Helium leak automatic detection method based on helium mass spectrum ultra-high vacuum exhaust platform

Abstract

The application provides a helium mass spectrum-based automatic helium leakage detection method for an ultrahigh vacuum exhaust station, relates to the technical field of gas leakage detection, and comprises the following steps: controlling the ultrahigh vacuum exhaust station to start a vacuum pumping system, pumping a cavity to be detected to a preset vacuum range; starting a helium filling device to inject helium into the outer surface area of the cavity to be detected, and recording the helium injection time sequence; obtaining a first group of helium mass spectrum response signals and a second group of helium mass spectrum response signals by a helium mass spectrum detection unit, the second group of helium mass spectrum response signals being delay sampling response signals; comparing response time curves, identifying delay response characteristics; analyzing the delay response characteristics according to a time delay analysis model, and outputting a helium leakage detection result according to a significance judgment result. The application solves the technical problem that the prior art usually assumes that helium leakage occurs instantaneously, ignores the hysteresis effect in the leakage path of a complex structure, and thus causes missed detection or misjudgment, thereby affecting the accuracy of helium leakage detection.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of gas leakage detection, in particular to a helium leakage automatic detection method based on helium mass spectrometry for an ultrahigh vacuum exhaust platform. BACKGROUND

[0002] In some complex devices, especially when there are complex structures such as sandwich structures, hidden cavities or micro cracks, helium leakage does not immediately respond but is affected by hysteresis effect. However, the prior art usually assumes that the response of helium leakage is instantaneous, that is, the helium concentration is basically synchronized with the time sequence of helium filling. For most simple leakage cases, the conventional method can effectively detect the leakage. However, in the above-mentioned complex structure, the diffusion process of helium leakage is usually limited by the structure, and the change of helium concentration does not immediately respond but has a certain time delay. This situation is particularly evident in sandwich-type leakage, because helium needs to pass through multiple layers of structure, cavities or micro channels to finally diffuse out, which causes the response time of helium concentration rise to lag. The conventional method cannot accurately judge these lag-type leakages, and finally leads to missed detection or misjudgment. SUMMARY

[0003] The application provides a helium leakage automatic detection method based on helium mass spectrometry for an ultrahigh vacuum exhaust platform, aiming to solve the technical problem that the prior art usually assumes that helium leakage occurs instantaneously, ignores the hysteresis effect in the leakage path of complex structure, and thus affects the accuracy of helium leakage detection.

[0004] The helium leakage automatic detection method based on helium mass spectrometry for an ultrahigh vacuum exhaust platform disclosed in the application comprises the following steps: controlling an ultrahigh vacuum exhaust platform to start a vacuum pumping system, pumping a cavity to be detected to a preset vacuum range; starting a helium filling device to inject helium into the outer surface area of the cavity to be detected, and recording the helium filling time sequence; obtaining a first group of helium mass spectrum response signals and a second group of helium mass spectrum response signals of the outer surface area by a helium mass spectrum detection unit, wherein the second group of helium mass spectrum response signals are delayed sampling response signals of the helium mass spectrum detection unit at the sampling position of the first group of helium mass spectrum response signals; comparing the response time curves according to the helium filling time sequence, the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals, identifying the delay response characteristics; analyzing the delay response characteristics according to a time delay analysis model, judging whether the delay response characteristics have significance, and outputting a helium leakage detection result according to the significance judgment result.

[0005] One or more technical solutions provided in the application have at least the following beneficial effects:

[0006] By controlling the ultra-high vacuum exhaust station to open the vacuum pumping system, the cavity to be detected is pumped to a preset vacuum range, ensuring that the helium filling and detection process are carried out in a preset vacuum environment, thereby reducing the influence of environmental fluctuations on the detection data and improving the detection accuracy; starting the helium filling device and accurately controlling the injection of helium into the outer surface area of the cavity, and recording the helium filling time sequence, which can provide detailed time information of the helium injection, making the subsequent signal comparison and delay response analysis more reliable; obtaining the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals through the helium mass spectrum detection unit, and specially designing a delay sampling mechanism, which can capture the concentration changes of helium at different time points, providing strong data support for judging the leakage type, such as delayed leakage and direct leakage; comparing the response time curves according to the helium filling time sequence and the mass spectrum response signals, clearly identifying the delay response characteristics of helium leakage, especially in complex structures, by comparing the responses at different time points, the time difference between the signals can be accurately analyzed, thereby judging whether the leakage exists significant delay; using a time delay analysis model to analyze the delay response characteristics, judging whether they are significant, and outputting the helium leakage detection result according to the significance judgment result, accurately distinguishing between significant and non-significant leakage, which improves the precision and sensitivity of helium leakage detection, especially for detecting delayed leakage (such as interlayer leakage, pore adsorption, etc.), avoiding the insensitivity of traditional methods to such leakage.

[0007] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the following specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described. BRIEF DESCRIPTION OF DRAWINGS

[0008] Figure 1 The present application provides a helium leakage automatic detection method based on helium mass spectrum of an ultra-high vacuum exhaust station.

[0009] Figure 2 The present application provides a time delay analysis model construction process diagram in a helium leakage automatic detection method based on helium mass spectrum of an ultra-high vacuum exhaust station. DETAILED DESCRIPTION

[0010] The present application provides a helium leakage automatic detection method based on helium mass spectrum of an ultra-high vacuum exhaust station, which solves the technical problem that the prior art usually assumes that helium leakage occurs instantaneously, ignores the hysteresis effect in the leakage path of complex structures, leading to missed detection or misjudgment, thereby affecting the accuracy of helium leakage detection.

[0011] After introducing the basic principles of the present application, the various non-limiting embodiments of the present application will be specifically introduced in combination with the drawings of the specification.

[0012] As Figure 1 shown, the embodiment of the present application provides a helium mass spectrum based automatic detection method for helium leakage of ultra-high vacuum exhaust platform, which comprises the following steps:

[0013] Controlling the ultra-high vacuum exhaust platform to start the vacuum pumping system, and pumping the cavity to be detected to a preset vacuum range.

[0014] Controlling the ultra-high vacuum exhaust platform to start the vacuum pumping system, which usually includes mechanical pump, molecular pump and other equipment, can quickly extract the gas in the cavity, thereby reducing the pressure in the cavity. The gas in the cavity to be detected is extracted by the vacuum pumping system. The cavity pressure is monitored during the extraction process until the pressure is adjusted to the preset vacuum range. The preset vacuum range depends on the experimental requirements, and is usually in an ultra-high vacuum state. Through this process, it is ensured that the internal environment of the cavity is clean and free of other contaminant gases, and reaches a controllable standard, so that the subsequent helium injection detection can accurately reflect the helium leakage situation.

[0015] Starting the helium filling device to inject helium into the outer surface area of the cavity to be detected, and recording the helium filling time sequence.

[0016] Starting the helium filling device, which is a device specially designed to fill helium into the cavity to be detected. The atom of helium is small and has very high penetrability, easy to leak and be detected by mass spectrum. By injecting helium into the outer surface area of the cavity to be detected, which is usually the area most susceptible to leakage, the relevant time sequence data of helium filling is recorded during this process, i.e. when to start injection, injection rate, duration of filling and other parameters, forming a helium filling time sequence for subsequent analysis.

[0017] The first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals of the outer surface area are obtained by the helium mass spectrum detection unit, wherein the second group of helium mass spectrum response signals is the delayed sampling response signal of the helium mass spectrum detection unit at the sampling position of the first group of helium mass spectrum response signals.

[0018] The helium mass spectrometry detection unit is a high-precision device for detecting and analyzing the concentration and distribution of helium. Based on the principle of mass spectrometry, it analyzes the mass spectrum response generated by helium molecules after being accelerated and ionized in an electric field, thereby quantifying the concentration of helium and the distribution of helium in the outer surface area. The first group of helium mass spectrometry response signals is the response of the helium concentration in the outer surface area after helium injection at a certain sampling time. By using the mass spectrometry detection unit, the instantaneous change of helium concentration can be monitored, and the first group of helium mass spectrometry response signals can be obtained. The second group of helium mass spectrometry response signals is obtained by delayed sampling. Specifically, the helium mass spectrometry detection unit delays for a certain period of time at the sampling position of the first group of signals and then samples the helium concentration again to obtain the second group of helium mass spectrometry response signals. This is to capture the subsequent response of the helium concentration change. Under normal circumstances, the second group of signals can reveal the delayed response characteristics caused by helium leakage. In this way, more comprehensive signal data can be obtained to help detect small changes in helium concentration, especially the lag effect of helium leakage.

[0019] According to the helium filling time sequence, the first group of helium mass spectrometry response signals, and the second group of helium mass spectrometry response signals, the response time curve comparison is performed to identify the delayed response characteristics.

[0020] Furthermore, according to the helium filling time sequence, the first group of helium mass spectrometry response signals, and the second group of helium mass spectrometry response signals, the response time curve comparison is performed to identify the delayed response characteristics. The method comprises:

[0021] The helium filling time sequence includes the start time and end time of helium filling. The first group of helium mass spectrometry response signals is the signal of the change of helium concentration over time within the first sampling time. The second group of helium mass spectrometry response signals is the signal of the change of helium concentration over time within the second sampling time. The second sampling time is the delayed sampling time of the first sampling time. The fusion response time curve of the helium filling time sequence, the first group of helium mass spectrometry response signals, and the second group of helium mass spectrometry response signals is constructed, and the delayed response characteristics are identified according to the fusion response time curve.

[0022] The helium filling time sequence includes the start time and end time of helium filling. The start time marks the starting time of helium injection, which is usually the time point of starting to record data. The end time marks the end time of helium injection, indicating the completion of the helium filling process.

[0023] The first group of helium mass spectrum response signals is a signal of the change of the helium concentration with time in the first sampling time, that is, the helium mass spectrum response signals collected at certain time intervals after the start of the filling, mainly reflecting the change of the helium concentration with time, which shows the immediate response of the helium concentration after the helium filling, helping the detection system to capture the diffusion process of the helium in the cavity, especially the signal characteristics of the direct leakage.

[0024] The second sampling time is a delayed sampling time of the first sampling time, which is usually a delay time after the collection of the first group of signals, and can be several seconds or longer, depending on the experimental setup. This delay time helps to capture the delayed response caused by the complexity of the helium leakage path (such as interlayer, pore adsorption, etc.). The second group of helium mass spectrum response signals is collected at the delay time based on the first group of helium mass spectrum response signals. The delayed sampling can capture the helium response after a period of time after the first sampling.

[0025] The helium filling time sequence is combined with the first group and the second group of helium mass spectrum response signals to construct a fusion response time curve, which can reflect the time sequence relationship between the entire helium filling process and the helium mass spectrum response. Specifically, the helium filling time sequence is corresponded to the response signal to ensure that the change of the helium concentration at each time point can be correctly reflected in the curve. The first group and the second group of helium mass spectrum response signals are fused respectively, so that the curve can show the complete change process of the helium concentration with time, especially the change of the delayed response characteristics.

[0026] In the constructed fusion response time curve, the delayed response characteristics are identified by comparing the changes of the helium concentration in different time periods. The delayed response characteristics are manifested as the lag of the second group of helium mass spectrum response signals, which is usually caused by the complexity of the leakage path (such as interlayer, pore adsorption, etc.).

[0027] Further, the method for identifying the delayed response characteristics according to the fusion response time curve comprises:

[0028] Extracting a set of key parameters in the fusion response time curve, the set of key parameters including a response start time, a maximum response time, a lag time, an upward slope, and a response duration; calculating a delay factor, a delay proportion, and a slow upward factor according to the set of key parameters; and outputting the delay factor, the delay proportion, and the slow upward factor as the delayed response characteristics.

[0029] In the constructed fusion response time curve, a series of key parameters reflecting the helium diffusion process are extracted through the analysis of the timing of the signal and the change of helium concentration. Among them, the response starting time represents the time point when the helium concentration starts to change significantly, that is, the time when the helium concentration starts to change from the baseline after the helium filling starts. The response starting time marks the time when the helium leakage event starts to affect the mass spectrum signal, and is an important reference for analyzing the time of leakage occurrence. The maximum response time refers to the time point when the helium concentration reaches the peak, that is, the time when the highest concentration of helium leakage diffusion occurs. This parameter can reflect the diffusion rate of the leakage, especially for large-scale leakage, the maximum response time is earlier, and for delayed leakage, the maximum response time may be delayed. The lag time is the time difference between the maximum response time and the response starting time of the helium concentration. The longer the lag time, the more the diffusion of helium leakage is affected by complex structures or obstacles, which may be caused by structural interlayers, small pores or complex channels, reflecting the characteristics of delayed leakage. The rising slope reflects the rate of change of helium concentration from the start to the maximum, indicating the slope of the increase in helium concentration. A higher rising slope means that the change in helium concentration is rapid, and a lower slope indicates that the helium diffusion is slow or the leakage point is small and the leakage amount is small. The response duration is the length of time that the helium concentration remains within a certain range. This parameter helps to identify the persistence of the leakage, especially for the concentration change after the leakage occurs. It can provide information on the duration of helium leakage. For delayed leakage, the response duration may be longer because the change in helium concentration is slower.

[0030] The delay factor is a quantitative measure of the delay characteristics of the leakage response by combining the lag time and the response duration. It represents the degree of time lag of the helium leakage signal. The larger the delay factor, the more obvious the structural or material obstacles to the diffusion of the leakage, and the slower the rise of the helium concentration. The delay factor can be calculated by the ratio of the lag time to the response duration.

[0031] The delay ratio refers to the proportion of the lag time in the response duration, which reflects the degree of delay of the leakage. The larger the ratio, the longer the delay time of the rise of the helium concentration during the leakage, which may be caused by complex leakage paths such as interlayers, pore adsorption, etc. The delay ratio can be simply calculated by the ratio of the lag time to the response duration. The larger the ratio, the more delayed the leakage characteristics, and there may be complex leakage paths.

[0032] The slow rise factor reflects the speed of change of the helium concentration. This factor can quantify the slow rise of the helium concentration, especially when the helium leakage path is restricted. The slow rise factor can be reflected by the rising slope of the helium concentration. If the rising slope is low, it means that the rise of the helium concentration is slow, and the corresponding slow rise factor is large.

[0033] The obtained delay factor, delay ratio, and slow rise factor are output as delay response features, which can be directly used for classification and judgment: if the delay factor, delay ratio, and slow rise factor are large, it indicates that the leakage is caused by complex structures (such as sandwich, pore adsorption, etc.) causing delayed leakage, rather than conventional direct leakage. Based on these features, the type of helium leakage can be accurately judged, especially when facing non-direct leakage paths, which can effectively improve the accuracy of detection.

[0034] The delay response features are analyzed according to the time delay analysis model, and it is judged whether the delay response features have significance, and a helium leakage detection result is output according to the significance judgment result.

[0035] The core goal of the time delay analysis model is to judge whether helium gas leakage exists and evaluate the degree of leakage by comparing and analyzing the delay response features. The time delay analysis model is trained based on historical data, labeled leakage samples, and delay response features to provide accurate judgment basis for detection. Direct leakage samples (no delay leakage) and delayed leakage samples (leakage with significant delay) are used in the training process.

[0036] The key of significance judgment lies in identifying abnormal patterns from the delay response features. For example, in microchannels or sandwich structures, the concentration rise of helium gas will show a delayed response due to the particularity of the leakage position, which usually shows a certain time lag compared to normal conditions. Through model training, the time delay analysis model automatically judges whether the delay response features have significance. If the delay response is greater than the set threshold, it indicates that the helium gas leakage event has significance. The significance judgment in this step means whether the change of the delay response features is large enough to be considered as caused by the leakage phenomenon. If the time delay of the delay response features meets the standard of leakage and this delay does not conform to the normal gas diffusion rules, it can be judged as significant leakage. According to the analysis result of the model, the significance of helium leakage is judged. If the delay response features have significance, the output detection result is leakage; if the delay response features do not have significance, the output is no leakage.

[0037] Traditional helium leakage detection methods are difficult to detect small leaks in complex structures, especially in sandwich or microchannel structures. Due to the slow leakage of helium, conventional detection methods are not sensitive enough to respond to this hysteresis effect. After introducing the time delay analysis mechanism, the hysteresis effect can be effectively identified according to the leakage position and the rules of helium diffusion, thereby improving the sensitivity of detection. Moreover, by establishing a model, automatic analysis and judgment can be performed, reducing human intervention and improving the efficiency and reliability of detection.

[0038] Further, asFigure 2 As shown, according to the time delay analysis model to analyze the delay response characteristics, the method for constructing the time delay analysis model comprises:

[0039] Collecting a labeled sample data set, the labeled sample data set comprising direct leakage data samples and delayed leakage data samples, and each group of data samples comprising helium filling timing samples and two groups of helium mass spectrum response signals; extracting delay response characteristic samples corresponding to the labeled sample data set; performing random forest training according to the delay response characteristic samples, a direct leakage label and a delayed leakage label, and constructing a time delay analysis model, wherein the direct leakage label is 0 and the delayed leakage label is 1.

[0040] Collecting a labeled sample data set, the labeled sample data set being real or simulated data with labels prepared for training the model, the labeled sample data set comprising two main types of leakage samples, namely direct leakage data samples and delayed leakage data samples, wherein the direct leakage data samples correspond to a situation where the increase in helium concentration is immediately visible when helium leakage occurs, and there is no significant time delay, that is, the helium mass spectrum response signal is highly synchronized with the helium filling timing and almost instantaneously reacts after helium injection; the delayed leakage data samples correspond to a situation where there is a significant time delay in the process of increasing the helium concentration when helium leakage occurs, which usually occurs in microchannels, interlayers or complex structures, and the leakage diffusion of helium does not immediately occur but experiences a lag phase, and in this sample, the helium mass spectrum response signal shows a significant time difference with the helium filling timing, that is, a delay response.

[0041] Each group of data samples comprises helium filling timing samples and two groups of helium mass spectrum response signals, wherein the helium filling timing samples record the time process of helium injection, including the start time, end time and injection rate, etc., so as to ensure that the model can analyze and compare the response characteristics according to the timing of helium filling; the two groups of helium mass spectrum response signals correspond to the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals, which record the concentration change of helium in the outer surface region of the cavity and provide data support.

[0042] In the labeled sample data set, the response time curve comparison is performed according to the helium filling timing samples and the two groups of helium mass spectrum response signals to identify the delay response characteristic samples, and this process is the same as the delay response characteristic identification in the foregoing steps, which will not be described herein for the sake of brevity of the description.

[0043] Random forest is an ensemble learning method that uses multiple decision trees for prediction and outputs the final result through a voting mechanism. During the training process, each decision tree learns on a different subset of data to ensure the model has high accuracy and good generalization ability. In the construction of the time delay analysis model, random forest is used to determine whether there is a delay leakage (label 1) or direct leakage (label 0) based on the delay response feature sample, direct leakage label and delay leakage label.

[0044] wherein the direct leakage label is 0, corresponding to the direct leakage data sample, indicating no delay leakage; the delay leakage label is 1, corresponding to the delay leakage data sample, indicating that there is a time delay in the leakage, i.e. the rise in helium concentration has a significant time lag, which may be due to reasons such as sandwich structure, pore adsorption, micro-channel blockage, etc.

[0045] Specifically, the labeled direct leakage data sample and delay leakage data sample are used as training data, divided into training set and validation set, the training set is used for model training, and the validation set is used for evaluating the performance of the model. The random forest algorithm automatically selects the best feature combination to build multiple decision trees, each decision tree uses a different subset of features for training, and the delay response feature sample in the training data is used to build the splitting nodes of the tree. In the learning process of each decision tree, the time delay of helium concentration is analyzed to determine whether it has a significant delay response. Through the learning of multiple trees, the model can obtain a comprehensive judgment of the leakage situation. The ultimate goal of training is to enable the model to accurately distinguish between direct leakage and delay leakage based on the delay response feature.

[0046] Further, the training of the random forest model is not only to identify simple direct leakage, but also to identify non-direct leakage paths caused by sandwich structure, pore adsorption or internal cavity blockage, etc. Specifically, delay leakage may be due to the following factors: sandwich structure, such as the presence of multiple air layers or gas gaps, causing helium to experience lag diffusion when leaking from one region to another; pore adsorption, in some small pores, helium may be adsorbed first, and then released after a period of time, causing a gradual rise in helium concentration; internal cavity blockage, the complexity of some microchannels or internal cavities may slow down the diffusion speed of the gas, resulting in a significant lag response. By using the random forest training model, these non-direct leakage paths can be more effectively identified, thereby improving the accuracy of helium leakage detection, especially when dealing with complex structures.

[0047] Further, to determine whether the delay response feature is significant, the helium leakage detection result is output according to the significance determination result, the method comprising:

[0048] a significant delay probability threshold is set; a significant delay probability is obtained by predicting the delay response feature according to the time delay analysis model, wherein the significant delay probability is the probability that the predicted output label is 1; and a helium leakage detection result is obtained by comparing the significant delay probability threshold and the significant delay probability.

[0049] A significant delay probability threshold is set to distinguish between significant delay leakage and direct leakage, in other words, the significant delay probability threshold is the criterion for helium leakage detection. The threshold can be determined by experimental data or performance evaluation during model training, for example, the threshold is adjusted according to the performance of the model during training (such as accuracy, F1 score, etc.) so that the system can make more accurate judgments between delay leakage and direct leakage.

[0050] The time delay analysis model is used to predict the delay response feature and determine whether there is a significant delay response. The significant delay probability is a probability value output by the model, indicating the likelihood that the delay response feature belongs to delay leakage (i.e., there is a significant delay response for helium leakage). The range is 0 to 1, and the closer the value is to 1, the higher the probability of delay leakage. The significant delay probability is calculated by the model based on the features and labeled data, reflecting the prediction of the current signal.

[0051] The significant delay probability is compared with the pre-set significant delay probability threshold. If the significant delay probability is greater than the threshold, it is considered that the signal represents a delay leakage event, and the detection result of delay leakage is output. If the significant delay probability is less than or equal to the threshold, it is considered that the signal represents direct leakage, and the detection result of direct leakage is output.

[0052] Further, if the significant delay probability is greater than the significant delay probability threshold, the helium leakage detection result is delay leakage; if the significant delay probability is less than or equal to the significant delay probability threshold, the helium leakage detection result is direct leakage.

[0053] If the significant delay probability is greater than the set significant delay probability threshold, it means that the delay response feature of the leakage is significant enough, and the model judges that the leakage occurs in a lag or delay period. In other words, when the significant delay probability is greater than the threshold, it means that the leakage does not occur directly, but due to factors such as interlayer structure, microchannel, pore adsorption, etc., there is a significant time lag in the diffusion and concentration rise of helium after leakage. In this case, the detection result of delay leakage is output, i.e., the helium leakage is caused by complex non-direct leakage paths, which may be caused by interlayer, microchannel blockage, or pore adsorption, etc.

[0054] If the significant delay probability is less than or equal to the set significant delay probability threshold, it means that the delay response characteristics of the helium leakage signal predicted by the model are not significant enough, the concentration change of the leaked helium is almost synchronized with the inflation timing, or the time difference of the delay response is not enough to be considered as significant. In other words, when the significant delay probability is less than or equal to the set threshold, it is considered that the leakage is a direct leakage, that is, the concentration change of helium is almost instantaneous, in this case, the detection result of direct leakage is output. Direct leakage refers to the rapid rise of helium concentration when helium leakage occurs, with almost no time delay, and the change of helium concentration is synchronized with the inflation timing. This type of leakage is usually caused by obvious defects such as leakage holes or cracks, and the leakage point directly affects the diffusion of helium, resulting in an immediate response of the signal.

[0055] Further, the helium mass spectrometry detection unit acquires the first set of helium mass spectrometry response signals and the second set of helium mass spectrometry response signals of the outer surface region, and the method further comprises:

[0056] A plurality of synchronous sampling points are arranged on the outer surface region, and the helium mass spectrometry detection unit acquires a plurality of first sets of helium mass spectrometry response signals and a plurality of second sets of helium mass spectrometry response signals at the plurality of synchronous sampling points.

[0057] A plurality of synchronous sampling points are arranged on the outer surface region of the cavity to be detected. The synchronous sampling points can be uniformly distributed on the outer surface region, or high-density sampling can be performed according to parts with higher leakage risk, for example, more intensive sampling points can be arranged at the joints, welds or vulnerable parts of the equipment. The positions of these synchronous sampling points are determined according to the structure of the equipment, the possible areas of gas leakage and the needs of experimental design, with the goal of fully covering the outer surface region and ensuring that the concentration changes of helium at different positions can be captured in time. Synchronous sampling can ensure that signals are obtained from different sampling points at the same time, thereby avoiding measurement errors caused by time differences.

[0058] The helium mass spectrometry detection unit synchronously collects the helium concentration change signals at the plurality of synchronous sampling points, and generates two sets of helium mass spectrometry response signals at each synchronous sampling point: the first set of helium mass spectrometry response signals and the second set of helium mass spectrometry response signals.

[0059] Further, the helium filling device comprises a pulse output unit for outputting a plurality of pulse injection signals; and the helium filling device is controlled to sequentially inject helium into the outer surface region of the cavity to be detected according to the plurality of pulse injection signals.

[0060] The helium filling device includes a pulsed output unit, which injects helium into the outer surface area of ​​the chamber under test in a pulsed manner. This pulsed injection method differs from continuous, stable gas injection; it can inject helium into the chamber at a certain periodicity or adjustable frequency, allowing for more precise control of the helium input time and resulting in a stronger correlation between each injection pulse and the signal response of the helium mass spectrometry detection unit. Through pulsed injection, changes in helium concentration can be monitored within each pulse cycle. The characteristics of the pulse (such as frequency and intensity) can help identify different types of leaks, especially when the leak occurs against a background of periodic changes.

[0061] The helium filling device injects helium sequentially into the outer surface of the chamber under test according to a predetermined set of pulsed injection signals. These pulsed injection signals have a certain time interval and adjustable pulse intensity, ensuring sufficient interval between each set of pulses to clearly distinguish the signal response after each pulse injection. Reasonable intervals avoid signal overlap, ensuring that the response data of each pulse is independent and accurate. After each pulse injection, the helium mass spectrometry detection unit collects the helium concentration response signal and records the helium concentration change within that period. Through multiple sets of pulsed injections, the leakage characteristics of helium on the outer surface of the chamber can be analyzed more comprehensively, especially minute delayed leaks. These characteristics help in further analyzing the leakage path and cause.

[0062] Furthermore, the pulsed injection signal includes a programmable pulse width, pulse interval, and pulse intensity gradient.

[0063] The pulsed injection signal includes programmable pulse width, pulse interval, and pulse intensity gradient. The pulse width, the duration of each pulse, controls the amount of helium injected within a specific time. By setting the pulse width, the duration of each helium injection can be controlled to simulate leaks of different scales. Longer pulse widths are suitable for detecting larger-scale leaks, while shorter pulse widths are suitable for detecting minute leaks. The time interval between pulses determines the interval between different pulses. Too short an interval may cause signal superposition, affecting detection accuracy, while too long an interval may reduce detection efficiency. A reasonable pulse interval helps to more clearly distinguish the signal response after different pulse injections. The pulse intensity gradient refers to the intensity variation of each pulse. Stepped pulse injection intensity can be set to simulate different leak amounts or leak rates. By adjusting the pulse intensity gradient, leaks of different sizes can be detected, especially for fine-tuning the detection of minute leaks.

[0064] Through programmable pulse width, pulse interval and intensity gradient, the detection strategy can be flexibly adjusted to adapt to the detection needs of different types of leaks, improving the accuracy and efficiency of detection. For example, when detecting small leaks, short pulses and higher intensity gradients are used; while detecting larger leaks, longer pulse width and lower pulse interval are used.

[0065] Further, the method of constructing a time delay analysis model further comprises:

[0066] Collecting a temperature data set, including the inner wall temperature of the cavity to be detected, the temperature data of the outer surface area and the temperature of the filled helium; establishing a mapping relationship between the temperature data set and the delay response, obtaining a regression correction coefficient according to the mapping relationship, and inputting the regression correction coefficient to optimize the training of the time delay analysis model.

[0067] Temperature has a significant impact on the diffusion process of helium leakage. The diffusion rate and concentration change of helium are often closely related to temperature, especially under different temperature conditions, the molecular motion and diffusion behavior of helium may be different. Therefore, collecting temperature data set can provide more environmental parameters for leakage analysis. The inner wall temperature of the cavity to be detected directly affects the diffusion characteristics of helium, and the temperature inhomogeneity in the cavity to be detected may cause different leakage responses in different regions. Obtaining the inner wall temperature data helps to identify the diffusion effect caused by temperature inhomogeneity; the temperature of the outer surface area is an important reference for leakage detection, and temperature change directly affects the adsorption and diffusion process of helium, especially in complex structures or interlayers. By monitoring the temperature change of the outer surface, more clues about the leakage location and mode can be provided to the model; the temperature of the filled helium affects the diffusion behavior of helium, especially in high temperature environment, the diffusion speed of helium is faster. By recording the temperature of the filled helium, the model can identify whether the diffusion change is caused by temperature anomaly, avoiding misjudgment.

[0068] Analyzing the correlation between the temperature data set and the delay response, and establishing a mapping relationship. Since temperature affects the diffusion rate of gas molecules, it will affect the lag time and diffusion pattern of helium concentration rise. When the temperature is high, the molecular motion of helium accelerates and the diffusion speed is faster; while at low temperature, the diffusion of helium is relatively slow and the delay response characteristics are more obvious. By analyzing a large number of temperature and delay response, a mapping relationship between the two is established. Through regression analysis, such as linear regression, polynomial regression, etc., a regression correction coefficient is obtained from the mapping relationship between temperature data and delay response. The regression correction coefficient reflects how temperature changes affect the characteristics of delay response. For example, if the temperature is high, a coefficient needs to be added to the model to adjust the prediction results of the delay response to adapt to the changes in helium diffusion under different temperature conditions.

[0069] The regression correction coefficient is input into the time delay analysis model to optimize the accuracy of the model under temperature changes, especially when analyzing complex leakage paths such as sandwich structures, pore adsorption, or internal cavity blockage. For example, the response time of a sandwich leakage at different temperatures may vary. Through the mapping relationship between temperature and delay response, the model can adjust the prediction of the delay response according to the actual temperature conditions, improving the detection capability for these complex path leaks. In this way, the model can better adapt to complex environments and improve the recognition capability for leaks caused by non-direct leakage paths (such as internal cavities and sandwich structures).

[0070] In summary, the helium mass spectrum-based automatic helium leakage detection method for the ultra-high vacuum exhaust station provided by the embodiments of the present application has the following technical effects:

[0071] By controlling the ultra-high vacuum exhaust station to start the vacuum pumping system, the detected cavity is pumped to a preset vacuum range, ensuring that the helium filling and detection process are carried out in a preset vacuum environment, thereby reducing the influence of environmental fluctuations on the detection data and improving the detection accuracy. Starting the helium filling device and accurately controlling the helium injection into the outer surface area of the cavity, and recording the helium filling timing, can provide detailed timing information for helium injection, making the subsequent signal comparison and delay response analysis more reliable. The first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals are obtained by the helium mass spectrum detection unit, and a delay sampling mechanism is specially designed, which can capture the concentration changes of helium at different time points, providing strong data support for judging the leakage type, such as delayed leakage and direct leakage. According to the response time curve comparison between the helium filling timing and the mass spectrum response signals, the delay response characteristics of helium leakage are clearly identified, especially in complex structures. By comparing the responses at different time points, the time difference between the signals can be accurately analyzed to determine whether there is a significant delay in the leakage. The time delay analysis model is used to analyze the delay response characteristics to determine whether they are significant, and the helium leakage detection result is output according to the significance determination result, accurately distinguishing between significant and non-significant leaks. This process improves the precision and sensitivity of helium leakage detection, especially for delayed leaks (such as sandwich leaks and pore adsorption), avoiding the insensitivity of traditional methods to such leaks.

[0072] The above description of disclosed embodiments enables one of ordinary skill in the art to make or use the application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A helium leak automatic detection method based on helium mass spectrum for an ultrahigh vacuum exhaust station, characterized in that, The method comprises: controlling the ultra-high vacuum exhaust station to start the vacuum pumping system to pump the cavity to be detected to a preset vacuum range; starting the helium filling device to inject helium into the outer surface area of the cavity to be detected, and recording the helium filling time sequence; obtaining a first group of helium mass spectrum response signals and a second group of helium mass spectrum response signals of the outer surface area by the helium mass spectrum detection unit, wherein the second group of helium mass spectrum response signals are delayed sampling response signals of the helium mass spectrum detection unit at the sampling position of the first group of helium mass spectrum response signals; comparing the response time curves according to the helium filling time sequence, the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals, and identifying the delay response characteristics; analyzing the delay response characteristics according to the time delay analysis model, judging whether the delay response characteristics have significance, and outputting the helium leakage detection result according to the significance judgment result; comparing the response time curves according to the helium filling time sequence, the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals, and identifying the delay response characteristics, which comprises: wherein the helium filling time sequence comprises the start time and the end time of the helium filling, the first group of helium mass spectrum response signals are signals of the change of helium concentration with time within the first sampling time, and the second group of helium mass spectrum response signals are signals of the change of helium concentration with time within the second sampling time, and the second sampling time is the delayed sampling time of the first sampling time; constructing a fusion response time curve of the helium filling time sequence, the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals, and identifying the delay response characteristics according to the fusion response time curve; identifying the delay response characteristics according to the fusion response time curve, which comprises: extracting a set of key parameters in the fusion response time curve, and the set of key parameters comprises the response start time, the maximum response time, the lag time, the rising slope and the response duration; calculating the delay factor, the delay proportion and the slow rising factor according to the set of key parameters; outputting the delay factor, the delay proportion and the slow rising factor as the delay response characteristics.

2. The method of claim 1, wherein, analyzing the delay response characteristics according to the time delay analysis model, and the method for constructing the time delay analysis model comprises: collecting a labeled sample data set, wherein the labeled sample data set comprises direct leakage data samples and delayed leakage data samples, and each group of data samples comprises a helium filling time sequence sample and two groups of helium mass spectrum response signals; extracting delay response characteristic samples corresponding to the labeled sample data set; performing random forest training according to the delay response characteristic samples, direct leakage labels and delayed leakage labels to construct a time delay analysis model, wherein the direct leakage label is 0 and the delayed leakage label is 1.

3. The method of claim 2, wherein, judging whether the delay response characteristics have significance, and outputting the helium leakage detection result according to the significance judgment result, which comprises: setting a significant delay probability threshold; predicting the delay response characteristics according to the time delay analysis model to obtain a significant delay probability, wherein the significant delay probability is the probability that the prediction output label is 1. The helium leakage detection result is obtained by comparing the significant delay probability threshold and the significant delay probability.

4. The method of claim 3, wherein, If the significant delay probability is greater than the significant delay probability threshold, the helium leakage detection result is a delayed leakage. If the significant delay probability is less than or equal to the significant delay probability threshold, the helium leakage detection result is a direct leakage.

5. The method of claim 1, wherein, The helium mass spectrum detection unit obtains the first group of helium mass spectrum response signals and the second group of helium mass spectrum response signals of the outer surface region, and the method further comprises: A plurality of synchronous sampling points are arranged on the outer surface region; The helium mass spectrum detection unit obtains a plurality of first group of helium mass spectrum response signals and a plurality of second group of helium mass spectrum response signals at the plurality of synchronous sampling points.

6. The method of claim 1, wherein, The helium filling device comprises a pulse output unit for outputting a plurality of pulse injection signals. The helium filling device is controlled to sequentially inject helium into the outer surface region of the cavity according to the plurality of pulse injection signals.

7. The method of claim 6, wherein, The pulse injection signal comprises a programmable pulse width, pulse interval and pulse intensity gradient.

8. The method of claim 2, wherein, The method for constructing the time delay analysis model further comprises: Collecting a temperature data set, including the inner wall temperature of the cavity, the temperature data of the outer surface region and the temperature of the filled helium; A mapping relationship between the temperature data set and the delay response is established, a regression correction coefficient is obtained according to the mapping relationship, and the time delay analysis model is optimized and trained by inputting the regression correction coefficient.

Images