Method, system and device for measuring thickness of heterogeneous composite component based on ultrasonic-magnetic method

Abstract

The invention discloses a heterogeneous composite component thickness measuring method, system and device based on an ultrasonic-magnetic method, and the method comprises the steps: carrying out the thickness collection of a sample medium based on ultrasonic sensing equipment and magnetic eddy current sensing equipment, and obtaining an original ultrasonic signal and an original electromagnetic signal; preprocessing an original ultrasonic signal and an original electromagnetic signal to respectively obtain a first ultrasonic signal and a first electromagnetic signal, and performing feature extraction on the first ultrasonic signal and the first electromagnetic signal to respectively obtain an ultrasonic signal feature and an electromagnetic signal feature; performing multi-modal feature fusion on the ultrasonic signal features and the electromagnetic signal features to obtain final features; and inputting the final features into an analysis model to obtain the predicted thickness of the sample medium, performing error analysis on the predicted thickness and the actual thickness of the sample medium, and adjusting parameters of the analysis model to obtain a final model. The device has the beneficial effect that the high-precision measurement of the sample medium is realized on the premise of ensuring that the sample medium is not contacted.

Description

Technical Field

[0001] The present invention relates to a structure detection technology, and in particular to a thickness measurement method, system and device for a heterogeneous composite component based on an ultrasonic-magnetic method. Background Art

[0002] At present, heterogeneous composite components are widely used in the field of aerospace, such as thin-walled rotating parts such as cabins, end frames, and engine disc parts. The inner layer material of the composite component has the characteristics of large ultrasonic attenuation coefficient and low material hardness. Characteristics Due to the large acoustic attenuation coefficient, low-frequency ultrasonic signals need to be used for thickness detection, but the detection accuracy corresponding to low-frequency ultrasonic signals is low, which does not meet the requirements of high-precision heterogeneous composite components; if high-frequency ultrasonic signals are used, due to the high attenuation characteristics of the measured medium, the pulse ultrasonic echo signal is weak and cannot detect thicker heterogeneous composite components; if the contact magnetic thickness measurement method is used, although the ultrasonic attenuation characteristics mentioned above are perfectly avoided, due to the soft characteristics of the inner layer material, the contact magnetic measurement will damage the component, which does not meet the requirements of non-destructive testing of heterogeneous composite components; and because the structure of heterogeneous composite components is complex and has the characteristics of thinness, curved surface, variable wall thickness, heterogeneous composite, etc., it poses a huge challenge to automated detection. Summary of the invention

[0003] The technical problem to be solved by the present invention is that the existing detection methods and devices are not suitable for measuring the thickness of heterogeneous composite components with complex structures. The purpose is to provide a method, system and device for measuring the thickness of heterogeneous composite components based on ultrasonic-magnetic method. The signals collected by ultrasonic sensing equipment and magnetic eddy current sensor equipment are subjected to feature extraction and fusion, and the thickness is predicted and optimized using an analytical model to achieve high-precision measurement of the sample medium.

[0004] The present invention is achieved through the following technical solutions:

[0005] A method for measuring the thickness of a heterogeneous composite component based on an ultrasonic-magnetic method comprises: S1, respectively collecting the thickness of a sample medium based on an ultrasonic sensing device and a magnetic eddy current sensing device to obtain an original ultrasonic signal and an original electromagnetic signal; S2, preprocessing the original ultrasonic signal and the original electromagnetic signal to obtain a first ultrasonic signal and a first electromagnetic signal respectively, performing feature extraction on the first ultrasonic signal and the first electromagnetic signal to obtain an ultrasonic signal feature and an electromagnetic signal feature respectively; S3, performing multimodal feature fusion on the ultrasonic signal feature and the electromagnetic signal feature to obtain a final feature; S4, inputting the final feature into an analysis model to obtain a predicted thickness of the sample medium, performing error analysis between the predicted thickness and the actual thickness of the sample medium, adjusting the parameters of the analysis model to obtain a final model.

[0006] The beneficial effect of the present invention is that the thickness of the sample medium is collected contactlessly by ultrasonic sensing equipment and magnetic eddy current sensing equipment respectively, the soft layer in the sample medium is protected, the original ultrasonic signal and the original electromagnetic signal are preprocessed, and then feature extraction is performed to obtain ultrasonic signal features and electromagnetic signal features, and then the obtained ultrasonic signal features and electromagnetic signal features are subjected to multimodal feature fusion to obtain final features, thereby improving the comprehensive information of the signal, and the final features are input into the analysis model to predict the thickness, and finally the predicted thickness and the actual thickness of the sample medium are subjected to error analysis, and the parameters of the analysis model are adjusted to optimize the analysis model, thereby ensuring that high-precision measurement of the sample medium (heterogeneous composite component) is achieved without contacting the sample medium.

[0007] In some embodiments, the preprocessing of the original ultrasonic signal and the original electromagnetic signal in step S2 comprises the following steps:

[0008] S21, based on the transmission conditioning unit, filtering and amplifying the original ultrasonic signal to obtain an amplified ultrasonic signal, and then separating the amplified ultrasonic signal into an amplified ultrasonic signal sequence according to time nodes;

[0009] S22, transforming the amplified ultrasonic signal sequence according to a preset modulation frequency with m as a factor signal, where m is an integer greater than 0, and setting a first calculation formula to obtain an amplified ultrasonic digital signal sequence, wherein the first calculation formula is:

[0010]

[0011] Wherein, X(k) represents the amplified ultrasonic digital signal sequence after the amplified ultrasonic signal sequence is converted, x(k) represents the kth signal value of the amplified ultrasonic signal sequence, a0 represents the first phase weight coefficient, a1 represents the second phase weight coefficient, a2 represents the third phase weight coefficient, step() represents the quantization step function, and t0 represents the standard step size of the amplified ultrasonic digital signal sequence;

[0012] S23, if the sampling frequency of the amplified ultrasonic digital signal sequence is different from the sampling frequency of the original electromagnetic signal, the signal with the lowest frequency is interpolated and then time-axis aligned; or if the sampling frequency of the amplified ultrasonic digital signal sequence is the same as the sampling frequency of the original electromagnetic signal, the two signals are directly time-axis aligned;

[0013] S24, using the cross-correlation function to obtain the time offset (correlation) of the two signals after alignment, and based on the time offset, performing translation on the standard time axis to achieve time axis alignment, and obtaining the first ultrasonic signal and the first electromagnetic signal. By filtering and amplifying the original ultrasonic signal, converting it into a digital signal, and then performing time alignment with the original electromagnetic signal, and using the cross-correlation function to accurately obtain the time offset, the synchronization of the ultrasonic and electromagnetic signals is achieved, the accuracy and reliability of signal processing is improved, and the accuracy of the measurement results is enhanced.

[0014] In some embodiments, feature extraction is performed on the first ultrasonic signal and the first electromagnetic signal, and the feature extraction comprises the following steps:

[0015] The digital signals of the first ultrasonic signal or the first electromagnetic signal are subjected to multi-layer filtering decomposition using preset approximate coefficients and preset detail coefficients, and the approximate coefficients and detail coefficients obtained after each layer of signal filtering decomposition are used as the input of the next layer of decomposition, until the iteration stops when the preset number of decomposition layers is reached, and the final approximate coefficients and final detail coefficients are obtained, and the final approximate coefficients and final detail coefficients are the ultrasonic signal characteristics or electromagnetic signal characteristics of the corresponding digital signal, wherein the final approximate coefficients are used as the low-frequency part of the last decomposition layer, and the final detail coefficients are used as the high-frequency part of the last decomposition layer. Through multi-level signal filtering decomposition, the low-frequency and high-frequency parts of the ultrasonic signal and the electromagnetic signal are extracted respectively, and the final signal characteristics are obtained by iterative processing. This method can effectively extract the multi-scale characteristics of the signal and improve the accuracy and reliability of signal analysis.

[0016] In some embodiments, the multimodal fusion in step S3 includes:

[0017] Building a feature gating network based on the ultrasonic signal features and the electromagnetic signal features, taking the final approximation coefficient and the final detail coefficient of the digital signal corresponding to each time node as a feature subspace set, and mapping the feature subspace set to the feature gating network;

[0018] The second calculation formula is set to obtain the final feature. The second calculation formula performs multimodal feature fusion based on the mutual mapping of feature subspace sets. The second calculation formula is as follows:

[0019]

[0020] in, represents the ith final feature of the ultrasonic signal feature and the electromagnetic signal feature, f use () represents the autoencoder function, R(a i ,b i) represents the feature subspace mapping function of the approximate coefficients of the ultrasonic signal feature and the electromagnetic signal feature, n represents that the ultrasonic signal feature and the electromagnetic signal feature have a total of n time nodes, a i represents the approximate coefficient of the ultrasonic signal characteristics at the i-th time node, b i Represents the approximate coefficient of the electromagnetic signal characteristics at the i-th time node, ω ult Represents the weight coefficient of the approximation coefficient, ω elc Represents the weight coefficient of the detail coefficient, g p () represents the feature gating network function, E(u i ,k i ) represents the characteristic subspace mapping function of the detail coefficients of the ultrasonic signal characteristics and electromagnetic signal characteristics, u i The detail coefficient of the ultrasonic signal feature at the i-th time node, k i Represents the detail coefficient of the electromagnetic signal feature at the i-th time node. By constructing a feature gating network, fusing the approximate coefficients and detail coefficients of the ultrasonic signal and the electromagnetic signal, and performing multimodal feature fusion, more accurate signal features can be extracted, thus improving the accuracy and robustness of signal analysis.

[0021] In some embodiments, the final feature is input into an analysis model to obtain a predicted thickness of the sample medium, an error analysis is performed with the actual thickness of the sample medium, and the parameters of the analysis model are adjusted to obtain a final model, including:

[0022] The final feature obtained by fusing the original ultrasonic signal and the original electromagnetic signal is used as the input layer of the analysis model, and the predicted thickness of the sample medium is obtained based on the output layer of the analysis model:

[0023]

[0024] Wherein, H represents the thickness of the sample medium, c max represents the maximum eigenvalue of the final feature, c min represents the minimum eigenvalue of the final feature, V() represents the analysis function of the analysis model, Δh represents the height difference between the magnetic eddy current probe and the ultrasonic probe in the vertical direction, and T0 represents the total time value of the final feature;

[0025] The predicted thickness is compared with the actual thickness of the sample medium. If the comparison error is greater than the preset standard threshold, the measurement result is judged to be inaccurate, and the model parameters of the input layer or output layer of the analysis model are adjusted to obtain the final model. The thickness of the sample medium is predicted by the analysis model using the fused ultrasonic and electromagnetic signal characteristics, and the model parameters are optimized in combination with the error feedback mechanism to improve the accuracy and reliability of thickness prediction and ensure the accuracy of the measurement results.

[0026] In some embodiments, an overall system architecture is designed, and the final model is used as an algorithm part of the overall system architecture.

[0027] The present invention also provides a heterogeneous composite component thickness measurement system based on an ultrasonic-magnetic method, which is used in a heterogeneous composite component thickness measurement method based on an ultrasonic-magnetic method, comprising:

[0028] A signal acquisition module is used to acquire the thickness of the sample medium and obtain the original ultrasonic signal and the original electromagnetic signal;

[0029] A feature extraction module is used to preprocess the original ultrasonic signal and the original electromagnetic signal to obtain a first ultrasonic signal and a first electromagnetic signal, and perform feature extraction respectively to obtain ultrasonic signal features and electromagnetic signal features;

[0030] A feature fusion module is used to perform multi-modal feature fusion of ultrasonic signal features and electromagnetic signal features to obtain the final features;

[0031] The prediction module is used to perform error analysis between the predicted thickness of the sample medium and the actual thickness of the sample medium, adjust the parameters of the analysis model, and obtain the final model. Multimodal signals are collected by ultrasonic and magnetic eddy current sensors, and after feature extraction and fusion, the analysis model is used to predict the thickness and optimize the model, which is finally integrated into the system architecture to improve the thickness measurement accuracy and system stability, with high accuracy and reliability.

[0032] The present invention also provides a heterogeneous composite component thickness measuring device based on an ultrasonic-magnetic method, which is used in a heterogeneous composite component thickness measuring method based on an ultrasonic-magnetic method, comprising:

[0033] A three-dimensional coordinate measuring machine, wherein the three-dimensional coordinate measuring machine comprises a three-dimensional Z-axis end effector and a three-dimensional coordinate measuring machine turntable, and the three-dimensional coordinate measuring machine turntable can be equipped with a sample medium;

[0034] A magnetic probe, wherein the magnetic probe is connected to the three-coordinate Z-axis end effector;

[0035] The ultrasonic pen probe is mounted on the three-coordinate Z-axis end effector and is arranged side by side with the magnetic probe. By connecting the magnetic probe and the ultrasonic pen probe to the lower end of the three-coordinate Z-axis end effector, the sample medium is mounted on the turntable of the three-coordinate measuring instrument, so as to facilitate the adjustment of the angle of the sample medium, and ensure that the probe maintains a stable position and a correct angle during the measurement process through the three-coordinate Z-axis end effector.

[0036] In some embodiments, a probe holder and a water wetting unit are also included, wherein the upper end of the probe holder is connected to the three-dimensional Z-axis end effector, and the ultrasonic pen-type probe and the water wetting unit are both clamped on the probe holder. The surface of the sample medium is wetted by the water wetting unit to form a water film with the surface of the sample medium, and then the pen-type ultrasonic probe is close to the surface of the sample medium to form a coupling of the ultrasonic probe to the surface of the sample medium, and the original ultrasonic signal of the sample medium is obtained.

[0037] In some embodiments, the ultrasonic pen probe and the magnetic probe are always kept perpendicular to the virtual section of the detection point of the sample medium during the scanning process, so as to improve the measurement accuracy.

[0038] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0039] 1. The pen-type ultrasonic probe and magnetic eddy current probe collect ultrasonic signals and electromagnetic signals respectively, and perform preprocessing and feature extraction. The comprehensive information of the signal is enhanced through multimodal feature fusion, which is input into the analysis model for thickness prediction, and error analysis is performed to optimize the model. Finally, through system architecture design, the optimized model is integrated into the master control module, which improves the accuracy and reliability of thickness measurement of heterogeneous composite components.

[0040] 2. The ultrasonic signal is obtained by coupling the pen-type ultrasonic probe with the water-wetting unit, and the magnetic eddy current probe excites the alternating magnetic field to obtain the electromagnetic signal. The two sensing technologies are combined for thickness measurement, which improves the measurement accuracy and reliability and meets the thickness detection needs of different materials. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0042] Figure 1 It is the principle diagram of the method of the present invention;

[0043] Figure 2 It is a structural diagram of the device of the present invention;

[0044] Figure 3 It is a detection schematic diagram of the present invention.

[0045] Marks and corresponding parts names in the attached drawings:

[0046] A magnetic probe 20 , an ultrasonic pen probe 21 , a probe holder 22 , a three-dimensional coordinate measuring machine 30 , a three-dimensional coordinate measuring machine turntable 31 , and a three-dimensional Z-axis end effector 32 . DETAILED DESCRIPTION

[0047] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0048] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "one embodiment," "an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or sub-combination. In addition, it will be appreciated by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0049] In the description of the present invention, the directions or positional relationships indicated by terms such as “front”, “rear”, “left”, “right”, “up”, “down”, “vertical”, “horizontal”, “high”, “low”, “inside” and “outside” are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, they should not be understood as limiting the scope of protection of the present invention.

[0050] The terms "first" and "second" used in the present invention are only used to distinguish the corresponding components for the sake of clarity of description, and are not intended to limit any order or emphasize importance, etc. In addition, the term "connected" used in this article can be directly connected or indirectly connected via other components without special explanation.

[0051] The names in the embodiments are explained as follows:

[0052] Sample medium refers to the material being tested and analyzed for thickness.

[0053] The original ultrasonic signal refers to the unprocessed sound wave signal received by the sensor when the ultrasonic sensing device measures the thickness of the sample medium.

[0054] The original electromagnetic signal refers to the unprocessed electromagnetic signal received by the device when the sample medium is measured through the magnetic eddy current sensing device.

[0055] Preprocessing refers to the process of preliminary processing and cleaning of raw data (raw ultrasonic signals and raw electromagnetic signals) to remove noise, standardize data format, improve signal quality, and ensure that subsequent analysis and feature extraction can be carried out more accurately.

[0056] The first ultrasonic signal refers to an original ultrasonic signal after preprocessing.

[0057] The first electromagnetic signal refers to the original electromagnetic signal after preprocessing.

[0058] Feature extraction refers to extracting representative, effective and useful signal information from raw signals or data.

[0059] Ultrasonic signal features are information extracted from the original signal collected by the ultrasonic sensor, representing the thickness of the sample medium or other related physical properties, such as amplitude, frequency, phase, etc.

[0060] Electromagnetic signal characteristics refer to information extracted from electromagnetic signals that can reflect the physical properties and thickness of the measured medium, such as amplitude, frequency, phase, etc.

[0061] Multimodal feature fusion refers to forming a more comprehensive and accurate feature representation by combining feature information from different signal sources.

[0062] The final model refers to the fused features of ultrasonic signals and electromagnetic signals obtained after multimodal feature fusion.

[0063] The analytical model is a mathematical model used to predict the thickness of the sample medium based on the features extracted from the ultrasonic and electromagnetic signals.

[0064] Predicting thickness means predicting the thickness of the sample medium based on the input ultrasonic signal characteristics and electromagnetic signal characteristics through an analytical model.

[0065] True thickness refers to the actual physical thickness of the sample medium and is an accurate value obtained through direct measurement.

[0066] The overall system architecture refers to the structure of the entire acquisition, processing, analysis and feedback process.

[0067] The water wetting unit provides a suitable coupling medium between the ultrasonic probe and the surface of the sample to be tested.

[0068] The three-dimensional Z-axis end effector is a mechanical component that can be precisely positioned and moved in three-dimensional space and is used in automated testing and measurement systems.

[0069] Coupling refers to the process by which the signal is transmitted from the sensor to the surface of the sample medium and interacts with the sample medium.

[0070] The magnetic eddy current probe is a non-destructive testing device used to inspect the surface and near-surface areas of materials.

[0071] Magnetic eddy current thickness tester is a device used to detect the thickness and surface defects of materials.

[0072] An alternating magnetic field refers to a magnetic field whose direction and intensity change periodically over time.

[0073] Eddy current change processing refers to the process of analyzing and processing the reverse electromagnetic signal generated by the alternating magnetic field during the magnetic eddy current detection process.

[0074] The main function of the transmit conditioning unit is to pre-process the ultrasonic signal, including filtering, amplification, modulation, adjustment and other operations, in order to improve signal quality, reduce noise and adapt to the input requirements of the system.

[0075] Filtering and amplification, where filtering is to remove irrelevant frequency components, such as noise or interference signals, and retain useful ultrasonic signal frequency components; amplification is to amplify the ultrasonic signal to a sufficient amplitude to ensure that it can be clearly identified and analyzed in subsequent processing.

[0076] The preset modulation frequency is used to combine the original ultrasonic signal sequence with the preset modulation frequency through a mixing operation in digital signal processing, thereby obtaining a frequency-adjusted signal.

[0077] The factor signal refers to a signal used for modulation, and its function is to transform the original first ultrasonic signal sequence into a new signal sequence.

[0078] The function of the ADC module is to convert the analog amplified ultrasonic signal into a digital signal for subsequent digital signal processing and analysis.

[0079] A mixing sequence is a sequence used to modulate a signal. In signal processing, it is used to mix the original signal with a signal of a specific frequency.

[0080] The amplified ultrasonic digital signal sequence is a digital signal sequence obtained by performing modulation, mixing and quantization processing on the amplified ultrasonic signal sequence.

[0081] The unit signal refers to the signal value corresponding to each sampling point in the amplified ultrasonic signal sequence.

[0082] A sampling point refers to the discrete numerical representation obtained after a continuous signal is digitized.

[0083] A digital value is a discrete value obtained after a series of signal processing, modulation, mixing, and quantization.

[0084] The phase weight coefficient is a parameter related to modulation mixing and phase adjustment in signal processing, and is used to indicate how to adjust the phase relationship of the signal.

[0085] The standard step size is a parameter used to represent the unit interval quantization increment at each time sampling point in the signal quantization digital signal processing process.

[0086] Sampling frequency refers to the number of times a continuous signal is sampled per unit time.

[0087] Interpolation refers to the process of inserting new data points between existing data points, inferring the value of unknown data through the rules of known data, and adding data to fill in missing data points.

[0088] Time axis alignment refers to adjusting the time series of two signals so that their time points are aligned.

[0089] Time offset refers to the difference in time between two signals, which means the time delay or advance of one signal relative to the other.

[0090] The first ultrasonic signal refers to an ultrasonic signal obtained after being processed through the various steps in the above process and has been aligned with the electromagnetic signal.

[0091] The first electromagnetic signal refers to an electromagnetic signal that has been aligned with the ultrasonic signal.

[0092] Signal waveform distribution refers to the performance and changes of the signal on the time axis.

[0093] The preset approximate coefficient is a parameter used for signal filtering and decomposition, which represents the characteristics of the low-frequency part of the signal during the decomposition process and is a preset parameter.

[0094] The preset detail coefficient is a parameter used for signal filtering and decomposition, which represents the characteristics of the high-frequency part of the signal during the decomposition process and is a preset parameter.

[0095] Signal filtering improves signal quality or extracts key information by removing or enhancing specific frequency components of the signal.

[0096] The low frequencies represent the slower changing components of the signal.

[0097] The high frequency part represents the faster changing components in the signal.

[0098] The first frequency band refers to the frequency range of the low-frequency part and the high-frequency part corresponding to the signal after the first layer of signal decomposition.

[0099] The final approximation coefficients represent the main features of the low-frequency signal in the final decomposition layer of the signal after multi-layer wavelet decomposition.

[0100] The final detail coefficient represents the main features of the high-frequency signal of the final decomposition layer of the signal after multi-layer wavelet decomposition.

[0101] The preset number of decomposition levels means that in the iteration loop, the signal has been decomposed to the preset number of levels and cannot be decomposed further.

[0102] A feature gating network refers to a neural network structure that is used to effectively fuse and select features from different signal sources in multimodal learning tasks. It selects and adjusts the contribution of different feature subspaces through a gating mechanism to better handle the fusion problem of multimodal data.

[0103] The feature subspace set refers to a set of feature representations extracted from different features of ultrasonic signals and electromagnetic signals, forming multiple high-dimensional spaces or low-dimensional spaces.

[0104] Mutual mapping refers to converting and fusing features from different modalities through a certain mapping function, so that features from different modalities can interact with each other in the same space, and finally achieve the purpose of multimodal feature fusion. The input layer is the first processing layer in the analysis model, which is responsible for receiving information from external data sources.

[0105] The output layer is the last layer in the analysis model and is responsible for generating the final output results.

[0106] The preset standard threshold is a measure of the error tolerance of the model's prediction accuracy.

[0107] The master control module is the core control unit of the system, responsible for coordinating the operation of each part.

[0108] The robot arm motion control unit is the core hardware and software system used to control the movement of the robot arm. It adjusts and controls the movement of each joint (or axis) of the robot arm by receiving instructions from the master control module or other upper control systems to ensure that the robot arm performs tasks smoothly according to the predetermined trajectory or action.

[0109] Motion instructions refer to specific instructions for controlling the movement of the robotic arm and turntable. They are used to specify the robotic arm's target position, target posture, motion trajectory, speed, acceleration and other parameters to achieve the required movement and scanning tasks.

[0110] The turntable motion control unit is a hardware control system responsible for controlling the rotation of the turntable. Its main task is to control the parameters of the turntable such as rotation angle, speed and acceleration.

[0111] Example 1

[0112] This embodiment provides a method for measuring the thickness of a heterogeneous composite component based on an ultrasonic-magnetic method. Figures 1 to 3 , including: S1, based on the ultrasonic sensor device and the magnetic eddy current sensor device, respectively collect the thickness of the sample medium to obtain the original ultrasonic signal and the original electromagnetic signal; S2, pre-process the original ultrasonic signal and the original electromagnetic signal to obtain the first ultrasonic signal and the first electromagnetic signal respectively, extract the features of the first ultrasonic signal and the first electromagnetic signal to obtain the ultrasonic signal feature and the electromagnetic signal feature respectively; S3, perform multi-modal feature fusion on the ultrasonic signal feature and the electromagnetic signal feature to obtain the final feature; S4, input the final feature into the analysis model to obtain the predicted thickness of the sample medium, perform error analysis on the predicted thickness and the actual thickness of the sample medium, adjust the parameters of the analysis model, and obtain the final model. S5: Design the overall system architecture, take the final model as the algorithm part of the overall architecture, and the information acquisition device and the industrial computer as the equipment part of the overall architecture, and integrate them into the general control module. The signals collected by the ultrasonic sensor device and the magnetic eddy current sensor device are subjected to feature extraction and fusion, and the thickness is predicted and the model is optimized using the analysis model, and finally integrated into the system architecture, so as to improve the thickness measurement accuracy and system stability, and have high accuracy and reliability.

[0113] Among them, the specific process after integration into the master control module includes: sending motion instructions to the robot arm motion control unit and the turntable motion control unit, implementing mechanical motion scanning, and synchronously sending instructions to the ultrasonic sensor device and the magnetic eddy current sensor device during the scanning process, and obtaining their respective signals; the information acquisition device collects the respective signals, calculates the wall thickness based on the algorithm part of the master control module, and displays it on the software interface. The overall system architecture realizes the collaborative work of the algorithm part and the equipment part. The industrial computer controls the synchronous movement of the robot arm and the sensor device, collects ultrasonic and electromagnetic signals, calculates and displays the wall thickness, optimizes the measurement process, and improves the equipment coordination efficiency and measurement accuracy.

[0114] See also Figure 1 and Figure 2Specifically, the ultrasonic sensing device is composed of a pen-type ultrasonic probe, a probe holder and a wetting unit. During the thickness measurement test, the surface of the sample medium is wetted by the wetting unit to form a water film on the surface of the sample medium. Then the pen-type ultrasonic probe is close to the surface of the sample medium to couple the ultrasonic probe to the surface of the sample medium and obtain the original ultrasonic signal of the sample medium. The magnetic eddy current sensing device is composed of a magnetic eddy current probe and a magnetic eddy current thickness detector. During the test, the magnetic eddy current probe is placed above the surface of the sample medium, and the exciting current passes through the coil in the magnetic eddy current probe to generate an alternating magnetic field. The magnetic eddy current thickness detector detects the reverse electromagnetic signal generated by the alternating magnetic field, and performs eddy current change processing on the reverse electromagnetic signal to obtain the original electromagnetic signal of the sample medium.

[0115] See also Figure 1 The preprocessing of the original ultrasonic signal and the original electromagnetic signal in step S2 comprises the following steps:

[0116] S21, based on the transmission conditioning unit, filtering and amplifying the original ultrasonic signal to obtain an amplified ultrasonic signal, and then separating the amplified ultrasonic signal into an amplified ultrasonic signal sequence according to time nodes;

[0117] S22, transforming the amplified ultrasonic signal sequence according to a preset modulation frequency with m as a factor signal, where m is an integer greater than 0, and setting a first calculation formula to obtain an amplified ultrasonic digital signal sequence, wherein the first calculation formula is:

[0118]

[0119] Wherein, X(k) represents the amplified ultrasonic digital signal sequence after the amplified ultrasonic signal sequence is converted, x(k) represents the kth signal value of the amplified ultrasonic signal sequence, a0 represents the first phase weight coefficient, a1 represents the second phase weight coefficient, a2 represents the third phase weight coefficient, step() represents the quantization step function, and t0 represents the standard step size of the amplified ultrasonic digital signal sequence;

[0120] S23, determining the sampling frequencies of the amplified ultrasonic digital signal sequence and the original electromagnetic signal, if the sampling frequencies are different, performing time axis alignment after interpolation processing on the signal with the lowest frequency, otherwise, directly aligning the two signals with the time axis;

[0121] S24, using the cross-correlation function to obtain the time offset (correlation) of the two signals after alignment, and based on the time offset, performing translation on the standard time axis to achieve time axis alignment, and obtaining the first ultrasonic signal and the first electromagnetic signal. By filtering and amplifying the original ultrasonic signal, converting it into a digital signal, and then performing time alignment with the original electromagnetic signal, and using the cross-correlation function to accurately obtain the time offset, the synchronization of the ultrasonic and electromagnetic signals is achieved, the accuracy and reliability of signal processing is improved, and the accuracy of the measurement results is enhanced.

[0122] See also Figure 1 , performing feature extraction on the first ultrasonic signal and the first electromagnetic signal, wherein the feature extraction comprises the following steps:

[0123] According to the signal waveform distribution of the first ultrasonic signal and the first electromagnetic signal, all digital signals of the first ultrasonic signal or the first electromagnetic signal (all digital signals on the time axis must be decomposed) are subjected to multi-layer signal filtering decomposition using preset approximate coefficients and preset detail coefficients, and the approximate coefficients and detail coefficients obtained after each layer of signal filtering decomposition are used as the input of the next layer of decomposition, until the iteration is stopped when the preset number of decomposition layers is reached, and the final approximate coefficients and final detail coefficients (the approximate coefficients and detail coefficients of the last decomposition layer of the digital signal) are obtained, and the final approximate coefficients and final detail coefficients are the ultrasonic signal features or electromagnetic signal features of the corresponding digital signal, wherein the final approximate coefficients are used as the low-frequency part of the last decomposition layer, and the final detail coefficients are used as the high-frequency part of the last decomposition layer. Through multi-level signal filtering decomposition, the low-frequency and high-frequency parts of the ultrasonic signal and the electromagnetic signal are extracted respectively, and the final signal features are obtained by iterative processing. This method can effectively extract the multi-scale features of the signal and improve the accuracy and reliability of signal analysis.

[0124] See also Figure 1 , the multimodal fusion in step S3 includes:

[0125] Building a feature gating network based on the ultrasonic signal features and the electromagnetic signal features, taking the final approximation coefficient and the final detail coefficient of the digital signal corresponding to each time node as a feature subspace set, and mapping the feature subspace set to the feature gating network;

[0126] The second calculation formula is set to obtain the final feature. The second calculation formula performs multimodal feature fusion based on the mutual mapping of feature subspace sets. The second calculation formula is as follows:

[0127]

[0128] in, represents the ith final feature of the ultrasonic signal feature and the electromagnetic signal feature, f use () represents the autoencoder function, R(ai ,b i ) represents the feature subspace mapping function of the approximate coefficients of the ultrasonic signal feature and the electromagnetic signal feature, n represents that the ultrasonic signal feature and the electromagnetic signal feature have a total of n time nodes, a i represents the approximate coefficient of the ultrasonic signal characteristics at the i-th time node, b i Represents the approximate coefficient of the electromagnetic signal characteristics at the i-th time node, ω ult Represents the weight coefficient of the approximation coefficient, ω elc Represents the weight coefficient of the detail coefficient, g p () represents the feature gating network function, E(u i ,k i ) represents the characteristic subspace mapping function of the detail coefficients of the ultrasonic signal characteristics and electromagnetic signal characteristics, u i The detail coefficient of the ultrasonic signal feature at the i-th time node, k i Represents the detail coefficient of the electromagnetic signal feature at the i-th time node. By constructing a feature gating network, fusing the approximate coefficients and detail coefficients of the ultrasonic signal and the electromagnetic signal, and performing multimodal feature fusion, more accurate signal features can be extracted, thus improving the accuracy and robustness of signal analysis.

[0129] join Figure 1 , inputting the final feature into the analysis model to obtain the predicted thickness of the sample medium, performing error analysis between the predicted thickness and the actual thickness of the sample medium, adjusting the parameters of the analysis model according to the error analysis results, and obtaining the final model, including:

[0130] The final feature obtained by fusing the original ultrasonic signal and the original electromagnetic signal is used as the input layer of the analysis model, and the predicted thickness of the sample medium is obtained based on the output layer of the analysis model:

[0131]

[0132] Wherein, H represents the thickness of the sample medium, c max represents the maximum eigenvalue of the final feature, c min represents the minimum eigenvalue of the final feature, V() represents the analysis function of the analysis model, Δh represents the height difference between the magnetic eddy current probe and the ultrasonic probe in the vertical direction, and T0 represents the total time value of the final feature;

[0133] The predicted thickness is compared with the actual thickness of the sample medium. If the comparison error is greater than the preset standard threshold, the measurement result is judged to be inaccurate, and the model parameters of the input layer or output layer of the analysis model are adjusted until the comparison error is less than the preset standard threshold to obtain the final model. The thickness of the sample medium is predicted by the analysis model using the fused ultrasonic and electromagnetic signal characteristics, and the model parameters are optimized in combination with the error feedback mechanism to improve the accuracy and reliability of thickness prediction and ensure the accuracy of the measurement results.

[0134] See also Figure 1 , design the overall system architecture, and use the final model as the algorithm part of the overall system architecture.

[0135] Example 2

[0136] This embodiment provides a heterogeneous composite component thickness measurement system based on an ultrasonic-magnetic method, which is used in a heterogeneous composite component thickness measurement method based on an ultrasonic-magnetic method, including:

[0137] A signal acquisition module is used to acquire the thickness of the sample medium and obtain the original ultrasonic signal and the original electromagnetic signal;

[0138] A feature extraction module is used to preprocess the original ultrasonic signal and the original electromagnetic signal to obtain a first ultrasonic signal and a first electromagnetic signal, and to extract features from the first ultrasonic signal and the first electromagnetic signal to obtain ultrasonic signal features and electromagnetic signal features;

[0139] A feature fusion module is used to perform multi-modal feature fusion of ultrasonic signal features and electromagnetic signal features to obtain the final features;

[0140] The prediction module is used to perform error analysis between the predicted thickness of the sample medium and the actual thickness of the sample medium, adjust the parameters of the analysis model, and obtain the final model. Multimodal signals are collected by ultrasonic and magnetic eddy current sensors, and after feature extraction and fusion, the analysis model is used to predict the thickness and optimize the model, which is finally integrated into the system architecture to improve the thickness measurement accuracy and system stability, with high accuracy and reliability.

[0141] Example 3

[0142] This embodiment provides a heterogeneous composite component thickness measurement device based on ultrasonic-magnetic method, see Figure 3 , used for the thickness measurement method of heterogeneous composite components based on ultrasonic-magnetic method, including:

[0143] A three-dimensional coordinate measuring machine 30, wherein the three-dimensional coordinate measuring machine 30 comprises a three-dimensional Z-axis end effector 32 and a three-dimensional coordinate measuring machine turntable 31, wherein the three-dimensional coordinate measuring machine turntable 31 can be equipped with a sample medium;

[0144] A magnetic probe 20, wherein the magnetic probe 20 is connected to the three-coordinate Z-axis end effector 32;

[0145] The ultrasonic pen probe 21 is mounted on the three-dimensional Z-axis end effector 32 and arranged side by side with the magnetic probe 20. By connecting the magnetic probe 20 and the ultrasonic pen probe 21 to the lower end of the three-dimensional Z-axis end effector 32, the sample medium is mounted on the three-dimensional coordinate measuring machine turntable 31, which facilitates the adjustment of the angle of the sample medium, and ensures that the probe maintains a stable position and a correct angle during the measurement process through the three-dimensional Z-axis end effector 32.

[0146] See also Figure 3 , further comprising a probe holder 22 and a wetting unit, wherein the upper end of the probe holder 22 is connected to the three-dimensional Z-axis end effector 32, and the ultrasonic pen-type probe 21 and the wetting unit are both clamped on the probe holder 22. The surface of the sample medium is wetted by the wetting unit to form a water film with the surface of the sample medium, and then the pen-type ultrasonic probe is close to the surface of the sample medium, forming a coupling of the ultrasonic probe 21 to the surface of the sample medium, and obtaining the original ultrasonic signal of the sample medium.

[0147] See also Figure 3 During the scanning process, the ultrasonic pen probe 21 and the magnetic probe 20 always remain perpendicular to the virtual section of the detection point of the sample medium, so as to improve the measurement accuracy.

[0148] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for measuring the thickness of a heterogeneous composite component based on an ultrasonic-magnetic method, characterized in that: include: S1, based on the ultrasonic sensing device and the magnetic eddy current sensing device, respectively collect the thickness of the sample medium to obtain the original ultrasonic signal and the original electromagnetic signal; S2, preprocessing the original ultrasonic signal and the original electromagnetic signal to obtain a first ultrasonic signal and a first electromagnetic signal respectively, and performing feature extraction on the first ultrasonic signal and the first electromagnetic signal to obtain ultrasonic signal features and electromagnetic signal features respectively; S3, performing multi-modal feature fusion on the ultrasonic signal feature and the electromagnetic signal feature to obtain a final feature; S4, inputting the final feature into the analysis model to obtain the predicted thickness of the sample medium, performing error analysis between the predicted thickness and the actual thickness of the sample medium, adjusting the parameters of the analysis model, and obtaining the final model.

2. The method for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method according to claim 1, characterized in that: The preprocessing of the original ultrasonic signal and the original electromagnetic signal in step S2 comprises the following steps: S21, based on the transmission conditioning unit, filtering and amplifying the original ultrasonic signal to obtain an amplified ultrasonic signal, and then separating the amplified ultrasonic signal into an amplified ultrasonic signal sequence according to time nodes; S22, transforming the amplified ultrasonic signal sequence according to a preset modulation frequency with m as a factor signal, where m is an integer greater than 0, and setting a first calculation formula to obtain an amplified ultrasonic digital signal sequence, wherein the first calculation formula is: Wherein, X(k) represents the amplified ultrasonic digital signal sequence after the amplified ultrasonic signal sequence is converted, x(k) represents the kth signal value of the amplified ultrasonic signal sequence, a0 represents the first phase weight coefficient, a1 represents the second phase weight coefficient, a2 represents the third phase weight coefficient, step() represents the quantization step function, and t0 represents the standard step size of the amplified ultrasonic digital signal sequence; S23, if the sampling frequency of the amplified ultrasonic digital signal sequence is different from the sampling frequency of the original electromagnetic signal, the signal with the lowest frequency is interpolated and then time-axis aligned; or if the sampling frequency of the amplified ultrasonic digital signal sequence is the same as the sampling frequency of the original electromagnetic signal, the two signals are directly time-axis aligned; S24, using a cross-correlation function to obtain the time offset of the two signals after alignment, and performing translation on the standard time axis based on the time offset to achieve time axis alignment, thereby obtaining a first ultrasonic signal and a first electromagnetic signal.

3. The method for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method according to claim 2, characterized in that: Performing feature extraction on the first ultrasonic signal and the first electromagnetic signal, wherein the feature extraction comprises the following steps: Perform multi-layer filtering decomposition of the digital signals of the first ultrasonic signal or the first electromagnetic signal using preset approximate coefficients and preset detail coefficients, use the approximate coefficients and detail coefficients obtained after each layer of signal filtering decomposition as input for the next layer of decomposition, and stop iterating when the preset number of decomposition layers is reached, to obtain final approximate coefficients and final detail coefficients, wherein the final approximate coefficients and final detail coefficients are ultrasonic signal features or electromagnetic signal features of the corresponding digital signal; Among them, the final approximation coefficient is used as the low-frequency part of the final decomposition layer, and the final detail coefficient is used as the high-frequency part of the final decomposition layer.

4. The method for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method according to claim 3 is characterized in that: The multimodal fusion in step S3 includes: Building a feature gating network based on the ultrasonic signal features and the electromagnetic signal features, taking the final approximation coefficient and the final detail coefficient of the digital signal corresponding to each time node as a feature subspace set, and mapping the feature subspace set to the feature gating network; The second calculation formula is set to obtain the final feature. The second calculation formula performs multimodal feature fusion based on the mutual mapping of feature subspace sets. The second calculation formula is as follows: in, represents the ith final feature of the ultrasonic signal feature and the electromagnetic signal feature, f use () represents the autoencoder function, R(a i ,b i ) represents the feature subspace mapping function of the approximate coefficients of the ultrasonic signal feature and the electromagnetic signal feature, n represents that the ultrasonic signal feature and the electromagnetic signal feature have a total of n time nodes, a i represents the approximate coefficient of the ultrasonic signal characteristics at the i-th time node, b i Represents the approximate coefficient of the electromagnetic signal characteristics at the i-th time node, ω ult Represents the weight coefficient of the approximation coefficient, ω elc Represents the weight coefficient of the detail coefficient, g p () represents the feature gating network function, E(u i ,k i ) represents the characteristic subspace mapping function of the detail coefficients of the ultrasonic signal characteristics and electromagnetic signal characteristics, u i The detail coefficient of the ultrasonic signal feature at the i-th time node, k i Represents the detail coefficient of the electromagnetic signal characteristics at the i-th time node.

5. The method for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method according to claim 4, characterized in that: Inputting the final feature into the analysis model to obtain the predicted thickness of the sample medium, performing error analysis between the predicted thickness and the actual thickness of the sample medium, adjusting the parameters of the analysis model, and obtaining the final model, including: The final feature is used as the input layer of the analysis model, and the predicted thickness of the sample medium is obtained based on the output layer of the analysis model: Wherein, H represents the thickness of the sample medium, c max represents the maximum eigenvalue of the final feature, c min represents the minimum eigenvalue of the final feature, V() represents the analysis function of the analysis model, Δh represents the height difference between the magnetic eddy current probe and the ultrasonic probe in the vertical direction, and T0 represents the total time value of the final feature; The predicted thickness is compared with the actual thickness of the sample medium to obtain a comparison error. If the comparison error is greater than a preset standard threshold, the measurement result is judged to be inaccurate, and the model parameters of the input layer or output layer of the analysis model are adjusted until the comparison error is less than the preset standard threshold to obtain the final model.

6. The method for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method according to claim 5, characterized in that: Design an overall system architecture, and use the final model as the algorithm part of the overall system architecture.

7. A heterogeneous composite component thickness measurement system based on ultrasonic-magnetic method, characterized in that: The method for measuring thickness of a heterogeneous composite component based on an ultrasonic-magnetic method as described in any one of claims 1 to 6 comprises: A signal acquisition module is used to acquire the thickness of the sample medium and obtain the original ultrasonic signal and the original electromagnetic signal; A feature extraction module is used to preprocess the original ultrasonic signal and the original electromagnetic signal to obtain a first ultrasonic signal and a first electromagnetic signal, and to extract features from the first ultrasonic signal and the first electromagnetic signal to obtain ultrasonic signal features and electromagnetic signal features; A feature fusion module is used to perform multi-modal feature fusion of ultrasonic signal features and electromagnetic signal features to obtain the final features; The prediction module is used to perform error analysis between the predicted thickness of the sample medium and the actual thickness of the sample medium, adjust the parameters of the analysis model, and obtain a final model.

8. A device for measuring thickness of heterogeneous composite components based on ultrasonic-magnetic method, characterized in that: The method for measuring thickness of a heterogeneous composite component based on an ultrasonic-magnetic method as described in any one of claims 1 to 6 comprises: A three-dimensional coordinate measuring machine, wherein the three-dimensional coordinate measuring machine comprises a three-dimensional Z-axis end effector and a three-dimensional coordinate measuring machine turntable, and the three-dimensional coordinate measuring machine turntable can be equipped with a sample medium; A magnetic probe, wherein the magnetic probe is connected to the three-coordinate Z-axis end effector; An ultrasonic pen probe is mounted on the three-coordinate Z-axis end effector and is arranged side by side with the magnetic probe.

9. The heterogeneous composite component thickness measuring device based on ultrasonic-magnetic method according to claim 8, characterized in that: It also includes a probe bracket and a water moistening unit. The upper end of the probe bracket is connected to the three-coordinate Z-axis end effector, and the ultrasonic pen probe and the water moistening unit are both clamped on the probe bracket.

10. The heterogeneous composite component thickness measuring device based on ultrasonic-magnetic method according to claim 8, characterized in that: The ultrasonic pen probe and the magnetic probe are always kept perpendicular to the virtual section of the detection point of the sample medium during the scanning process.

Images