Packaging device leak detection method, system and device based on intelligent sensor

Abstract

The invention provides a packaging device leakage detection method, system and device based on an intelligent sensor, and belongs to the technical field of device packaging, and the method comprises the steps: determining the combination type of a device of a to-be-detected packaging sensor, and obtaining a plurality of leakage detection technologies matched with the to-be-detected packaging sensor from a type-packaging technology comparison table; extracting a historical leak detection information set matched with each leak detection process from the historical database, constructing a leak detection effective function, and screening a final process; sequentially obtaining a parameter set for carrying out leak detection on the packaged sensor to be subjected to leak detection according to the final process, and carrying out quantization processing on the parameter set; and sequentially filling the quantitative processing results into a parameter analysis table to obtain the influence degree of each process step on the leakage detection result, and obtaining the final leakage detection result of the packaged sensor to be subjected to leakage detection based on all the influence degrees and the correlation between the adjacent process steps. The leakage detection result is effectively determined, and the reliability of the intelligent sensor is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of device packaging, and in particular to a packaging device leakage detection method, system and device based on an intelligent sensor. Background Art

[0002] The architecture of the Industrial Internet relies on integrated circuit technology. The accelerated development of the Industrial Internet has promoted the development of the integrated circuit industry, especially in terms of precision assurance, process optimization, and the increase in the number of demand sides during the design and manufacturing process. The Industrial Internet has become an important demand side and testing ground for integrated circuits. At the same time, integrated circuits are the foundation of the electronic information industry. Integrated circuits are integrated on a single chip by integrating a large number of electronic components and circuits, including microprocessors, memory, and sensors. Therefore, packaging is inseparable from integrated circuits, and packaging technology can provide protection, heat dissipation, signal connection and other support for the application of integrated circuits. Packaging is to encapsulate different types of electronic components or integrated circuit chips in a shell with certain processes and materials to facilitate the use and protection of electronic components or integrated circuit chips. However, when semiconductors and higher-level electronic devices are combined together, the problems associated with the use of basic packaging will become more serious. In order to achieve smarter sensors, the basic assembly differences between sensors and microelectronic device packaging are being resolved. However, due to the different device combinations involved in smart sensors, if the packaging leak detection method of a certain device is used for leak detection, it will undoubtedly reduce the leak detection efficiency and reduce the reliability of smart sensors.

[0003] Therefore, the present invention proposes a packaging device leak detection method, system and device based on intelligent sensors. Summary of the invention

[0004] The present invention provides a packaging device leak detection method, system and device based on intelligent sensors, which are used to screen the final process based on the sensing type matching of several leak detection processes, effectively improve the leak detection efficiency, and subsequently determine the leak detection result by combining quantitative analysis of the parameter set and the influence degree analysis after the result is filled with the correlation, so as to effectively ensure the reliability of the intelligent sensor.

[0005] The present invention provides a packaging device leak detection method based on an intelligent sensor, comprising: Step 1: Determine the component combination type of the packaged sensor to be leak tested, and obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaging process comparison table; Step 2: Extract the historical leak detection information set matching each leak detection process from the historical database, construct the leak detection effective function, and select the final process; Step 3: sequentially obtaining parameter sets for leak testing the packaged sensor to be leak tested according to the final process, and quantifying the parameter sets, wherein the parameter sets include values ​​of several leak testing parameters in each process step in the final process; Step 4: Fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

[0006] Preferably, constructing an effective leak detection function includes: Determine the first loss coefficient of each historical step based on the operation time of each historical step involved in the same leak detection and in combination with the historical leak detection rate; Determine the leak detection density function based on the first loss coefficients of all historical steps under the same leak detection; Each leak detection information in the historical leak detection information set of each leak detection process is placed in the order of leak detection to construct a leak detection matrix; Constructing a first distribution function of each column vector in the leakage detection matrix, and obtaining a second loss coefficient based on each leakage detection parameter; A historical curve is constructed based on all the second loss coefficients, and parameter alignment is performed with the standard curve of the leak detection process to obtain an amount to be adjusted for each leak detection parameter; According to the to-be-adjusted amounts of all leak detection parameters involved in the same historical step, the leak detection density function is optimized to obtain a leak detection related function of the corresponding leak detection process based on the historical leak detection information set; The leak detection related function is reset and adjusted to obtain a leak detection effective function.

[0007] Preferably, the leak detection density function is optimized, including: Fill the first variable and the leak detection parameter corresponding to each historical step in the leak detection density function into the variable comparison table respectively, obtain a number of leak detection parameters matching each first variable, and establish a control group; Based on the control relationship of each control group and in combination with the amount to be adjusted of the leak detection parameter involved in the control group, the variable coefficient adjustment value for the corresponding first variable in the leak detection density function and the additional variable are obtained to obtain the adjustment density function; Obtain the adjustment density function of each leak detection involved in the corresponding leak detection process, and obtain all value coefficients involved in the same second variable; Extract the strong correlation coefficient and the weak correlation coefficient from all the value coefficients involved in the same second variable, and calculate the absolute value array of the coefficient difference between the strong correlation coefficient, the weak correlation coefficient and each value coefficient in turn, to determine whether it is necessary to add an auxiliary coefficient for the same second variable; If not necessary, the trend status is determined based on all value coefficients involved in the same second variable; If necessary, determining the additional number of auxiliary coefficients to be interpolated based on the coefficient difference absolute value array to determine the trend state of all the interpolated coefficients; Based on the trend state of each second variable, a leak detection related function corresponding to the leak detection process is determined.

[0008] Preferably, quantizing the parameter set includes: Obtaining step attributes of each process step in the final process and constructing an attribute vector; Acquire a quantization model matching the attribute vector from an attribute-model database, wherein each network layer in the quantization model corresponds to a process step, and each network layer contains parameter quantization weights and quantization change information; The parameter set is quantized according to the quantization model to obtain a quantization result.

[0009] Preferably, obtaining a quantization model matching the attribute vector from an attribute-model database includes: When there is a first model in the attribute-model database that is completely consistent with the attribute vector, the first model is regarded as a quantization model; When there is no first model in the attribute-model database that is completely consistent with the attribute vector, cluster analysis is performed on all step attributes under the final process to obtain a number of clusters, and an initial cluster vector is constructed based on cluster centers of the clusters; Determine the distance between each cluster center and the remaining attributes in the corresponding clustering result in turn, and perform weakening analysis on the corresponding step attributes according to the distance; ; in, represents the weakening analysis result of the jth remaining attribute in the corresponding clustering result; n1 represents the total number of attributes involved in the corresponding clustering result; sum1 represents the total number of attributes involved in the final process; represents the distance between the cluster center of the corresponding clustering result and the jth remaining attribute; a1 represents the set threshold; The attributes whose weakening analysis results are 0 are replaced with 0 and added to the initial cluster vector, and the attributes whose weakening analysis results are non-0 are replaced with the original attributes and added to the initial cluster vector to obtain the current cluster vector; Comparing the current cluster vector with the attribute-model database, if there is a second model that is completely consistent with the current cluster vector, then considering the second model as a quantization model; Otherwise, obtaining from the attribute-model database a third model having the largest number of attributes matching the attribute vector, and at the same time, obtaining from the attribute-model database a fourth model having the largest number of attributes matching the current cluster vector; The parameter quantization weights of the same-layer networks in the third model and the fourth model are respectively obtained, and the required network layer is locked according to max(y1i,y2i), wherein y1i represents the parameter quantization weight of the i-th layer network in the third model; y2i represents the parameter quantization weight of the i-th layer network in the fourth model; All required network layers are reintegrated to obtain a quantized model.

[0010] Preferably, obtaining the influence of each process step on the leak detection result includes: Compare the filling results based on each process step in the filling analysis table with the corresponding standard results; Determine the extent of the impact based on the comparison results.

[0011] Preferably, based on all the influence degrees and the correlation between the adjacent process steps, the final leak detection result of the packaged sensor to be leak detected is obtained, including: Obtaining an initial detection result of the last process step, and adjusting the initial detection result once based on all impact levels and correlations between adjacent process steps; Determine the characteristic parameters of the packaged sensor to be leak tested, and determine in turn the coefficient of influence of the last process step on each characteristic parameter; According to all the influencing coefficients, the first leak detection result is adjusted to obtain the final leak detection result.

[0012] The present invention provides a packaging device leak detection system based on an intelligent sensor, comprising: A process matching module, used to determine the combination type of components of the packaged sensor to be leak tested, and to obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaged process comparison table; A function building module is used to extract a historical leak detection information set matching each leak detection process from a historical database, build a leak detection effective function, and select the final process; A quantification module, used to sequentially obtain a parameter set for leak testing the packaged sensor to be leak tested according to the final process, and perform quantification processing on the parameter set, wherein the parameter set includes values ​​of several leak testing parameters under each process step; The result acquisition module is used to fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

[0013] The present invention provides a packaging device leak detection device based on an intelligent sensor, which is used to execute any of the packaging device leak detection methods based on an intelligent sensor.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The final process is screened by matching several leak detection processes based on the sensing type, which effectively improves the leak detection efficiency. The leak detection result is subsequently determined by combining quantitative analysis of the parameter set and the impact degree analysis after the result is filled with the correlation, effectively ensuring the reliability of the smart sensor.

[0015] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0016] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings: Figure 1 It is a flow chart of a packaging device leakage detection method based on an intelligent sensor in an embodiment of the present invention; Figure 2 It is a structural diagram of a packaging device leakage detection system based on an intelligent sensor in an embodiment of the present invention; Figure 3 4 is a structural diagram of a packaging device in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0019] The present invention provides a packaging device leak detection method based on an intelligent sensor, such as Figure 1 As shown, including: Step 1: Determine the component combination type of the packaged sensor to be leak tested, and obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaging process comparison table; Step 2: Extract the historical leak detection information set matching each leak detection process from the historical database, construct the leak detection effective function, and select the final process; Step 3: sequentially obtaining parameter sets for leak testing the packaged sensor to be leak tested according to the final process, and quantifying the parameter sets, wherein the parameter sets include values ​​of several leak testing parameters in each process step in the final process; Step 4: Fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

[0020] In this embodiment, the packaged sensor to be leak-tested refers to an intelligent sensor that needs to be leak-tested, and the intelligent sensor includes at least one device. Therefore, the type of each device in the intelligent sensor is obtained to obtain the combination type. For example, the packaged sensor to be leak-tested includes: device 1 and device 2, device 1 is type A, and device 2 is type B. At this time, the combination type is A+B.

[0021] In this embodiment, the type-packaging process comparison table includes several leak detection processes under different combination types, and there may be multiple leak detection processes under each combination type, which are all stored in advance for easy direct matching and acquisition. For example, the leak detection process is: S1: placing packaged electronic components into a pre-treatment tank and pressurizing and filling the packaged electronic components with helium; S2: After the helium filling is completed, the pretreatment tank is subjected to pressure maintenance treatment; S3: Exhaust the helium in the pretreatment tank, and then introduce dry hot nitrogen for purging; S4: Take out the packaged electronic components in the pre-treatment tank and place them in the work station for detailed inspection. The work station is sealed by a fluororubber sealing ring and negative pressure suction.

[0022] The parameters involved in S1 are described as follows: the time of pressurization and helium filling and the pressure when pressurization and helium filling are stopped; The parameters involved in S2 are described as follows: the time of pressure holding treatment and the helium leakage during the pressure holding treatment; The parameters involved in S3 are described as follows: helium discharge time, time to start passing dry hot nitrogen, helium concentration before starting passing dry hot nitrogen, purge speed and duration of dry hot nitrogen; The parameter description for S4 is as follows: Detailed inspection results.

[0023] In this embodiment, the historical database includes leak detection results based on different steps generated by leak detection on several smart sensors of the same type using different leak detection processes. Therefore, a historical leak detection information set can be obtained by matching the leak detection process. It should be noted that the historical leak detection information set includes leak detection information of several historical leak detections based on the same leak detection process, that is, one leak detection process corresponds to one historical leak detection information set.

[0024] In this embodiment, the leak detection effectiveness function is for a corresponding historical leak detection information set, and is constructed based on basic variables for each process step involved in the corresponding leak detection process.

[0025] In this embodiment, the final process refers to a process obtained by screening a process from a plurality of leak detection processes according to a leak detection effectiveness function.

[0026] In this embodiment, the leak detection parameters refer to the parameters of the leak detection operation involved in the corresponding process steps.

[0027] In this embodiment, after the final process is determined, the parameter description involved is matched from the process-parameter database, that is, the database contains different processes and the parameters involved in each step in the process, which are pre-set. Therefore, the values ​​of the parameter descriptions involved in the leak detection process are captured to construct a parameter set.

[0028] In this embodiment, the quantization processing result is obtained by processing the parameter set based on the quantization model in order to standardize the numerical value and facilitate the subsequent determination of the degree of influence.

[0029] In this embodiment, the degree of influence is obtained based on the comparison between the values ​​involved in the corresponding process steps in the filled parameter analysis table and the standard results, and the standard results are pre-set. Because each smart sensor has its corresponding evaluation standard before leaving the factory for use, abnormalities of the smart sensor are avoided as much as possible.

[0030] In this embodiment, the final leak detection result is based on the last process step and is obtained with the assistance of the previously obtained influence degree and correlation, and may be a leak detection rate.

[0031] In this embodiment, the coefficients of each variable in the leak detection effective function are superimposed, and the process corresponding to the function with the largest superposition sum is regarded as the final process.

[0032] In this embodiment, the packaging device for the smart sensor, such as Figure 3 As shown, 1 is a leak detection tank, 2 is a placement table, 3 is a handle, 4 is a device, and 5 is a material tray.

[0033] The beneficial effects of the above technical solution are: the final process is screened based on the sensing type matching of several leak detection processes, which effectively improves the leak detection efficiency, and subsequently determines the leak detection result by combining quantitative analysis of the parameter set and the impact degree analysis after the result filling with the correlation, thereby effectively ensuring the reliability of the smart sensor.

[0034] The present invention provides a packaging device leak detection method based on an intelligent sensor, and constructs an effective leak detection function, including: Determine the first loss coefficient of each historical step based on the operation time of each historical step involved in the same leak detection and in combination with the historical leak detection rate; Determine the leak detection density function based on the first loss coefficients of all historical steps under the same leak detection; Each leak detection information in the historical leak detection information set of each leak detection process is placed in the order of leak detection to construct a leak detection matrix; Constructing a first distribution function of each column vector in the leakage detection matrix, and obtaining a second loss coefficient based on each leakage detection parameter; A historical curve is constructed based on all the second loss coefficients, and parameter alignment is performed with the standard curve of the leak detection process to obtain an amount to be adjusted for each leak detection parameter; According to the to-be-adjusted amounts of all leak detection parameters involved in the same historical step, the leak detection density function is optimized to obtain a leak detection related function of the corresponding leak detection process based on the historical leak detection information set; The leak detection related function is reset and adjusted to obtain a leak detection effective function.

[0035] In this embodiment, the operation time refers to the total time of the corresponding historical steps, and the historical leak detection rate refers to the leak detection rate under the corresponding historical leak detection.

[0036] In this embodiment, the first loss coefficient = , where tz is the operation time of the corresponding historical step; t0 is the setting time of the corresponding historical step; is the corresponding historical leak detection rate; ts is the sum of the set time of each historical step in the process adopted by the corresponding leak detection; In this embodiment, the leak detection density function = G (the first loss coefficient of each historical step, the coefficient distribution density of all first loss coefficients), and is obtained based on a density analysis model. The model is obtained by training a neural network model based on the first loss coefficients of different historical steps and the analysis results of this combination as samples, and thus the leak detection density function can be directly obtained.

[0037] In this embodiment, the leak detection matrix = , where the leak detection information itself includes parameter descriptions and values ​​under different steps, and the leak detection sequence is the leak detection sequence.

[0038] In this embodiment, the first distribution function is obtained by determining the position distribution of all values ​​described by the corresponding parameters in the numerical coordinate system.

[0039] In this embodiment, the second loss coefficient = the first distribution function × the function loss variable, wherein the function loss variable is obtained by matching from a distribution-variable comparison table, which contains different distribution functions and loss variables corresponding to the function, mainly for calculating the second loss coefficient.

[0040] In this embodiment, the horizontal axis of the history curve is the corresponding parameter description, and the vertical axis is the corresponding second loss coefficient.

[0041] In this embodiment, the horizontal axis of the standard curve is the corresponding parameter description, and the vertical axis is the corresponding standard loss coefficient, and the standard loss coefficient is pre-set and is for different parameter descriptions.

[0042] In this embodiment, the quantity to be adjusted is: the second loss coefficient of the corresponding parameter description (i.e., the leak detection parameter) is greater than the standard loss coefficient, which is calculated based on (second loss coefficient-standard loss coefficient)×the set quantity of the corresponding parameter, wherein the set quantity of the corresponding parameter is obtained by matching from a parameter-quantity comparison table, which includes different parameter descriptions and the set quantities that match them.

[0043] In this embodiment, each historical step includes at least one leak detection parameter, that is, the parameters involved in each step are well-corresponding and known.

[0044] In this embodiment, the reset adjustment refers to setting the coefficients of the parameter variables involved in the leak detection correlation function to 1 or 0. If the coefficient is less than 0.05, it is set to 0, and if the coefficient is greater than 0.95, it is set to 1.

[0045] The beneficial effects of the above technical solution are: the first loss coefficient is preliminarily determined based on the operation time and the historical leak detection rate, and then the density function is constructed, and subsequently the adjusted amount of each parameter is obtained by constructing a matrix and analyzing the second loss coefficient of each parameter, thereby realizing the optimization of the function, and combining with the reset to ensure the rationality of the function, providing a basis for the screening of leak detection processes.

[0046] The present invention provides a packaging device leak detection method based on an intelligent sensor, which optimizes the leak detection density function, including: Fill the first variable and the leak detection parameter corresponding to each historical step in the leak detection density function into the variable comparison table respectively, obtain a number of leak detection parameters matching each first variable, and establish a control group; Based on the control relationship of each control group and in combination with the amount to be adjusted of the leak detection parameter involved in the control group, the variable coefficient adjustment value for the corresponding first variable in the leak detection density function and the additional variable are obtained to obtain the adjustment density function; Obtain the adjustment density function of each leak detection involved in the corresponding leak detection process, and obtain all value coefficients involved in the same second variable; Extract the strong correlation coefficient and the weak correlation coefficient from all the value coefficients involved in the same second variable, and calculate the absolute value array of the coefficient difference between the strong correlation coefficient, the weak correlation coefficient and each value coefficient in turn, to determine whether it is necessary to add an auxiliary coefficient for the same second variable; If not necessary, the trend status is determined based on all value coefficients involved in the same second variable; If necessary, determining the additional number of auxiliary coefficients to be interpolated based on the coefficient difference absolute value array to determine the trend state of all the interpolated coefficients; Based on the trend state of each second variable, a leak detection related function corresponding to the leak detection process is determined.

[0047] In this embodiment, the leakage detection density function = u1x1 + u2x2 + u3x3 + ... + unxn, wherein x1, x2, x3, ..., xn represent the corresponding first variables, u1, u2, u3, ..., un represent the corresponding value coefficients (related to density, the larger the density, the higher the corresponding value coefficient, and the value is between 0 and 1), and n represents the number of historical steps. Since each step contains at least one parameter, which are all known and set, a direct comparison between the steps and the parameters can be obtained based on the variable comparison table. Furthermore, the first variable corresponding to each step involves a corresponding number of leakage detection parameters, and the control group = {corresponding to the first variable, a number of leakage detection parameters matching the first variable}.

[0048] In this embodiment, the control relationship is the relationship between the variable and the parameter.

[0049] In this embodiment, the variable coefficient adjustment value = the amount to be adjusted of the leak detection parameter involved in the corresponding first variable × the sum of the parameter weights. It should be noted that the sum of the weights of all parameters involved in the process is 1, and the parameters and parameter weights involved in each step are pre-planned and can be directly used for calculation.

[0050] In this embodiment, the additional variables refer to the new variables added after the leakage detection density function is adjusted, for example, the adjusted density function = u1x1 + u2x2 + u3x3 + ... + unxn + un1xn1 + un2xn2, wherein un1 and un2 are the value coefficients of the new variables xn1 and xn2, and the additional variables are obtained by inputting the vector composed of the variable coefficient adjustment values ​​of all the first variables involved in the leakage detection density function into the vector analysis model, and the vector analysis model is based on the vector composed of different values ​​and the additional results of the vector as samples to train the neural network model, so the additional variables can be directly obtained.

[0051] In this embodiment, each variable in the adjustment density function is collectively referred to as a second variable.

[0052] In this embodiment, linear fitting is performed on all value coefficients of the same second variable, and the values ​​on the fitting line are averaged. The result is regarded as a strong correlation coefficient, the discrete points not on the fitting line are locked, and the straight-line distance between each discrete point and the corresponding fitting line is calculated, and the value of the discrete point with the largest straight-line distance is regarded as a weak correlation coefficient.

[0053] In this embodiment, the array of value coefficients = {the absolute value of the difference between the strong correlation coefficient and the corresponding value coefficient, the absolute value of the difference between the weak correlation coefficient and the corresponding value coefficient}; If the absolute value of the difference between the weak correlation coefficient and the corresponding value coefficient / the absolute value of the difference between the strong correlation coefficient and the corresponding value coefficient>N1, it is considered that an auxiliary coefficient needs to be added, where the value of N1 is 10, and the auxiliary coefficient = the absolute value of the difference between the weak correlation coefficient and the corresponding value coefficient - the absolute value of the difference between the strong correlation coefficient and the corresponding value coefficient.

[0054] Otherwise, it is deemed that there is no need to add the auxiliary coefficient.

[0055] In this embodiment, a curve is drawn for all value coefficients involved in the second variable to analyze the trend of the curve, and the trend states include: horizontal trend and fluctuating trend.

[0056] In this embodiment, the result of rounding up (the absolute value of the difference between the weak correlation coefficient and the corresponding value coefficient / the absolute value of the difference between the strong correlation coefficient and the corresponding value coefficient-N1) / N1 is the additional quantity.

[0057] In this embodiment, coefficient interpolation refers to randomly inserting coefficients at different positions among all value coefficients corresponding to the second variable, and then redrawing the curve to determine the trend status.

[0058] In this embodiment, if the trend state is a horizontal state, the average value of all value coefficients involved is used as the correlation coefficient of the second variable. If the trend state is not a horizontal state, the average value of the remaining coefficients is obtained after removing the maximum and minimum values ​​in the value coefficients, and the variance of the remaining coefficients is subtracted to obtain the result as the correlation coefficient of the second variable. At this time, the leak detection correlation function based on the variables and the correlation coefficients is: r1x1+r2x2+r3x3+...+rnxn+rn1xn1+rn2xn2, wherein r1, r2, r3,..., rn, rn1, rn2 are regarded as the correlation coefficients corresponding to the second variable.

[0059] The beneficial effect of the above technical solution is: by establishing a control relationship between variables and parameters, it is convenient to determine the adjustment density function, and then by analyzing the value coefficient of each variable, the trend state is effectively judged to obtain the leak detection related function, which provides a basis for subsequent screening of the final process.

[0060] The present invention provides a packaging device leak detection method based on an intelligent sensor, which quantifies the parameter set, including: Obtaining step attributes of each process step in the final process and constructing an attribute vector; Acquire a quantization model matching the attribute vector from an attribute-model database, wherein each network layer in the quantization model corresponds to a process step, and each network layer contains parameter quantization weights and quantization change information; The parameter set is quantized according to the quantization model to obtain a quantization result.

[0061] In this embodiment, the process step refers to the execution attributes of the corresponding step, such as helium filling attributes, nitrogen filling attributes, etc., and the attribute vector = {step attributes of each process step}.

[0062] In this embodiment, the attribute-model database contains vectors combined with different attributes and quantization models matching the combined vectors, in order to standardize the values ​​of the parameters in the parameter set to facilitate subsequent analysis.

[0063] In this embodiment, the number of network layers included in the quantization model is consistent with the number of process steps.

[0064] In this embodiment, the parameter quantization weights and quantization change information of each network layer are known, in order to reasonably standardize the values ​​of the corresponding parameters.

[0065] The beneficial effect of the above technical solution is that: by constructing the attribute vector, it is easy to match the quantitative model from the database, realize the quantization of the parameter value, and facilitate the subsequent analysis.

[0066] The present invention provides a packaging device leak detection method based on an intelligent sensor, which obtains a quantitative model matching the attribute vector from an attribute-model database, comprising: When there is a first model in the attribute-model database that is completely consistent with the attribute vector, the first model is regarded as a quantization model; When there is no first model in the attribute-model database that is completely consistent with the attribute vector, cluster analysis is performed on all step attributes under the final process to obtain a number of clusters, and an initial cluster vector is constructed based on cluster centers of the clusters; Determine the distance between each cluster center and the remaining attributes in the corresponding clustering result in turn, and perform weakening analysis on the corresponding step attributes according to the distance; ; in, represents the weakening analysis result of the jth remaining attribute in the corresponding clustering result; n1 represents the total number of attributes involved in the corresponding clustering result; sum1 represents the total number of attributes involved in the final process; represents the distance between the cluster center of the corresponding clustering result and the jth remaining attribute; a1 represents the set threshold; The attributes whose weakening analysis results are 0 are replaced with 0 and added to the initial cluster vector, and the attributes whose weakening analysis results are non-0 are replaced with the original attributes and added to the initial cluster vector to obtain the current cluster vector; Comparing the current cluster vector with the attribute-model database, if there is a second model that is completely consistent with the current cluster vector, then considering the second model as a quantization model; Otherwise, obtaining from the attribute-model database a third model having the largest number of attributes matching the attribute vector, and at the same time, obtaining from the attribute-model database a fourth model having the largest number of attributes matching the current cluster vector; The parameter quantization weights of the same-layer networks in the third model and the fourth model are respectively obtained, and the required network layer is locked according to max(y1i,y2i), wherein y1i represents the parameter quantization weight of the i-th layer network in the third model; y2i represents the parameter quantization weight of the i-th layer network in the fourth model; All required network layers are reintegrated to obtain a quantized model.

[0067] In this embodiment, cluster analysis is implemented based on the K-means algorithm, thereby obtaining clusters.

[0068] In this embodiment, the threshold value is set to 0.25.

[0069] In this embodiment, for example, the initial cluster vector = {cluster 1 cluster 2 cluster 3 blank element 01 blank element 02}, at this time, the current cluster vector obtained according to the weakening analysis result = {cluster 1 cluster 2 cluster 3 0 1}, and the attributes 0 and 1 are replaced and supplemented to be consistent with the number of elements involved in the original attribute vector.

[0070] In this embodiment, the number of attributes refers to the number of attribute elements in the corresponding vector.

[0071] In this embodiment, after obtaining the third model and the fourth model, and the parameter quantization weights of each network layer in the model are known, a more reliable network layer can be screened out by size comparison, and the quantization model can be obtained by reintegrating the network layer, that is, the network layer weights are reconfigured according to the current quantization weights of the screened network layer, and the original quantization change information remains unchanged.

[0072] The beneficial effects of the above technical solution are: by constructing an attribute vector to match the quantization model, when there is no completely matching model, cluster analysis and replacement of 0 and 1 are performed to ensure the rationality of model acquisition and the symmetry of processing each step, and further by screening the model with the most matching number of attributes under the two vectors, and combining the comparison of weights to obtain the final network layer, the reliability of the fusion model is guaranteed to ensure the accuracy of the value of the quantization parameter.

[0073] The present invention provides a packaging device leak detection method based on an intelligent sensor, which obtains the influence of each process step on the leak detection result, including: Compare the filling results based on each process step in the filling analysis table with the corresponding standard results; Determine the extent of the impact based on the comparison results.

[0074] In this embodiment, the degree of influence=(1-sim(filling result, standard result)) / N2, where N2 represents the number of parameters involved in the corresponding process step.

[0075] The beneficial effect of the above technical solution is that the degree of impact can be effectively determined by comparing the filling result with the standard result.

[0076] The present invention provides a packaging device leak detection method based on an intelligent sensor, which obtains the final leak detection result of the packaging sensor to be leak detected based on all the influence degrees and the correlation between adjacent process steps, including: Obtaining an initial detection result of the last process step, and adjusting the initial detection result once based on all impact levels and correlations between adjacent process steps; Determine the characteristic parameters of the packaged sensor to be leak tested, and determine in turn the coefficient of influence of the last process step on each characteristic parameter; According to all the influencing coefficients, the first leak detection result is adjusted to obtain the final leak detection result.

[0077] In this embodiment, the initial detection result refers to the initial leak detection rate of the smart sensor determined based on the last process; The result of one adjustment = initial leak detection rate × (1 + ln (1 + each degree of influence × the sum of the correlation coefficients of the corresponding adjacent process steps involved / the number of process steps)), where the correlation coefficient ranges from 0 to 1. It should be noted that the correlation between adjacent process steps is set in advance. For example, step 1 will affect the result of step 2, but step 2 will not affect the result of step 3. At this time, the correlation coefficient of step 2 and step 3 is 0.

[0078] In this embodiment, the characteristic parameter refers to the size, sensitivity, integration level, sensitivity, etc. of the sensor.

[0079] In this embodiment, the influence coefficient of each characteristic parameter affected by the last process step is obtained from the characteristic-relationship comparison table, and the value range is (0, 0.05).

[0080] Final leak detection result = one-time adjustment result × (1 + sum of all influencing coefficients).

[0081] The beneficial effect of the above technical solution is: adjusting the result once based on the degree of influence and the correlation, and then adjusting the result again in combination with the influence on the characteristic parameters, so as to ensure the accuracy of the final leak detection result.

[0082] The present invention provides a packaging device leak detection system based on an intelligent sensor, such as Figure 2 As shown, including: A process matching module, used to determine the sensing type of the packaged sensor to be leak tested, and obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaged process comparison table; A function building module is used to extract a historical leak detection information set matching each leak detection process from a historical database, build a leak detection effective function, and select the final process; A quantification module, used to sequentially obtain a parameter set for leak testing the packaged sensor to be leak tested according to the final process, and quantify the parameter set, wherein the parameter set includes a number of leak testing parameters under each process step; The result acquisition module is used to fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

[0083] The beneficial effects of the above technical solution are: the final process is screened based on the sensing type matching of several leak detection processes, which effectively improves the leak detection efficiency, and subsequently determines the leak detection result by combining quantitative analysis of the parameter set and the impact degree analysis after the result filling with the correlation, thereby effectively ensuring the reliability of the smart sensor.

[0084] The present invention provides a packaging device leak detection device based on an intelligent sensor, which is used to execute any of the packaging device leak detection methods based on an intelligent sensor.

[0085] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A packaging device leak detection method based on intelligent sensors, characterized in that: include: Step 1: Determine the component combination type of the packaged sensor to be leak tested, and obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaging process comparison table; Step 2: Extract the historical leak detection information set matching each leak detection process from the historical database, construct the leak detection effective function, and select the final process; Step 3: sequentially obtaining parameter sets for leak testing the packaged sensor to be leak tested according to the final process, and quantifying the parameter sets, wherein the parameter sets include values ​​of several leak testing parameters in each process step in the final process; Step 4: Fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

2. The packaging device leak detection method based on intelligent sensor according to claim 1 is characterized in that: Construct a leak detection effective function, including: Determine the first loss coefficient of each historical step based on the operation time of each historical step involved in the same leak detection and in combination with the historical leak detection rate; Determine the leak detection density function based on the first loss coefficients of all historical steps under the same leak detection; Each leak detection information in the historical leak detection information set of each leak detection process is placed in the order of leak detection to construct a leak detection matrix; Constructing a first distribution function of each column vector in the leakage detection matrix, and obtaining a second loss coefficient based on each leakage detection parameter; A historical curve is constructed based on all the second loss coefficients, and parameter alignment is performed with the standard curve of the leak detection process to obtain an amount to be adjusted for each leak detection parameter; According to the to-be-adjusted amounts of all leak detection parameters involved in the same historical step, the leak detection density function is optimized to obtain a leak detection related function of the corresponding leak detection process based on the historical leak detection information set; The leak detection related function is reset and adjusted to obtain a leak detection effective function.

3. The packaging device leak detection method based on intelligent sensor according to claim 2 is characterized in that: Optimizing the leak detection density function includes: Fill the first variable and the leak detection parameter corresponding to each historical step in the leak detection density function into the variable comparison table respectively, obtain a number of leak detection parameters matching each first variable, and establish a control group; Based on the control relationship of each control group and in combination with the amount to be adjusted of the leak detection parameter involved in the control group, the variable coefficient adjustment value for the corresponding first variable in the leak detection density function and the additional variable are obtained to obtain the adjustment density function; Obtain the adjustment density function of each leak detection involved in the corresponding leak detection process, and obtain all value coefficients involved in the same second variable; Extract the strong correlation coefficient and the weak correlation coefficient from all the value coefficients involved in the same second variable, and calculate the absolute value array of the coefficient difference between the strong correlation coefficient, the weak correlation coefficient and each value coefficient in turn, to determine whether it is necessary to add an auxiliary coefficient for the same second variable; If not necessary, the trend status is determined based on all value coefficients involved in the same second variable; If necessary, determining the additional number of auxiliary coefficients to be interpolated based on the coefficient difference absolute value array to determine the trend state of all the interpolated coefficients; Based on the trend state of each second variable, a leak detection related function corresponding to the leak detection process is determined.

4. The packaging device leak detection method based on intelligent sensor according to claim 1, characterized in that: Quantifying the parameter set includes: Obtaining step attributes of each process step in the final process and constructing an attribute vector; Acquire a quantization model matching the attribute vector from an attribute-model database, wherein each network layer in the quantization model corresponds to a process step, and each network layer contains parameter quantization weights and quantization change information; The parameter set is quantized according to the quantization model to obtain a quantization result.

5. The packaging device leak detection method based on intelligent sensor according to claim 4 is characterized in that: Acquiring a quantitative model matching the attribute vector from an attribute-model database, including: When there is a first model in the attribute-model database that is completely consistent with the attribute vector, the first model is regarded as a quantization model; When there is no first model in the attribute-model database that is completely consistent with the attribute vector, cluster analysis is performed on all step attributes under the final process to obtain a number of clusters, and an initial cluster vector is constructed based on cluster centers of the clusters; Determine the distance between each cluster center and the remaining attributes in the corresponding clustering result in turn, and perform weakening analysis on the corresponding step attributes according to the distance; ; in, represents the weakening analysis result of the jth remaining attribute in the corresponding clustering result; n1 represents the total number of attributes involved in the corresponding clustering result; sum1 represents the total number of attributes involved in the final process; represents the distance between the cluster center of the corresponding clustering result and the jth remaining attribute; a1 represents the set threshold; The attributes whose weakening analysis results are 0 are replaced with 0 and added to the initial cluster vector, and the attributes whose weakening analysis results are non-0 are replaced with the original attributes and added to the initial cluster vector to obtain the current cluster vector; Comparing the current cluster vector with the attribute-model database, if there is a second model that is completely consistent with the current cluster vector, then considering the second model as a quantization model; Otherwise, obtaining from the attribute-model database a third model having the largest number of attributes matching the attribute vector, and at the same time, obtaining from the attribute-model database a fourth model having the largest number of attributes matching the current cluster vector; The parameter quantization weights of the same-layer networks in the third model and the fourth model are respectively obtained, and the required network layer is locked according to max(y1i,y2i), wherein y1i represents the parameter quantization weight of the i-th layer network in the third model; y2i represents the parameter quantization weight of the i-th layer network in the fourth model; All required network layers are reintegrated to obtain a quantized model.

6. The method for leak detection of packaged devices based on intelligent sensors according to claim 1, characterized in that: Get the impact of each process step on the leak detection results, including: Compare the filling results based on each process step in the filling analysis table with the corresponding standard results; Determine the extent of the impact based on the comparison results.

7. The packaging device leak detection method based on intelligent sensor according to claim 6, characterized in that: Based on all the influence degrees and the correlation between the adjacent process steps, the final leak detection result of the packaged sensor to be leak detected is obtained, including: Obtaining an initial detection result of the last process step, and adjusting the initial detection result once based on all impact levels and correlations between adjacent process steps; Determine the characteristic parameters of the packaged sensor to be leak tested, and determine in turn the coefficient of influence of the last process step on each characteristic parameter; According to all the influencing coefficients, the first leak detection result is adjusted to obtain the final leak detection result.

8. A packaging device leak detection system based on intelligent sensors, characterized in that: include: A process matching module, used to determine the combination type of components of the packaged sensor to be leak tested, and to obtain several leak detection processes matching the packaged sensor to be leak tested from a type-packaged process comparison table; A function building module is used to extract a historical leak detection information set matching each leak detection process from a historical database, build a leak detection effective function, and select the final process; A quantification module, used to sequentially obtain a parameter set for leak testing the packaged sensor to be leak tested according to the final process, and perform quantification processing on the parameter set, wherein the parameter set includes values ​​of several leak testing parameters under each process step; The result acquisition module is used to fill the quantitative processing results into the parameter analysis table in sequence to obtain the influence of each process step on the leak detection result, and based on all influence degrees and the correlation between adjacent process steps, obtain the final leak detection result of the packaged sensor to be leak tested.

9. A packaging device leak detection device based on an intelligent sensor, characterized in that: Used to execute the packaging device leak detection method based on intelligent sensor as described in any one of claims 1-7.

Images