] The method for obtaining the peak position is as follows: within the wavelength interval corresponding to the spectral peak, locate the sampling point with the maximum spectral intensity, and the wavelength value corresponding to it is the central wavelength of the spectral peak, which is used as the peak position; when there are multiple consecutive sampling points with similar intensities, the centroid method can be used for fine-tuning to improve the peak position positioning accuracy.

[0078] The method for obtaining the peak height is as follows: take the spectral intensity value at the peak position as the peak top intensity, and estimate the baseline according to the background intensities in the left and right neighborhoods of the spectral peak (for example, take the average of the left / right 5 points each), and finally calculate the peak height after baseline correction to avoid the direct influence of energy fluctuations or background drift on the intensity.

[0079] The method for obtaining the spectral peak width (full width at half maximum) is as follows: first determine half of the peak height as the "half height" reference line, and search for the two points where the spectral curve first intersects this half height value along the wavelength axis to the left and right, and accurately calculate the wavelength positions corresponding to these two intersection points through linear interpolation. The difference between the two is the full width at half maximum of the spectral peak.

[0080] The method for obtaining the spectral peak area is as follows: use the left and right boundaries of the spectral peak (the peak position ± several sampling points can be taken) to form an integration interval, and perform numerical integration on the spectral intensity values within this interval. Commonly used methods include the trapezoidal integration method or the Simpson method to obtain the total intensity of the area under the peak envelope, which is used as the spectral peak area.

[0081] The peak contrast is obtained by calculating the ratio of the peak vertex intensity to the background intensity in the far adjacent regions on both sides of the spectral peak. The background intensity is the average intensity of two background intervals selected on the left and right outside the set interval from the center wavelength of the spectral peak, which is used to eliminate the interference of the peak skirt on the baseline.

[0082] The terminal uniformly encodes the above five characteristic parameters into a structured feature vector, and arranges them in the natural order of the target spectral peaks in the spectrum, and combines them to form the complete spectral feature information for subsequent vacuum degree modeling.

[0083] In this embodiment, through the multi-dimensional quantitative feature extraction and structured encoding of the target spectral peak, not only the physical properties and spectral morphological features of the spectral peak are retained, but also the effective conversion of the original spectral data into the input of numerical modeling is realized, improving the recognizability, input stability of the feature data and the correlation expression ability with the change of the vacuum degree, providing a high-quality and information-intensive input basis for the subsequent prediction model, and helping to improve the overall accuracy and robustness of the vacuum degree detection.

[0084] In an exemplary embodiment, after obtaining the target spectral peaks by performing validity screening on the spectral peaks according to the preset spectral intensity threshold and the minimum peak spacing in step S203, the following steps are further included: obtaining the theoretical spectral peaks corresponding to the known elements contained in the vacuum switch to be detected as the theoretical reference spectral peaks; identifying the reference target spectral peaks corresponding to the theoretical reference spectral peaks from the target spectral peaks; performing offset correction processing on the wavelength position of the target spectral peaks based on the deviation between the theoretical reference spectral peaks and the reference target spectral peaks, and performing normalization processing on the intensity of the target spectral peaks based on the intensity of the reference target spectral peaks to obtain the processed target spectral peaks;

[0085] In step S204 of extracting the spectral peak features of the target spectral peaks to obtain the spectral feature information, the following steps are further included: extracting the spectral peak features of the processed target spectral peaks to obtain the spectral feature information.

[0086] The theoretical reference spectral peaks refer to the emission spectral lines of elements with known device structures or material compositions, such as the characteristic spectral lines of elements such as aluminum (Al) and iron (Fe) at specific wavelengths, and their theoretical wavelength values can be obtained from public spectral databases. The reference target spectral peaks refer to the spectral peaks that are identified as the closest to the theoretical reference spectral peaks from the target spectral peaks in actual detection and are used as the internal control reference for the current spectral data.

[0087] Exemplarily, the terminal presets the corresponding theoretical characteristic spectral line wavelengths of elements (such as aluminum, calcium, silicon, etc.) contained in the vacuum switch insulation structure as reference values, and searches in the selected target spectral peak set to see if there is a spectral peak with a central wavelength close to the theoretical wavelength. The similarity judgment can be based on whether the wavelength difference is less than a set tolerance range (such as ±0.2 nm). If the match is successful, the target spectral peak is used as the reference target spectral peak, and the difference between its central wavelength and the theoretical wavelength is calculated as the wavelength offset. The terminal applies this offset to the central wavelengths of all target spectral peaks, that is, corrects the wavelength parameters of each spectral peak. Through this correction process, the wavelength axis of the entire set of target spectral peaks can be "aligned" to the standard coordinate system as a whole, eliminating the overall spectral line offset problem caused by changes in the laser excitation position, thermal drift of the optical path system, or equipment errors, thereby improving the spectral position stability.

[0088] After the wavelength correction is completed, the terminal further performs intensity normalization processing. The terminal uses the peak intensity of the reference target spectral peak as the standard factor, and divides the peak height of all target spectral peaks by the standard factor, and divides the area of the target spectral peak by the square of the standard factor for normalization processing. The normalization operation can offset the overall spectral intensity offset caused by unstable laser energy, plasma concentration fluctuations, or sample surface differences, making the spectral peak characteristics more consistent and comparable. After the processing is completed, the terminal performs subsequent spectral peak feature extraction operations on the spectral peak set that has been wavelength-corrected and intensity-normalized, and the obtained feature data constitutes the spectral feature information finally used for vacuum degree modeling.

[0089] In this embodiment, by introducing the reference mechanism of the theoretical spectral peak and using the spectral lines of known elements in the vacuum switch material as internal references, not only the precise correction of the possible global wavelength offset in the actual detected spectrum is realized, but also the intensity drift caused by inconsistent laser excitation intensity or sample response differences is effectively compensated, thereby improving the stability and physical correspondence of the spectral peak characteristics, providing a high-confidence input basis for subsequent vacuum degree modeling, and enhancing the adaptability and detection accuracy under high vacuum or low signal-to-noise ratio conditions.

[0090] In an exemplary embodiment, the above step S101 of obtaining the one-dimensional spectral signal of the to-be-detected vacuum switch based on laser-induced plasma further includes: performing multiple laser excitation samplings on the same detection position of the to-be-detected vacuum switch to obtain multiple candidate one-dimensional spectral signals; performing average fusion processing on the multiple candidate one-dimensional spectral signals to obtain the one-dimensional spectral signal.

[0091] Among them, the candidate one-dimensional spectral signal refers to multiple independent spectral response data respectively obtained by a spectrum acquisition device after the terminal performs multiple laser excitations on the same detection area of the vacuum switch. Each spectral signal corresponds to the original wavelength-intensity sequence generated by one excitation event. The average fusion process refers to, on the premise that the wavelengths of these multiple acquired spectral signals are the same, performing point-to-point average calculation on the intensity values at the same wavelength position to generate a representative one-dimensional fusion spectrum.

[0092] Exemplarily, during the detection process, the terminal applies multiple laser excitations to the target detection position of the vacuum switch, and a complete spectral curve is collected each time. Due to uncontrollable factors such as plasma state, laser energy fluctuation, and material response difference, there may be slight differences in each sampling. Therefore, the terminal aligns the multiple candidate spectral signals collected according to the wavelength dimension and performs average processing on the intensity values at the same wavelength point to obtain the fused one-dimensional spectral signal. This fusion process can effectively suppress random noise and abnormal signal offset, improve the signal-to-noise ratio and stability of the spectral data, and lay a higher-quality signal foundation for subsequent spectral peak identification and feature extraction.

[0093] In this embodiment, by introducing the mechanism of multiple laser excitations and spectral average fusion processing, when facing working conditions such as laser power fluctuation, inconsistent sample response, or insufficient signal intensity under low pressure, a stable and reliable one-dimensional spectral input can still be obtained, enhancing the adaptability of the entire detection system to real working condition changes and improving the accuracy and robustness of subsequent vacuum degree detection.

[0094] In an exemplary embodiment, after obtaining the target vacuum degree detection result in the above step S103, it further includes: comparing the target vacuum degree detection result of the current detection cycle with the target vacuum degree detection results of several previous detection cycles of the current detection cycle to determine whether the vacuum degree change rate exceeds a preset change threshold; and generating a trend warning message when the vacuum degree change rate exceeds the preset change threshold.

[0095] Among them, the vacuum degree change rate refers to the relative change amplitude calculated by the terminal based on the numerical difference between the current detection result and the past detection results, and is usually dynamically calculated based on multiple historical detection points by using a sliding window method; the preset change threshold is a judgment boundary configured according to the actual application scenario for identifying abnormal trends, and is used to distinguish normal fluctuations from possible vacuum deterioration risks.

[0096] Exemplarily, after each complete vacuum degree detection, the terminal compares the current detection value with the vacuum degree values in the past M consecutive detection cycles (for example, the past 5 times), and calculates its average change rate or maximum slope value; if the change rate exceeds a preset threshold (for example, set to 5% or 10%), the terminal determines that there is a continuous deterioration trend in the current detection result, and automatically generates a trend warning message, which can be used to prompt the operation and maintenance personnel to pay attention to the possible risks of airtight aging, leakage, or vacuum decline of the switch. The warning message can include content such as the current detection value, the change rate value, and the historical curve segment, and is uploaded to the monitoring system or the alarm interface through the communication module.

[0097] In this embodiment, by introducing a trend analysis mechanism based on the historical results of the target vacuum degree, not only can an instantaneous anomaly at a certain time be identified, but also potential hidden dangers of the gradual decline of the vacuum degree can be captured in advance, realizing the expansion from single-point detection to trend perception, and improving the early warning ability and operation and maintenance response efficiency of the system in the actual operating environment.

[0098] In another exemplary embodiment, as Figure 3 shown, the present application provides a method for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules, including:

[0099] Performing operations such as data preprocessing, automatic peak searching, and calculating the integral of the spectral peak intensity on the one-dimensional spectral data of the laser plasma to generate spectral features;

[0100] Inputting the spectral features into support vector regression (SVR), random forest (RF), multi-layer perceptron regression (MLP), and gradient boosting regression (GBR) models respectively to obtain pre-detection results P1, P2, P3, and P4 of the vacuum degree;

[0101] Inputting the pre-detection results P1, P2, P3, and P4 of the vacuum degree into a linear regression model. Specifically, stacking P1, P2, P3, and P4 into a prediction vector, such as (P1, P2, P3, P4), and inputting it into the linear regression model for the final vacuum degree detection to obtain the final vacuum degree P.

[0102] As Figure 4 shown, it is a schematic diagram of the model framework of a method for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules in an embodiment, including: a data preprocessing module, a vacuum degree pre-detection module, and a comprehensive detection module.

[0103] Among them, the data preprocessing module automatically identifies spectral peaks through an automatic peak searching algorithm, and extracts features from the automatically identified characteristic peaks.

[0104] Among them, in the vacuum degree pre-detection module, it is necessary to synchronously optimize the penalty factor C and kernel width γ of the support vector regression model, the number of trees N and depth D of the random forest model, the number of neurons K and regularization coefficient α in the hidden layer of the multi-layer perceptron regression model, and the learning rate L and minimum sample number M for splitting of the gradient boosting regression model in a preset space in advance according to the training data set including spectral features and vacuum degree labels through a parameter optimization algorithm.

[0105] This embodiment mainly aims at vacuum switch products with known and stable material types, such as aluminum shells, ceramic shells, etc. By constructing a training model corresponding to a specific material, high-precision vacuum degree detection is achieved.

[0106] In this embodiment, by integrating four heterogeneous models of SVR, RF, MLP, and GBR, using SVR to capture non-linear relationships, RF to suppress overfitting, MLP to learn high-dimensional features, GBR to correct residuals, and combining linear regression to dynamically allocate weights, making full use of the characteristics of different machine learning models, high-precision detection of the vacuum degree of the vacuum switch is achieved, thus having higher detection accuracy and stability. In addition, by synchronously optimizing model parameters through grid search, the influence of spectral fluctuations is dynamically offset at the algorithm level, and the detection robustness and stability can be improved without modifying the hardware, solving the problem of poor working condition adaptability caused by static parameters.

[0107] It should be understood that although the steps in the flowcharts involved in the above-described embodiments are shown sequentially according to the indications of the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps does not have a strict order limit, and these steps can be executed in other orders. Moreover, at least some of the steps in the flowcharts involved in the above-described embodiments may include multiple steps or multiple stages. These steps or stages do not necessarily have to be executed at the same time, but can be executed at different times. The execution order of these steps or stages does not necessarily have to be sequential, but can be executed alternately or alternately with at least some of the steps or stages in other steps or other steps.

[0108] Based on the same inventive concept, the embodiment of the present application also provides a vacuum switch vacuum degree detection device based on multiple pre-detection modules for implementing the vacuum switch vacuum degree detection method based on multiple pre-detection modules described above. The implementation solutions for solving problems provided by this device are similar to the implementation solutions described in the above method. Therefore, the specific limitations in one or more embodiments of the vacuum switch vacuum degree detection device based on multiple pre-detection modules provided below can refer to the limitations on the vacuum switch vacuum degree detection method based on multiple pre-detection modules in the above text, and will not be repeated here.

[0109] In an exemplary embodiment, asFigure 5 As shown in Figure 5 , a vacuum degree detection device for a vacuum switch based on multiple pre-detection modules is provided, including: a spectral feature acquisition module 501, a vacuum degree pre-detection module 502, and a vacuum degree determination module 503, where:

[0110] The spectral feature acquisition module 501 is configured to acquire a one-dimensional spectral signal of the to-be-detected vacuum switch generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal;

[0111] The vacuum degree pre-detection module 502 is configured to input the spectral feature information into multiple vacuum degree pre-detection modules respectively to obtain multiple vacuum degree prediction values;

[0112] The vacuum degree determination module 503 is configured to input the multiple vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

[0113] In one embodiment, the above spectral feature acquisition module 501 is further configured to perform smoothing and noise reduction processing on the one-dimensional spectral signal to obtain a noise-reduced spectral signal; determine spectral peaks in the noise-reduced spectral signal according to the first derivative and the second derivative of the noise-reduced spectral signal; perform validity screening on the spectral peaks according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks; extract spectral peak features of the target spectral peaks to obtain spectral feature information.

[0114] In one embodiment, the above spectral feature acquisition module 501 is further configured to extract the peak position, peak height, spectral peak width, spectral peak area, and peak contrast of each target spectral peak; encode the peak position, peak height, spectral peak width, spectral peak area, and peak contrast corresponding to each target spectral peak into a feature vector, and combine the feature vectors corresponding to each target spectral peak to form spectral feature information.

[0115] In one embodiment, the above spectral feature acquisition module 501 is further configured to obtain theoretical spectral peaks corresponding to known elements contained in the to-be-detected vacuum switch as theoretical reference spectral peaks; identify reference target spectral peaks corresponding to the theoretical reference spectral peaks from the target spectral peaks; perform offset correction processing on the wavelength position of the target spectral peaks based on the deviation between the theoretical reference spectral peaks and the reference target spectral peaks, and perform normalization processing on the intensity of the target spectral peaks based on the intensity of the reference target spectral peaks to obtain processed target spectral peaks.

[0116] In one embodiment, the above spectral feature acquisition module 501 is further configured to extract spectral peak features of the processed target spectral peaks to obtain spectral feature information.

[0117] In one embodiment, the above-mentioned spectral feature acquisition module 501 is further configured to perform multiple laser excitation samplings on the same detection position of the vacuum switch to be detected, obtain multiple candidate one-dimensional spectral signals; and perform average fusion processing on the multiple candidate one-dimensional spectral signals to obtain one-dimensional spectral signals.

[0118] In one embodiment, the above-mentioned vacuum degree detection device of the vacuum switch based on multiple pre-detection modules further includes a warning information generation module, configured to compare the target vacuum degree detection result of the current detection period with the target vacuum degree detection results of several previous detection periods of the current detection period, determine whether the vacuum degree change rate exceeds a preset change threshold; and generate a trend warning information when the vacuum degree change rate exceeds the preset change threshold.

[0119] Each module in the above-mentioned vacuum degree detection device of the vacuum switch based on multiple pre-detection modules can be implemented in whole or in part by software, hardware, and their combinations. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0120] In an exemplary embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as Figure 6As shown in the figure. The computer device includes a processor, a memory, an input / output interface, a communication interface, a display unit, and an input device. Among them, the processor, the memory, and the input / output interface are connected through a system bus, and the communication interface, the display unit, and the input device are connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals in a wired or wireless manner, and the wireless manner can be implemented through WIFI, a mobile cellular network, near field communication (NFC), or other technologies. When the computer program is executed by the processor, it implements a method for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules. The display unit of the computer device is used to form a visually visible picture, which can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0121] Those skilled in the art can understand that Figure 6 the structure shown in the figure is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0122] In one embodiment, a computer device is further provided, including a memory and a processor. A computer program is stored in the memory, and when the processor executes the computer program, the steps in the above method embodiments are implemented.

[0123] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0124] In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by the processor, the steps in the above method embodiments are implemented.

[0125] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data that have been authorized by the user or fully authorized by all parties, and the collection, use, and processing of relevant data need to comply with relevant regulations.

[0126] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include Read-Only Memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, Resistive Random Access Memory (ReRAM), Magnetoresistive Random Access Memory (MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (PCM), graphene memory, etc. Volatile memory can include Random Access Memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), etc. The databases involved in the embodiments provided in this application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in this application can be general-purpose processors, central processors, graphics processors, digital signal processors, programmable logic devices, data processing logics based on quantum computing, Artificial Intelligence (AI) processors, etc., without limitation.

[0127] The technical features of the above embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this application.

[0128] The above-described embodiments merely represent several implementation manners of this application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation on the patent scope of this application. It should be noted that for those of ordinary skill in the art, without departing from the concept of this application, several modifications and improvements can still be made, and these all belong to the protection scope of this application. Therefore, the protection scope of this application shall be subject to the appended claims.

Claims

1. A method for detecting the vacuum degree of a vacuum switch based on multiple pre-detection modules, characterized in that The method includes: Obtaining a one-dimensional spectral signal of a vacuum switch to be detected generated based on laser-induced plasma, and inputting the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal; Respectively inputting the spectral feature information into a plurality of vacuum degree pre-detection modules to obtain a plurality of vacuum degree prediction values; Inputting the plurality of vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

2. The method according to claim 1, wherein The inputting the one-dimensional spectral signal into the feature extraction module to obtain the spectral feature information of the one-dimensional spectral signal includes: Performing smoothing and noise reduction processing on the one-dimensional spectral signal to obtain a noise-reduced spectral signal; Determining spectral peaks in the noise-reduced spectral signal according to the first derivative and the second derivative of the noise-reduced spectral signal; Performing validity screening on the spectral peaks according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks; Extracting spectral peak features of the target spectral peaks to obtain the spectral feature information.

3. The method according to claim 2, wherein The extracting spectral peak features of the target spectral peaks to obtain the spectral feature information includes: Extracting the peak position, peak height, spectral peak width, spectral peak area, and peak contrast of each target spectral peak; Encoding the peak position, peak height, spectral peak width, spectral peak area, and peak contrast corresponding to each target spectral peak into a feature vector, and combining the feature vectors corresponding to each target spectral peak to form the spectral feature information.

4. The method according to claim 2, wherein After performing validity screening on the spectral peaks according to a preset spectral intensity threshold and a minimum peak spacing to obtain target spectral peaks, it further includes: Obtaining theoretical spectral peaks corresponding to known elements contained in the vacuum switch to be detected as theoretical reference spectral peaks; Identifying reference target spectral peaks corresponding to the theoretical reference spectral peaks from the target spectral peaks; Performing offset correction processing on the wavelength position of the target spectral peaks based on the deviation between the theoretical reference spectral peaks and the reference target spectral peaks, and performing normalization processing on the intensity of the target spectral peaks based on the intensity of the reference target spectral peaks to obtain processed target spectral peaks; The extracting spectral peak features of the target spectral peaks to obtain the spectral feature information includes: Extracting spectral peak features of the processed target spectral peaks to obtain the spectral feature information.

5. The method according to claim 1, wherein The obtaining a one-dimensional spectral signal of a vacuum switch to be detected generated based on laser-induced plasma includes: Performing multiple laser excitation samplings on the same detection position of the vacuum switch to be detected to obtain a plurality of candidate one-dimensional spectral signals; Performing average fusion processing on the plurality of candidate one-dimensional spectral signals to obtain the one-dimensional spectral signal.

6. The method according to any one of claims 1 to 5, characterized in that After obtaining the target vacuum degree detection result, it further includes: Comparing the target vacuum degree detection result of the current detection cycle with the target vacuum degree detection results of the previous several detection cycles of the current detection cycle to determine whether the vacuum degree change rate exceeds a preset change threshold; Generating a trend warning message when the vacuum degree change rate exceeds the preset change threshold.

7. A vacuum degree detection device for a vacuum switch based on multiple pre-detection modules, characterized in that, The device includes: A spectral feature acquisition module, configured to acquire a one-dimensional spectral signal of a vacuum switch to be detected generated based on laser-induced plasma, and input the one-dimensional spectral signal into a feature extraction module to obtain spectral feature information of the one-dimensional spectral signal; A vacuum degree pre-detection module, configured to respectively input the spectral feature information into a plurality of vacuum degree pre-detection modules to obtain a plurality of vacuum degree prediction values; A vacuum degree determination module, configured to input the plurality of vacuum degree prediction values into a comprehensive detection module to obtain a target vacuum degree detection result.

8. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images