Shell air tightness testing system and method

Abstract

The present invention relates to the field of gas leakage detection technology, specifically a shell air tightness testing system and method, the shell air tightness testing system including a pressure monitoring module, a gradient feature module, an anomaly screening module, a leakage path module, and a sealing assessment module. The present invention sets equidistant sampling points within the structure and extracts pressure time series differences to construct multi-segment continuous pressure data, enhances the ability to capture details of pressure changes, and achieves high-precision identification of stable sections. Combined with the difference in the rate of change of adjacent sections, local abnormal points are accurately extracted, and the sensitivity and positioning accuracy of abnormal identification are improved. Combining periodic pressure ratio analysis with path trend judgment, a directional and continuous pressure change path chain can be constructed. The sealing state is evaluated based on the path pressure variation value and the gradient difference of the adjacent area, the abnormal path segment is accurately identified, and the local sealing integrity judgment is supported, thereby improving the efficiency of shell local leakage identification and the depth of result analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of gas leakage detection, and in particular to a shell air tightness testing system and method. Background Art

[0002] Gas leak detection technology includes methods and equipment for detecting gas leaks in sealed structures. Its core content is to identify and judge the behavior of gas leakage from closed structures through pressure differences, changes in gas concentration, acoustic wave responses, etc., and to classify, analyze, and locate the leaks based on the leakage parameters. This detection technology systematically covers detection solutions for various closed structures such as pressure vessels, piping systems, automotive parts, electronic devices, and sealed packaging, and is widely used in industrial manufacturing, safety monitoring, and quality control. Commonly used gas leak detection methods on the market include differential pressure method, helium mass spectrometry, acoustic emission method, pressure maintenance method, and bubble method. Each detection method selects the appropriate detection method and detection device according to the characteristics of the structure to be tested and the sealing requirements.

[0003] Among them, the shell air tightness test scheme refers to the device and its supporting detection method used to perform air tightness detection on shells with closed structures and strict sealing requirements. It mainly covers the detection process of measuring pressure changes to determine whether there is a leak after compressed air is filled into the shell in a sealed state. It usually includes a gas filling mechanism, a pressure detection component, and a supporting pressure stabilization control component and a test chamber. The pressure data changes before and after the test are compared through standardized test procedures to complete the judgment of the shell sealing performance.

[0004] Existing shell airtightness testing processes generally rely on static data comparison to determine and analyze the sealing status of the shell. This lacks dynamic modeling of the entire pressure evolution process, resulting in results that favor overall trends while ignoring local details. In multi-segment structures, single-point pressure acquisition struggles to capture spatially distributed anomalies, especially when the sealing state fluctuates slightly, leading to missed detections due to insufficient data coverage. Data processing often relies on overall pressure differential analysis, ignoring the inherent connections between the direction of pressure variation, temporal structure, and local characteristics, making it difficult to identify trends and path tracing. Anomaly identification criteria are primarily empirically defined, lacking quantitative analysis of local fluctuations, limiting the ability to locate anomalies. Regarding path determination, traditional methods are unable to construct pressure variation pathways based on data evolution, resulting in unclear identification of leak origins and diffusion directions, making it difficult to effectively capture areas of sealing failure. The result evaluation criteria are overly broad, lacking detailed delineation based on pressure gradients and contrast differences between adjacent areas. This hinders the accurate identification of weak sealing areas, resulting in limited detection range, ambiguous spatial judgment, and delayed fault identification. This reduces the adaptability of testing solutions for complex structures or high-precision scenarios. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose a shell air tightness testing system that can improve the efficiency of identifying local leakage of the shell and the depth of result analysis.

[0006] In order to achieve the above object, the present invention adopts the following technical solutions:

[0007] A shell air tightness testing system includes a pressure monitoring module, a gradient characteristic module, an abnormality screening module, a leakage path module, and a sealing assessment module, wherein:

[0008] The pressure monitoring module is set at equidistant sampling points in each functional section of the shell structure, and is used to collect the pressure sequence of equidistant points in each functional section of the shell structure, calculate the pressure difference between adjacent points and adjacent moments, and generate partition pressure difference data;

[0009] The gradient feature module is used to extract the frequency of sliding window extreme values ​​based on the partition pressure difference data, screen the low-value fluctuation amplitude segment by comparing the gradient stability threshold, and generate a stable pressure evolution segment identification result;

[0010] The anomaly screening module is used to compare the segment change rate difference with the anomaly threshold based on the stable pressure evolution segment identification result, screen out abnormal points, and record the corresponding segment number and time tag to obtain a local pressure anomaly identification list;

[0011] The leakage path module is used to calculate the periodic pressure ratio between each abnormal point and other sections based on the local pressure anomaly identification list, select the spatial position between the two points with the largest ratio difference, determine the pressure value change trend on the path segment, and establish a leakage trend path tracking chain;

[0012] The sealing assessment module is used to track the chain according to the leakage trend path, compare the path pressure difference with the gradient average on both sides, mark the path exceeding the sealing threshold value, and obtain the local sealing integrity judgment result of the shell.

[0013] As a further solution of the present invention, the partition pressure difference data includes a multi-segment pressure time series structure, sampling point number information, and a pressure change amplitude per unit time;

[0014] The stable pressure evolution segment identification result includes the segment stability label, time period continuity parameter and characteristic vector fluctuation amplitude;

[0015] The local pressure anomaly identification list includes the abnormal point number, time mark information and local change rate difference;

[0016] The leakage trend path tracking chain includes the direction of pressure change, the spatial position relationship of the path segments, and the periodic pressure accumulation ratio;

[0017] The shell local sealing integrity determination result includes the pressure gradient value, the mean difference of the pressure gradient of the adjacent areas, and the abnormal path segment identification record.

[0018] As a further solution of the present invention, the pressure monitoring module includes an equidistant sampling submodule, a pressure collection submodule and a pressure difference calculation submodule, wherein:

[0019] The equidistant sampling submodule is set at equidistant sampling points in each functional section of the shell structure, and is used to calibrate the sampling position number of each section at a fixed interval according to the total length of the shell structure, use a high-frequency sensor to collect the pressure change value per unit time at each sampling number, obtain a bound data set of the section position number and the pressure change value, and generate section pressure sampling data;

[0020] The pressure aggregation submodule is used to arrange the pressure change values ​​at multiple moments under the same segment number in chronological order based on the segment pressure sampling data, in combination with the segment number and the sampling time sequence value, to construct a mapping relationship data set between the segment number and the pressure value of the corresponding continuous time period, and to establish a partition pressure data sequence;

[0021] The pressure difference calculation submodule is used to calculate the absolute difference between the pressure values ​​of two consecutive moments in each sequence based on the partition pressure data sequence, according to the time series pressure value and the segment number value, and classify and summarize the difference results according to the segment number to generate partition pressure difference data.

[0022] As a further solution of the present invention, the gradient feature module includes a sliding window construction submodule, a frequency extraction submodule and a gradient screening submodule, wherein:

[0023] The sliding window construction submodule is used to obtain the partition pressure difference data, combine the time series value and the segment number, construct an adjacent interval sliding window according to the time index of the pressure difference sequence of each segment, set the window width and sliding step value, divide the window segments into time series groups under each segment, and generate a partition sliding window sequence;

[0024] The frequency extraction submodule is used to count the number of repetitions of the maximum value in each window based on the set of pressure difference values ​​in each window based on the partitioned sliding window sequence, establish a corresponding relationship sequence between the window time index and the frequency value, and generate a window feature vector sequence;

[0025] The gradient screening submodule is used to calculate the amplitude change and threshold judgment of the feature sequence of each segment in chronological order based on the window feature vector sequence, combined with the window index value, the maximum frequency sequence and the gradient stability threshold, using the formula:

[0026] ;

[0027] Calculate the characteristic fluctuation amplitude value of each segment , judge whether it is continuously lower than the set gradient stability threshold, obtain a set of continuous time segments, and establish the stable pressure evolution segment identification result, where, For the Section No. The maximum frequency within a window, For the Section No. The maximum frequency within a window, is the total number of segment windows.

[0028] As a further solution of the present invention, the abnormality screening module includes a change rate extraction submodule, a difference calculation submodule and an abnormality identification submodule, wherein:

[0029] The change rate extraction submodule is used to calculate the difference between the pressure values ​​and time values ​​at two consecutive moments in each segment based on the stable pressure evolution segment identification result and the pressure sequence and time tag data of each identified segment, and calculate the pressure change rate per unit time in chronological order to obtain the segment pressure change rate sequence;

[0030] The difference calculation submodule is used to read the change rate of each segment and the change rates of its two adjacent segments based on the segment pressure change rate sequence, calculate the absolute difference between the change rates of the current segment and the segments on both sides in time sequence and calculate the average, obtain the change difference sequence between adjacent segments, and establish the pressure change difference sequence;

[0031] The anomaly identification submodule is used to call the current difference data and the local anomaly change judgment threshold according to the pressure change difference sequence, and combine the segment number and time label corresponding to each point to adopt the formula:

[0032] ;

[0033] Calculate the The abnormal intensity value of the segment at the current time point , and screen, retain the data points with abnormal intensity values ​​greater than zero, extract the time label and segment number, and establish a local pressure anomaly identification list, where, For the The pressure change difference of the section, is the threshold for judging local abnormal changes.

[0034] As a further solution of the present invention, the leakage path module includes a cycle accumulation submodule, a ratio screening submodule and a path tracking submodule, wherein:

[0035] The cycle accumulation submodule is used to perform time aggregation on the pressure change values ​​of each abnormal section in consecutive cycles based on the local pressure anomaly identification list, in combination with the abnormal section number and the pressure change value per unit time, and to obtain the pressure accumulation value by periodic accumulation to obtain the total pressure of the abnormal section cycle;

[0036] The ratio screening submodule is used to read the corresponding segment number and the total periodic pressure of other segments in the same period based on the total periodic pressure of the abnormal segment, calculate the periodic pressure ratio pair by pair, extract the two segment numbers with the largest ratio difference, obtain the corresponding spatial position distance value, and obtain the spatial segment pair with the largest pressure ratio difference;

[0037] The path tracking submodule is used to determine the start and end segment numbers based on the maximum pressure ratio difference spatial segment pair, retrieve the pressure change sequence of all segments between the two points in a continuous cycle, and judge the direction of the pressure value change trend using the formula:

[0038] ;

[0039] Calculate the total amount of pressure gradient trend in the path segment , and combined with the sign change trend to judge the overall pressure trend, establish a leakage trend path tracking chain, where For the path The total periodic pressure of the segment, For the path The total periodic pressure of the segment, is the total number of path segments, is a sign function, indicating positive or negative.

[0040] As a further solution of the present invention, the sealing assessment module includes a pressure gradient submodule, a neighboring gradient submodule and a sealing determination submodule, wherein:

[0041] The pressure variation submodule is used to calculate the pressure variation per unit length of each path segment based on the leakage trend path tracking chain, combining the periodic pressure value and path length parameter within the path segment, determine the pressure variation level along the path segment, and generate a path segment pressure gradient sequence;

[0042] The adjacent area gradient submodule is used to calculate the unit length pressure gradients of adjacent segments on both sides of the path segment based on the path segment pressure gradient sequence and the periodic pressure values ​​and segment spacing of the adjacent primary segments on the left and right sides of the path segment, and group them by lateral position to obtain the average value of the adjacent area pressure gradient group;

[0043] The sealing judgment submodule is used to count the pressure gradient of each path segment and the gradient mean of its left and right adjacent areas based on the adjacent area pressure gradient mean group, compare the degree of gradient difference, and determine whether it exceeds the sealing performance threshold value, mark the path segments that meet the conditions, and establish the local sealing integrity judgment result of the shell.

[0044] Another object of the present invention is to provide a method for testing the air tightness of a housing, the method being used to implement the housing air tightness testing system, comprising the following steps:

[0045] S1: Obtain the pressure values ​​of equidistant sampling points in each section, calculate the pressure difference at consecutive moments, group them by number, and generate partition pressure difference data;

[0046] S2: constructing a sliding window based on the partition pressure difference data, extracting the maximum frequency, establishing a feature vector, calculating the frequency variation, screening the stable segment, and generating a stable pressure evolution segment identification result;

[0047] S3: Calculate the difference in change rate between the current and adjacent sections based on the stable pressure evolution section identification result, screen abnormal points, and obtain a local pressure anomaly identification list;

[0048] S4: Based on the local pressure anomaly identification list, the periodic pressure ratio is counted, the path trend is determined, the pressure value change trend on the path segment is determined, and a leakage trend path tracking chain is established;

[0049] S5: Calculate the path pressure variation value according to the leakage trend path tracing chain, compare it with the gradient average of the adjacent area, mark the path that exceeds the sealing boundary value, and form a local sealing integrity judgment result of the shell.

[0050] Compared with the prior art, the advantages and positive effects of the present invention are:

[0051] In the present invention, by setting equidistantly distributed sampling points, the pressure changes at each point per unit time are continuously recorded, and partitioned pressure difference data are established to enhance the precision of expression of dynamic pressure changes. The maximum frequency in the pressure difference is extracted based on the sliding time window, and a feature vector sequence is constructed to achieve continuous characterization of the pressure fluctuation state. The change amplitude screening of the feature sequence can eliminate the interference of random disturbances, accurately define the section with stable pressure trend, perform local comparison of pressure change rate in the stable section, extract abnormal points, and achieve differential identification of subtle disturbances. By comparing the periodic pressure ratios of the abnormal points, the direction of pressure change is judged, and a spatial path chain with continuous trend characteristics is established. The pressure variation value of the path segment and the pressure gradient difference of the adjacent area are further calculated to identify the abnormal change interval, refine the quantitative evaluation of the local sealing state of the structure, improve the recognition depth of local leakage risk in complex structures and the structural response accuracy, and enhance the structural resolution and fault tracing capabilities of the detection data. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 is a system flow chart of the present invention;

[0053] Figure 2 This is a flow chart of the pressure monitoring module of the present invention;

[0054] Figure 3 This is a flow chart of the gradient feature module of the present invention;

[0055] Figure 4 This is a flow chart of the abnormality screening module of the present invention;

[0056] Figure 5 This is a flow chart of the leakage path module of the present invention;

[0057] Figure 6 This is a flow chart of the sealing assessment module of the present invention. DETAILED DESCRIPTION

[0058] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0059] In the description of the present invention, it should be understood that the terms "length," "width," "up," "down," "front," "back," "left," "right," "vertical," "horizontal," "top," "bottom," "inside," "outside," and the like, indicating positions or relationships, are based on the positions or relationships shown in the accompanying drawings and are intended only to facilitate the description of the present invention and simplify the description. They do not indicate or imply that the devices or elements referred to must have a specific orientation, be constructed, or operate in a specific orientation. Therefore, they should not be construed as limiting the present invention. Furthermore, in the description of the present invention, "plurality" means two or more, unless otherwise expressly and specifically defined.

[0060] See also Figure 1 The shell air tightness testing system provided by the present invention includes a pressure monitoring module, a gradient characteristic module, an abnormality screening module, a leakage path module and a sealing assessment module, wherein:

[0061] The pressure monitoring module is set at equidistant sampling points in each functional section of the shell structure. It is used to use high-frequency sensors to record the pressure change value per unit time at each sampling point in real time, group all time-series pressure data into a continuous pressure data sequence according to the section number, and perform absolute difference calculation on the pressure values ​​at two consecutive moments to generate partition pressure difference data;

[0062] The gradient feature module is used to construct a sliding window for adjacent time periods based on the pressure difference set of each functional segment in the partitioned pressure difference data, using the time series as the benchmark. The maximum frequency of the pressure difference in each window is extracted to form a window feature vector sequence. The feature vector sequence of each segment is screened for change amplitude, and the time period where the pressure fluctuation amplitude is continuously lower than the gradient stability threshold is extracted to generate the stable pressure evolution segment identification result.

[0063] The anomaly screening module is used to obtain the current pressure change rate in each identified segment based on the stable pressure evolution segment identification results, compare it with the change rate of the two adjacent segments, calculate the absolute difference between them, compare the difference with the local anomaly change judgment threshold, extract points that are greater than the local anomaly change judgment threshold, and record the corresponding segment number and time tag to obtain a local pressure anomaly identification list;

[0064] The leakage path module is used to identify all abnormal section numbers in the list based on local pressure anomalies, accumulate the pressure change value per unit time in continuous cycles, calculate the periodic pressure ratio between each abnormal point and other sections, select the spatial position between the two points with the largest ratio difference, determine whether the pressure value on the path segment shows a unidirectional increasing or decreasing trend, and establish a leakage trend path tracking chain;

[0065] The sealing assessment module is used to track the pressure values ​​of the path segments in the chain according to the leakage trend path, calculate the pressure variation value per unit path length, and simultaneously obtain the average pressure gradient of the first-level adjacent areas on the left and right sides of the path, compare the degree of difference, and mark the path segments with a difference greater than the sealing performance boundary value as abnormal sealing paths, thereby forming a local sealing integrity judgment result for the shell.

[0066] In an embodiment of the present invention, the partitioned pressure difference data includes a multi-segment pressure time series structure, sampling point number information, and pressure change amplitude per unit time; the stable pressure evolution segment identification result includes the segment stability label, time period continuity parameter, and characteristic vector fluctuation amplitude; the local pressure anomaly identification list includes the abnormal point number, time identification information, and local change rate difference; the leakage trend path tracing chain includes the pressure change direction, the spatial position relationship of the path segment, and the periodic pressure accumulation ratio; the shell local sealing integrity judgment result includes the pressure gradient value, the mean difference of the adjacent area pressure gradient, and the abnormal path segment identification record.

[0067] See also Figure 2 The pressure monitoring module includes an equidistant sampling submodule, a pressure collection submodule, and a pressure difference calculation submodule, wherein:

[0068] The equidistant sampling submodule is set at equidistant sampling points in each functional section of the shell structure. It is used to calibrate the sampling position number of each section at a fixed interval according to the total length of the shell structure, use a high-frequency sensor to collect the pressure change value per unit time at each sampling number, obtain a bound data set of the section position number and the pressure change value, and generate the section pressure sampling data;

[0069] In this embodiment, when setting equidistant sampling points for each functional section in the shell structure, it is first necessary to obtain the structural dimensions of the shell and the functional section division information. For example, the total length is 2.0 meters, divided into sections A and B, and the length of each section is 1.0 meters. The sampling interval is set to 0.5 meters, and two sampling points are arranged in each section, numbered P1, P2, P3, and P4, respectively, and the corresponding spatial positions are 0.0 meters, 0.5 meters, 1.0 meters, and 1.5 meters. This setting is based on the shell length and the range of the functional sections. The number and position sequence are obtained by dividing the coordinates of the section boundary position by the sampling interval value. After the sampling points are arranged, the section attributes and number identifiers are assigned respectively, and the pressure sampling values ​​in the initial state are recorded at the same time. For example, a high-frequency pressure sensor is used to read the initial pressure value of each position at a sampling frequency of 1 second. During the reading process, it is necessary to ensure that the sampling environment is constant and there is no instantaneous impact change. The pressure value corresponding to each point is bound by time and position to obtain the following data:

[0070] Table 1 Initial pressure gauge of monitoring points

[0071] Sampling point number Functional section Sampling location (m) Initial pressure (Pa) P1 Section A 0.0 101325 P2 Section A 0.5 101330 P3 Segment B 1.0 101328 P4 Segment B 1.5 101332

[0072] As shown in Table 1, the pressure sampling values ​​at all sampling points are close to normal pressure, with a difference of no more than 10 Pa, reflecting the basic uniformity of the initial air pressure distribution inside the shell. The pressure value record of each sampling point is established based on its position number and time stamp. After the sampling data is completed, it is uniformly packaged into a data set, and a binding data structure containing fields such as number, segment, position, time, and pressure value is constructed in a structured form to finally generate the segment pressure sampling data.

[0073] The pressure aggregation submodule is used to arrange the pressure change values ​​at multiple moments under the same segment number in chronological order based on the segment pressure sampling data, combined with the segment number and sampling time series value, to construct a mapping relationship dataset between the segment number and the corresponding continuous time period pressure value, and establish a partition pressure data sequence;

[0074] Among them, the above-mentioned segment-based pressure sampling data requires calling the number of each sampling point, the functional segment attributes and the corresponding time series pressure value, and classifying and integrating the pressure data with the same segment number. For example, segment A in Table 1 is aggregated, all time series pressure values ​​of P1 and P2 are extracted, arranged in chronological order, and a mapping array of time and pressure values ​​is constructed. For the data integration operation within each segment, the time step of the pressure data must be consistent to form a standardized data sequence format. During the data call process, the pressure value sequence of each sampling point is read in sequence according to the segment number, and a time index is constructed internally. The corresponding sampling point data is expanded in ascending order. For example, there are sampling points P1 and P2 in segment A. If the pressure value recorded by point P1 from t0 to t4 is The corresponding values ​​for point P1 and point P2 are 101325, 101326, 101324, 101327, and 101325 Pa, and those for point P2 are 101330, 101331, 101329, 101330, and 101332 Pa, respectively. The two sets of data constitute the pressure time series of segment A. During the pressure value aggregation process, these multi-point data are integrated into a two-dimensional matrix, where each column is the pressure time series of a sampling point, and the rows represent the pressures of each point at the same moment. The matrix is ​​then bound to the segment number to form a data block. By performing the above operations on multiple segments, the pressure data of all segments are merged into several independent time series. Each sequence has a sampling point number index, a segment number label, and a time-ordered pressure value, and finally a partitioned pressure data sequence is obtained.

[0075] The pressure difference calculation submodule is used to calculate the absolute difference between two consecutive pressure values ​​in each sequence based on the partition pressure data sequence, according to the time series pressure value and the segment number value, and classify and summarize the difference results according to the segment number to generate partition pressure difference data;

[0076] Among them, based on the partition pressure data sequence, it is necessary to perform difference processing on the pressure sequence of each segment, call the pressure values ​​of the continuous sampling moments in each segment, and obtain the absolute difference of the pressure change at adjacent moments by point-by-point difference. In the difference calculation process, each time series is traversed according to the segment number, and the pressure values ​​between each two consecutive moments are difference processed. For example, the continuous pressure records of point P1 in segment A are 101325, 101326, 101324, 101327, and 101325Pa, and the difference sequences are 1, 2, 3, and 2Pa respectively. Then, the difference sequences of each sampling point are averaged and merged, and the pressure values ​​of the same segment are calculated on this basis. The difference sequence of all sampling points in the segment is matched and averaged by time point, and finally the average pressure difference of the segment at each time step is obtained. At the same time, the corresponding time index and segment number are established. Through this series of calculation and processing processes, the pressure change trend between different sampling points is integrated into a data set reflecting the degree of pressure fluctuation of the entire segment. For example, if the average difference of segment A from t1 to t4 is 2.0, 2.5, 2.0, and 2.5 Pa, then the average amplitude of its pressure change per unit time is stable within ±2.5 Pa. After the data is collected, structured data is formed to record fields such as segment number, time node, and average pressure change amplitude, and finally the partition pressure difference data is generated.

[0077] See also Figure 3 The gradient feature module includes a sliding window construction submodule, a frequency extraction submodule, and a gradient screening submodule, where:

[0078] The sliding window construction submodule is used to obtain the partition pressure difference data, combine the time series value with the segment number, construct the adjacent interval sliding window according to the time index of the pressure difference sequence of each segment, set the window width and sliding step value, and divide the window segments according to the time series grouping under each segment to generate the partition sliding window sequence;

[0079] In this embodiment, the partition pressure difference data is obtained, and the time series pressure difference values ​​of each partition are grouped according to the segment number. Each group contains the pressure difference data corresponding to multiple consecutive time points. For example, in segment A, if the sampling frequency is once per second within 10 seconds;

[0080] The data sequence can be expressed as: (Unit: kPa), to construct a sliding window for this sequence, you need to set the window width and step size. If the window width is 4 and the step size is 2, the sliding window division results are window 1 (1st to 4th second), window 2 (3rd to 6th second), window 3 (5th to 8th second) and window 4 (7th to 10th second). During the window division process, first call the index information of each sampling point in the time series and the pressure difference data for position matching, and then extract the continuous data values ​​within the set window length according to the set window length. For example, the data corresponding to window 1 is , window 2 is , and so on, each window is indexed by the start time, and a mapping data table between the window number and its pressure difference sequence is constructed, as shown in Table 2.

[0081] Table 2 Sliding window pressure difference data table

[0082]

[0083] As shown in Table 2, the construction of the sliding window depends on the original time series data of the partition pressure difference and the set window width and step size. After the data structure is divided, it is automatically numbered according to the time index. The pressure difference sequence is loaded into each sliding window and stored, and finally a partition sliding window sequence is generated.

[0084] The frequency extraction submodule is used to count the number of repetitions of the maximum value in each window based on the set of pressure difference values ​​in each window based on the partitioned sliding window sequence, establish a corresponding relationship sequence between the window time index and the frequency value, and generate a window feature vector sequence;

[0085] Among them, based on the partition sliding window sequence, it is necessary to analyze the pressure difference sequence in each sliding window, extract its maximum value, and determine whether the maximum value appears more than once in the current window. If it appears twice or more, it is recorded as a maximum value repetition event. For example, for the window 1 sequence , the maximum value is 0.6, occurs once, and the frequency is 1; for window 2 , the maximum value 0.6 still only appears once; and if the sequence in a window such as window 5 is , the maximum value 0.6 appears 2 times, and the frequency is 2. After executing the same logical statistical process on all windows, an ordered pair is established according to the window number and the frequency result, such as , forming a frequency sequence under each segment , this sequence is the window feature vector sequence, and its data structure is used to characterize the local characteristics of the fluctuation behavior of the segment under the pressure difference angle.

[0086] The gradient screening submodule is used to calculate the amplitude change and threshold judgment of the feature sequence of each segment in chronological order based on the window feature vector sequence, combined with the window index value, maximum frequency sequence and gradient stability threshold, using the formula:

[0087] ;

[0088] Calculate the characteristic fluctuation amplitude value of each segment , judge whether it is continuously lower than the set gradient stability threshold, obtain a set of continuous time segments, and establish the stable pressure evolution segment identification result, where, For the Section No. The maximum frequency within a window, For the Section No. The maximum frequency within a window, is the total number of segment windows;

[0089] In this embodiment, according to the window feature vector sequence, the frequency sequence in each segment is first called, set as , as obtained in the previous paragraph , set the frequency stability threshold The setting basis is based on the expected fluctuation range of the maximum frequency in the sliding window of each segment. The threshold is determined by the joint analysis of the standard deviation and mean of the frequency series in multiple different segments of the same structure. Its value setting depends on the overall mean level and standard deviation variation range of the maximum frequency series in the window. Specifically, when the overall mean of the maximum frequency series is concentrated between 1.0 and 2.0, and the standard deviation does not exceed 0.5, in order to avoid misjudging a small sudden increase as an unstable state, the stability judgment threshold of this type of segment is set to 1.2. When the structural stress response maintains a single peak, the frequency difference fluctuates within 1 in most cases, so 1.2 is set as the fluctuation limit. This value is slightly adjusted as the sequence mean increases. Usually, when the mean exceeds 2.5, the threshold is raised to above 1.8. It is necessary to calculate whether the fluctuation value of the segment frequency series is lower than the threshold. In order to achieve quantitative judgment, the sequence Substitute into the calculation, we get:

[0090] ;

[0091] At this time, the frequency stability threshold For comparison, , indicating that the fluctuation amplitude of the current segment is in a stable range, and within the time period where the fluctuation is continuously lower than the threshold, that is, windows 1 to 5 continuously meet the conditions. Therefore, it is judged that the pressure change in this time period as a whole is stable, and the stable pressure evolution segment identification result is obtained.

[0092] See also Figure 4 The anomaly screening module includes a change rate extraction submodule, a difference calculation submodule, and an anomaly identification submodule, wherein:

[0093] The change rate extraction submodule is used to calculate the difference between the pressure values ​​and time values ​​at two consecutive moments in each segment based on the stable pressure evolution segment identification results and the pressure sequence and time tag data of each identified segment, and calculate the pressure change rate per unit time in chronological order to obtain the segment pressure change rate sequence;

[0094] Among them, based on the identification results of the stable pressure evolution segment, the pressure time series data and corresponding time tags recorded in the identified segment are extracted. For the time series data in each segment, the pressure values ​​and their corresponding times at two consecutive time points are extracted point by point, and the pressure change rate per unit time is calculated. In the actual scenario, taking segment A1 as an example, if the pressure values ​​of a monitoring point at t1=10s and t2=11s are P1=120kPa and P2=121.2kPa respectively, then its change rate is (121.2−120) / 1= 1.2kPa / s. In this way, the pressure change rates of segments A2 and A3 are calculated to be 2.8kPa / s and 1.6kPa / s respectively. Next, the previous segment (A0) and the next segment (A2) of segment A1 are extracted respectively. If A1 has no previous segment, it is filled with adjacent known values. Similarly, for A2, the previous segment is A1 and the next segment is A3. If the next segment of A3 is missing, it is filled with A4. Through these operations, the current change rate of each marked segment and the change rate of its adjacent segments can be obtained, which can be summarized as follows:

[0095] Table 3 Local pressure change calculation table

[0096]

[0097] As shown in Table 3, each section lists the difference between the change rate and the adjacent section, which prepares for the subsequent judgment of anomalies and finally obtains the section pressure change rate sequence.

[0098] The difference calculation submodule is used to read the change rate of each segment and the change rates of the two adjacent segments before and after it according to the segment pressure change rate sequence, calculate the absolute difference between the change rates of the current segment and the segments on both sides in time sequence and calculate the average, obtain the change difference sequence between adjacent segments, and establish the pressure change difference sequence;

[0099] According to the records in each row of Table 3, the current pressure change rate of each marked segment and the change rate of its adjacent segments are extracted. The absolute difference between the three is calculated and averaged to obtain the pressure change difference between the segment and its adjacent segments. For example, the current change rate of segment A2 is 2.8 kPa / s, and the change rates of its adjacent segments are 1.2 kPa / s and 1.6 kPa / s. The pressure difference between the three is calculated as follows: , The average of the two is (1.6+1.2) / 2=1.4kPa / s, so A2 , similarly, we can calculate A1 as follows: , The average difference is (0.2+0.4) / 2=0.3kPa / s, and the difference of A3 is (|1.6−2.8|+|1.6−1.1|) / 2=(1.2+0.5) / 2=0.85kPa / s. This calculation process makes the difference between each segment and the adjacent segment quantifiable, thereby establishing a pressure change difference sequence.

[0100] The anomaly identification submodule is used to call the current difference data and the local anomaly change judgment threshold based on the pressure change difference sequence, and combine the segment number and time label corresponding to each point, using the formula:

[0101] ;

[0102] Calculate the The abnormal intensity value of the segment at the current time point , and screen, retain the data points with abnormal intensity values ​​greater than zero, extract the time label and segment number, and establish a local pressure anomaly identification list, where, For the The pressure change difference of the section, The threshold for judging local abnormal changes;

[0103] Among them, according to the above calculated pressure change difference sequence, each difference is called and local judgment threshold , and combine the segment number and time label of each point to make point-by-point anomaly judgment, and use the formula to calculate the anomaly intensity value. For the A1 segment, enter the parameters: , , calculated as:

[0104] ;

[0105] For the A2 section, enter: , :

[0106] ;

[0107] For the A3 section, enter: , :

[0108] ;

[0109] Compare each segment and the size of 0, if , then the point is considered an outlier, and its time tag and segment number are recorded. Ultimately, a list of local pressure anomaly identification is established. This result shows that the anomaly intensity in segment A2 is significantly higher than in the other two segments, making it more likely to be identified as an anomaly. This indicates that the pressure fluctuations within this segment exhibit inconsistent characteristics within its spatial neighborhood, making it valuable for identification.

[0110] See also Figure 5 The leakage path module includes a cycle accumulation submodule, a ratio screening submodule, and a path tracking submodule, where:

[0111] The cycle accumulation submodule is used to combine the abnormal section number and the pressure change value per unit time based on the local pressure anomaly identification list, and to perform time aggregation on the pressure change values ​​of each abnormal section in consecutive cycles. The pressure accumulation value is accumulated by cycle to obtain the total pressure of the abnormal section cycle.

[0112] Among them, based on the abnormal section numbers listed in the local pressure anomaly identification list, for each monitoring area corresponding to the number, the pressure change value per unit time is collected. For example, in a continuous cycle with a time interval of 1 second, the pressure value sequence of sections S1 to S5 at each second is collected in sequence and recorded as ,in is the segment number, The time index is used. When performing the cycle accumulation operation, the pressure change amplitude within each second must be calculated first, and then the change amplitudes for 10 consecutive seconds are summed to obtain the total pressure of the cycle. For example, the pressure change values ​​of the S1 segment from t = 1 to t = 10 seconds are: [180.2, 182.5, 183.0, 181.9, 182.1, 183.4, 184.0, 183.5, 183.8, 184.1]. The sum of the change values ​​is 1820.5 Pa, which is the pressure accumulation of S1 in the current cycle. Similarly, the cycle total pressure of S2 to S5 can be calculated to obtain their respective cycle total pressures, and then the cycle total pressure record table shown in Table 1 is generated. This process is suitable for the preliminary quantitative assessment of the gas leakage trend in a certain area in the field pressure monitoring system and for obtaining the cycle total pressure of the abnormal section.

[0113] Table 4 Cycle pressure total amount table

[0114]

[0115] As shown in Table 4 , each segment corresponds to a set of periodic pressure values, which are used for subsequent ratio comparison and trend identification.

[0116] The ratio screening submodule is used to read the corresponding segment number and the total periodic pressure of other segments in the same period according to the total periodic pressure of the abnormal segment, calculate the periodic pressure ratio pair by pair, extract the two segment numbers with the largest ratio difference, obtain the corresponding spatial position distance value, and obtain the spatial segment pair with the maximum pressure ratio difference;

[0117] According to the results of the total pressure of the cycle, the ratio of the total pressure of each abnormal section to the remaining non-abnormal sections in the same cycle is calculated. That is, the S3 section is set as the target abnormal section, and its cycle pressure is 1955.2Pa. The ratios with S1 (1820.5Pa), S2 (1750.0Pa), S4 (1890.7Pa), and S5 (1785.9Pa) are calculated as follows: 1955.2 / 1820.5≈1.074, 1955.2 / 1750.0≈1.117, ... .2 / 1890.7≈1.034, 1955.2 / 1785.9≈1.095, and the maximum ratio difference is 1.117 between S2 and other points. Then reversely calculate the ratio difference of S2 to other points, and the maximum difference is formed by S2 to S3. It is recorded as the maximum pressure ratio difference segment pair. Combined with the segment number and the corresponding spatial coordinate information, such as S2 at x=150m, S3 at x=290m, and the spatial distance is 140m, this section area is used as the leakage trend interval to be judged, and the maximum pressure ratio difference spatial segment pair is obtained.

[0118] The path tracking submodule is used to determine the start and end segment numbers based on the maximum pressure ratio difference spatial segment pair, retrieve the pressure change sequence of all segments between the two points in a continuous cycle, and judge the direction of the pressure value change trend using the formula:

[0119] ;

[0120] Calculate the total amount of pressure gradient trend in the path segment , and combined with the sign change trend to judge the overall pressure trend, establish a leakage trend path tracking chain, where For the path The total periodic pressure of the segment, For the path The total periodic pressure of the segment, is the total number of path segments, is a sign function, indicating positive or negative;

[0121] Based on the obtained maximum pressure ratio difference spatial segment pair, taking the S2 to S3 segment as an example, the total periodic pressure of each segment on the path segment is extracted. If there are multiple continuous segments on the path, their pressure values ​​are recorded in the order of numbering as , there are 5 continuous sections, the pressure values ​​are: 1750.0, 1790.5, 1825.3, 1885.1, 1955.2Pa, and the directional trend formula is used to calculate as follows:

[0122] ;

[0123] ;

[0124] ;

[0125] ;

[0126] ;

[0127] The results show that the total pressure on the path segment shows a continuous increasing trend, with a total trend intensity of 205.2 Pa. Based on this result, it can be judged that the pressure direction is a unidirectional increase from S2 to S3, and then a leakage trend path tracking chain is established.

[0128] See also Figure 6 The sealing assessment module includes a pressure gradient submodule, a neighboring gradient submodule, and a sealing determination submodule, wherein:

[0129] The pressure gradient submodule is used to track the chain according to the leakage trend path, combine the periodic pressure value and path length parameter within the path segment, calculate the pressure change per unit length on each path segment, determine the pressure change level along the path segment, and generate the path segment pressure gradient sequence;

[0130] Among them, according to the leakage trend path tracking chain, the spatial position information and periodic pressure value of each path segment in the chain are extracted, the path segment pressure data are matched with the corresponding path length data, and the path segment start and end numbers, segment center position, length range and periodic pressure measurement value are recorded. The path length is directly measured through the structural CAD data, for example, the length of the P1 segment is 3.5 meters, the length of the P2 segment is 4.0 meters, and the length of the P3 segment is 2.5 meters. The periodic pressure is extracted from the average pressure data of each cycle through a high-frequency sampling device, such as P1. The periodic pressure difference is 12.3Pa, P2 is 16.5Pa, and P3 is 9.8Pa. The pressure change value of each path segment is divided by the path length to obtain the pressure gradient per unit length. The pressure gradient per unit length of P1 is calculated in this way as 12.3 / 3.5=3.51Pa / m, P2 is 4.13Pa / m, and P3 is 3.92Pa / m. The obtained path segment pressure change results are recorded as the path segment pressure gradient sequence, which is used for subsequent comparison and judgment with adjacent areas to obtain the path segment pressure gradient sequence.

[0131] The neighboring gradient submodule is used to calculate the unit length pressure gradient of the adjacent segments on both sides of the path segment based on the path segment pressure gradient sequence, according to the periodic pressure values ​​and segment spacing of the adjacent first-level segments on the left and right sides of the path segment, and group them by lateral position to obtain the average value of the neighboring pressure gradient group;

[0132] Among them, based on the path segment pressure gradient sequence, the periodic pressure values ​​of the left and right first-level neighboring areas adjacent to each path segment and the corresponding segment spacing are called, the neighboring area number, spatial position and path segment are associated, the neighboring area pressure change value is extracted and uniformly processed with quantitative parameters such as path segment length, and calculations are performed in a unified dimension. For example, the pressure difference of the left neighboring area of ​​P1 is 10.5Pa, and the right neighboring area is 9.9Pa. The path length is 3.5m, then the unit length gradient is calculated to be 3.0Pa / m and 2.83Pa / m respectively, and the left neighboring area of ​​P2 is 13.2Pa , the right neighboring area is 12.7Pa, then the gradient per unit length is 3.3Pa / m and 3.18Pa / m, and the mean gradients on the left and right sides are calculated respectively. The mean of the P1 neighboring area is (3.0+2.83) / 2=2.915Pa / m, and that of P2 is (3.3+3.18) / 2=3.24Pa / m. The left and right neighboring areas of P3 are 8.1Pa and 7.6Pa respectively, so the mean is (3.24+3.04) / 2=3.14Pa / m. In this way, the mean pressure gradient group of the neighboring areas corresponding to each path segment is obtained to form the mean pressure gradient group of the neighboring areas.

[0133] The sealing determination submodule is used to calculate the pressure gradient of each path segment and the gradient mean of its left and right adjacent areas based on the adjacent area pressure gradient mean group, compare the degree of gradient difference, and determine whether it exceeds the sealing performance threshold. The path segments that meet the conditions are marked and the local sealing integrity determination result of the shell is established;

[0134] According to the mean pressure gradient group of the adjacent areas, the path segment pressure gradient sequence is called to compare with the mean gradients on both sides of the adjacent areas, and the difference between the pressure gradient of each path segment and the mean gradient of its adjacent areas is calculated one by one. For example, the pressure gradient of the P1 path segment is 3.51Pa / m, the corresponding adjacent area mean is 2.915Pa / m, the difference is 0.595Pa / m, the P2 difference is 4.13-3.24=0.89Pa / m, and the P3 difference is 3.92-3.14=0.78Pa / m. The sealing performance boundary value of 1.5Pa / m is set based on the maximum allowable leakage gradient change value of the gas channel inside the structural shell. This value is obtained by measuring the maximum pressure change range per unit length in the isobaric test of multiple groups of different structural shells under standard closed conditions for 10 minutes, combined with the shell unit thickness (generally between 35mm), the average length of the path segment (about 2.55m) and the type of gas (such as commonly used nitrogen The study comprehensively evaluated the changes in the sensitive range of pressure to the sealing performance of the structural wall surface by considering physical factors such as air or dry air. It was ultimately determined that when the pressure difference between the path segment and the adjacent area is greater than 1.5 Pa / m, the pressure difference between the path segment and the adjacent area can cause the pressure in the structural gap to increase or decrease unidirectionally, thereby triggering problems such as crack expansion or sealing ring detachment. This threshold value shows a weak increasing trend with the shortening of the path length and the increase in the degree of pressure fluctuation. It is suitable for the pressure assessment standard of medium-strength shells of typical precision equipment and has high rationality and versatility. Referring to this threshold, the gradient difference of each path segment is compared with the threshold value. As shown in Table 5, the difference values ​​of all path segments do not exceed the threshold value and are therefore not marked as abnormal path segments. When the difference value of a path segment exceeds the threshold value, it is determined to be a sealing abnormal path segment. The path segment number, interval range, and gradient information are integrated and written into the judgment record to form the local sealing integrity judgment result of the shell.

[0135] Table 5 Path segment and adjacent area pressure data

[0136]

[0137] As shown in Table 5, the pressure data of the path segments and adjacent areas are collected periodically and uniformly converted into pressure differences per unit length. If the difference between each path segment and the adjacent area is below the preset threshold, it is not yet marked as an abnormal sealed path segment. If the subsequent periodic fluctuation increases and the difference exceeds the threshold, the judgment result will be synchronously updated to an abnormal path.

[0138] In an embodiment of the present invention, the shell air tightness testing method includes the following steps:

[0139] S1: Obtain the pressure values ​​of equidistant sampling points in each section, calculate the pressure difference at consecutive moments, group them by number, and generate partition pressure difference data;

[0140] S2: Construct a sliding window based on the partition pressure difference data, extract the maximum frequency, establish a feature vector, calculate the frequency change amplitude, screen the stable segment, and generate the stable pressure evolution segment identification result;

[0141] S3: Based on the identification results of the stable pressure evolution section, the difference in change rate between the current section and the adjacent section is calculated, abnormal points are screened, and a local pressure anomaly identification list is obtained;

[0142] S4: Based on the local pressure anomaly identification list, the periodic pressure ratio is counted to determine the path trend, the pressure value change trend on the path segment, and a leakage trend path tracking chain is established;

[0143] S5: Calculate the path pressure variation value based on the leakage trend path tracking chain, compare it with the gradient average of the adjacent area, mark the path that exceeds the sealing boundary value, and form the local sealing integrity judgment result of the shell.

[0144] In the embodiment of the present invention, the detailed steps and parameters involved in the above-mentioned shell air tightness test method refer to the above-mentioned Figures 1 to 6 The description of the shell air tightness test system shown will not be repeated here, but it is not intended to limit the present invention.

[0145] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A shell air tightness testing system, characterized by: The system comprises: The pressure monitoring module sets equidistant sampling points in each functional section of the shell structure, collects the pressure sequence of equidistant points in each section of the shell structure, calculates the pressure difference between adjacent points and adjacent moments, and generates partitioned pressure difference data; The gradient feature module extracts the frequency of extreme values ​​of the sliding window based on the partition pressure difference data, screens the low-value segments of the fluctuation amplitude by comparing with the gradient stability threshold, and generates a stable pressure evolution segment identification result; The anomaly screening module compares the segment change rate difference with the anomaly threshold based on the stable pressure evolution segment identification result, filters out anomalies, and records the corresponding segment number and time tag to obtain a local pressure anomaly identification list; The leakage path module, based on the local pressure anomaly identification list, counts the periodic pressure ratios of each abnormal point and other sections, selects the spatial position between the two points with the largest ratio difference, determines the pressure value change trend on the path segment, and establishes a leakage trend path tracking chain; The sealing assessment module tracks the chain according to the leakage trend path, compares the path pressure difference with the gradient average on both sides, marks the path that exceeds the sealing threshold value, and obtains the local sealing integrity judgment result of the shell; The partitioned pressure difference data includes the multi-segment pressure time series structure, sampling point number information, and the pressure change amplitude per unit time. The stable pressure evolution segment identification result includes the segment stability label, time period continuity parameter, and characteristic vector fluctuation amplitude. The local pressure anomaly identification list includes the abnormal point number, time identification information, and local change rate difference. The leakage trend path tracing chain includes the pressure change direction, the spatial position relationship of the path segment, and the periodic pressure accumulation ratio. The shell local sealing integrity judgment result includes the pressure variation value, the mean difference of the pressure gradient of the adjacent areas, and the abnormal path segment identification record.

2. The housing air tightness testing system according to claim 1, characterized in that: The pressure monitoring module includes: The equidistant sampling submodule sets equidistant sampling points for each functional section in the shell structure, calibrates the sampling position number of each section at fixed intervals according to the total length of the shell structure, uses a high-frequency sensor to collect the pressure change value per unit time at each sampling number, obtains a bound data set of the section position number and the pressure change value, and generates the section pressure sampling data; The pressure aggregation submodule arranges the pressure change values ​​at multiple moments under the same segment number in chronological order based on the segment pressure sampling data, combined with the segment number and the sampling time sequence value, constructs a mapping relationship data set between the segment number and the corresponding continuous time period pressure value, and establishes a partition pressure data sequence; The pressure difference calculation submodule is based on the partition pressure data sequence, calculates the absolute difference between the pressure values ​​of two consecutive moments in each sequence according to the time series pressure value and the segment number value, and classifies and summarizes the difference results according to the segment number to generate partition pressure difference data.

3. The housing air tightness testing system according to claim 1, characterized in that: The gradient feature module includes: The sliding window construction submodule obtains the partition pressure difference data, combines the time series value with the segment number, constructs an adjacent interval sliding window based on the time index of the pressure difference sequence of each segment, sets the window width and sliding step value, and divides the window segments into time series groups under each segment to generate a partition sliding window sequence; The frequency extraction submodule is based on the partitioned sliding window sequence, and according to the pressure difference value set in each window, counts the number of repetitions of the maximum value in each window, establishes a corresponding relationship sequence between the window time index and the frequency value, and generates a window feature vector sequence; The gradient screening submodule calculates the amplitude change and threshold judgment of the feature sequence of each segment in chronological order based on the window feature vector sequence, combined with the window index value, the maximum frequency sequence and the gradient stability threshold, using the formula: ； Calculate the characteristic fluctuation amplitude value of each segment , judge whether it is continuously lower than the set gradient stability threshold, obtain a set of continuous time segments, and establish the stable pressure evolution segment identification result, where, For the Section No. The maximum frequency within a window, For the Section No. The maximum frequency within a window, is the total number of segment windows.

4. The housing air tightness testing system according to claim 1, characterized in that: The abnormality screening module includes: The change rate extraction submodule calculates the difference between the pressure values ​​and time values ​​at two consecutive moments in each segment based on the stable pressure evolution segment identification results and the pressure sequence and time tag data of each identified segment, and calculates the pressure change rate per unit time in chronological order to obtain the segment pressure change rate sequence; The difference calculation submodule reads the change rate of each segment and the change rates of its two adjacent segments based on the segment pressure change rate sequence, calculates the absolute difference between the change rates of the current segment and the segments on both sides in time sequence, and calculates the average, obtains the change difference sequence between adjacent segments, and establishes the pressure change difference sequence; The anomaly identification submodule calls the current difference data and the local anomaly change judgment threshold based on the pressure change difference sequence, and combines the segment number and time label corresponding to each point, using the formula: ； Calculate the The abnormal intensity value of the segment at the current time point , and screen, retain the data points with abnormal intensity values ​​greater than zero, extract the time label and segment number, and establish a local pressure anomaly identification list, where, For the The pressure change difference of the section, It is the threshold for judging local abnormal changes.

5. The housing air tightness testing system according to claim 1, characterized in that: The leakage path module includes: The cycle accumulation submodule, based on the local pressure anomaly identification list, combines the abnormal section number and the pressure change value per unit time, and performs time aggregation on the pressure change values ​​of each abnormal section in consecutive cycles, and obtains the pressure accumulation value by periodic accumulation to obtain the total pressure of the abnormal section period; The ratio screening submodule reads the corresponding segment number and the total periodic pressure of other segments in the same period based on the total periodic pressure of the abnormal segment, calculates the periodic pressure ratio pair by pair, extracts the two segment numbers with the largest ratio difference, obtains the corresponding spatial position distance value, and obtains the spatial segment pair with the maximum pressure ratio difference; The path tracking submodule determines the start and end segment numbers based on the maximum pressure ratio difference spatial segment pair, retrieves the pressure change sequence of all segments between the two points in a continuous cycle, and determines the direction of the pressure value change trend using the formula: ； Calculate the total amount of pressure gradient trend in the path segment , and combined with the sign change trend to judge the overall pressure trend, establish a leakage trend path tracking chain, where For the path The total periodic pressure of the segment, For the path The total periodic pressure of the segment, is the total number of path segments, is a sign function, indicating positive or negative.

6. The housing air tightness testing system according to claim 1, characterized in that: The seal assessment module includes: The pressure gradient submodule calculates the pressure variation per unit length of each path segment based on the leakage trend path tracking chain and combines the periodic pressure value and path length parameter within the path segment, determines the pressure variation level along the path segment, and generates a path segment pressure gradient sequence; The neighboring gradient submodule calculates the pressure gradient per unit length of the adjacent segments on both sides of the path segment based on the path segment pressure gradient sequence and the periodic pressure values ​​and segment spacing of the adjacent first-level segments on the left and right sides of the path segment, and groups them according to the lateral position to obtain the average value of the neighboring pressure gradient. The sealing judgment submodule calculates the pressure gradient of each path segment and the gradient mean of its left and right adjacent areas based on the adjacent area pressure gradient mean group, compares the degree of gradient difference, and determines whether it exceeds the sealing performance threshold value. The path segments that meet the conditions are marked to establish the local sealing integrity judgment result of the shell.

7. A shell air tightness testing method, characterized in that: The method is used to implement the housing air tightness testing system according to any one of claims 1 to 6, comprising the following steps: S1: Obtain the pressure values ​​of equidistant sampling points in each section, calculate the pressure difference at consecutive moments, group them by number, and generate partition pressure difference data; S2: constructing a sliding window based on the partition pressure difference data, extracting the maximum frequency, establishing a feature vector, calculating the frequency variation, screening the stable segment, and generating a stable pressure evolution segment identification result; S3: Calculate the difference in change rate between the current and adjacent sections based on the stable pressure evolution section identification result, screen abnormal points, and obtain a local pressure anomaly identification list; S4: Based on the local pressure anomaly identification list, the periodic pressure ratio is counted, the path trend is determined, the pressure value change trend on the path segment is determined, and a leakage trend path tracking chain is established; S5: Calculate the path pressure variation value according to the leakage trend path tracing chain, compare it with the gradient average of the adjacent area, mark the path that exceeds the sealing boundary value, and form a local sealing integrity judgment result of the shell.

Images