Vacuum box helium leak detection equipment data processing method

Abstract

The invention relates to the technical field of data processing, in particular to a vacuum box helium leak detection equipment data processing method. The method comprises the following steps: acquiring a helium concentration signal, filtering the helium concentration signal to obtain a preprocessed concentration signal, setting a reference point of each data point, constructing a reference point set of each data point, and calculating the credibility of the reference points in the reference point set; determining the confidence weight of each reference point based on the credibility of the reference points; and taking the product of each confidence weight and the preprocessed concentration data of the corresponding reference point as a local contribution value, and taking the sum of a plurality of local contribution values as a final filtering result. The helium concentration data processing method and device have the effect of improving the helium concentration data processing accuracy.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly to a data processing method for a helium leak detection device in a vacuum chamber. Background Art

[0002] In the field of precision manufacturing, it is usually necessary to monitor and evaluate the sealing performance of products. As an efficient leak detection tool, the helium leak detection device in a vacuum chamber is applied in many fields such as aerospace and automotive manufacturing. The leak detection principle of the helium leak detection device in a vacuum chamber is mainly based on the unique properties of helium gas. Helium gas has the characteristics of small molecular size and fast diffusion speed. During the product detection process, the product is placed in the vacuum chamber of the helium detection device in a vacuum chamber, a clamping cavity is formed between the product and the vacuum chamber, the gas in the clamping cavity is pumped out to construct a vacuum environment. Then helium gas is introduced into the product. Under the action of air pressure, helium gas can enter the vacuum chamber from the leakage point, and then the helium gas concentration information in the vacuum chamber is collected by a sensor for detecting the helium gas concentration in the vacuum chamber. Whether the product leaks can be judged by analyzing the helium gas concentration information. When collecting the helium gas concentration signal in the vacuum chamber, electromagnetic interference (power lines, industrial equipment or wireless signals) around the device may affect the signal collection of the sensor, and then noise data is generated in the helium gas concentration signal. Therefore, it is usually necessary to perform denoising processing on the helium gas concentration signal before analyzing the data of the helium gas concentration.

[0003] The NLM (Non-Local Means) non-local mean filtering algorithm is a common filtering algorithm. It can not only be applied to the denoising of two-dimensional images, but also be used for the processing of one-dimensional signals. During the filtering process, the algorithm first defines a search window and a reference segment based on the data point to be processed, searches for similar segments to the reference segment in the window, and calculates the weighted average of the segments based on the similarity to update the value of the data point to be processed. The distribution of weights depends on the size of the similarity. The filtering result of the data point only depends on the locally similar neighboring data, which may lead to a greater influence of noise data on the filtering result, especially when the noise is concentrated or there are many outliers, resulting in the distortion of the subsequent similarity measurement, and then it may wrongly assign high weights to actually irrelevant segments, ultimately resulting in inaccurate filtering results. Summary of the Invention

[0004] In order to reduce the occurrence of inaccurate filtering during the data processing process, this application provides a data processing method for a helium leak detection device in a vacuum chamber.

[0005] This application provides a data processing method for a helium leak detection device in a vacuum chamber, adopting the following technical solutions: A data processing method for a helium leak detection device in a vacuum chamber, comprising the steps of: obtaining a helium concentration signal, filtering the helium concentration signal to obtain a preprocessed concentration signal, setting reference points for each data point, constructing a reference point set for each data point, and calculating the credibility of the reference points in the reference point set; determining the confidence weight of each reference point based on the credibility of the reference point; taking the product of each confidence weight and the preprocessed concentration data of the corresponding reference point as a local contribution value, and taking the sum of multiple local contribution values as the final filtering result; Among them, the steps of calculating the credibility of the reference points in the reference point set include: constructing a reference segment based on the reference points; constructing a neighborhood segment corresponding to the reference segment based on the position of the reference segment; calculating the noise probability of the reference segment and the noise probability of the neighborhood segment; determining a time weight based on the time interval between the data point and the corresponding reference point, and taking the product of the noise probability of the helium concentration data in the reference segment, the noise probability of the neighborhood segment, and the time weight as the degree of abnormality, and processing the degree of abnormality through an exponential function to obtain the credibility of the reference segment for the data point.

[0006] The beneficial effects are as follows: First, the collected helium concentration signal is initially filtered to obtain a preprocessed concentration signal. Since the traditional filtering method is affected by noise and there is a situation where the filtering result is inaccurate. Therefore, in this application, the data points in the preprocessed concentration signal are adjusted again to obtain the final filtering result of each data point, thereby improving the accuracy of the final processing of the helium concentration data.

[0007] During the process of adjusting the data points in the preprocessed signal, on the one hand, it is based on its neighboring reference points. During the initial filtering of the helium concentration signal to obtain the preprocessed concentration signal, if there is a lot of noise in the reference data segment corresponding to a data point, then the preprocessed concentration data corresponding to the reference point is also inaccurate. Therefore, during the process of adjusting the preprocessed concentration data of the data point to obtain the final filtering result, the credibility of its corresponding reference point is obtained, and the reference point is weighted based on the credibility, thereby further improving the accuracy of the final filtering result calculation.

[0008] During the calculation process of the credibility, analyze the probability that the helium concentration data is a noise signal, and finally obtain the credibility of the reference point based on the noise probability of the data in the reference segment corresponding to the reference point.

[0009] Optionally, the steps of calculating the noise probability of the helium concentration data in the reference segment include: obtaining the pressure signal in the vacuum chamber, calculating the noise factor of each data point in the helium concentration signal based on the pressure signal and the helium concentration signal, and taking the mean value of the noise factors corresponding to multiple data points in the reference segment as the noise probability of the reference segment.

[0010] The beneficial effects are as follows: During the airtightness detection of the product, if the product has poor sealing, that is, there are leakage points. When helium leaks, on the one hand, it will cause a change in the helium concentration in the vacuum chamber, and on the other hand, it will also cause a change in the pressure in the vacuum chamber. Therefore, in this method, the pressure signal in the vacuum chamber is obtained, and the noise factor of the data points in the helium concentration signal is analyzed by combining the changes in the pressure signal and the changes in the helium concentration signal. For a reference segment, there are multiple corresponding helium concentration data. Therefore, the mean value of the noise factors corresponding to the multiple helium concentration data is used as the noise probability of the reference segment.

[0011] Optionally, the steps of calculating the noise factor of each data point in the helium concentration signal based on the pressure signal and the helium concentration signal include: for any data point in the ammonia concentration signal, construct a comparison data segment; take the standard deviation of all data in the comparison data segment as the overall difference, and determine the contribution degree of the data point corresponding to the comparison data segment to the overall difference based on the overall difference; obtain the data ranking of the pressure data in the pressure signal corresponding to the data point in the comparison data points, and take the product of the data ranking, contribution degree, and concentration data of the pressure data at the same moment as the noise degree, and take the normalized result of the noise degree as the noise factor.

[0012] The beneficial effects are as follows: For a noise signal, it usually shows data that stands out from the surrounding. Therefore, in this method, a comparison data segment corresponding to the data point is constructed based on the data point, and the helium concentration data corresponding to the data point is compared with the helium concentration data in the comparison data segment. Calculate the standard deviation of all data in the comparison data segment. The standard deviation represents the degree of dispersion of the data. At the same time, the value of the standard deviation is based on the value of each data. If a data contributes more to the standard deviation, it means that the data is more prominent in the comparison data segment. Finally, calculate the noise factor of the data point based on the contribution degree, pressure data, and helium concentration data corresponding to the current data point.

[0013] Optionally, the steps of obtaining the data ranking of the pressure data in the pressure signal corresponding to the data point in the comparison data points include: for any moment, obtain multiple pressure data in the pressure signal corresponding to the comparison data segment, and arrange them in ascending order based on the numerical size, and take the position of the pressure data corresponding to this moment as the data ranking of the pressure data at this moment.

[0014] The beneficial effects are as follows: Here, the data in the comparison data segment is sorted to highlight the overall level of the helium concentration data of the current data point in the comparison data segment.

[0015] Optionally, the steps of constructing the comparison data segment include: determining the length of the comparison data segment , and obtaining the data segment centered on the current data point with a length of as the comparison data segment.

[0016] The beneficial effect is that an equal amount of data is selected on both sides of the current data point to form a comparison data segment.

[0017] Optionally, the step of determining the contribution degree of a data point to the overall difference based on the overall difference includes: for any data point, obtaining the standard deviation of all data except itself in the corresponding comparison data segment as the reference difference, and taking the normalized result of the difference between the overall difference and the reference difference as the contribution degree of the data point to the overall difference.

[0018] The beneficial effect is that in this method, the standard deviation of all data except the helium concentration data corresponding to the data point is calculated, and the contribution degree of the current data point to the standard deviation is determined by comparing the differences between the two standard deviations.

[0019] Optionally, the step of constructing a neighborhood segment corresponding to the reference segment based on the position of the reference segment includes: determining a sampling time range based on the reference segment, and obtaining a preset number of data segments within the sampling time range as the neighborhood segment.

[0020] Optionally, the step of setting reference points for each data point and constructing a reference point set for each data point includes: determining the number of reference point sets , taking the data points that are adjacent and continuous on both sides of the data point as the reference points; taking the set composed of multiple reference points as the reference point set.

[0021] The beneficial effect is that under normal circumstances, the overall helium concentration signal changes linearly, so the current data point can be optimized and adjusted based on adjacent data points. Furthermore, in this method, the data points adjacent to both sides of the current data point are used as reference points to adjust the preprocessed concentration data of the current data point.

[0022] Optionally, the step of constructing a reference segment based on the reference point includes taking a data segment with a preset length centered on the reference point as the reference segment corresponding to the reference point.

[0023] Optionally, the step of obtaining a preset number of data segments within the sampling time range includes: obtaining multiple data segments of the same length as the reference segment, calculating the mean value of the corresponding helium concentration data in the data segment, obtaining the absolute difference between the mean value and the mean value of the helium concentration data in the reference segment, and selecting the data segments with the smallest absolute difference as the neighborhood segment.

[0024] This application has the following technical effects: The helium concentration signal is initially processed to obtain a preprocessed concentration signal, based on the noise corresponding to each data point in the helium concentration signal. Therefore, the data in the preprocessed concentration signal is adjusted again to reduce the influence of noise on the final filtering result and improve the accuracy of helium concentration signal processing. Brief Description of the Drawings

[0025] Figure 1 is a flowchart of a method for processing data of a helium leak detection device for a vacuum chamber according to an embodiment of the present application.

[0026] Figure 2 is a flowchart of step S2 in a method for processing data of a helium leak detection device for a vacuum chamber. Detailed Embodiment

[0027] An embodiment of the present application discloses a method for processing data of a helium leak detection device for a vacuum chamber, which obtains a helium concentration signal, performs primary filtering processing on the helium concentration signal to obtain a preprocessed concentration signal. For each data point in the concentration signal, a corresponding reference point set of each data point is constructed, and each data point is subjected to secondary optimization adjustment according to the credibility of the reference points in the reference point set to complete the final filtering of the data. In this method, the credibility reflects whether the primary filtering result corresponding to the data point is reliable. Weighted filtering based on the credibility of the reference points enables the final filtering result to more truly reflect the potential trend of the data. This method effectively reduces the influence of noise on the filtering result and improves the accuracy after data processing.

[0028] Referring to Figure 1 , a method for processing data of a helium leak detection device for a vacuum chamber includes steps S1 - S3.

[0029] S1: Obtain a helium concentration signal, filter the ammonia concentration signal to obtain a preprocessed concentration signal, set reference points for each data point, and construct a reference point set for each data point.

[0030] During the process of performing a sealing detection on the product, a helium concentration signal in the vacuum chamber is collected by a helium sensor arranged in the vacuum chamber. In this embodiment, the collection frequency is once per second, and in other embodiments, it can be twice per second, three times per second, etc.

[0031] Perform primary filtering processing on the helium concentration signal using the NLM algorithm to obtain a preprocessed concentration signal.

[0032] For any one data point, construct a reference point set for this data point.

[0033] For any one data point, optimize itself through its adjacent data points, so select the adjacent data points as reference points.

[0034] Determine that the data volume of the reference points is 。During the initial filtering process using the NLM algorithm, a corresponding reference data segment is determined for each data point. By searching for segments similar to the reference data point in the search box, the values of the data points to be processed are updated by weighted averaging of the segments based on similarity. The steps of this NLM algorithm are conventional technical means in the art and will not be elaborated in this embodiment. Here is equal to the length of the reference data segment. In this embodiment, the length of the reference data segment is 21. During the specific process of selecting reference points, for any data point, 10 data points are selected on each side of the data point to form reference points. Multiple reference points and the data point itself form a reference point set.

[0035] S2: Calculate the credibility of the reference points in the reference point set.

[0036] Referring to Figure 2 , the steps of calculating the credibility of the reference points in the reference point set include: Step S21 - Step S23.

[0037] S21: Construct a reference segment based on the reference points; construct a neighborhood segment corresponding to the reference segment based on the position of the reference segment.

[0038] For the credibility of any reference point, first construct a reference segment based on the reference point. Specifically: Obtain a data segment with a length of centered on the reference point as the reference segment. It can also be understood as obtaining data on both sides of the reference point centered on the reference point, so as to form the reference segment corresponding to the reference point.

[0039] Based on the position of the reference point and the length of the reference segment, it can be known that for any data point to be processed, there are reference points on both sides of it, each reference point corresponds to a reference segment, and at the same time each reference segment includes the current data point (the data point to be processed).

[0040] For the reference segment corresponding to each reference point, construct a neighborhood segment corresponding to the reference segment.

[0041] Construct a data segment adjacent to the reference segment as the neighborhood segment of the reference segment. The credibility of the data inside the reference segment is jointly reflected by the neighborhood segment and the reference segment. Specifically, based on the reference segment, determine the sampling time range, and obtain a preset number of data segments as the neighborhood segments within the sampling time range.

[0042] In this embodiment, the sampling time range is 10 minutes before and after the sampling moment of the reference point corresponding to the reference segment. Search for multiple data segments of the same length as the reference segment within this time period, calculate the mean value of the corresponding helium concentration data in the data segments, obtain the absolute difference between the mean value and the mean value of the helium concentration data in the reference segment, and select the data segments with the smallest absolute difference as the neighborhood segments.

[0043] The smaller the absolute difference here, the more similar the mean values of the corresponding helium concentration data in the reference segment and the neighborhood segment are. It can also be understood that the two data segments are more similar, thereby being able to further reflect the situation of the data in the reference segment.

[0044] S22: Calculate the noise probability of the reference segment and the noise probability of the neighborhood segment.

[0045] The steps for calculating the noise probability of the helium concentration data in the reference segment include: obtaining the pressure signal in the vacuum chamber and calculating the noise factor of each data point in the helium concentration signal based on the pressure signal and the helium concentration signal.

[0046] For any data point in the ammonia concentration signal, construct a comparison data segment; take the standard deviation of all data in the comparison data segment as the overall difference, and determine the contribution degree of the data point corresponding to the comparison data segment to the overall difference based on the overall difference; obtain the data ranking of the pressure data in the pressure signal corresponding to the data point in the comparison data points, and take the product of the data ranking, contribution degree, and concentration data of the pressure data at the same moment as the noise degree, and take the normalized result of the noise degree as the noise factor. Take the mean value of the noise factors corresponding to multiple data points in the reference segment as the noise probability of the reference segment.

[0047] For any data point in the helium concentration signal, analyze the relationship between this data point and the adjacent data points. Therefore, construct a comparison data segment corresponding to the data point. During the construction of the comparison data segment, first determine the length of the comparison data segment , and obtain the data segment centered on the current data point and with a length of as the comparison data segment. In this embodiment, the length of the comparison data segment is 31. Select 15 adjacent and continuous data points on each side of the data point, and the 15 adjacent data points and the data point itself form the comparison data segment. Subsequently, analyze the relationship between the helium concentration corresponding to the data point in the comparison data segment.

[0048] In the comparison data segment, the standard deviation of the data in the comparison data segment represents the degree of dispersion of the data. During the calculation of the standard deviation, the value of each concentration data contributes to the final standard deviation to a certain extent. If the helium concentration data of a data point is larger, then the gain of the standard deviation of this data in the comparison data segment is greater. This further indicates that the concentration data is more prominent among the neighboring concentration data, that is, the concentration data of this data point is much larger or much smaller than the neighboring concentration data, and then the possibility that it belongs to noise is greater. Based on this, the contribution degree of the data point to the whole can reflect the possibility that the data point is noise to a certain extent. Therefore, in this step, the standard deviation of the overall data in the comparison data segment is calculated, and this standard deviation is used as the overall difference. At the same time, for the comparison data segment corresponding to a data point, the standard deviation of all data points except the data point itself in the comparison data segment is calculated, and this standard deviation is used as the reference difference, and the absolute difference between the overall difference and the reference difference is used as the contribution degree of this data point to the overall difference.

[0049] There are also some fluctuations in the true helium concentration signal itself. When the data point is the peak of the helium concentration signal, it is also likely to cause a large gain in the standard deviation of the comparison data segment corresponding to this data. Therefore, in this application, the pressure signal in the vacuum chamber is obtained. Based on the actual scenario, when the helium concentration changes, helium enters the interior of the vacuum chamber, which will in turn cause the air pressure in the vacuum chamber to change. In the true helium concentration signal, the change in helium concentration is usually accompanied by the change in the pressure in the vacuum chamber. If a data point has a large helium concentration, but the corresponding pressure data does not change at the same time, then the possibility that the helium concentration at this data point belongs to noise is greater. Therefore, in this step, the noise probability of the data point is calculated based on the contribution degree, the helium concentration data, and the pressure data in the vacuum chamber.

[0050] Specifically, the calculation formula for the noise probability of the data point can be expressed as: ; where represents the noise factor of a data point of the helium concentration signal; represents the concentration data corresponding to the data point; represents the overall difference; represents the reference difference; represents the first hyperparameter, mainly used to prevent the situation where it is zero from occurring; represents the data ranking of the pressure data corresponding to the data point at the same moment in the multiple pressure data sequences corresponding to the comparison data segment; represents the linear normalization function.

[0051] Data ranking refers to the order of the pressure data corresponding to a data point among multiple pressure data (arranged from small to large) in the comparison data segment. For example, there is a set of pressure data: 56.5, 55.5, 57.5, 59.5, 63.5, 62.5, 52.5, and the value corresponding to the current data point is 63.5. After sorting the above multiple data from large to small: 63.5, 62.5, 59.5, 57.5, 56.5, 55.5, 52.5, then the data ranking of the current data point is 1.

[0052] In the formula, represents the contribution degree of the concentration data corresponding to the data point to the overall difference. The greater the contribution degree, the more discrete the concentration data corresponding to the data point, and further the greater the possibility that the helium concentration data corresponding to the data point is noise.

[0053] represents the helium concentration corresponding to the data point. The larger this value, the more prominent it is in the comparison data segment, and further the more likely it is to belong to noise data. represents the position of the pressure data corresponding to the data point at the same moment in the sequence sorted from small to large among all pressure data in the comparison data segment. The smaller this value, the smaller the pressure data.

[0054] S23: Determine the time weight based on the time interval between the data point and the corresponding reference point, take the product of the noise probability of the helium concentration data in the reference segment, the noise probability of the neighborhood segment, and the time weight as the degree of abnormality, and process the degree of abnormality through an exponential function to obtain the credibility of the reference segment for the data point.

[0055] In the preprocessed concentration signal, a data point corresponds to multiple reference segments, and each reference segment corresponds to a reference point. If the noise probabilities of multiple data in the helium concentration signal corresponding to the reference segment are all relatively large, it indicates that the filtering result based on this reference segment, that is, the accuracy of the data value corresponding to the data point in the preprocessed concentration signal is lower. At the same time, the credibility of the data value of the reference point is also affected by the data in the neighborhood segment corresponding to the reference segment itself. If the noise probabilities of the data points in all neighborhood segments corresponding to a reference segment are all relatively large, it also indicates that the preprocessed concentration value of this reference point is inaccurate.

[0056] In addition, for a data point, during the process of readjusting the preprocessed concentration data corresponding to the data point, if a reference segment is far from the data point, the reference value of the reference segment is lower. Therefore, the time interval between the data point and the reference segment should be considered simultaneously when calculating the credibility. For this purpose, the time weight corresponding to the reference segment is calculated based on the time interval between the data point and the reference point in this method. Finally, the credibility of the final reference segment is calculated by combining the noise probability of the helium concentration corresponding to the reference segment, the noise probability of the helium concentration corresponding to the neighborhood segment, and the time weight corresponding to the reference segment.

[0057] Specifically, the calculation formula for the credibility of the reference segment corresponding to a data point can be expressed as ; in the formula, represents the credibility of the th reference segment corresponding to the data point; represents the time interval between the data point and the th reference point; represents the second hyperparameter, mainly used to prevent the situation where is 0; represents the noise probability of the th reference segment; represents the noise probability of the th reference segment corresponding to the th neighborhood segment; represents the number of neighborhood segments.

[0058] In the formula, The larger it is, the larger the sampling time interval between the data point and the reference point, and the lower the reference value of the corresponding reference segment, which in turn leads to a lower credibility of the reference segment. The larger it is, the larger the average value of the noise probability of the helium concentration data corresponding to the reference segment. There is a large amount of noise in the helium concentration data corresponding to the reference segment, and this noise will affect the filtering of the helium concentration signal, which in turn leads to low accuracy of the preprocessed concentration data corresponding to the reference point. Similarly, The larger it is, the larger the average value of the noise probability of the helium concentration data in the neighborhood segment corresponding to the reference segment, which in turn indicates a lower accuracy of the corresponding reference point.

[0059] S3: Determine the confidence weights of each reference point based on the credibility of the reference point; multiply each confidence weight by the preprocessed concentration data of the corresponding reference point as the local contribution value, and take the sum of multiple local contribution values as the final filtering result.

[0060] For the preprocessed concentration data of a data point, it is adjusted through multiple reference points around it. Different reference points have different credibility levels. Based on the credibility, the weights corresponding to each reference point are determined, and finally, a weighted average is performed to obtain the final filtering result corresponding to this data point.

[0061] For the final filtering result corresponding to any data point, its calculation formula can be expressed as: ; In the formula, represents the final filtering result of the data point; represents the credibility of the th reference point corresponding to the data point; represents the th reference point corresponding to the preprocessed concentration data; represents the number of reference segments.

[0062] represents the credibility of the reference point. The greater the credibility, the smaller the noise probability of the helium concentration data in the reference segment corresponding to this reference point, and thus the higher the accuracy of the preprocessed concentration data corresponding to the reference point, and a higher weight is given accordingly.

[0063] Repeat the above steps to calculate the final filtering result of each data point in the helium concentration signal, obtain the final filtering signal, and monitor the airtightness of the product based on the final filtering signal. Specifically, the box plot algorithm can be used to detect anomalies in the helium concentration data. For example, 120 consecutive data points in the time series can be selected to form a box plot for anomaly detection. If it is detected that there are 5 consecutive helium concentration anomaly data exceeding the upper limit of the anomaly within the box in the time series, it can be determined that the airtightness of the device being leak-tested is unqualified. The box plot algorithm is a conventional technical means in the art and will not be described in this application.

[0064] The embodiment of the present application also discloses a data processing method for a vacuum chamber helium leak detection device, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, a data processing method for a vacuum chamber helium leak detection device according to the present application is implemented.

[0065] The above system also includes other components well-known to those skilled in the art such as a communication bus and a communication interface. Their settings and functions are known in the art, so they will not be described in detail here.

[0066] The above are all the preferred embodiments of the present application. The protection scope of the present application is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of the present application should be covered within the protection scope of the present application.

Claims

1. A method for processing data of a vacuum chamber helium leak detection device, characterized in that, It includes the steps of: obtaining a helium concentration signal, filtering the helium concentration signal to obtain a preprocessed concentration signal, setting reference points for each data point, and constructing a reference point set for each data point; calculating the credibility of the reference points in the reference point set; determining the confidence weights of each reference point based on the credibility of the reference point; taking the product of each confidence weight and the preprocessed concentration data of the corresponding reference point as the local contribution value, and taking the sum of multiple local contribution values as the final filtering result; Among them, the steps of calculating the credibility of the reference points in the reference point set include: constructing a reference segment based on the reference point; constructing a neighborhood segment corresponding to the reference segment based on the position of the reference segment; calculating the noise probability of the reference segment and the noise probability of the neighborhood segment; determining the time weight based on the time interval between the data point and the corresponding reference point, and taking the product of the noise probability of the helium concentration data in the reference segment, the noise probability of the neighborhood segment, and the time weight as the degree of abnormality, and obtaining the credibility of the reference segment to the data point by processing the degree of abnormality through an exponential function.

2. The data processing method of a vacuum chamber helium leak detection device according to claim 1, characterized in that, The steps of calculating the noise probability of the helium concentration data in the reference segment include: obtaining the pressure signal in the vacuum chamber, calculating the noise factor of each data point in the helium concentration signal based on the pressure signal and the helium concentration signal, and taking the mean value of the noise factors corresponding to multiple data points in the reference segment as the noise probability of the reference segment.

3. A data processing method for a vacuum chamber helium leak detection device according to claim 2, characterized in that, The steps of calculating the noise factor of each data point in the helium concentration signal based on the pressure signal and the helium concentration signal include: for any data point in the ammonia concentration signal, constructing a comparison data segment; taking the standard deviation of all data in the comparison data segment as the overall difference, and determining the contribution degree of the data point corresponding to the comparison data segment to the overall difference based on the overall difference; obtaining the data ranking of the pressure data in the pressure signal corresponding to the data point in the comparison data points, and taking the product of the data ranking, contribution degree, and concentration data of the pressure data at the same moment as the noise degree, and taking the normalized result of the noise degree as the noise factor.

4. A data processing method for a vacuum chamber helium leak detection device according to claim 3, characterized in that, The steps of obtaining the data ranking of the pressure data in the pressure signal corresponding to the data point in the comparison data points include: for any moment, obtaining multiple pressure data in the pressure signal corresponding to the comparison data segment, arranging them in ascending order based on the numerical size, and taking the position of the pressure data corresponding to this moment as the data ranking of the pressure data at this moment.

5. A data processing method for a vacuum chamber helium leak detection device according to claim 3, characterized in that, The steps for constructing a comparison data segment include: determining the length of the comparison data segment , obtaining a data segment centered on the current data point and with a length of as the comparison data segment.

6. A data processing method for a vacuum chamber helium leak detection device according to claim 3, characterized in that The steps of determining the contribution degree of the data point to the overall difference based on the overall difference include: for any data point, taking the standard deviation of all data except itself in the corresponding comparison data segment as the reference difference, and taking the normalized result of the difference between the overall difference and the reference difference as the contribution degree of the data point to the overall difference.

7. A data processing method for a vacuum chamber helium leak detection device according to claim 1, characterized in that, The steps of constructing a neighborhood segment corresponding to the reference segment based on the position of the reference segment include: determining the sampling time range based on the reference segment, and obtaining a preset number of data segments as the neighborhood segment within the sampling time range.

8. A data processing method for a vacuum chamber helium leak detection device according to claim 1, characterized in that Steps for setting reference points for each data point and constructing a reference point set for each data point include: determining the number of reference point sets , and using the adjacent and continuous data points on both sides of the data point as reference points; using the set composed of multiple reference points as the reference point set.

9. A data processing method for a vacuum chamber helium leak detection device according to claim 1, characterized in that, The steps of constructing a reference segment based on a reference point include using a data segment with a preset length centered on the reference point as the reference segment corresponding to the reference point.

10. A method for processing data of a vacuum chamber helium leak detection device according to claim 7, characterized in that The steps of obtaining a preset number of data segments within the sampling time range include: obtaining a plurality of data segments of the same length as the reference segment, calculating the mean value of the corresponding helium concentration data in the data segments, obtaining the absolute difference between the mean value and the mean value of the helium concentration data in the reference segment, and selecting the data segments with the smallest absolute difference as the neighborhood segments.

Images