Full Wavefield Data Research Articles

A probabilistic machine learning framework for stiffness tensor estimation of carbon composite laminate

The potential uses of carbon composite material are vast, particularly in the civil, mechanical and aero-structures. Nonetheless, the practical utilization of carbon composite faces challenges due to complex experimental methods and imprecise algorithms for inverting material properties. Characterization of full set of stiffness tensor for an anisotropic material in particular orthotropic material can be considered as a complex non-linear inversion problem. The inversion process requires a robust optimization algorithm and forward models. The complexity of this inversion process impedes the practical utility in large-scale automation and in real-time structural health monitoring (SHM) application. To cater this, in this work an advanced probabilistic machine learning framework based on multi-output Gaussian process regression (moGPR) is proposed as an inversion algorithm. The inversion algorithm proposed is based on probabilistic framework that utilises the dispersion of Lamb waves as an input for optimal estimation of stiffness tensor. Experimental measurements of full wavefield data of propagating waves are conducted by scanning laser Doppler vibrometer. Further, an optimal training dataset is generated by employing a forward model based on stiffness matrix method (SMM). In the presence of measurement uncertainties, the effectiveness of the optimal prediction algorithm is showcased through a comparison between the estimated parameter and its ground truth. This comparison reveals that the proposed inversion algorithm is both efficient and robust, delivering satisfactory performance even in the presence of substantial measurement uncertainties.

Full wavefield inversion with multiples: Nonlinear Bayesian inverse multiple scattering theory beyond the Born approximation

Imaging earth’s structures is fundamental to the earth’s interior, yet how to reconstruct earth’s heterogeneities using full-waveform inversion of full wavefield data that includes primary reflections and multiples remains enigmatic. I develop a nonlinear Bayesian approach for full-waveform inversion method with multiple scattering. Instead of using single scattering Born approximation to formulate the sensitivity kernels, I develop multiple scattering sensitivity kernels using multiple scattering-based Green’s functions. It is based on the Lippmann-Schwinger integral and Marchenko methods, for which the Green’s functions are retrieved from reflection data by solving a Marchenko equation. To estimate the uncertainty of velocities, I apply a Bayesian framework to the inverse problem. Our results indicate that the method does not depend on the earth’s scattering potential and the full-waveform inversion method with multiple scattering is a good alternative approach to image the earth’s structures.

The lithosphere–asthenosphere boundary structure at 11–21 Ma from wide-angle seismic data in the equatorial Atlantic Ocean

SUMMARYThe lithosphere–asthenosphere boundary (LAB) separates the rigid lithospheric plate above with the ductile and convective asthenosphere below and plays a fundamental role in plate tectonic processes. The LAB has been imaged using passive geophysical methods, but these methods only provide low-resolution images. Recently, seismic reflection imaging method has provided high-resolution images of the LAB, but imaging of the LAB at younger ages has been difficult. Here, we present the image of the LAB using wide-angle seismic reflection data covering 11–21 Ma old lithosphere in the equatorial Atlantic Ocean. Using ocean bottom seismometers (OBSs), we have observed wide-angle reflections between 150 and 400 km offsets along with crustal and mantle refraction arrivals. We first performed traveltime tomography to obtain the velocity in the crust and upper mantle. The Pn arrivals provide the information about P-wave velocity down to 4 km below the Moho. The disappearance of Pn arrivals beyond 130 km offset suggests that vertical P-wave velocity gradient is negligible or negative below this depth. We extended these velocities down to 90 km depth and then applied two imaging techniques to wide-angle reflection data, namely traveltime mapping of picked reflection arrivals and pre-stack depth migration of full wavefield data. We find that these reflections originate between 34 and 67 km depth, possibly from the LAB system. We have carried out extensive modelling to show that these reflections are real and not artefacts of imaging. Comparison of our results with coincident passive seismological and magnetotelluric results suggests that wide-angle imaging technique can be successfully used to study the lithosphere and the LAB system. We find that the LAB gradually deepens with age, but becomes very deep at 17–19 Ma, which we interpret to be due to the anomalous geology along this part of the profile.

Surface-related multiple attenuation based on a self-supervised deep neural network with local wavefield characteristics

Multiple suppression is a very important step in seismic data processing. To suppress surface-related multiples, we develop a self-supervised deep neural network method based on a local wavefield characteristic loss function (SDNN-LWCLF). The first and second input data and the output data of the self-supervised deep neural network (SDNN) are the predicted surface-related multiples, the full-wavefield data, and the estimated true surface-related multiples, respectively. The role of the SDNN is to replace the convolutional filter part of adaptive subtraction. Although there are differences in amplitudes and phases between the predicted and true surface-related multiples, the predicted surface-related multiples correspond kinematically to the true surface-related multiples and can be mapped to the estimated true surface-related multiples by the SDNN. The SDNN-LWCLF uses a local wavefield characteristic (LWC) loss function with physical properties to constrain the nonlinear optimization process. The LWC loss function is composed of the mean-absolute-error (MAE) and local normalized crosscorrelation (LNCC) loss functions. LNCC can measure the local similarity between the estimated multiples and the estimated primaries. By minimizing the LWC loss function, the MAE loss function corrects amplitudes and phases of the predicted surface-related multiples to their true values, and the LNCC loss function automatically checks and reduces the leaked multiples and residual primaries in the estimated true surface-related multiples. Our SDNN-LWCLF method does not need label data, such as true primaries and true surface-related multiples, which are usually unavailable in real-world applications. Therefore, the SDNN-LWCLF solves the problem of missing training data. Synthetic and field data examples demonstrate that our method can well suppress the surface-related multiples, and its suppression effect is better than the traditional L1-norm adaptive subtraction method and the SDNN method based on only the MAE loss function.

Pushing seismic resolution to the limit with FWI imaging

Although the resolution of a seismic image is ultimately bound by the spatial and temporal sampling of the acquired seismic data, the seismic images obtained through conventional imaging methods normally fall very short of this limit. Conventional seismic imaging methods take a piecemeal approach to imaging problems with many steps designed in preprocessing, velocity model building, migration, and postprocessing to solve one or a few specific issues at each step. The inefficacies of each step and the disconnects between them lead to various issues such as velocity errors, residual noise and multiples, illumination holes, and migration swings that prevent conventional imaging methods from obtaining a high-resolution image with good signal-to-noise (S/N) and well-focused details. In contrast, full-waveform inversion (FWI) imaging models and uses the full-wavefield data including primaries and multiples and reflection and transmission waves to iteratively invert for the velocity and reflectivity in one go. It is a systemic approach to address imaging issues. FWI imaging has proven to be a superior method over conventional imaging methods because it provides seismic images with greatly improved illumination, S/N, focusing, and resolution. We demonstrate with a towed-streamer data set and an ocean-bottom-node (OBN) data set that FWI imaging with a frequency close to the temporal resolution limit of seismic data (100 Hz or higher) can provide seismic images with unprecedented resolution from the acquired seismic data. This has been impossible to achieve with conventional imaging methods. Moreover, incorporating more accurate physics into FWI imaging (e.g., upgrading the modeling engine from acoustic to elastic) can further improve the seismic resolution substantially. Elastic FWI imaging can further reduce the mismatch between modeled and recorded data, especially around bodies of large impedance contrast such as salt. It appreciably improves the S/N and resolution of the inverted images. We show with an OBN data set in the Gulf of Mexico that elastic FWI imaging further improves the resolution of salt models and subsalt images over its acoustic counterpart.

Deep learning super-resolution for the reconstruction of full wavefield of Lamb waves

Scanning laser Doppler vibrometer is a popular tool for the acquisition of the full wavefield of propagating guided waves, in particular Lamb waves. Signal processing of such a wavefield enables us to reveal damage in the inspected structure. However, the process of acquiring the full wavefield of guided waves is time-consuming. One possible solution to tackle this problem is to acquire the Lamb waves in a low-resolution form and then apply a compressive sensing (CS) or deep learning-based super-resolution approach to that low-resolution form of full wavefield data. In this work, we applied two deep learning-based super-resolution approaches on a large synthetic low-resolution dataset of full wavefields of propagating Lamb waves in a plate of CFRP. The developed deep learning approaches are elaborated, and the results obtained from both models are compared with the conventional CS approach. Furthermore, the validation of the proposed deep learning models in a real-world scenario is performed by an experiment on a CFRP plate with embedded Teflon inserts simulating delaminations. It is concluded that the performance of both models outperforms the conventional CS approach.

Seismic multiple suppression based on a deep neural network method for marine data

Seismic multiples in marine seismic data can affect the identification of oil and gas reservoirs. The efficiency of traditional multiple suppression methods, such as the Radon transform, depends on the accuracy of the velocity model for primaries and multiples, and the assumption of random background noise. To attenuate multiples with background noise, a method of primary reconstruction using a deep neural network based on data augmentation training is proposed. The designed deep neural network (DNN) includes convolutional encoding and decoding processes. The convolutional encoding process uses convolutional layers and maximum pooling layers to learn the features of the primaries, multiples, and background noise in a seismic data set. The convolutional decoding process uses these features to reconstruct the primaries and suppress the multiples and background noise. Using the data augmentation training method in the training phase, the full-wavefield data and the predicted multiples are added to background noise and then rotated to constitute the augmented data sets. This allows the DNN to have a better multiple suppression effect and better robustness. A well-trained DNN can be used for primary reconstruction in the same work area directly or in other work areas with a transfer learning method. Three examples of synthetic data with two simple models and a Pluto model verify the effectiveness, efficiency, stability, and good generalization of the proposed method for primary reconstruction and multiple suppression. Another example from field data finds that the proposed method can efficiently suppress seismic multiples under complex conditions.

Unsupervised Learning for Seismic Internal Multiple Suppression Based on Adaptive Virtual Events

Seismic internal multiples are the key factors affecting the accuracy and reliability of velocity analysis and migration. The removal of internal multiples is a challenging direction. To effectively remove the internal multiples from the seismic data, we propose the unsupervised deep neural network (DNN) combined with the adaptive virtual events (AVEs) method. First, we use the AVE method to get the predicted internal multiples, which can calibrate the true internal multiples in the original data, also called the full wavefield data. Second, the unsupervised learning with the DNN is used as a nonlinear operator to minimize the difference between the estimated internal multiples and original data. The trained DNN can obtain the estimated internal multiples through the predicted internal multiples, thereby completing the suppression of the internal multiples. Since our proposed unsupervised learning is essentially an optimization process, it does not require true primaries as the label data to participate in the training process for the DNN. Therefore, our proposed method can deal with the problem of lack of training set and would have some good practical application value with low computational cost. The effectiveness and efficiency of our proposed method are verified through two sets of synthetic data and one land field data examples.

An Unsupervised Deep Neural Network Approach Based on Ensemble Learning to Suppress Seismic Surface-Related Multiples

Surface-related multiples are generally removed as noise. To suppress surface-related multiples, we propose an unsupervised deep neural network approach based on ensemble learning (UDNNEL). The unsupervised deep neural network (UDNN) has excellent nonlinear mapping ability, which maps the predicted surface-related multiples to true surface-related multiples, thereby completing the separation and estimation of multiples and primaries. UDNN consists of three deep neural networks (DNNs), one input data, six outputs, and six pseudo-labels (PLs). In practical use, input data are the predicted surface-related multiples, PLs consist of full-wavefield data and 0 matrices, and outputs are the desired results of estimated true surface-related multiples and differences between these desired results. Input data, one DNN, and corresponding output data are combined into a single base learner. Each base learner corrects amplitudes and phases of the predicted surface-related multiples and maps predicted multiples to true surface-related multiples under the minimization of the total loss function. The principle ensures that our UDNNEL method does not need true primaries or true multiples as training data and solves the problem of missing training datasets. Ensemble learning combines the advantages of three base learners and integrates the nonlinear optimization capabilities of three DNNs to achieve better multiple suppression effectiveness than a single base learner. Therefore, UDNNEL is better than UDNN based on a single base learner (UDNNSBL). Two synthetic data examples verify that our proposed method has good surface-related multiple suppression effectiveness. Another field data example demonstrates that our proposed method can efficiently suppress multiples under complex conditions.

3D adaptive internal multiples modeling based on globally optimised Fourier finite-difference method

Abstract Numerical methods have been widely applied to simulate seismic wave propagation. However, few studies have focused on internal multiples modeling. The formation mechanism and response of internal multiples are still unclear. Therefore, we develop a weighted-optimised-based internal multiples simulation method under 3D conditions. Using a one-way wave equation and full-wavefield method, the different-order internal multiples are computed numerically in a recursive manner. The traditional Fourier finite-difference (FFD) method has low numerical accuracy in a horizontal direction. A globally optimised FFD (OFFD) method is used to improve the lateral propagation accuracy of the seismic waves. Meanwhile, we adopt an adaptive variable-step technique to improve computational efficiency. The 3D internal multiples modeling technique is capable of calculating the different-order multiple reflections in complex structures. We use the present method to simulate internal multiples in several models. Theoretical analyses are consistent with the numerical results. Numerical examples demonstrate that the 3D internal multiples modeling technique has superior performance when adapting to lateral velocity changes and steep dip. This also implies that our method is fit for the simulation of internal multiples propagation in a 3D complex medium and can assist in identifying the internal multiples from full-wavefield data.

Accelerating Seismic Redatuming Using Tile Low-Rank Approximations on NEC SX-Aurora TSUBASA

With the aim of imaging subsurface discontinuities, seismic data recorded at the surface of the Earth must be numerically re-positioned inside the subsurface where reflections have originated, a process referred to as redatuming. The recently developed Marchenko method is able to handle full-wavefield data including multiple arrivals. A downside of this approach is that a multi-dimensional convolution operator must be repeatedly evaluated to solve an expensive inverse problem. As such an operator applies multiple dense matrix-vector multiplications (MVM), we identify and leverage the data sparsity structure for each frequency matrix and propose to accelerate the MVM step using tile low-rank (TLR) matrix approximations. We study the TLR impact on time-to-solution for the MVM using different accuracy thresholds whilst at the same time assessing the quality of the resulting subsurface seismic wavefields and show that TLR leads to a minimal degradation in terms of signal-to-noise ratio on a 3D synthetic dataset. We mitigate the load imbalance overhead and provide performance evaluation on two distributed-memory systems. Our MPI+OpenMP TLR-MVM implementation reaches up to 3X performance speedup against the dense MVM counterpart from NEC scientific library on 128 NEC SX-Aurora TSUBASA cards. Thanks to the second generation of high bandwidth memory technology, it further attains up to 67X performance speedup compared to the dense MVM from Intel MKL when running on 128 dual-socket 20-core Intel Cascade Lake nodes with DDR4 memory. This corresponds to 110 TB/s of aggregated sustained bandwidth for our TLR-MVM implementation, without suffering deterioration in the quality of the reconstructed seismic wavefields.

Elastic constants identification of fibre-reinforced composites by using guided wave dispersion curves and genetic algorithm for improved simulations

There is a great potential for model-based approaches in which guided waves are utilised for damage detection, localisation and size estimation. However, to be effective, the model must accurately represent guided wave propagation behaviour, especially wave velocity dependence on the angle of propagation. Unfortunately, improperly assumed or derived through homogenisation elastic material properties can lead to larger wave velocity changes than caused by damage. In order to account for this problem, the elastic constants are estimated based on experimental measurements of Lamb waves and optimisation techniques.In the proposed approach, experimental measurements of full wavefield data of propagating Lamb waves are conducted by scanning laser Doppler vibrometer. The obtained wavefield data is processed by using the 3D Fourier transform. Slicing the data at selected frequencies yields wavenumber profiles. These profiles can be directly translated to wave propagation velocity dependence on the angle of propagation. Data can also be sliced at selected angles giving dispersion curve images. At the same time, wavenumber profiles or dispersion curves can be calculated by using a semi-analytical model for the range of material properties. A genetic algorithm is used in which the best fit between experimental data and the semi-analytical model is optimised.Experiments were conducted on a unidirectional CFRP panel and the above-mentioned procedure was applied. Identified material properties were used in the in–house code of the time domain spectral element method for simulations of the propagating waves. Finally, numerical results were compared to the experimental full wavefield data showing much better accuracy than in case of application of homogenisation techniques.Furthermore, the correctness of elastic constants determined by the proposed method was validated against the elastic constants obtained through static test standards.

Full-waveform inversion for full-wavefield imaging: Decades in the making

Seismic imaging using full-wavefield data that includes primary reflections, transmitted waves, and their multiples has been the holy grail for generations of geophysicists. To be able to use the full-wavefield data effectively requires a forward-modeling process to generate full-wavefield data, an inversion scheme to minimize the difference between modeled and recorded data, and, more importantly, an accurate velocity model to correctly propagate and collapse energy of different wave modes. All of these elements have been embedded in the framework of full-waveform inversion (FWI) since it was proposed three decades ago. However, for a long time, the application of FWI did not find its way into the domain of full-wavefield imaging, mostly owing to the lack of data sets with good constraints to ensure the convergence of inversion, the required compute power to handle large data sets and extend the inversion frequency to the bandwidth needed for imaging, and, most significantly, stable FWI algorithms that could work with different data types in different geologic settings. Recently, with the advancement of high-performance computing and progress in FWI algorithms at tackling issues such as cycle skipping and amplitude mismatch, FWI has found success using different data types in a variety of geologic settings, providing some of the most accurate velocity models for generating significantly improved migration images. Here, we take a step further to modify the FWI workflow to output the subsurface image or reflectivity directly, potentially eliminating the need to go through the time-consuming conventional seismic imaging process that involves preprocessing, velocity model building, and migration. Compared with a conventional migration image, the reflectivity image directly output from FWI often provides additional structural information with better illumination and higher signal-to-noise ratio naturally as a result of many iterations of least-squares fitting of the full-wavefield data.

Refraction waves full waveform inversion of deep reflection seismic profiles in the central part of Lhasa Terrane

The deep reflection seismic method is an effective technique for detecting the fine structures of the lithosphere and solving deep geological problems. At present, the methods for obtaining the underground velocity from land deep reflection seismic profile mainly rely on the travel-time information based velocity analysis and tomography, which theoretically limit the resolution of the imaging results. To achieve the goal of using deep reflection seismic profile for crustal-scale high-precision velocity imaging, we must finally overcome the problem of full waveform inversion (FWI) using full wavefield data. In this paper, we successfully conduct FWI using the early-arrival wavefields (mainly primary and multiple refraction wavefields) in the deep reflection seismic profile, which can be seen as the first important step towards the goal of crustal-scale high-precision velocity imaging. We propose a robust refraction waves FWI workflow with global normalized cross-correlation, multiple refraction matching, reverse time source estimation and wavenumber domain gradient filtering as its core algorithm. The application of the proposed FWI method to the refraction waves of the deep reflection seismic profiles in the central part of Lhasa Terrane shows that the refraction FWI images have a higher resolution to underground structures both in horizontal and vertical directions than the first-arrival tomography images. The lateral velocity change can be well-matched with the typical geological characteristics of the work area. The FWI result can better describe the lateral boundaries of the strata in different eras. At a depth of less than 3 km from the surface, the refraction FWI can correct the interface morphology of the tomography results and recover some small-scale velocity structures and high-velocity anomalies.

Sparse Reconstruction and Damage Imaging Method Based on Uniform Sparse Sampling

The full wavefield detection method based on guided waves can efficiently detect and locate damages relying on the collection of large amounts of wavefield data. The acquisition process by scanning laser Doppler vibrometer (SLDV) is generally time-consuming, which is limited by Nyquist sampling theorem. To reduce the acquisition time, full wavefield data can be reconstructed from a small number of random sampling point signals combining with compressed sensing. However, the random sampling point signals need to be obtained by adding additional components to the SLDV system or offline processing. Because the random sparse sampling is difficult to achieve via the SLDV system, a new uniform sparse sampling strategy is proposed in this paper. By using the uniform sparse sampling coordinates instead of the random spatial sampling point coordinates, sparse sampling can be applied to SLDV without adding additional components or offline processing. The simulation and experimental results show that the proposed strategy can reduce the measurement locations required for accurate signal recovery to less than 90% of the Nyquist sampling grid, and the damage location error is within the minimum half wavelength. Compared with the conventional jittered sampling strategy, the proposed sampling strategy can directly reduce the sampling time of the SLDV system by more than 90% without adding additional components and achieve the same accuracy of guided wavefield reconstruction and damage location as the jittered sampling strategy. The research results can greatly improve the efficiency of damage detection technology based on wavefield analysis.

Parallel spectral element method for guided wave based structural health monitoring

Parallel implementation of the spectral element method is developed in which flat shell spectral elements are utilized for spatial domain representation. The implementation is realised by using Matlab Parallel Computing Toolbox and optimized for Graphics Processing Unit (GPU) computation. In this way, considerable computation speed-up can be achieved in comparison to computation on conventional processors. The implementation includes an interpolation of wavefield on a uniform grid. The method was tested on experimental data set available on the Open Guided Waves platform. The qualitative comparison was performed on full wavefield data, whereas quantitative comparison was made directly on signals of propagating Lamb waves registered by piezoelectric transducers. In both cases, good agreement between numerical and experimental results was achieved. The proposed method is particularly useful for structural health monitoring algorithms in which signal parameters are required such as wave velocity dependence on the angle of propagation. Moreover, it enables model–assisted damage size estimation in which the damage influence curve is estimated based on large numerical data sets and sparse experimental data. Due to relatively short computation times, large data sets can be generated and used for machine learning or other soft computing methods opening up new possibilities in health monitoring of metallic and composite structures.

Elastic constants identification of woven fabric reinforced composites by using guided wave dispersion curves and genetic algorithm

Typically, material properties are estimated by destructive tests and used in computational models in the design and analysis of structures. This approach is well-established in relation to isotropic homogenous materials. However, if this approach is used for composite laminates, inaccuracies can arise that lead to vastly different stress distributions, strain rates, natural frequencies, and velocities of propagating elastic waves. In order to account for this problem, the alternative method is proposed, which utilise Lamb wave propagation phenomenon and optimisation technique. Propagating Lamb waves are highly sensitive to changes in material parameters and are often used for structural health monitoring of structures. In the proposed approach, the elastic constants, which are utilised to determine dispersion curves of Lamb waves, are optimised to achieve a good correlation between model predictions and experimental observations. The dispersion curves of Lamb waves were calculated by using the semi-analytical spectral element method. The resulting dispersion curves were compared with experimental measurements of full wavefield data conducted by scanning laser Doppler vibrometer and processed by 3D Fourier transform. Next, elastic constants were determined by using a genetic algorithm which resulted in a good correlation between numerical and experimental dispersion curves.

Statistical evaluation of damage size based on amplitude mapping of damage-induced ultrasonic wavefield

The inability to statistically evaluate the damage size based on amplitude mapping has long plagued the ultrasound propagation imaging (UPI) system and related non-destructive evaluation (NDE) community. This paper proposes a damage visualization method called the Statistically Thresholded Anomaly Mapping (STAM) method to solve this problem. It could isolate the damage-induced anomalous waves from an ultrasonic full wavefield data, map their amplitudes into an image and statistically differentiate damage from the background noise and pristine region of a specimen, without prior knowledge or reference data of the specimen being inspected. The proposed method was demonstrated through the visualization and evaluation of a thermal damage inflicted in a glass fibre reinforced composite specimen. The effects of two parameters on result quality and accuracy of damage size evaluation were studied by varying their values independently. The optimized result showed that 99.9% of the background noises could be removed while maintaining clear visualization of damage, hence allowing the users to evaluate the presence, location, shape, size and severity of the damage accurately.

Baseline-free guided wave damage detection with surrogate data and dictionary learning.

In guided wave structural health monitoring, damage detection is often accomplished by comparing measurements before damage (i.e., baseline data) and after damage (i.e., test data). Yet, in practical scenarios, baseline data is often unavailable. Data from surrogate structures (structures similar to the test structure) could replace baseline data, but due to small differences in material properties, such as thickness, temperature, and other effects, this data is often unreliable. In this paper, a dictionary learning framework overcomes this challenge and detects damage with surrogate information. The framework combines wave propagation characteristics of a test structure with geometric information from surrogate structures to create a synthetic damage-free baseline. The test data is compared with the synthetic baseline to detect damage. The framework is evaluated with four 108 mm ×108 mm plates: two 1.6-mm thick aluminum plates, one 1.6-mm thick steel plate, and one 6.25 mm thick aluminum plate. The framework is applied to each test structure after learning from plates with different material properties and thicknesses. With full wavefield data, this paper successfully isolates reflections from a mass without using explicit baseline data. Furthermore, with sparse guided wave data, this paper shows that a drop in a correlation coefficient can detect a mass without using explicit baseline data.

Lamb Wave Phase Velocity Imaging of Concrete Plates with 2D Arrays

In the nondestructive evaluation of concrete structures, ultrasonic techniques are considered to be more capable than low-frequency techniques such as the impact-echo method. This is especially true with the recent development of ultrasonic transducers, synthetic apertures, and results in an image form, and because low-frequency techniques are usually limited in their evaluation to the frequency of one single resonant mode. With the aim of reducing this gap in capabilities, we present a 2D array and wide-frequency bandwidth technique for Lamb wave phase velocity imaging. The presentation involves a measurement on a newly cast concrete plate using a hammer and an accelerometer as an example. The key concept of the technique is the use of 2D arrays that record a full wave field response over a limited surface subdomain within the complete measurement domain. Through a discrete Fourier transform, a spectral estimate is obtained for the 2D array in the frequency-phase velocity domain. The variation of the phase velocity is then mapped using a stepwise movement of the 2D array within the complete measurement domain. With two different types of 2D arrays, the variation of the phase velocity for the A0 Lamb mode is mapped and displayed in a polar and image plot, and low variation is observed for both cases. This result verifies the expected condition of a homogenous material and plate thickness and, more importantly, highlights the potential of wide-frequency bandwidth techniques based on full wave field data.