A cloud platform-based nondestructive testing data processing method

Abstract

This invention relates to the field of nondestructive testing (NDT) technology, and more particularly to a cloud-based NDT data processing method. The invention integrates a spatial positioning device and a multi-dimensional sensor module to construct a testing terminal that communicates with the cloud platform. This terminal performs a full-domain scan of the target component, acquiring raw ultrasonic data and spatial point cloud coordinates in real time. Furthermore, by analyzing the curvature difference characteristics of local curved surface regions, non-morphological interference regions are identified and periodically sampled for monitoring. For morphological interference regions, anomalous measurement points are screened based on the peak variation characteristics of the raw ultrasonic data, and it is determined whether these are local interference regions. In addition, by identifying abrupt change points and their regional spatial correlation, signal types are distinguished. Finally, based on the determination results, an early warning signal is issued, and relevant data is uploaded to the cloud platform for real-time monitoring. This improves the recognition rate and reliability of early, weak damage signals under complex background noise.

Description

Technical Field

[0001] This invention relates to the field of nondestructive testing technology, and in particular to a nondestructive testing data processing method based on a cloud platform. Background Technology

[0002] As industrial infrastructure rapidly develops towards larger scale, greater complexity, and higher intelligence, the health status monitoring of components is not only closely related to production safety and operational efficiency, but the high-dimensional, massive amounts of data generated by the testing are also closely linked to subsequent structural integrity assessments, remaining life predictions, and predictive maintenance decisions. Therefore, efficient and accurate non-destructive testing data processing has become an indispensable part of ensuring modern industrial safety and achieving intelligent manufacturing transformation and upgrading.

[0003] Chinese Patent Publication No. CN120214078A discloses a non-destructive testing method, system, and data processing method, relating to the field of non-destructive testing technology. The method includes acquiring initial testing data of the component under test; establishing at least one reference model based on the initial testing data and acquiring simulated data corresponding to each reference model; determining target data based on the initial testing data and simulated data; acquiring subsequent testing data and comparing it with the target data to form an evaluation result. This invention improves the penetration capability and resolution of the detection by employing a high- and low-frequency mixed excitation signal, enabling simultaneous detection of surface defects and buried defects, significantly enhancing the comprehensiveness and accuracy of the detection. By using an orthogonal phase-locked loop method to decouple the acquired complex response signal, the magnetic field signal and electric field signal are effectively separated, improving the signal-to-noise ratio and detection accuracy. By establishing reference models and acquiring simulated data, accurate discrimination of defect type, depth, and orientation is achieved.

[0004] However, the following problems still exist in the existing technology.

[0005] In nondestructive testing (NDT), the complex internal geometry of the component being tested, such as cavities, curved surfaces, and interfaces between heterogeneous materials (e.g., composite laminate interfaces and adhesive interfaces), can affect the incident test wave. These complex internal structures and material interfaces can lead to strong scattering, reflection, and mode conversion of the test wave, generating complex background interference signals. These background interference signals are highly superimposed on early damage signals, potentially masking weak anomalous signals generated by early damage, thus reducing the NDT's ability to identify early damage and the reliability of the test results. Summary of the Invention

[0006] To address this, the present invention provides a cloud-based nondestructive testing (NDT) data processing method to overcome the challenges posed by complex internal geometries of the tested components in existing technologies. These geometries include cavities, curved surfaces, and interfaces between heterogeneous materials, such as composite laminate interfaces and adhesive interfaces, which can affect the incident detection wave. These complex internal structures and material interfaces can lead to strong scattering, reflection, and mode conversion of the detection wave, generating complex background interference signals. These background interference signals are highly superimposed on early damage signals, potentially masking weak anomalous signals generated by early damage, thus reducing the NDT's recognition rate for early damage and the reliability of the detection results.

[0007] To achieve the above objectives, the present invention provides a cloud-based nondestructive testing data processing method, comprising:

[0008] Step S1: Integrate the spatial positioning device and the multi-dimensional sensor module to build a detection terminal that communicates with the cloud platform. Based on the detection terminal, perform a full-domain scan of the target detection component and collect the original ultrasonic data of several target measurement points and the corresponding spatial point cloud coordinates in real time.

[0009] Step S2: Determine the curvature difference characteristics of the local surface region corresponding to each target measurement point, and determine the non-morphological interference region corresponding to the target measurement point based on the curvature difference characteristics, and perform periodic sampling monitoring.

[0010] Step S3: Analyze the morphological interference region corresponding to the remaining target measurement points, including determining the peak variation characteristics of each target measurement point based on the original ultrasonic data, screening out anomalous measurement points, determining the waveform offset characteristics corresponding to the anomalous measurement points, and determining whether it is a local interference region.

[0011] Step S4, to identify and analyze the local interference area, including: identifying abrupt change points based on the original ultrasonic data; determining the regional spatial correlation degree of each abrupt change point based on the spatial point cloud coordinates of each abrupt change point; and identifying the signal type based on the regional spatial correlation degree.

[0012] Step S5: Based on the judgment result, issue an early warning signal and upload the data of the area where the target measuring point corresponding to the early warning signal is located to the cloud platform for real-time monitoring.

[0013] Furthermore, the target measuring points are arranged along the surface of the target detection component at a predetermined density, wherein each target measuring point corresponds to a local curved surface region.

[0014] Furthermore, determining the curvature difference characteristics of the local surface region corresponding to each target measurement point includes,

[0015] Determine the neighborhood point set of the target measurement point based on the spatial point cloud coordinates corresponding to the target measurement point;

[0016] Determine the average curvature value of each point within the neighborhood point set;

[0017] Determine the curvature difference ratio between the curvature value and the average curvature value of the target measurement point;

[0018] The curvature difference ratio is defined as the curvature difference feature;

[0019] In the neighborhood point set, the spatial distance between each point and the target measuring point is less than a preset spatial distance threshold.

[0020] Furthermore, the process of determining the non-morphological interference region corresponding to the target measurement point includes,

[0021] If the curvature difference feature is less than the preset curvature difference threshold, then the local surface region corresponding to the target measurement point is determined as a non-morphological interference region.

[0022] Furthermore, the process of determining the peak value variation characteristics of each target measuring point includes,

[0023] Extract a predetermined number of waveform segments with complete cycles from the original ultrasonic data;

[0024] Determine the peak amplitude corresponding to the waveform segment;

[0025] The ratio of the average peak amplitude of the target measurement point waveform segment to the average peak amplitude of the preset reference waveform segment is calculated as the peak variation characteristic.

[0026] The process of further screening for anomalous and specific measurement points includes,

[0027] If the peak variation characteristic corresponding to the target measurement point is greater than or equal to the preset peak variation threshold, then the target measurement point is determined as a variation-specific measurement point.

[0028] Furthermore, the process of determining the waveform offset characteristics corresponding to the anomalous measurement points and judging whether they are local interference regions includes,

[0029] Extract the waveform segments corresponding to the anomalous and specific measurement points;

[0030] The overall time offset between the waveform segment and the preset waveform segment is determined as the waveform offset feature;

[0031] If the waveform offset feature is greater than or equal to the waveform offset threshold, then the local surface region corresponding to the abnormal measurement point is determined to be a local interference region.

[0032] Furthermore, the process of identifying mutation points includes,

[0033] For each target measurement point within the local interference area, calculate the amplitude change rate of the signal waveform;

[0034] If the amplitude change rate is greater than or equal to the preset amplitude change threshold, the target measurement point is identified as a sudden change point.

[0035] Furthermore, the process of determining whether a signal is a candidate abnormal signal includes,

[0036] The longest side length of the enclosed region is used as the expansion width of the enclosed region;

[0037] If the spread is greater than or equal to a preset spread threshold, it is determined to be a candidate abnormal signal.

[0038] Furthermore, the process of determining the regional spatial correlation degree of each mutation point and identifying the signal type based on the regional spatial correlation degree includes,

[0039] Calculate the average spatial distance between each mutation point and its nearest mutation point, and use it as the spatial correlation degree of the first region;

[0040] Calculate the variance of the spatial distance between each mutation point and its nearest neighbor mutation point, and use it as the spatial correlation degree of the second region;

[0041] The regional spatial correlation degree includes the first regional spatial correlation degree and the second regional spatial correlation degree;

[0042] If the spatial correlation of the first region is less than a predetermined spatial distance threshold and the spatial correlation of the second region is less than a regional spatial variance threshold, then the candidate abnormal signal is identified as a real defect signal.

[0043] If the spatial correlation degree of the first region is greater than or equal to a predetermined spatial distance threshold, or the spatial correlation degree of the second region is greater than or equal to a regional spatial variance threshold, then the candidate abnormal signal is identified as a random interference signal.

[0044] Compared with existing technologies, this invention integrates a spatial positioning device and a multi-dimensional sensor module to construct a detection terminal that communicates with a cloud platform. This terminal performs a full-domain scan of the target component, acquiring raw ultrasonic data and spatial point cloud coordinates in real time. Furthermore, by analyzing the curvature difference characteristics of local curved surfaces, non-morphological interference regions are identified and periodically sampled for monitoring. For morphological interference regions, anomalous measurement points are screened based on the peak variation characteristics of the raw ultrasonic data, and it is determined whether these are local interference regions. In addition, by identifying abrupt change points and their regional spatial correlation, signal types are distinguished. Finally, based on the determination results, an early warning signal is issued, and relevant data is uploaded to the cloud platform for real-time monitoring. This improves the recognition rate and reliability of early, weak damage signals under complex background noise.

[0045] In particular, this invention determines the curvature difference characteristics of the local curved surface regions corresponding to each target measurement point, and determines the non-morphological interference region corresponding to the target measurement point based on the curvature difference characteristics, and performs periodic sampling monitoring. In non-destructive testing, large and complex industrial components, such as aircraft skin, pressure vessel heads, and turbine blades, generally have complex free-form surface geometries. Curvature is a significant feature characterizing abrupt changes in local geometry. In regions where curvature changes are gradual, when ultrasonic waves propagate in local regions where curvature changes are gradual, the wavefront expansion law, energy attenuation characteristics, and mode conversion behavior remain relatively stable, and it is not easy to generate complex morphological interference noise such as scattered waves and edge diffraction waves. The detection signals in these areas are characterized by low background noise levels, high signal-to-noise ratio, and good waveform consistency, significantly reducing the probability of geometric pseudo-anomalies that may be confused with early damage signals. Conversely, in areas with drastic curvature changes, such as weld edges, around screw holes, and at the junctions of curved surfaces, ultrasonic waves undergo strong scattering, mode conversion, and energy redistribution. These physical effects induced by the geometry itself generate complex background interference signals. Therefore, periodically sampling and monitoring the non-morphological interference areas corresponding to the target measurement points, and implementing higher-density, higher-frequency real-time monitoring of morphological interference areas with drastic curvature changes and significant interference, can improve the overall efficiency of the detection system and enhance the ability to identify early weak damage signals and the reliability of detection in complex interference environments.

[0046] In particular, this invention analyzes the morphological interference regions corresponding to the remaining target measurement points, including determining the peak variation characteristics of each target measurement point based on the original ultrasonic data, screening for anomalous measurement points, determining the waveform shift characteristics corresponding to the anomalous measurement points, and determining whether they are local interference regions. In nondestructive testing, due to the internal geometric features of complex structures and the inhomogeneity of material interfaces, ultrasonic waves are affected by various factors during propagation, leading to signal distortion and interference. For example, when ultrasonic waves propagate in a component, their interaction with the internal structure of the material follows specific physical laws. When sound waves encounter discrete defects such as cracks and pores, or complex material discontinuities such as composite material lamination interfaces and residual stress zones, the sound waves undergo strong scattering, reflection, and energy absorption at these discontinuous interfaces, manifesting as a variation in peak amplitude. Simultaneously, during ultrasonic wave propagation, due to local changes in the material's elastic modulus or abrupt changes in acoustic impedance caused by microscopic defects, the signal propagation time may change, potentially leading to an overall time shift in the waveform. This strategy, based on the combined analysis of energy domain and time domain characteristics, first performs initial screening based on peak variation characteristics and then performs precise verification by combining waveform offset characteristics. This can improve the ability to identify early weak damage signals and the reliability of detection in complex interference environments.

[0047] In particular, this invention analyzes and identifies the local interference regions by: identifying abrupt change points based on the original ultrasonic data; constructing an enclosing region based on the spatial point cloud coordinates of each abrupt change point; determining whether the enclosing region is a candidate anomalous signal based on its span; determining the regional spatial correlation of each abrupt change point within the enclosing region; and identifying the signal type based on the regional spatial correlation. In nondestructive testing, detection signals are often contaminated by various random interferences and harmless surface anomalies. Such signals, such as electronic noise, minor undulations in local materials, or anomalous waveform characteristics of isolated surface scratches, do not pose a substantial threat to structural integrity and are generally considered acceptable background noise in engineering. Because these types of signals have significant limitations in terms of spatial extension scale, they usually appear as isolated point distributions or tiny clusters. Therefore, by constructing an enclosing region, and based on the expansion of the enclosing region, we can eliminate those local interferences that, although abnormal in waveform, are small in spatial scale and lack engineering significance. This can shield a large number of false alarm sources from the source. At this point, further discrimination and analysis are performed on the candidate abnormal signals. Since real structural damage, such as fatigue cracks and stress corrosion cracks, is constrained by the mechanical properties of materials and stress distribution during its initiation and propagation, it exhibits obvious continuity and clustering characteristics in space. This physical essence is reflected in the detection signal, manifesting as a regular spatial distribution of abrupt change points: the distance between each abrupt change point and its nearest neighbor remains stable, the average spatial distance is small and the distribution is uniform, forming a spatially meaningful correlation pattern. Conversely, interference signals caused by random scattering or poor instantaneous coupling do not have a physical continuity in their spatial distribution, exhibiting discrete and random distribution characteristics of abrupt change points in space, with significant differences in the distance between each point and its neighbor. This spatial correlation-based discrimination method more accurately distinguishes between physically plausible real damage signals and spatially randomly distributed interference signals, thereby improving the ability to identify and detect early weak damage signals in complex interference environments. Attached Figure Description

[0048] Figure 1 This is a schematic diagram illustrating the steps of a cloud-based nondestructive testing data processing method according to an embodiment of the invention.

[0049] Figure 2 A logic decision diagram for screening anomalous and specific test points in an embodiment of the invention;

[0050] Figure 3 This is a logic diagram for determining whether a region is a local interference area according to an embodiment of the invention.

[0051] Figure 4 This is a logic diagram for identifying signal types in an embodiment of the invention. Detailed Implementation

[0052] To make the objectives and advantages of the present invention clearer, the present invention will be further described below with reference to embodiments; it should be understood that the specific embodiments described herein are merely for explaining the present invention and are not intended to limit the present invention.

[0053] Preferred embodiments of the present invention will now be described with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are merely illustrative of the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0054] Furthermore, it should be noted that, in the description of this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0055] Please see Figure 1 The diagram illustrates the steps of a cloud-based nondestructive testing data processing method according to an embodiment of the invention. The cloud-based nondestructive testing data processing method of this invention includes:

[0056] Step S1: Integrate the spatial positioning device and the multi-dimensional sensor module to build a detection terminal that communicates with the cloud platform. Based on the detection terminal, perform a full-domain scan of the target detection component and collect the original ultrasonic data of several target measurement points and the corresponding spatial point cloud coordinates in real time.

[0057] Step S2: Determine the curvature difference characteristics of the local surface region corresponding to each target measurement point, and determine the non-morphological interference region corresponding to the target measurement point based on the curvature difference characteristics, and perform periodic sampling monitoring.

[0058] Step S3: Analyze the morphological interference region corresponding to the remaining target measurement points, including determining the peak variation characteristics of each target measurement point based on the original ultrasonic data, screening out anomalous measurement points, determining the waveform offset characteristics corresponding to the anomalous measurement points, and determining whether it is a local interference region.

[0059] Step S4, performing identification and analysis on the local interference area, including: identifying abrupt change points based on the original ultrasonic data; constructing an enclosing region based on the spatial point cloud coordinates of each abrupt change point; determining whether the enclosing region is a candidate abnormal signal based on the span of the enclosing region; determining the regional spatial correlation of each abrupt change point within the enclosing region; and identifying the signal type based on the regional spatial correlation.

[0060] Step S5: Based on the judgment result, issue an early warning signal and upload the data of the area where the target measuring point corresponding to the early warning signal is located to the cloud platform for real-time monitoring.

[0061] In implementation, the detection terminal includes a spatial positioning device and a multi-dimensional sensor module. The type of spatial positioning device is not limited and can be a spatial positioning system based on visual markers, a lidar SLAM positioning system, or a high-precision GPS / IMU integrated navigation system, etc. It only needs to be able to obtain the three-dimensional spatial coordinates and attitude angles of the detection terminal in the component coordinate system in real time, which will not be elaborated further.

[0062] In implementation, there are no restrictions on the specific selection and configuration of the multi-dimensional sensor module. It can be a combination of ultrasonic probe array and eddy current sensor, an integrated module of phased array ultrasonic and pulsed eddy current, or a fusion system of laser ultrasonic and microwave sensor, etc. It only needs to meet the requirement of being able to collect the original ultrasonic data of the target measurement point and the corresponding spatial point cloud coordinates in real time and communicate with the cloud platform. This will not be elaborated further.

[0063] Specifically, the target measuring points are arranged along the surface of the target detection component at a predetermined density, wherein each target measuring point corresponds to a local curved surface region.

[0064] In practice, the predetermined density is determined in advance. Preferably, the distance between adjacent target measurement points is set as a predetermined ratio of the feature side length of the target detection component. Typically, the predetermined ratio is selected within the range of [1 / 150, 1 / 50], and in practice, it is preferably 1 / 80.

[0065] Specifically, determining the curvature difference characteristics of the local surface region corresponding to each target measurement point includes,

[0066] Determine the neighborhood point set of the target measurement point based on the spatial point cloud coordinates corresponding to the target measurement point;

[0067] Determine the average curvature value of each point within the neighborhood point set;

[0068] Determine the curvature difference ratio between the curvature value and the average curvature value of the target measurement point;

[0069] The curvature difference ratio is defined as the curvature difference feature;

[0070] In the neighborhood point set, the spatial distance between each point and the target measuring point is less than a preset spatial distance threshold.

[0071] In practice, there are no restrictions on the method for determining the curvature value. Common curvature calculation methods in this field, such as differential geometry calculation methods based on local surface fitting and discrete curvature estimation algorithms based on point cloud normal vector changes, can be used. As long as they can accurately reflect the geometric bending characteristics of the local surface corresponding to the neighborhood point set, they will not be elaborated further.

[0072] In practice, the purpose of the preset spatial distance threshold is to characterize the geometric features of the local curved surface and the distribution of point cloud density. Those skilled in the art can determine the statistical mean of the distance between adjacent measuring points by statistically analyzing the spatial distribution characteristics of the point cloud data on the surface of the target detection component, so as to represent the spatial distance under normal conditions. In order to adapt to the errors in actual measurement and the non-uniformity of point cloud distribution, the preset spatial distance threshold is set as the product of the statistical mean and the distance error coefficient. Under normal circumstances, the distance error coefficient is selected within [0.85, 1.35], and is preferably 1.25 in practice.

[0073] Specifically, the process of determining the non-morphological interference region corresponding to the target measurement point includes,

[0074] If the curvature difference feature is less than the preset curvature difference threshold, then the local surface region corresponding to the target measurement point is determined as a non-morphological interference region.

[0075] In implementation, the purpose of the curvature difference threshold is to characterize the critical condition under which the geometric shape of the structural surface undergoes a significant change. Those skilled in the art can determine the statistical mean of the curvature difference ratio by statistically analyzing the curvature difference ratio data of each measuring point on the surface of the target detection component. This means the curvature difference ratio represents the situation where the curvature of the component surface changes continuously and gradually under normal conditions. To distinguish between regions of abrupt geometric changes and regions of gradual change, the curvature difference threshold is set as the product of the statistical mean and the difference accuracy coefficient. Typically, the difference accuracy coefficient is selected within the range of [1.15, 1.45], and is preferably 1.3 in implementation.

[0076] This invention determines the curvature difference characteristics of the local curved surface regions corresponding to each target measurement point, and determines the non-morphological interference region corresponding to the target measurement point based on the curvature difference characteristics, and performs periodic sampling monitoring. In non-destructive testing, large and complex industrial components, such as aircraft skin, pressure vessel heads, and turbine blades, generally have complex free-form surface geometries. Curvature is a significant feature characterizing abrupt changes in local geometry. In regions where curvature changes are gradual, when ultrasonic waves propagate in local regions where curvature changes are gradual, the wavefront expansion law, energy attenuation characteristics, and mode conversion behavior remain relatively stable, and it is not easy to generate complex morphological interference noise such as scattered waves and edge diffraction waves. The detection signals in these areas are characterized by low background noise levels, high signal-to-noise ratio, and good waveform consistency, significantly reducing the probability of geometric pseudo-anomalies that may be confused with early damage signals. Conversely, in areas with drastic curvature changes, such as weld edges, around screw holes, and at the junctions of curved surfaces, ultrasonic waves undergo strong scattering, mode conversion, and energy redistribution. These physical effects induced by the geometry itself generate complex background interference signals. Therefore, periodically sampling and monitoring the non-morphological interference areas corresponding to the target measurement points, and implementing higher-density, higher-frequency real-time monitoring of morphological interference areas with drastic curvature changes and significant interference, can improve the overall efficiency of the detection system and enhance the ability to identify early weak damage signals and the reliability of detection in complex interference environments.

[0077] Specifically, the process of determining the peak value variation characteristics of each target measuring point includes,

[0078] Extract a predetermined number of waveform segments with complete cycles from the original ultrasonic data;

[0079] Determine the peak amplitude corresponding to the waveform segment;

[0080] The ratio of the average peak amplitude of the target measurement point waveform segment to the average peak amplitude of the preset reference waveform segment is calculated as the peak variation characteristic.

[0081] In practice, the predetermined quantity is determined in advance. Typically, the predetermined quantity is selected within the range [3, 8], and is preferably 5 in practice.

[0082] Please see Figure 2 As shown, it is a logic decision diagram for screening anomalous and specific test points according to an embodiment of the invention. Specifically, the process of screening anomalous and specific test points includes,

[0083] If the peak variation characteristic corresponding to the target measurement point is greater than or equal to the preset peak variation threshold, then the target measurement point is determined as a variation-specific measurement point.

[0084] In practice, the purpose of the peak variation threshold is to characterize the critical condition of significant abnormal change in the amplitude of the ultrasonic signal. Those skilled in the art can determine the statistical mean of the peak variation characteristic by statistically analyzing the reference waveform data of the defect-free area of ​​the target detection component, so as to represent the stable state of the ultrasonic signal under normal conditions. In order to represent the situation of ultrasonic signal amplitude variation, the peak variation threshold is set to a predetermined multiple of the statistical mean. Typically, the predetermined multiple is selected within the range of [1.25, 1.45], and is preferably 1.4 in practice.

[0085] Please see Figure 3 As shown, this is a logic diagram for determining whether a region is a local interference region according to an embodiment of the invention. Specifically, the process of determining the waveform offset characteristics corresponding to the abnormal measurement points and determining whether a region is a local interference region includes...

[0086] Extract the waveform segments corresponding to the anomalous and specific measurement points;

[0087] The overall time offset between the waveform segment and the preset waveform segment is determined as the waveform offset feature;

[0088] If the waveform offset feature is greater than or equal to the waveform offset threshold, then the local surface region corresponding to the abnormal measurement point is determined to be a local interference region.

[0089] In practice, the preset waveform segment is predetermined. A typical waveform segment of a defect-free area is selected as the preset waveform segment and used as a reference. This will not be elaborated further.

[0090] In practice, there are no restrictions on the method for determining the overall time offset. Common signal alignment methods in this field, such as the cross-correlation function peak positioning method or the dynamic time warping distance minimization method, can be used. As long as the relative displacement of the two waveform segments in the time domain can be accurately quantified, this will not be elaborated further.

[0091] In practice, the purpose of the waveform offset threshold is to characterize the critical condition of significant changes in the ultrasonic signal in the time domain. The waveform offset threshold is predetermined. Those skilled in the art can determine the statistical mean of the time offset by statistically analyzing the overall time offset of the waveform segment in the defect-free area of ​​the target detection component, so as to represent the stability of the ultrasonic wave propagation time in the material under normal conditions. In order to represent the situation that can distinguish abnormal sound path changes, the waveform offset threshold is set as the product of the statistical mean and the offset error coefficient. Under normal circumstances, the offset error coefficient is selected in the range of [1.3, 1.6], and is preferably 1.5 in practice.

[0092] This invention analyzes the morphological interference regions corresponding to the remaining target measurement points, including determining the peak variation characteristics of each target measurement point based on the original ultrasonic data, screening for anomalous measurement points, determining the waveform shift characteristics corresponding to the anomalous measurement points, and determining whether they are local interference regions. In nondestructive testing, due to the internal geometric features of complex structures and the inhomogeneity of material interfaces, ultrasonic waves are affected by various factors during propagation, leading to signal distortion and interference. For example, when ultrasonic waves propagate in a component, their interaction with the internal structure of the material follows specific physical laws. When sound waves encounter discrete defects such as cracks and pores, or complex material discontinuities such as composite material lamination interfaces and residual stress zones, the sound waves undergo strong scattering, reflection, and energy absorption at these discontinuous interfaces, manifesting as peak amplitude variations. Simultaneously, during ultrasonic wave propagation, due to local changes in the material's elastic modulus or abrupt changes in acoustic impedance caused by microscopic defects, the signal propagation time may change, potentially leading to an overall time shift in the waveform. This strategy, based on the combined analysis of energy domain and time domain characteristics, first performs initial screening based on peak variation characteristics and then performs precise verification by combining waveform offset characteristics. This can improve the ability to identify early weak damage signals and the reliability of detection in complex interference environments.

[0093] Specifically, the process of identifying mutation points includes,

[0094] For each target measurement point within the local interference area, calculate the amplitude change rate of the signal waveform;

[0095] If the amplitude change rate is greater than or equal to the preset amplitude change threshold, the target measurement point is identified as a sudden change point.

[0096] In implementation, the purpose of the amplitude change threshold is to characterize the critical condition of abrupt changes in the amplitude of the ultrasonic signal. The amplitude change threshold is predetermined. Those skilled in the art can determine the mean value of the amplitude change rate by statistically analyzing the amplitude change rate of the signal waveform in the defect-free area of ​​the target detection component. This means the mean value represents the normal signal fluctuation of ultrasonic waves propagating in the material under normal conditions. To distinguish abrupt signal changes, the amplitude change threshold is set to a predetermined multiple of the mean value. Typically, the predetermined multiple is selected within the range of [1.15, 1.45], and is preferably 1.35 in implementation.

[0097] Specifically, the process of determining whether a signal is a candidate abnormal signal includes,

[0098] The longest side length of the enclosed region is used as the expansion width of the enclosed region;

[0099] If the spread is greater than or equal to a preset spread threshold, it is determined to be a candidate abnormal signal.

[0100] In implementation, there is no limitation on the construction method of the enclosing region based on the spatial point cloud coordinates of each mutation point. Preferably, it can be achieved by calculating the minimum outer rectangle or minimum outer convex hull of the mutation point set. It is only necessary to ensure that the enclosing region can completely cover the spatial distribution range of all mutation points and accurately reflect their geometric distribution characteristics. This will not be elaborated further.

[0101] In implementation, the purpose of the spread threshold is to characterize the minimum spatial scale necessary for a defect of engineering significance. The spread threshold is predetermined. Those skilled in the art can statistically analyze the size data of permissible defects found in historical inspections of the target component to determine its mean, representing the maximum harmless defect allowed in the component under normal conditions. To indicate the ability to filter out signals that pose a potential threat to structural safety, the spread threshold is set as the product of the mean and a safety factor. Typically, the safety factor is selected within the range [1.25, 1.5], and is preferably 1.4 in implementation.

[0102] Please see Figure 4 As shown, this is a logic diagram for identifying signal types according to an embodiment of the invention. Specifically, the process of determining the regional spatial correlation of each mutation point and identifying the signal type based on the regional spatial correlation includes:

[0103] Calculate the average spatial distance between each mutation point and its nearest mutation point, and use it as the spatial correlation degree of the first region;

[0104] Calculate the variance of the spatial distance between each mutation point and its nearest neighbor mutation point, and use it as the spatial correlation degree of the second region;

[0105] The regional spatial correlation degree includes the first regional spatial correlation degree and the second regional spatial correlation degree;

[0106] If the spatial correlation of the first region is less than a predetermined spatial distance threshold and the spatial correlation of the second region is less than a regional spatial variance threshold, then the candidate abnormal signal is identified as a real defect signal.

[0107] If the spatial correlation degree of the first region is greater than or equal to a predetermined spatial distance threshold, or the spatial correlation degree of the second region is greater than or equal to a regional spatial variance threshold, then the candidate abnormal signal is identified as a random interference signal.

[0108] In implementation, the purpose of the spatial distance threshold is to characterize the minimum degree of aggregation of the spatial distribution of mutation points in real defects. The spatial distance threshold is predetermined. Those skilled in the art can determine the mean of the nearest neighbor distance between mutation points by statistically analyzing the spatial distribution of mutation points in the known real defect area of ​​the target detection component. This means that the spatial aggregation characteristics of real defects under normal conditions are represented by the measurement system error and the inherent fluctuation of the material. The spatial distance threshold is set as the product of the mean and the spatial distance error coefficient. Typically, the spatial distance error coefficient is selected within the range of [0.65, 0.95], and is preferably 0.85 in implementation.

[0109] In implementation, the purpose of the regional spatial variance threshold is to characterize the uniformity of the spatial distribution of mutation points in real defects. The regional spatial variance threshold is predetermined. Those skilled in the art can determine the mean variance of the nearest neighbor distance between mutation points by statistically analyzing the spatial distribution of mutation points in the known real defect area of ​​the target detection component, so as to represent the typical uniformity characteristics of the spatial distribution of real defects under normal conditions. In order to represent the situation that can distinguish between the uniform distribution of real defects and the disordered distribution of random interference, the regional spatial variance threshold is set as the product of the mean variance and the spatial variance error coefficient. Under normal circumstances, the spatial variance error coefficient is selected within [0.75, 1.15], and is preferably 0.8 in implementation.

[0110] This invention analyzes and identifies the local interference region, including: identifying abrupt change points based on the original ultrasonic data; constructing an enclosing region based on the spatial point cloud coordinates of each abrupt change point; determining whether the enclosing region is a candidate anomalous signal based on its span; determining the regional spatial correlation of each abrupt change point within the enclosing region; and identifying the signal type based on the regional spatial correlation. In nondestructive testing, the detection signal is often contaminated by various random interferences and harmless surface anomalies. Such signals, such as electronic noise, minor undulations in local materials, or anomalous waveform characteristics of isolated surface scratches, do not pose a substantial threat to structural integrity and are generally considered acceptable background noise in engineering. Because these types of signals have significant limitations in terms of spatial extension scale, they usually appear as isolated point distributions or tiny clusters. Therefore, by constructing an enclosing region, and based on the expansion of the enclosing region, we can eliminate those local interferences that, although abnormal in waveform, are small in spatial scale and lack engineering significance. This can shield a large number of false alarm sources from the source. At this point, further discrimination and analysis are performed on the candidate abnormal signals. Since real structural damage, such as fatigue cracks and stress corrosion cracks, is constrained by the mechanical properties of materials and stress distribution during its initiation and propagation, it exhibits obvious continuity and clustering characteristics in space. This physical essence is reflected in the detection signal, manifesting as a regular spatial distribution of abrupt change points: the distance between each abrupt change point and its nearest neighbor remains stable, the average spatial distance is small and the distribution is uniform, forming a spatially meaningful correlation pattern. Conversely, interference signals caused by random scattering or poor instantaneous coupling do not have a physical continuity in their spatial distribution, exhibiting discrete and random distribution characteristics of abrupt change points in space, with significant differences in the distance between each point and its neighbor. This spatial correlation-based discrimination method more accurately distinguishes between physically plausible real damage signals and spatially randomly distributed interference signals, thereby improving the ability to identify and detect early weak damage signals in complex interference environments.

[0111] The technical solution of the present invention has been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will all fall within the scope of protection of the present invention.

Claims

1. A nondestructive testing data processing method based on a cloud platform, characterized in that, include, Step S1: Integrate the spatial positioning device and the multi-dimensional sensor module to build a detection terminal that communicates with the cloud platform. Based on the detection terminal, perform a full-domain scan of the target detection component and collect the original ultrasonic data of several target measurement points and the corresponding spatial point cloud coordinates in real time. Step S2: Determine the curvature difference characteristics of the local surface region corresponding to each target measurement point, and determine the non-morphological interference region corresponding to the target measurement point based on the curvature difference characteristics, and perform periodic sampling monitoring. Step S3: Analyze the morphological interference region corresponding to the remaining target measurement points, including determining the peak variation characteristics of each target measurement point based on the original ultrasonic data, screening out anomalous measurement points, determining the waveform offset characteristics corresponding to the anomalous measurement points, and determining whether it is a local interference region. Step S4, performing identification and analysis on the local interference area, including: identifying abrupt change points based on the original ultrasonic data; constructing an enclosing region based on the spatial point cloud coordinates of each abrupt change point; determining whether the enclosing region is a candidate abnormal signal based on the span of the enclosing region; determining the regional spatial correlation of each abrupt change point within the enclosing region; and identifying the signal type based on the regional spatial correlation. Step S5: Based on the judgment result, issue an early warning signal and upload the data of the area where the target measuring point corresponding to the early warning signal is located to the cloud platform for real-time monitoring.

2. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The target measuring points are arranged along the surface of the target detection component at a predetermined density, wherein each target measuring point corresponds to a local curved surface region.

3. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The determination of the curvature difference characteristics of the local surface region corresponding to each target measurement point includes... Determine the neighborhood point set of the target measurement point based on the spatial point cloud coordinates corresponding to the target measurement point; Determine the average curvature value of each point within the neighborhood point set; Determine the curvature difference ratio between the curvature value and the average curvature value of the target measurement point; The curvature difference ratio is defined as the curvature difference feature; In the neighborhood point set, the spatial distance between each point and the target measuring point is less than a preset spatial distance threshold.

4. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The process of determining the non-morphological interference region corresponding to the target measurement point includes: If the curvature difference feature is less than the preset curvature difference threshold, then the local surface region corresponding to the target measurement point is determined as a non-morphological interference region.

5. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The process of determining the peak value variation characteristics of each target measurement point includes: Extract a predetermined number of waveform segments with complete cycles from the original ultrasonic data; Determine the peak amplitude corresponding to the waveform segment; The ratio of the average peak amplitude of the target measurement point waveform segment to the average peak amplitude of the preset reference waveform segment is calculated as the peak variation characteristic.

6. The nondestructive testing data processing method based on a cloud platform according to claim 5, characterized in that, The process of screening for anomalous and specific test points includes: If the peak variation characteristic corresponding to the target measurement point is greater than or equal to the preset peak variation threshold, then the target measurement point is determined as a variation-specific measurement point.

7. The nondestructive testing data processing method based on a cloud platform according to claim 6, characterized in that, The process of determining the waveform shift characteristics corresponding to the anomalous measurement points and judging whether they are local interference regions includes: Extract the waveform segments corresponding to the anomalous and specific measurement points; The overall time offset between the waveform segment and the preset waveform segment is determined as the waveform offset feature; If the waveform offset feature is greater than or equal to the waveform offset threshold, then the local surface region corresponding to the abnormal measurement point is determined to be a local interference region.

8. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The process of identifying mutation points includes, For each target measurement point within the local interference area, calculate the amplitude change rate of the signal waveform; If the amplitude change rate is greater than or equal to the preset amplitude change threshold, the target measurement point is identified as a sudden change point.

9. The nondestructive testing data processing method based on a cloud platform according to claim 1, characterized in that, The process of determining whether a signal is a candidate abnormal signal includes: The longest side length of the enclosed region is used as the expansion width of the enclosed region; If the spread is greater than or equal to a preset spread threshold, it is determined to be a candidate abnormal signal.

10. The nondestructive testing data processing method based on a cloud platform according to claim 9, characterized in that, The process of determining the regional spatial correlation of each mutation point and identifying the signal type based on the regional spatial correlation includes: Calculate the average spatial distance between each mutation point and its nearest mutation point, and use it as the spatial correlation degree of the first region; Calculate the variance of the spatial distance between each mutation point and its nearest neighbor mutation point, and use it as the spatial correlation degree of the second region; The regional spatial correlation degree includes the first regional spatial correlation degree and the second regional spatial correlation degree; If the spatial correlation of the first region is less than a predetermined spatial distance threshold and the spatial correlation of the second region is less than a regional spatial variance threshold, then the candidate abnormal signal is identified as a real defect signal. If the spatial correlation degree of the first region is greater than or equal to a predetermined spatial distance threshold, or the spatial correlation degree of the second region is greater than or equal to a regional spatial variance threshold, then the candidate abnormal signal is identified as a random interference signal.

Images