Monitoring system and method for glass surface flatness detection equipment

Abstract

The invention discloses a monitoring system and method for glass surface flatness detection equipment, and particularly relates to the technical field of equipment monitoring, and the monitoring system comprises a data acquisition module, a detection method selection evaluation module, a risk evaluation module and an algorithm processing module. According to the method, on the basis of monitoring and analyzing detection precision information, error correction capability information and atom attachment information of a traditional detection method for the target miniature rubidium atomic clock air chamber glass, the applicability of the traditional detection method for the target miniature rubidium atomic clock air chamber glass is evaluated, and when it is judged that the target miniature rubidium atomic clock air chamber glass is not suitable for detection through the traditional method, the target miniature rubidium atomic clock air chamber glass is detected. The method comprises the following steps: constructing an equipment performance change model, obtaining an equipment performance change index of a modern optical detection method on target miniature rubidium atomic clock air chamber glass, and constructing a support vector machine model, so that parameter setting of monitoring equipment is optimized according to an evaluation result of the support vector machine model, and the detection precision is improved.

Description

Technical Field

[0001] This invention relates to the field of equipment monitoring technology, and more specifically, to a monitoring system and method for equipment used to detect the flatness of glass surfaces. Background Technology

[0002] A miniature rubidium atomic clock is a high-precision time standard device. One of its core components is a gas cell, typically made of glass. Rubidium atoms within the gas cell interact with a specific laser beam, generating a stable frequency reference through a physical process. To ensure the accuracy of the rubidium atomic clock, the quality of the gas cell glass is crucial, especially the flatness of the glass surface, which directly affects the distribution of rubidium atoms and the propagation path of light, thus impacting the stability and accuracy of the frequency output.

[0003] The flatness of the glass in the cell of a miniature rubidium atomic clock directly affects the propagation of light and its interaction with rubidium atoms. Any surface defects, such as unevenness or localized irregularities, can lead to scattering, diffraction, and even shifts in the reflection angle of the light beam, thus affecting the frequency stability of the rubidium atomic clock. To ensure a high degree of flatness on the glass surface and thus guarantee the accuracy of the physical processes within the cell, mechanical and optical methods are typically used to detect the flatness of the glass surface in traditional industrial production.

[0004] Mechanical inspection: This method uses probes or contact instruments to detect minute deformations on the glass surface. While it can directly inspect surface morphology, its contact nature can introduce minute scratches or other damage, making it unsuitable for gas chamber glass requiring extremely high precision.

[0005] Interferometry: Interferometry utilizes the interaction of light waves to determine surface flatness by observing changes in interference fringes. This method offers high precision and can be performed non-contactly, but it may introduce errors for complex surfaces or under non-ideal conditions (such as ambient light interference).

[0006] To overcome the limitations of traditional inspection methods, some new optical inspection technologies have been developed in recent years. These methods can inspect glass surfaces non-contactly and with high precision, and are suitable for rapid inspection in mass production, such as white light interferometers, laser confocal scanning, and optical profilometers. However, they still have problems in terms of inspection accuracy, non-contact inspection effect, inspection speed, data processing, environmental adaptability, and cost control, such as interference fringe resolution limitations, the contradiction between inspection speed and real-time monitoring, the complexity of data processing and analysis, and environmental sensitivity and equipment stability. Summary of the Invention

[0007] To overcome the aforementioned deficiencies of the prior art, embodiments of the present invention provide a monitoring system and method for a device for detecting the flatness of a glass surface. The system monitors and analyzes the detection accuracy, error correction capability, and atomic adhesion information of the target miniature rubidium atomic clock gas chamber glass using traditional detection methods. It evaluates the applicability of traditional detection methods for the target miniature rubidium atomic clock gas chamber glass. When it is determined that traditional methods are unsuitable for detection, a device performance change model is constructed to obtain the device performance change index of the target miniature rubidium atomic clock gas chamber glass caused by modern optical detection methods. A support vector machine model is then constructed, and the parameter settings of the monitoring device are optimized based on the evaluation results of the support vector machine model to improve detection accuracy.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] A monitoring method for a device used to detect the flatness of a glass surface includes the following steps:

[0010] Step S1: Detect the target miniature rubidium atomic clock gas cell glass, and collect information on detection accuracy, error correction capability, atomic attachment, thermal expansion, spin polarization change, and optical pumping.

[0011] Step S2: Construct an evaluation model for the selection of detection methods to evaluate traditional detection methods for the target miniature rubidium atomic clock gas cell glass;

[0012] Step S3: For the target miniature rubidium atomic clock gas cell glass that is not suitable for detection using traditional methods, a device performance change model is constructed by combining thermal expansion information, spin polarization change information and optical pump information to obtain the device performance change index of the target miniature rubidium atomic clock gas cell glass by modern optical detection methods.

[0013] Step S4: Establish a support vector machine model using the detection method selection evaluation index, equipment performance change index, and real-time monitoring data as factors, and optimize the parameter settings of the monitoring equipment based on the evaluation results of the support vector machine model to improve detection accuracy.

[0014] Specifically, in step S1, modern optical inspection methods, utilizing various monitoring devices, achieve non-contact, high-precision inspection of glass surfaces despite high equipment costs. Examples include: White light interferometer: White light interferometry can acquire surface profiles and analyze flatness through high-precision scanning of the surface. Compared to traditional laser interferometry, white light interferometry is more suitable for large-area inspection and has stronger anti-interference capabilities; Laser confocal scanning: Laser confocal scanning technology uses a laser beam to accurately measure minute morphological changes on the glass surface and acquires surface height data through three-dimensional imaging. This technology can achieve sub-micron level inspection accuracy and is suitable for inspecting high-requirement micro-cell glass; Optical profilometer: An optical profilometer measures surface flatness by projecting a beam and analyzing its reflection. This method can quickly acquire surface information and is unaffected by the transparency of the glass material, making it suitable for online inspection in mass production. In practical use, due to factors such as equipment cost and the complexity of real-time monitoring, this invention is based on the premise that the glass detection equipment is operating normally and the environment in which the rubidium atoms are located is relatively stable, and the glass data of the micro rubidium atomic clock gas cell is continuously collected. This may lead to the phenomenon of risk changes at a certain monitoring moment.

[0015] In step S2, a detection method selection evaluation model is constructed. Specifically, this involves collecting information on the detection accuracy, error correction capability, and atomic adhesion of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods. The detection accuracy information includes a detection accuracy coefficient, calibrated as J. jc The error correction capability information includes the error correction capability coefficient, calibrated as N. wc Atomic adhesion information includes the atomic adhesion coefficient, denoted as F. yz ;

[0016] Detection accuracy coefficient J jc The target miniature rubidium atomic clock gas cell glass was tested using traditional detection methods, and the measured value was calibrated as X. i Let i be the measurement number, i = 1, 2, 3, ..., n. The average value of the measurements is obtained and calibrated as... Then the detection accuracy coefficient

[0017] Error correction capability coefficient N wc The maximum error value generated during the statistical detection process is calibrated as w. m Obtain the final error value after the detection is completed, and calibrate it as W. z Then the error correction capability coefficient N wc =W m –W z ;

[0018] Atomic adhesion coefficient F yzThe number of rubidium atoms adhering to the surface of the target miniature rubidium atomic clock's gas cell glass was collected and calibrated as S within ten hours of use. r The concentration of rubidium atoms in the gas chamber of the miniature rubidium atomic clock was obtained and calibrated as N. r Then F yz =S r ÷N r ;

[0019] The evaluation model for the selection of detection methods is constructed by weighting three aspects: the detection accuracy, error correction capability, and atomic adhesion information of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods. This generates the evaluation index FX for the selection of detection methods. pg The corresponding coefficients are the detection accuracy coefficient J. jc Error correction capability coefficient N wc Atomic adhesion coefficient F yz The formula is FX Pg =α1×J jc -α2×N wc +α3×F yz ;

[0020] Meanwhile, α1, α2, and α3 are all greater than 0 and are set according to the actual situation. For example, the expert weighting method can be used, which involves inviting experts in relevant fields to determine the weights of each indicator through professional opinion surveys and comprehensive evaluations, to ensure that the weight coefficients accurately reflect the importance of each indicator in the evaluation of the selection of testing methods. In addition, methods such as the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation can also be considered to determine the weight coefficients to ensure their objectivity and scientific validity. These will not be elaborated upon here.

[0021] In step S2, the detection method selection evaluation index obtained from the detection method selection evaluation model is used to reflect the detection method suitable for the target miniature rubidium atomic clock gas cell glass. The smaller the value, the higher the detection accuracy of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods, the stronger the error correction capability of the detection method, the smaller the influence of rubidium atom adhesion, and the more suitable it is to use traditional methods for detection.

[0022] In step S2, when the detection method selects the evaluation index FX pg When the threshold is less than or equal to the set threshold, it indicates that the target miniature rubidium atomic clock gas cell glass is suitable for detection using traditional methods, and the obtained results are suitable for user needs. No optimization of the detection method is required, no subsequent steps are needed, and the evaluation results are directly output.

[0023] When the detection method selects the evaluation index FX pgIf the threshold value exceeds the set threshold, it indicates that the target miniature rubidium atomic clock gas cell glass is not suitable for detection using traditional methods and needs to be improved using modern optical detection methods. This involves sending optimized demand data signals and optimizing the improved data.

[0024] In step S3, after receiving the optimization requirement data signal, the target miniature rubidium atomic clock gas cell glass is inspected using modern optical detection methods such as white light interferometer, laser confocal scanning, and optical profilometer. If the properties of rubidium atoms in the gas cell do not change during the inspection, the precision and rationality of the inspection method can be reflected to a certain extent. The surface flatness of the inspected glass is monitored and evaluated by the equipment performance change evaluation index.

[0025] The equipment performance change index is determined by constructing a model of equipment performance change. Specifically, this involves using modern optical detection methods to obtain information on thermal expansion, spin polarization change, and optical pumping from the target miniature rubidium atomic clock's gas cell glass. The thermal expansion information includes the thermal expansion coefficient, denoted as R0. pz The spin polarization change information includes the spin polarization change coefficient, denoted as Z. gb The optical pump information includes the optical pump coefficient, denoted as G. bp ;

[0026] Thermotropic expansion coefficient R pz The temperature change during the detection of the target miniature rubidium atomic clock's gas cell glass was calibrated as T. b Obtain the initial volume of the selected region, calibrated as V. c The coefficient of thermal expansion of glass material is calibrated as R. x Then the coefficient of thermal expansion R pz =T b ×V c ×R x ;

[0027] Spin polarization change coefficient Z gb By collecting the Landau factor and Bohr magneton of rubidium atoms, respectively calibrated as L y and B c The external magnetic field strength generated by the detection equipment is calibrated as C. w The spin polarization change coefficient Z gb =L y ×B c ×C w ;

[0028] Optical pump coefficient G bp The refractive index of air when the incident angle of light is perpendicular, as measured by the glass of a miniature rubidium atomic clock chamber, is calibrated as Z. g Obtain the refractive index of the glass and calibrate it as Z.b Then the optical pump coefficient

[0029] The device performance change model is constructed by weighting three aspects: thermal expansion information, spin polarization change information, and optical pumping information collected during the detection of the glass chamber of the miniature rubidium atomic clock. This generates the device performance change index SX. pg The corresponding coefficients are the thermal expansion coefficients R. pz Spin polarization change coefficient Z gb Optical pump information G bp The formula formed is SX pg =β1×R pz +β2×Z gb -β3×G bp ;

[0030] Meanwhile, β1, β2, and β3 are all greater than 0 and are set according to actual conditions. For example, an expert weighting method can be used, which involves inviting experts in relevant fields to determine the weights of each indicator through professional opinion surveys and comprehensive evaluations, ensuring that the weight coefficients accurately reflect the importance of each indicator in the equipment performance change assessment. In addition, methods such as the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation can also be considered to determine the weight coefficients to ensure their objectivity and scientific validity. These will not be elaborated upon here.

[0031] In step S3, the equipment performance change index obtained from the equipment performance change model is used to reflect the degree of influence of the equipment on the results when modern optical detection methods detect the target miniature rubidium atomic clock gas chamber glass. The larger the value, the greater the influence of the detection equipment on the degree of thermal expansion inside the miniature rubidium atomic clock gas chamber glass, resulting in a higher degree of change in the spin polarization of rubidium atoms, a greater impact on the efficiency and stability of optical pumping, and a smaller degree of precision and rationality of equipment monitoring.

[0032] In step S4, the detection method selection evaluation index, equipment performance change index, and real-time detection data are used as construction factors for the support vector machine model to optimize the parameter settings of the monitoring equipment.

[0033] Furthermore, the method for constructing the support vector machine model of this invention is as follows:

[0034] Step S4.1, data preparation and standardization: Input the flatness of the glass surface, the ambient temperature, pressure, stress, and light reflectivity of the testing equipment as input variables, and standardize or normalize the input data. A common standardization method is to subtract the mean of each feature and divide by the standard deviation.

[0035] Step S4.2 involves selecting a kernel function and constructing the optimal classification hyperplane. A suitable kernel function is chosen to handle nonlinear data. Common kernel functions include linear kernel functions, polynomial kernel functions, and Gaussian radial basis function (RBF) kernel functions. Taking the Gaussian RBF kernel function as an example, it is defined as K(x... i x j ), where x i x j Let represent the values ​​of the input variables. Since the goal of the support vector machine is to find a hyperplane that maximizes the inter-class margin, the hyperplane can be defined by the following formula: Where w is the normal vector of the hyperplane, which determines the direction of the hyperplane, b is the bias term of the hyperplane, which determines the position of the hyperplane, and y i Indicates the type to which sample i belongs, x i The input feature vector;

[0036] Step S4.3, Soft Margin and Regularization: In practical applications, it may be impossible to find a hyperplane that can correctly classify all samples, especially when the data contains noise or outliers. To address this, Support Vector Machines (SVMs) introduce the "soft margin" technique, which improves the model's generalization ability by allowing a small number of classification errors. The optimization objective of the soft margin problem is to minimize the loss function with a regularization term. Where, ξ i is a slack variable representing the classification error of sample i, and C is a penalty parameter used to balance the classification error and the margin.

[0037] Step S4.4, Model Training and Solving: Using the Lagrange multiplier method and the dual problem, the above optimization problem can be transformed into a dual problem, which can then be solved using existing optimization algorithms. The objective function of the dual problem is: Where, α i It is a Lagrange multiplier, kernel function K(x) i x j This is used to map input data to a high-dimensional space. When training a support vector machine model, an optimization algorithm is used to solve for the Lagrange multipliers α. i This allows us to determine the normal vector w and the bias term b of the hyperplane.

[0038] Step S4.5, Model Evaluation and Result Optimization: After training, cross-validation and confusion matrix can be used to evaluate the model's performance, with accuracy, precision, and recall as evaluation metrics. After obtaining the classification results, the operating conditions of the monitoring equipment can be optimized by analyzing the contribution of different features to the results. For example, the parameter settings of the monitoring equipment can be optimized by analyzing the influence of temperature, pressure, and light reflectivity on the flatness detection results, thereby improving its detection accuracy.

[0039] A monitoring system for a device for detecting the flatness of a glass surface, used in the aforementioned monitoring method for a device for detecting the flatness of a glass surface, comprising:

[0040] The data acquisition module detects the target miniature rubidium atomic clock gas cell glass and collects information on detection accuracy, error correction capability, atomic adhesion, thermal expansion, spin polarization change, and optical pumping.

[0041] The detection method selection and evaluation module constructs a detection method selection and evaluation model to evaluate traditional detection methods for the target miniature rubidium atomic clock gas cell glass;

[0042] The device performance change assessment module, for target miniature rubidium atomic clock gas cell glass that is not suitable for detection using traditional methods, constructs a device performance change model by combining thermal expansion information, spin polarization change information and optical pump information, and obtains the device performance change index of target miniature rubidium atomic clock gas cell glass by modern optical detection methods.

[0043] The algorithm optimization module establishes a support vector machine model using the detection method selection evaluation index, equipment performance change index, and real-time monitoring data as factors. Based on the evaluation results of the support vector machine model, it optimizes the parameter settings of the monitoring equipment to improve detection accuracy.

[0044] The technical effects and advantages of this invention are as follows:

[0045] This invention monitors and analyzes the detection accuracy, error correction capability, and atomic adhesion information of traditional detection methods for the target miniature rubidium atomic clock gas chamber glass. It evaluates the applicability of traditional detection methods for the target miniature rubidium atomic clock gas chamber glass. When it is determined that traditional methods are not suitable for detection, a device performance change model is constructed to obtain the device performance change index of modern optical detection methods on the target miniature rubidium atomic clock gas chamber glass. Then, a support vector machine model is constructed, and the parameter settings of the monitoring device are optimized based on the evaluation results of the support vector machine model to improve the detection accuracy. Attached Figure Description

[0046] Figure 1 This is a flowchart of a monitoring method for a device used to detect the flatness of a glass surface according to the present invention. Detailed Implementation

[0047] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0048] This invention monitors and analyzes the detection accuracy, error correction capability, and atomic adhesion information of the target miniature rubidium atomic clock gas chamber glass using traditional detection methods. It evaluates the applicability of traditional detection methods for the target miniature rubidium atomic clock gas chamber glass. When it is determined that traditional methods are unsuitable for detection, a device performance change model is constructed to obtain the device performance change index of the target miniature rubidium atomic clock gas chamber glass caused by modern optical detection methods. Then, a support vector machine model is constructed, and the parameter settings of the monitoring equipment are optimized based on the evaluation results of the support vector machine model to improve detection accuracy.

[0049] Example 1

[0050] like Figure 1 As shown, the monitoring method for a device used to detect the flatness of a glass surface includes the following steps:

[0051] Step S1: Detect the target miniature rubidium atomic clock gas cell glass, and collect information on detection accuracy, error correction capability, atomic attachment, thermal expansion, spin polarization change, and optical pumping.

[0052] Step S2: Construct an evaluation model for the selection of detection methods to evaluate traditional detection methods for the target miniature rubidium atomic clock gas cell glass;

[0053] Step S3: For the target miniature rubidium atomic clock gas cell glass that is not suitable for detection using traditional methods, a device performance change model is constructed by combining thermal expansion information, spin polarization change information and optical pump information to obtain the device performance change index of the target miniature rubidium atomic clock gas cell glass by modern optical detection methods.

[0054] Step S4: Establish a support vector machine model using the detection method selection evaluation index, equipment performance change index, and real-time monitoring data as factors, and optimize the parameter settings of the monitoring equipment based on the evaluation results of the support vector machine model to improve detection accuracy.

[0055] Specifically, in step S1, modern optical inspection methods, utilizing various monitoring devices, achieve non-contact, high-precision inspection of glass surfaces despite high equipment costs. Examples include: White light interferometer: White light interferometry can acquire surface profiles and analyze flatness through high-precision scanning of the surface. Compared to traditional laser interferometry, white light interferometry is more suitable for large-area inspection and has stronger anti-interference capabilities; Laser confocal scanning: Laser confocal scanning technology uses a laser beam to accurately measure minute morphological changes on the glass surface and acquires surface height data through three-dimensional imaging. This technology can achieve sub-micron level inspection accuracy and is suitable for inspecting high-requirement micro-cell glass; Optical profilometer: An optical profilometer measures surface flatness by projecting a beam and analyzing its reflection. This method can quickly acquire surface information and is unaffected by the transparency of the glass material, making it suitable for online inspection in mass production. In practical use, due to factors such as equipment cost and the complexity of real-time monitoring, this invention is based on the premise that the glass detection equipment is operating normally and the environment in which the rubidium atoms are located is relatively stable, and the glass data of the micro rubidium atomic clock gas cell is continuously collected. This may lead to the phenomenon of risk changes at a certain monitoring moment.

[0056] In step S2, a detection method selection evaluation model is constructed. Specifically, this involves collecting information on the detection accuracy, error correction capability, and atomic adhesion of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods. The detection accuracy information includes a detection accuracy coefficient, calibrated as J. jc The error correction capability information includes the error correction capability coefficient, calibrated as N. wc Atomic adhesion information includes the atomic adhesion coefficient, denoted as F. yz ;

[0057] Detection accuracy coefficient J jc The target miniature rubidium atomic clock gas cell glass was tested using traditional detection methods, and the measured value was calibrated as X. i Let i be the measurement number, i = 1, 2, 3, ..., n. The average value of the measurements is obtained and calibrated as... Then the detection accuracy coefficient

[0058] Error correction capability coefficient N wc The maximum error value generated during the statistical detection process is calibrated as W. m Obtain the final error value after the detection is completed, and calibrate it as W. z Then the error correction capability coefficient N wc =W m -W z ;

[0059] Atomic adhesion coefficient F yzThe number of rubidium atoms adhering to the surface of the target miniature rubidium atomic clock's gas cell glass was collected and calibrated as S within ten hours of use. r The concentration of rubidium atoms in the gas chamber of the miniature rubidium atomic clock was obtained and calibrated as N. r Then F yz =S r ÷N r ;

[0060] The evaluation model for the selection of detection methods is constructed by weighting three aspects: the detection accuracy, error correction capability, and atomic adhesion information of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods. This generates the evaluation index FX for the selection of detection methods. pg The corresponding coefficients are the detection accuracy coefficient J. jc Error correction capability coefficient N wc Atomic adhesion coefficient F yz The formula is FX pg =α1×J jc -α2×N wc +α3×F yz ;

[0061] Meanwhile, α1, α2, and α3 are all greater than 0 and are set according to the actual situation. For example, the expert weighting method can be used, which involves inviting experts in relevant fields to determine the weights of each indicator through professional opinion surveys and comprehensive evaluations, to ensure that the weight coefficients accurately reflect the importance of each indicator in the evaluation of the selection of testing methods. In addition, methods such as the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation can also be considered to determine the weight coefficients to ensure their objectivity and scientific validity. These will not be elaborated upon here.

[0062] In step S2, the detection method selection evaluation index obtained from the detection method selection evaluation model is used to reflect the detection method suitable for the target miniature rubidium atomic clock gas cell glass. The smaller the value, the higher the detection accuracy of the target miniature rubidium atomic clock gas cell glass when using traditional detection methods, the stronger the error correction capability of the detection method, the smaller the influence of rubidium atom adhesion, and the more suitable it is to use traditional methods for detection.

[0063] In step S2, when the detection method selects the evaluation index FX pg When the threshold is less than or equal to the set threshold, it indicates that the target miniature rubidium atomic clock gas cell glass is suitable for detection using traditional methods, and the obtained results are suitable for user needs. No optimization of the detection method is required, no subsequent steps are needed, and the evaluation results are directly output.

[0064] When the detection method selects the evaluation index FX pgIf the threshold value exceeds the set threshold, it indicates that the target miniature rubidium atomic clock gas cell glass is not suitable for detection using traditional methods and needs to be improved using modern optical detection methods. This involves sending optimized demand data signals and optimizing the improved data.

[0065] In step S3, after receiving the optimization requirement data signal, the target miniature rubidium atomic clock gas cell glass is inspected using modern optical detection methods such as white light interferometer, laser confocal scanning, and optical profilometer. If the properties of rubidium atoms in the gas cell do not change during the inspection, the precision and rationality of the inspection method can be reflected to a certain extent. The surface flatness of the inspected glass is monitored and evaluated by the equipment performance change evaluation index.

[0066] The equipment performance change index is determined by constructing a model of equipment performance change. Specifically, this involves using modern optical detection methods to obtain information on thermal expansion, spin polarization change, and optical pumping from the target miniature rubidium atomic clock's gas cell glass. The thermal expansion information includes the thermal expansion coefficient, denoted as R0. pz The spin polarization change information includes the spin polarization change coefficient, denoted as Z. gb The optical pump information includes the optical pump coefficient, denoted as G. bp ;

[0067] Thermotropic expansion coefficient R pz The temperature change during the detection of the target miniature rubidium atomic clock's gas cell glass was calibrated as T. b Obtain the initial volume of the selected region, calibrated as V. c The coefficient of thermal expansion of glass material is calibrated as R. x Then the coefficient of thermal expansion R pz =T b ×V c ×R x ;

[0068] Spin polarization change coefficient Z gb By collecting the Landau factor and Bohr magneton of rubidium atoms, respectively calibrated as L y and B c The external magnetic field strength generated by the detection equipment is calibrated as C. w The spin polarization change coefficient Z gb =L y ×B c ×C w ;

[0069] Optical pump coefficient G bp The refractive index of air when the incident angle of light is perpendicular, as measured by the glass of a miniature rubidium atomic clock chamber, is calibrated as Z. g Obtain the refractive index of the glass and calibrate it as Z.b Then the optical pump coefficient

[0070] The device performance change model is constructed by weighting three aspects: thermal expansion information, spin polarization change information, and optical pumping information collected during the detection of the glass chamber of the miniature rubidium atomic clock. This generates the device performance change index SX. pg The corresponding coefficients are the thermal expansion coefficients R. pz Spin polarization change coefficient Z gb Optical pump information G bp The formula formed is SX pg =β1×R pz +β2×Z gb -β3×G bp ;

[0071] Meanwhile, β1, β2, and β3 are all greater than 0 and are set according to actual conditions. For example, an expert weighting method can be used, which involves inviting experts in relevant fields to determine the weights of each indicator through professional opinion surveys and comprehensive evaluations, ensuring that the weight coefficients accurately reflect the importance of each indicator in the equipment performance change assessment. In addition, methods such as the analytic hierarchy process (AHP) and fuzzy comprehensive evaluation can also be considered to determine the weight coefficients to ensure their objectivity and scientific validity. These will not be elaborated upon here.

[0072] In step S3, the equipment performance change index obtained from the equipment performance change model is used to reflect the degree of influence of the equipment on the results when modern optical detection methods detect the target miniature rubidium atomic clock gas chamber glass. The larger the value, the greater the influence of the detection equipment on the degree of thermal expansion inside the miniature rubidium atomic clock gas chamber glass, resulting in a higher degree of change in the spin polarization of rubidium atoms, a greater impact on the efficiency and stability of optical pumping, and a smaller degree of precision and rationality of equipment monitoring.

[0073] In step S4, the detection method selection evaluation index, equipment performance change index, and real-time detection data are used as construction factors for the support vector machine model to optimize the parameter settings of the monitoring equipment.

[0074] Furthermore, the method for constructing the support vector machine model of this invention is as follows:

[0075] Step S4.1, data preparation and standardization: Input the flatness of the glass surface, the ambient temperature, pressure, stress, and light reflectivity of the testing equipment as input variables, and standardize or normalize the input data. A common standardization method is to subtract the mean of each feature and divide by the standard deviation.

[0076] Step S4.2 involves selecting a kernel function and constructing the optimal classification hyperplane. A suitable kernel function is chosen to handle nonlinear data. Common kernel functions include linear kernel functions, polynomial kernel functions, and Gaussian radial basis function (RBF) kernel functions. Taking the Gaussian RBF kernel function as an example, it is defined as K(x... i ,x j ), where x i x j Let represent the values ​​of the input variables. Since the goal of the support vector machine is to find a hyperplane that maximizes the inter-class margin, the hyperplane can be defined by the following formula: Where w is the normal vector of the hyperplane, which determines the direction of the hyperplane, b is the bias term of the hyperplane, which determines the position of the hyperplane, and y i Indicates the type to which sample i belongs, x i The input feature vector;

[0077] Step S4.3, Soft Margin and Regularization: In practical applications, it may be impossible to find a hyperplane that can correctly classify all samples, especially when the data contains noise or outliers. To address this, Support Vector Machines (SVMs) introduce the "soft margin" technique, which improves the model's generalization ability by allowing a small number of classification errors. The optimization objective of the soft margin problem is to minimize the loss function with a regularization term. Where, ξ i is a slack variable representing the classification error of sample i, and C is a penalty parameter used to balance the classification error and the margin.

[0078] Step S4.4, Model Training and Solving: Using the Lagrange multiplier method and the dual problem, the above optimization problem can be transformed into a dual problem, which can then be solved using existing optimization algorithms. The objective function of the dual problem is: Where, α i It is a Lagrange multiplier, kernel function K(x) i x j This is used to map input data to a high-dimensional space. When training a support vector machine model, an optimization algorithm is used to solve for the Lagrange multipliers α. i This allows us to determine the normal vector w and the bias term b of the hyperplane.

[0079] Step S4.5, Model Evaluation and Result Optimization: After training, cross-validation and confusion matrix can be used to evaluate the model's performance, with accuracy, precision, and recall as evaluation metrics. After obtaining the classification results, the operating conditions of the monitoring equipment can be optimized by analyzing the contribution of different features to the results. For example, the parameter settings of the monitoring equipment can be optimized by analyzing the influence of temperature, pressure, and light reflectivity on the flatness detection results, thereby improving its detection accuracy.

[0080] Example 2 is a system implementation of Example 1, used to implement a monitoring method for a glass surface flatness detection device described in Example 1, specifically including:

[0081] The data acquisition module detects the target miniature rubidium atomic clock gas cell glass and collects information on detection accuracy, error correction capability, atomic adhesion, thermal expansion, spin polarization change, and optical pumping.

[0082] The detection method selection and evaluation module constructs a detection method selection and evaluation model to evaluate traditional detection methods for the target miniature rubidium atomic clock gas cell glass;

[0083] The device performance change assessment module, for target miniature rubidium atomic clock gas cell glass that is not suitable for detection using traditional methods, constructs a device performance change model by combining thermal expansion information, spin polarization change information and optical pump information, and obtains the device performance change index of target miniature rubidium atomic clock gas cell glass by modern optical detection methods.

[0084] The algorithm optimization module establishes a support vector machine model using the detection method selection evaluation index, equipment performance change index, and real-time monitoring data as factors. Based on the evaluation results of the support vector machine model, it optimizes the parameter settings of the monitoring equipment to improve detection accuracy.

[0085] The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0086] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0087] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0088] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0089] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0090] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0091] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0092] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0093] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A monitoring method for detecting glass surface flatness equipment, characterized in that: The steps include: Step S1, testing the target micro-rubidium atomic clock gas chamber glass, collecting detection accuracy information, error correction capability information, atomic attachment information, thermal expansion information, spin polarization change information and optical pumping information; Step S2, constructing a detection method selection evaluation model to evaluate the traditional detection method of the target micro-rubidium atomic clock gas chamber glass; Step S3, for the target micro-rubidium atomic clock gas chamber glass that is not suitable for detection using traditional methods, a device performance change model is constructed by combining thermal expansion information, spin polarization change information, and optical pumping information to obtain a device performance change index for the target micro-rubidium atomic clock gas chamber glass using modern optical detection methods; Step S4, establishing a support vector machine model using the detection method selection evaluation index, the equipment performance change index and the real-time monitoring data as factors, and optimizing the parameter settings of the monitoring equipment according to the evaluation results of the support vector machine model to improve the detection accuracy.

2. The monitoring method for detecting the flatness of a glass surface according to claim 1, characterized in that: In step S2, a detection method selection evaluation model is constructed. The specific process is to collect the detection accuracy information, error correction capability information and atomic attachment information of the target micro-rubidium atomic clock gas chamber glass when using the traditional detection method, wherein the detection accuracy information includes the detection accuracy coefficient, which is calibrated as J jc , the error correction capability information includes the error correction capability coefficient, calibrated as N wc , the atomic attachment information includes the atomic attachment coefficient, calibrated as F yz ; Detection accuracy coefficient J jc The target micro-rubidium atomic clock gas chamber glass is tested by traditional detection methods, and the measured value is calibrated as X i , i is the number of measurements, i = 1, 2, 3, ..., n, get the average value of the measured value, calibrated as The detection accuracy coefficient Error correction capability coefficient N wc The maximum error value generated during the statistical detection process is calibrated as W m , obtain the final error value after the test, calibrated as W z , then the error correction capability coefficient N wc =W m -W z ; Atomic adhesion coefficient F yz The number of rubidium atoms attached to the surface of the target micro-rubidium atomic clock gas chamber glass within ten hours of use is calibrated as S r , obtain the rubidium atomic concentration in the gas chamber of the micro-rubidium atomic clock, calibrated as N r , then F yz =S r ÷N r .

3. The monitoring method for detecting the flatness of a glass surface according to claim 2, characterized in that: The detection method selection evaluation model is constructed by weighting the detection accuracy information, error correction capability information and atomic attachment information of the target micro-rubidium atomic clock gas chamber glass when using the traditional detection method, and generates the detection method selection evaluation index FX pg , the corresponding coefficients are the detection accuracy coefficient J jc , Error correction capability coefficient N wc , atomic adhesion coefficient F yz , the formula is FX pg =α1×J jc -α2×N wc +α3×F yz , α1, α2, and α3 are the weight coefficients of the corresponding indicators, all of which are greater than 0.

4. The monitoring method for detecting the flatness of a glass surface according to claim 3, characterized in that: In step S2, when the detection method selects the evaluation index FX pg When the method changes the threshold when the value is less than or equal to the set value, there is no need to optimize the detection method, and no subsequent steps are performed, and the evaluation results are directly output; When the detection method is selected, the evaluation index FX pg When it is greater than the set method change threshold, it is necessary to use modern optical detection methods for improvement, send optimization requirement data signals, and optimize the improvement data.

5. The monitoring method for detecting the flatness of a glass surface according to claim 1, characterized in that: The device performance change index is obtained by constructing a device performance change model. The specific process is to use modern optical detection methods to detect the target micro-rubidium atomic clock gas chamber glass to obtain thermal expansion information, spin polarization change information and optical pump information. The thermal expansion information includes the thermal expansion coefficient, which is calibrated as R pz , the spin polarization change information includes the spin polarization change coefficient, calibrated as Z gb , the optical pump information includes the optical pump coefficient, denoted as G pb ; Thermal expansion coefficient R pz The temperature change of the target micro-rubidium atomic clock gas chamber glass during detection is calibrated as T b , get the initial volume of the selected area, calibrated as V c , the thermal expansion coefficient of the glass material is calibrated as R x , then the thermal expansion coefficient R pz =T b ×V c ×R x ; Spin polarization change coefficient Z gb By collecting the Landau factor and Bohr magneton of the rubidium atom, they are calibrated as L y and B c , the external magnetic field strength generated by the detection equipment is calibrated as C w , then the spin polarization change coefficient Z gb =L y ×B c ×C w ; Optical pump coefficient G bp The refractive index of air when the incident angle of light is vertical during the detection process through the glass of the micro-rubidium atomic clock gas chamber is calibrated as Z g , get the refractive index of the glass, calibrated as Z b , then the optical pump coefficient 6. A monitoring method for detecting glass surface flatness equipment according to claim 5, characterized in that: The device performance change model is constructed by weighting the thermal expansion information, spin polarization change information and optical pump information collected during the detection of the micro-rubidium atomic clock gas chamber glass to generate the device performance change index SX pg , the corresponding coefficients are the thermal expansion coefficient R pz , spin polarization change coefficient Z gb , Optical pump information G bp , the formula is SX pg =β1×R pz +β2×z gb -β3×G bp , β1, β2, and β3 are the weight coefficients of the corresponding indicators, all of which are greater than 0.

7. A monitoring method for detecting glass surface flatness equipment according to claim 6, characterized in that: In step S4, the detection method selection evaluation index, the equipment performance change index and the real-time detection data are used as construction factors of the support vector machine model, and the support vector machine model is established to optimize the parameter settings of the monitoring equipment; Furthermore, the construction method of the support vector machine model is as follows: Step S4.1, data preparation and standardization, inputting the acquired flatness of the glass surface, the ambient temperature, pressure, stress, and light reflectivity of the testing equipment as input variables, and performing standardization or normalization processing on the input data; Step S4.2, selecting kernel function and constructing optimal classification hyperplane; Step S4.3, soft margin and regularization; Step S4.4, model training and solving; Step S4.5, model evaluation and result optimization.

8. A monitoring system for detecting a glass surface flatness device, used to implement a monitoring method for detecting a glass surface flatness device according to any one of claims 1 to 7, characterized in that: include: The data acquisition module detects the target micro-rubidium atomic clock gas chamber glass and collects information on detection accuracy, error correction capability, atomic attachment, thermal expansion, spin polarization change, and optical pumping. The detection method selection and evaluation module builds a detection method selection and evaluation model to evaluate the traditional detection methods of the target micro-rubidium atomic clock gas chamber glass; The device performance change evaluation module builds a device performance change model for target micro-rubidium atomic clock gas chamber glass that is not suitable for detection using traditional methods by combining thermal expansion information, spin polarization change information, and optical pumping information, and obtains the device performance change index of the target micro-rubidium atomic clock gas chamber glass by modern optical detection methods; The algorithm optimization module establishes a support vector machine model based on the detection method selection evaluation index, equipment performance change index and real-time monitoring data, and optimizes the parameter settings of the monitoring equipment according to the evaluation results of the support vector machine model to improve the detection accuracy.

Images