Adaptive spatio-temporal complementary coded-phase congruency all-in-focus method for non-destructive inspection

Abstract

The application discloses a kind of self-adapting space-time complementary coding-phase consistency all-focusing methods of nondestructive testing, establishes ROI and sound velocity model;Construct time-domain three-value complementary gap coding sequence;Grouping to each array element, apply space code using Hadamard matrix to realize in-group synchronous transmission and can be de-modulated and separated;Two-stage decoding at receiving end: first spatial de-modulation obtains equivalent single-array element excitation-full-array element receiving response, then complementary pulse pressure decoding obtains equivalent full-matrix capture matrix;Extract amplitude and phase information, construct all-focusing amplitude chart and phase consistency chart and adaptively fuse;Dynamically select code length, gap duty cycle and grouping scale, according to the material attenuation, coupling state and near-field dead zone constraint estimated online;The application constructs time-domain-space-phase collaborative framework, realizes deep high signal-to-noise ratio, near-field dead zone controllable, efficiency improvement, and compatible standard FMC / TFM / PCI process.

Description

Technical Field

[0001] This invention belongs to the field of industrial nondestructive testing and structural health monitoring technology, and particularly relates to an adaptive spatiotemporal complementary coding-phase consistent full focusing method for nondestructive testing. Background Technology

[0002] In industrial nondestructive testing, ultrasonic phased array testing (PAUT) technology, due to its flexible electronic scanning and focusing capabilities, has been widely applied to defect detection and dimensional assessment of welds, thick-walled components, pressure vessels, and critical equipment. In recent years, full matrix acquisition (FMC) combined with total focusing imaging (TFM) has become an important evolutionary direction for phased array technology. By acquiring the A-scan signals from all transmit-receive channels of the array and performing focusing summation on each pixel within the region of interest, high-resolution two-dimensional / three-dimensional imaging can be achieved. Meanwhile, phase-consistent imaging (PCI), as a complementary method, utilizes echo phase information rather than relying solely on amplitude, exhibiting higher sensitivity for small diffractors, crack tips, and weak defects against grain noise backgrounds. The increasing demands for reliability, efficiency, and deep defect detection capabilities in industrial settings are driving the exploration of applications of FMC / TFM / PCI methods in difficult-to-detect materials such as thick-walled metals, austenitic stainless steel, composite materials, and large forgings.

[0003] To overcome the problems of insufficient energy and low signal-to-noise ratio (SNR) of traditional short-pulse excitation in deep defect detection, existing technologies have introduced coded excitation methods. Common code types include chirp, Barker codes, and Golay complementary sequences. Longer code transmissions increase average acoustic power, and pulse compression at the receiver restores axial resolution. To improve detection efficiency, some studies employ spatial coding strategies, such as orthogonal coding based on the Hadamard matrix, to achieve synchronous transmission of multiple array elements. Decoding separates the responses of each channel, reducing the N transmissions required for FMC to the logarithmic level. To address the near-field dead zone expansion caused by long-code excitation, existing schemes insert receiving gaps into the coded sequence, allowing near-field echo reception during transmission intervals, balancing long-code gain and near-field visibility. Furthermore, to reduce data throughput, some technologies propose phase-dominant quantization or binarization compression of the echoes and image reconstruction using phase information. These methods have demonstrated their potential to improve SNR, accelerate acquisition, or compress data in medical ultrasound and some industrial applications.

[0004] However, existing technologies still have significant shortcomings in terms of complete system integration for industrial non-destructive testing. First, most solutions employ pure time-domain coding, pure spatial coding, or pure phase compression, lacking a systematic framework that unifies and coordinates complementary time-domain coding, receiver gap design, spatial orthogonal spread spectrum, and phase-consistent imaging. This makes it difficult to simultaneously address the three major contradictions of improving SNR for deep defects, increasing detection efficiency, and controlling near-field dead zones. Second, existing methods generally lack adaptive code selection and partitioning strategies for industrial field conditions (such as coupling fluctuations, material attenuation variations, rough surfaces, and regions of varying thicknesses), typically relying on fixed parameters, leading to unstable performance under complex conditions. Third, most coding or compression methods fail to provide an "equivalent FMC matrix output" interface seamlessly compatible with industrial standard FMC / TFM / PCI workflows, making it difficult for existing detection systems to directly achieve the combined benefits of high SNR, low data volume, and high detection efficiency without altering the backend imaging process. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide an adaptive spatiotemporal complementary coding-phase consistent full-focusing method for non-destructive testing that can alleviate the shortcomings of existing technologies, such as lack of a time-space-phase collaborative framework, lack of industrial operating condition adaptability, and difficulty in compatibility with standard FMC / TFM / PCI workflows. This method achieves high signal-to-noise ratio for deep defects, controlled near-field dead zone, optimized detection efficiency and data throughput, and supports pattern and partition adaptation.

[0006] Technical solution: The adaptive spatiotemporal complementary coding-phase consistent full focusing method for nondestructive testing described in this invention includes the following steps:

[0007] S1. Establish the ROI and sound velocity model of the region to be detected, and determine the number of array elements N, sampling rate Fs, and maximum detection depth D of the array transducer. max and minimum near-field distance D min ;

[0008] S2. Construct a time-domain ternary complementary gap coding sequence family {c_k(t)}, where the coding symbols in the sequence family {c_k(t)} take values ​​of {+1, -1, 0}, where the value 0 corresponds to a receive gap, used to allow callback reception through the receiver of the transmit-receive front end during the coding transmission period. The position and length of the receive gap are determined by the minimum near-field distance D. min The corresponding round-trip propagation time is determined; the sequence family has a preset code length and gap duty cycle;

[0009] S3. Implement grouped orthogonal spread spectrum for each array element, including: dividing N array elements into several groups, the number of groups and the number of array elements in each group constitute the group size; applying a spatial code to the array elements in each group using the Hadamard matrix, so that multiple array elements in the same group can transmit synchronously, and their echoes can be separated by despreading.

[0010] S4. At the receiving end of the transmitting and receiving front end, the echo generated by the synchronous transmission is decoded in two stages. First, spatial despreading is performed to obtain the equivalent intermediate response of single-element excitation-full-element reception. Then, complementary pulse compression decoding is performed on each intermediate response to obtain the equivalent full-matrix capture response matrix K(t). Time discretization and time delay correction are performed based on the sampling rate Fs.

[0011] S5. The amplitude and phase information are extracted from the equivalent full matrix capture response matrix K(t) by the acquisition and calculation unit, at the maximum detection depth D. max Within a defined range, construct a fully focused amplitude image A(x,z) and a phase-consistent image P(x,z), and generate a fused image F(x,z) according to an adaptive fusion criterion;

[0012] S6. Dynamically select the code length and gap duty cycle of the sequence family in step S2 and the group size in step S3, wherein the dynamic selection is based on online estimated material attenuation, coupling state and near-field dead zone constraints.

[0013] This invention effectively alleviates the shortcomings of existing methods by constructing a complete technical chain from parameter modeling, temporal ternary complementary gap coding, spatial group spread spectrum to two-stage decoding, phase-amplitude fusion imaging, and dynamic adaptive adjustment. This chain addresses the lack of a temporal-spatial-phase collaborative framework, insufficient adaptability to industrial conditions, and incompatibility with standard full-matrix acquisition / full-focusing / phase-coherent imaging workflows. Specifically, step S1 lays the foundation for the detection region and sound velocity model, ensuring precise matching of subsequent processing with detection depth and near-field constraints. Step S2 introduces ternary complementary coding with a receiving gap, dynamically configuring the gap position and length based on the minimum near-field distance to actively control the near-field dead zone at the coding level, while utilizing complementary characteristics to ensure pulse compression performance. Step S3 implements grouped orthogonal spread spectrum using a Hadamard matrix, supporting simultaneous transmission of multiple array elements and echo separation, significantly improving detection efficiency and data throughput. Step S4 performs two-stage decoding, first spatially despreading to obtain an equivalent single array. The intermediate response of the element excitation-full array element reception is then completed with complementary pulse compression decoding to generate an equivalent full matrix capture response matrix compatible with standard workflows. Step S5 extracts amplitude and phase information jointly under the full focusing framework to generate an adaptive fusion image, achieving high signal-to-noise ratio imaging for deep defects. Step S6 dynamically selects the code length, gap duty cycle, and group size online based on material attenuation, coupling state, and near-field dead zone constraints, enabling the entire method to have industrial working condition adaptive capabilities, realize flexible adjustment of code pattern and partition, and comprehensively optimize near-field control, detection efficiency, and data acquisition quality.

[0014] Preferably, the sequence family {c_k(t)} described in step S2 consists of one or more pairs of complementary sequences, such that after complementary pulse compression decoding, their sidelobe energies cancel each other out or are suppressed to below a set threshold within a preset delay window.

[0015] This preferred scheme introduces complementary sequence pairs, which enable the sidelobe energies after decoding to cancel each other out or significantly suppress each other within a specified delay window. This effectively reduces artifact interference, improves the clarity of the main lobe, enhances the identifiability of weak defect signals, and ensures the waveform fidelity in the near field, providing a purer impulse response for subsequent high-contrast fusion imaging.

[0016] Preferably, the code length and gap duty cycle of the sequence family {c_k(t)} in step S2 are configured according to the detection depth partition. The detection depth is divided into at least a near-field region, a mid-field region, and a far-field region. The near-field region uses short codes or no codes, the mid-field region uses medium code lengths, and the far-field region uses long codes. The number of complementary code pairs and the gap duty cycle are configured for different depth partitions respectively.

[0017] This preferred scheme adaptively configures the code length, number of complementary code pairs, and gap duty cycle according to the different requirements of signal-to-noise ratio and resolution in the near-field, mid-field, and far-field regions: short codes or no coding are used in the near-field region to prioritize near-field resolution and dead zone control; medium code lengths are used in the mid-field region to balance signal-to-noise ratio and imaging accuracy; and long codes and more complementary code pairs are used in the far-field region to enhance the echo energy and noise resistance of deep defects. This achieves refined zoning optimization of detection performance across the entire depth range, improving overall imaging consistency and adaptability to different operating conditions.

[0018] Preferably, the grouped orthogonal spreading in step S3 satisfies that the total number of array elements N is equal to the product of the number of groups G and the number of array elements in each group, i.e., N=G×M; Hadamard spreading is applied to each group and despreading is achieved on the receiving side by matrix multiplication or addition and subtraction operations, so that the echo of synchronous transmission is separated into M equivalent transmission channel responses.

[0019] This preferred scheme sets the total number of array elements to the product of the number of groups and the number of elements in each group, and applies orthogonal spreading codes to each group using the Hadamard matrix. At the receiving side, only matrix multiplication or addition and subtraction operations are required for efficient despreading, thereby accurately separating the mixed echoes of synchronous transmission into multiple equivalent transmission channel responses. This not only significantly reduces the total number of transmissions and improves the data acquisition throughput, but also maintains the integrity of the response information equivalent to standard full matrix acquisition, laying a low-complexity and high-efficiency data foundation for subsequent high real-time full-focus imaging.

[0020] Preferably, the complementary pulse compression decoding in step S4 is implemented using matched filtering or its equivalent form, and combined with a window function to suppress sidelobes.

[0021] This preferred scheme employs matched filtering or its equivalent form for complementary pulse compression decoding, and combines it with a window function to suppress side lobes. This effectively reduces side lobe interference in the decoding output while maintaining the advantages of main lobe width and signal-to-noise ratio, further improving the time resolution and signal fidelity of the equivalent full matrix capture response, and providing a cleaner pulse response waveform for the accurate extraction of subsequent amplitude and phase information.

[0022] Preferably, the equivalent full-matrix capture response matrix K(t) obtained in step S4 is consistent with the full-matrix capture matrix obtained by matrix-by-matrix excitation in terms of data structure.

[0023] This preferred scheme ensures that the equivalent full-matrix acquisition response matrix obtained after grouped synchronous transmission and two-level decoding is completely consistent with the standard full-matrix acquisition matrix obtained by element-by-element excitation in terms of data structure. This allows for seamless compatibility with existing full-matrix acquisition-based full-focus imaging and phase-coherent imaging processing workflows without the need for additional modifications to the underlying algorithm or data organization method. This reduces the complexity of technology migration and integration, while maintaining the universality and maturity of the post-processing method.

[0024] Preferably, the echo is low-bit quantized or edge-indexed at the receiving end of the transmitting and receiving front-end to compress the data volume, and the instantaneous phase or equivalent symbol phase sequence is reconstructed in the acquisition and calculation unit for focusing in phase-consistent imaging and full-focus imaging under phase-consistency constraints.

[0025] This preferred scheme performs low-bit quantization or edge indexing on the echo at the receiving end to compress the data volume, while reconstructing the instantaneous phase or equivalent symbol phase sequence in the acquisition and computing unit. Thus, without significantly increasing the storage and transmission burden, it effectively extracts the phase information required for phase-consistent imaging and phase-consistent constrained full-focus imaging, taking into account both the optimization of data throughput and the performance requirements of phase-sensitive imaging. It is especially suitable for high-precision non-destructive testing in large-scale arrays and long-term detection scenarios.

[0026] Preferably, the fused image F(x,z) in step S5 is constructed in the following manner or in its equivalent form:

[0027]

[0028] Wherein, α(x,z) is the fusion weighting factor, which is adaptively determined by the local coherence factor, coupling confidence, material attenuation estimation and defect type prior.

[0029] This preferred scheme constructs a fusion weighting factor, which incorporates local coherence factor, coupling confidence, material attenuation estimation, and defect type prior into an adaptive fusion model. This allows the fused image to dynamically adjust the contribution ratio of amplitude image and phase consistency image at different spatial locations. As a result, it prioritizes the robustness of phase information in areas with strong scattering or low signal-to-noise ratio, while retaining the resolution advantage of amplitude information in high-contrast areas, thus achieving optimal imaging performance that adapts to defect type and detection conditions.

[0030] Preferably, the online estimation of material attenuation and coupling state in step S6 is obtained based on the front wall echo, back wall echo, or reference reflector echo, and the code length, gap duty cycle, and packet size are adaptively adjusted accordingly.

[0031] This preferred scheme estimates the material attenuation and coupling state online using front wall echo, back wall echo, or reference reflector echo, and dynamically adjusts the code length, gap duty cycle, and packet size accordingly. This allows the transmission coding strategy and reception timing to match the acoustic attenuation characteristics and interface coupling conditions in real time, avoiding near-field dead zone deterioration or insufficient signal-to-noise ratio at depth due to fixed parameters. It achieves closed-loop adaptive optimization of detection parameters and medium characteristics, improving detection reliability and stability across the entire depth range.

[0032] The industrial non-destructive testing system of the present invention includes:

[0033] Array transducers are used to transmit and receive ultrasonic signals;

[0034] A programmable multi-channel transmitter, connected to the array transducer, is used to generate coded excitation signals;

[0035] A programmable receiving and quantization module, connected to the array transducer, is used to receive echoes and perform quantization;

[0036] A time-space coding controller is connected to the programmable multi-channel transmitter and the programmable receiver and quantization module, respectively, and is used to generate a time-domain ternary complementary gap coding sequence and control the spatial grouping orthogonal spread spectrum of the array.

[0037] The decoding and imaging processor, connected to the programmable receiving and quantization module, is used to perform spatial despreading, complementary pulse compression decoding, and the construction and fusion of full-focus imaging amplitude and phase-consistent images.

[0038] The display and storage module, connected to the decoding and imaging processor, is used to display the fused image and store the detection data.

[0039] Beneficial Effects: Compared with the prior art, the present invention has the following significant advantages: 1. The present invention constructs a family of time-domain ternary complementary coding sequences with a receiving gap, and performs grouped orthogonal spreading and two-level decoding on the pair elements. Based on obtaining the equivalent full matrix acquisition response, amplitude and phase information are extracted and adaptively fused, thus constructing a time-domain-space-phase collaborative framework for the first time. This effectively alleviates the shortcomings of the prior art, such as lack of industrial operating condition adaptive capability and difficulty in compatibility with standard FMC / TFM / PCI workflows. It achieves high signal-to-noise ratio for deep defects, controlled near-field dead zone, optimized detection efficiency and data throughput, and supports lossless adaptive coding pattern and partitioning. 1. Flaw detection; 2. This invention employs complementary pulse compression decoding, utilizing complementary sequence pairs to cancel out or suppress sidelobe energy to below a set threshold within a preset delay window, significantly improving imaging contrast and defect resolution, especially suitable for detecting weak defects in complex industrial scenarios with strong reflections from the front / back walls; 3. This invention dynamically selects code length, gap duty cycle, and group size by estimating material attenuation, coupling state, and near-field dead zone constraints online, enabling the system to adaptively adjust transmission and reception parameters to match different detection depth partitions (near field, mid field, far field) and industrial conditions, eliminating the need for repeated manual parameter adjustments and greatly improving the robustness and efficiency of detection. Attached Figure Description

[0040] Figure 1 This is a hardware block diagram of the ST-Focus™ system of the present invention;

[0041] Figure 2 This is a schematic diagram of the timing of the time-domain ternary complementary gap coding of the present invention;

[0042] Figure 3 This is a schematic diagram of the spatial Hadamard grouping spread spectrum and despreading method of the present invention;

[0043] Figure 4 This is a flowchart of the two-level decoding and equivalent FMC matrix construction of the present invention;

[0044] Figure 5 This is a schematic diagram of TFM / PCI reconstruction and fusion imaging according to the present invention;

[0045] Figure 6 This is a schematic diagram of a typical workpiece and scanning arrangement according to the present invention;

[0046] Figure 7 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0047] The technical solution of the present invention will be further described below with reference to the accompanying drawings.

[0048] This invention provides an adaptive spatiotemporal complementary coding-phase consistent full-focusing method for non-destructive testing, applied to an ultrasonic testing system comprising an array transducer, a transmitting and receiving front-end, and an acquisition and computing unit. (Refer to...) Figure 1 The ST-Focus™ system hardware block diagram shown below includes:

[0049] Array transducer (101): used for transmitting and receiving ultrasonic signals.

[0050] Wedge / coupling layer (102): Located between the array transducer (101) and the workpiece to be tested, it is used to achieve effective coupling of acoustic waves and can provide a specific refraction angle (such as 55° for transverse waves) according to the testing needs.

[0051] Multichannel transmitter (103): connected to array transducer (101), used to generate coded excitation signal under the control of spatiotemporal coding controller (105) to drive array transducer (101) to emit ultrasonic waves.

[0052] Multi-channel receiving and quantization module (104): connected to array transducer (101), used to receive echo signals and perform quantization processing (optional low-bit quantization or edge indexing to achieve data compression).

[0053] Spatiotemporal coding controller (105): It is connected to the multi-channel transmitter (103) and the multi-channel receiver and quantization module (104) respectively. Its core functions are: ① generating time-domain three-value complementary gap coding sequence; ② controlling the spatial grouping orthogonal spread spectrum of the array (such as Hadamard grouping).

[0054] Spatial despreading module (106): Spatial despreading is performed on the received aliased signal, and the equivalent intermediate response of single-element transmission and full-element reception is separated by orthogonal code (such as Hadamard matrix).

[0055] Complementary pulse compression decoding module (107): Performs complementary pulse compression decoding (matched filtering) on ​​each intermediate response after spatial despreading to obtain equivalent full matrix capture response data.

[0056] FMC matrix cache (108): Used to store the equivalent full matrix capture response matrix K(t) for subsequent imaging processing.

[0057] TFM reconstruction module (109): Based on the equivalent FMC matrix, perform full focusing algorithm to construct the amplitude image A(x,z) of each pixel in the ROI.

[0058] PCI phase consistency imaging module (110): Based on the equivalent FMC matrix, phase information is extracted and a phase consistency image P(x,z) is constructed. It is particularly sensitive to targets such as small diffractors and crack tips.

[0059] Fusion and Defect Assessment Module (111): The TFM amplitude image A(x,z) and the PCI phase consistency image P(x,z) are fused according to the adaptive fusion criterion to generate a fused image F(x,z), and defect detection, size assessment and classification are completed based on this image.

[0060] Display and storage module (112): Connected to fusion and defect assessment module (111), used to display the final fused image and store all detection data.

[0061] like Figure 7 As shown, the method flow of the present invention specifically includes the following steps:

[0062] S1. Establish the detection model and initialize parameters.

[0063] Establish the ROI and sound velocity model of the region to be detected, and determine the number of array elements N, sampling rate Fs, and maximum detection depth D of the array transducer. max and minimum near-field distance D min .

[0064] S2. Construct a family of time-domain ternary complementary gap coding sequences.

[0065] See Figure 2 A family of time-domain ternary complementary gap coding sequences {c_k(t)} is constructed by the spatiotemporal coding controller (105), with coding symbols taking values ​​of {+1, -1, 0}. The value 0 corresponds to the "reception gap" ( Figure 2(202, 203) are used to allow the receiver to receive echoes during the coded transmission period. The position and length of the receive gap are determined by the round-trip propagation time corresponding to the minimum near-field distance Dmin, in order to avoid near-surface information loss. This sequence family consists of one or more pairs of complementary sequences (such as Golay codes), such that after complementary pulse compression decoding, the sidelobe energy cancels each other out or is suppressed to below a set threshold within a preset delay window.

[0066] Preferably, the code length and gap duty cycle are configured according to the detection depth partition ( Figure 2 Lower part: The near field region (0-Dmin) uses short pulses or no coding; the mid field region uses medium code length Lm; and the far field region uses long code length Lf. The near field protection window (204) further ensures that the near field echo is not overwhelmed by the transmitted pulse.

[0067] S3, Spatial Grouping Orthogonal Spread Spectrum

[0068] See Figure 3 Divide the N array elements into G groups, with M elements in each group, satisfying N = G × M. Apply spatial orthogonal codes to the elements within each group using the Hadamard matrix. Figure 3 (303), enabling multiple array elements within the same group to launch synchronously. Figure 3 (304). The receiving end performs matrix despreading through despreader (305). Figure 3 (301), separating the equivalent intermediate response of "single element excitation - full element reception" ( Figure 3 (302).

[0069] S4, Two-level Decoding and Construction of Equivalent FMC Matrix

[0070] See Figure 4 At the receiving end, the echo generated by the synchronous transmission is decoded in two stages:

[0071] First stage: Spatial despreading. Spatial despreading is performed on the received aliased signal R(t) (all array elements). Figure 4 (401) Utilizing the orthogonality of the Hadamard matrix, the intermediate response D_mn(t) is obtained through matrix multiplication or addition / subtraction operations.

[0072] Level 2: Complementary pulse compression decoding. Complementary pulse compression decoding is performed on each D_mn(t) channel. Figure 4 (402), using matched filtering or its equivalent form (which can be combined with window functions to suppress sidelobes), the equivalent full-matrix capture response matrix K_ij(t) is obtained. Figure 4 (403).

[0073] Finally, K(t) is time-discrete and time-delay corrected based on the sampling rate Fs, and then processed through the cache / data interface ( Figure 4Output 404. The equivalent FMC matrix K(t) is completely identical in data structure to the FMC matrix obtained by conventional matrix-by-matrix excitation and can be directly input into the standard TFM / PCI process.

[0074] S5, TFM / PCI fusion imaging

[0075] See Figure 5 From the equivalent FMC matrix K(t) ( Figure 5 (501) Extract amplitude and phase information:

[0076] TFM reconstruction: Focus and sum the values ​​at each grid point (x,z) within the ROI to construct a full-focus amplitude image A(x,z). Figure 5 (502).

[0077] PCI Reconstruction: Calculate the phase consistency value of each grid point within the ROI and construct a phase consistency image P(x,z). Figure 5 (503). PCI is particularly sensitive to small diffractors and crack tips.

[0078] Adaptive fusion: Generating a fused image F(x,z) according to fusion criteria ( Figure 5 (504). Preferably, the following formula or its equivalent form is used:

[0079]

[0080] Where α(x,z) is a fusion weight factor adaptively determined by the local coherence factor, coupling confidence, material attenuation estimation, and defect type prior. The fused image is used for defect detection, size assessment, and classification. Figure 5 (505).

[0081] S6, Adaptive Dynamic Selection

[0082] Based on online estimations of material attenuation, coupling state, and near-field dead zone constraints, the code length and gap duty cycle in step S2 and the group size in step S3 are dynamically selected. Material attenuation and coupling state are estimated through front wall echo, back wall echo, or reference reflector echo, and parameters are adaptively adjusted accordingly without changing the peak voltage.

[0083] The following three specific embodiments further illustrate the beneficial effects of the present invention.

[0084] Example 1: Detection of deep cracks / lack of fusion in thick-walled steel welds (see...) Figure 6 )

[0085] This embodiment is for Figure 6 The thick-walled steel weld specimen shown. Figure 6In the diagram, 601 represents the workpiece to be inspected (40mm thick carbon steel plate, V-groove butt weld); 602 represents the region of interest (ROI); 603 represents the wedge (55° angle); 604 and 605 are the front and rear reference positions of the array transducer, respectively, indicating the placement range of the probe on the workpiece surface; 606 indicates a depth of 28mm (corresponding to the location of the non-fusion defect); 607 indicates a depth of 32mm (corresponding to the location of the side-drilled hole); and 608 indicates a depth of 35mm (corresponding to the location of the crack tip). The total thickness of the workpiece is 40mm, and the lateral range of the ROI is also indicated in the diagram. Artificial defects include: non-fusion EDM groove (28mm deep, corresponding to number 606); side-drilled SDH hole (Φ2mm, 32mm deep, corresponding to number 607); and crack tip (35mm deep, corresponding to number 608).

[0086] The probe uses a 64-element linear array with a center frequency of 5MHz, an element spacing of 0.6mm, a wedge shear wave of 55°, a sampling rate of 100MHz, and a ROI depth of 0-45mm.

[0087] Configuration of this invention:

[0088] Zones: Near field 0-10mm: no coding; Mid field 10-25mm: code length L=256, gap duty cycle 0.25; Far field 25-45mm: code length L=1024, gap duty cycle 0.35.

[0089] Spatial spread spectrum: G=8 groups, M=8 elements per group, according to Figure 3 The method is to perform concurrent transmission of Hadamard spread spectrum.

[0090] Receive: 8-bit amplitude is reserved for TFM, while a symbol phase sequence is generated for PCI.

[0091] according to Figure 4 The two-stage decoding process shown outputs an equivalent FMC matrix, according to... Figure 5 The process shown generates the TFM amplitude map, PCI phase map, and fusion map.

[0092] Results: Compared with traditional short-pulse FMC / TFM, the SNR of unfused defects increased from 12.8dB to 22.6dB (25.1dB in fusion graph); the SNR of crack tips increased from 6-8dB to 18.7dB, and the positioning deviation was ≤0.6mm. The number of emission events decreased from 64 to 16, and the efficiency was improved by 4 times.

[0093] Example 2: Detection of delamination / debonding within strong attenuation composite materials

[0094] This embodiment is for a 30mm thick CFRP laminate (such as...). Figure 6(Similar arrangement shown), material frequency-dependent attenuation is significant. Defects include: delamination D1 (18mm deep, 10mm Φ), D2 (24mm deep, 6mm Φ), and debonding D3 (26mm deep, 20×8mm).

[0095] The probe uses a 128-element linear array with a center frequency of 3.5MHz, a sampling rate of 80MHz, and a ROI depth of 0-35mm.

[0096] Configuration of this invention:

[0097] Zones: Near field 0-8mm short pulse; Mid field 8-18mm: code length L=512, gap duty cycle 0.30; Far field 18-35mm: code length L=2048, gap duty cycle 0.40.

[0098] Spatial spread spectrum: G=16 groups, M=8 elements in each group.

[0099] The fusion weights are biased towards amplitude dominance (α≈0.75), and PCI is used to suppress random noise hotspots.

[0100] Results: For delamination with a depth of 24 mm and a diameter of 6 mm, the traditional method achieved an SNR of 5-7 dB, while the present invention achieved 16.9 dB; for debonding with a depth of 26 mm, the boundary continuity was significantly improved, and the area measurement error was ≤8%.

[0101] Example 3: Small defect enhancement in coarse-grained materials (PCI-dominated)

[0102] This embodiment focuses on an austenitic stainless steel specimen (with strong structural noise). The targets are a micropore (Φ1mm) with a depth of 15mm and a crack tip with a depth of 20mm.

[0103] The probe uses 64 elements and has a center frequency of 7.5MHz.

[0104] Configuration of this invention:

[0105] Time-domain coding: L=1024 is used for mid-to-far field, with a gap duty cycle of 0.35.

[0106] The receiver enables phase-driven compressed acquisition: low-bit quantization / edge indexing, and instantaneous phase reconstruction is dedicated to PCI.

[0107] according to Figure 5 , Figure 6 Process reconstruction, primarily using PCI images.

[0108] Results: The micropore SNR increased from 7.9dB to 15.8dB; the crack tip showed a stable coherent peak in the PCI plot, and the standard deviation of the localization repeatability was ≤0.4mm.

[0109] The above three embodiments fully demonstrate that the "adaptive spatiotemporal complementary coding-phase consistent full focusing method" and system proposed in this invention effectively solves the core bottlenecks of traditional methods in terms of deep signal-to-noise ratio, detection efficiency, near-field dead zone control and small defect detection, while maintaining compatibility with the industry standard FMC / TFM process. It has significant creative and industrial application value.

Claims

1. An adaptive spatiotemporal complementary coding-phase consistent full-focusing method for non-destructive testing, applied to an ultrasonic testing system comprising an array transducer, a transmitting and receiving front-end, and an acquisition and computing unit, characterized in that, Includes the following steps: S1. Establish the ROI and sound velocity model of the region to be detected, and determine the number of array elements N, sampling rate Fs, and maximum detection depth D of the array transducer. max and minimum near-field distance D min ; S2. Construct a time-domain ternary complementary gap coding sequence family {c_k(t)}, where the coding symbols in the sequence family {c_k(t)} take values ​​of {+1, -1, 0}, where the value 0 corresponds to a receive gap, used to allow callback reception through the receiver of the transmit-receive front end during the coding transmission period. The position and length of the receive gap are determined by the minimum near-field distance D. min The corresponding round-trip propagation time is determined; the sequence family has a preset code length and gap duty cycle; S3. Implement grouped orthogonal spread spectrum for each array element, including: dividing N array elements into several groups, the number of groups and the number of array elements in each group constitute the group size; applying a spatial code to the array elements in each group using the Hadamard matrix, so that multiple array elements in the same group can transmit synchronously, and their echoes can be separated by despreading. S4. At the receiving end of the transmitting and receiving front end, the echo generated by the synchronous transmission is decoded in two stages. First, spatial despreading is performed to obtain the equivalent intermediate response of single-element excitation-full-element reception. Then, complementary pulse compression decoding is performed on each intermediate response to obtain the equivalent full-matrix capture response matrix K(t). Time discretization and time delay correction are performed based on the sampling rate Fs. S5. The amplitude and phase information are extracted from the equivalent full matrix capture response matrix K(t) by the acquisition and calculation unit, at the maximum detection depth D. max Within a defined range, construct a fully focused amplitude image A(x,z) and a phase-consistent image P(x,z), and generate a fused image F(x,z) according to an adaptive fusion criterion; S6. Dynamically select the code length and gap duty cycle of the sequence family in step S2 and the group size in step S3, wherein the dynamic selection is based on online estimated material attenuation, coupling state and near-field dead zone constraints.

2. The method according to claim 1, characterized in that, The sequence family {c_k(t)} mentioned in step S2 consists of one or more pairs of complementary sequences, such that after complementary pulse compression decoding, the sidelobe energy cancels each other out or is suppressed to below a set threshold within a preset delay window.

3. The method according to claim 1, characterized in that, The code length and gap duty cycle of the sequence family {c_k(t)} mentioned in step S2 are configured according to the detection depth partition. The detection depth is divided into at least a near field region, a mid field region, and a far field region. The near field region uses short codes or no coding, the mid field region uses medium code length, and the far field region uses long codes. The number of complementary code pairs and the gap duty cycle are configured for different depth partitions respectively.

4. The method according to claim 1, characterized in that, The grouped orthogonal spread spectrum described in step S3 satisfies the condition that the total number of array elements N is equal to the product of the number of groups G and the number of array elements M in each group, i.e., N=G×M; Hadamard spread spectrum is applied to each group and despreading is performed on the receiving side by matrix multiplication or addition and subtraction operations, so that the echo of synchronous transmission is separated into M equivalent transmission channel responses.

5. The method according to claim 1, characterized in that, The complementary pulse compression decoding described in step S4 is implemented using matched filtering or its equivalent form, and combined with a window function to suppress sidelobes.

6. The method according to claim 1, characterized in that, The equivalent full-matrix capture response matrix K(t) obtained in step S4 is consistent with the full-matrix capture matrix obtained by matrix-by-matrix excitation in terms of data structure.

7. The method according to claim 1, characterized in that, At the receiving end of the transmitter-receiver front end, the echo is quantized with low bits or edge-indexed to compress the data volume, and the instantaneous phase or equivalent symbol phase sequence is reconstructed in the acquisition and calculation unit for focusing in phase-consistent imaging and full-focus imaging under phase-consistency constraints.

8. The method according to claim 1, characterized in that, The fused image F(x,z) described in step S5 is constructed as follows or in its equivalent form: ; where α(x,z) is the fusion weighting factor, which is adaptively determined by the local coherence factor, coupling confidence, material attenuation estimation and defect type prior.

9. The method according to claim 1, characterized in that, The online estimation of material attenuation and coupling state in step S6 is obtained based on the front wall echo, back wall echo, or reference reflector echo, and the code length, gap duty cycle, and packet size are adaptively adjusted accordingly.

10. An industrial non-destructive testing system for implementing the method according to any one of claims 1 to 9, characterized in that, include: Array transducers are used to transmit and receive ultrasonic signals; A programmable multi-channel transmitter, connected to the array transducer, is used to generate coded excitation signals; A programmable receiving and quantization module, connected to the array transducer, is used to receive echoes and perform quantization; A time-space coding controller is connected to the programmable multi-channel transmitter and the programmable receiver and quantization module, respectively, and is used to generate a time-domain ternary complementary gap coding sequence and control the spatial grouping orthogonal spread spectrum of the array. The decoding and imaging processor, connected to the programmable receiving and quantization module, is used to perform spatial despreading, complementary pulse compression decoding, and the construction and fusion of full-focus imaging amplitude and phase-consistent images. The display and storage module, connected to the decoding and imaging processor, is used to display the fused image and store the detection data.

Images