Lamb Wave Signals Research Articles

Bayesian compressive sensing for recovering the time-frequency representation of undersampled Lamb wave signals

The characteristic extraction of ultrasonic Lamb wave is the prerequisite for its efficient utilization in the structural health monitoring. In the situation of intentional signal compression or unexpected data missing, the accurate recovery of contained information is challenging. To address this problem, this work proposes the time-frequency representation (TFR) reconstruction scheme for undersampled Lamb wave signal. Unlike the conventional method, both the sparse prior and structural sparse prior in the two-dimensional plane are considered in the design of Bayesian compressive sensing. The simulated signal is adopted to validate the effectiveness of the proposed method. Furthermore, different ratios of available samples are investigated to analyze the recovery ratio of TFR. Even if the available samples are smaller than those from the Nyquist rate, the TFR recovery ratio can reach 70%. The experiments using array transducers in noisy environments are also conducted. The time-of-flight information extracted from the recovered TFR is accurate and the relative error is smaller than 3%. Besides, the comparisons with conventional schemes for compressive sensing are carried out to demonstrate its superiority.

Lamb wave damage severity estimation using ensemble-based machine learning method with separate model network

Lamb wave-based damage estimation have great potential for structural health monitoring. However, designing a generalizable model that predicts accurate and reliable damage quantification result is still a practice challenge due to complex behavior of waves with different damage severities. In the recent years, machine learning (ML) algorithms have been proven to be an efficient tool to analyze damage-modulated Lamb wave signals. In this study, ensemble-based ML algorithms are employed to develop a generalizable crack quantification model for thin metallic plates. For this, the scattering of Lamb wave signals due to different configuration of crack dimension and orientation is extensively studied. Various finite element simulations signals, representing distinct crack severities in term of crack length, penetration and orientation are acquired. Realizing that both temporal and spectral information of signal is extremely important to damage quantification, three time-frequency (TF) based damage sensitive indices namely energy concentration, TF flux and coefficient of energy variance are proposed. These damage features are extracted by employing smoothed-pseudo Wigner–Ville distribution. After that data augmentation technique based on the spline-based interpolation is applied to enhance the size of the dataset. Eventually, these fully developed damage dataset is deployed to train ensemble-based models. Here we propose separate model network, in which different models are trained and then link together to predict new and unseen datasets. The performance of the proposed model is demonstrated by two cases: first simulated data incorporated with high artificial noises are employed to test the model and in the second scenario, experimental data in raw form are used. Results indicate that the proposed model has the potential to develop a general model that yields reliable answer for crack quantification.

Interpretable convolutional sparse coding method of Lamb waves for damage identification and localization

Lamb wave-based damage identification and localization methods hold the potential for nondestructive evaluation and structural health monitoring. Dispersive and multimodal characteristics lead to complicated Lamb wave signals that are challenging to be analyzed. Deep learning architectures could identify damage-related features effectively. Convolutional neural network (CNN) is a promising architecture that has been widely applied in recent years. However, this data-driven approach still lacks a certain degree of physical interpretability and requires a large number of parameters. In this article, an interpretable Lamb wave convolutional sparse coding (LW-CSC) method is proposed for structural damage identification and localization. First, toneburst signals at different center frequencies are considered in the first convolutional layer. The network convolves the waveforms with a set of parametrized functions that implement band-pass filters, which performs more physical interpretability compared to conventional CNN model. Subsequently, the damage features are extracted according to the multi-layer iterative soft thresholding algorithm for multi-layer CSC model, which could realize a deeper network without adding parameters unlike CNN. Finally, Lamb wave-based damage localization is visualized using an imaging algorithm. The experimental results demonstrate that the proposed method not only enables improvement of the classification accuracy but also achieves structural damage localization accurately.

Damage characterization using CNN and SAE of broadband Lamb waves

The method based on Lamb wave shows great potential for structural health monitoring (SHM) and nondestructive testing (NDT). Deep learning algorithms including convolutional neural networks (CNN) and stack autoencoder (SAE) are promising to extract features from Lamb wave signals that can be linked with damage for subsequent localization and quantification. Generally, narrowband Lamb wave with purified mode and suppressed dispersion is used because of clear relationship model between damage features and recorded signals. However, model performance is limited because contained damage information of narrowband Lamb wave is inadequate. To overcome this limitation, a broadband Lamb wave deep learning algorithm is proposed for damage localization and quantification. Compared with narrowband, broadband Lamb wave generated at a large frequency range contains richer information of structural damage. In this study, different mode selections, different signal processing methods and different deep learning algorithms are applied to extract damage features from different perspectives, and fusion of all extraction results facilitates the full utilization of rich broadband information. An experiment is given to demonstrate the effectiveness and high-accuracy of proposed method.

Prognosis of fatigue cracks in an aircraft wing using an adaptive tunable network and guided wave based structural health monitoring

Due to the stochastic behaviour of fatigue crack propagation resulting from the multifarious uncertainties such as aleatory uncertainty and epistemic uncertainty, a synthetic scheme integrating heterogeneous uncertainty resources is needed to make the prognosis of fatigue crack propagation be accurate. In this paper, an on-line updating strategy combining an adaptively tunable hybrid radial basis function (RBF) network and active Lamb waves based structural health monitoring is proposed for prognosis of fatigue cracks in an aircraft wing. A particle filter and Brownian motion with drift and proportionality coefficients are put forward to dynamically access a posteriori estimation of crack state and model parameters. Because of the advantages of flexibility, electrically stabilized and suitable for complex structure, the sensor layer with an embedded network of distributed piezoelectric transducers is utilized to collect Lamb wave signals describing crack states on-line. A hybrid spatial phase difference-based damage index (DI) is employed to capture the characteristics of Lamb wave signals to quantitatively determine the crack length. Furthermore, fatigue tests on the outboard wing of a real airplane are conducted to substantiate the proposed method. The experimental results demonstrate that the hybrid RBF model can timely trap the dynamic characteristics of sensor signals, and effectively predict the varying tendency of cracks growth. The uncertainties during the prognosis of fatigue cracking propagation can be effectively dealt with by combining the dynamic adjustability and the DIs-based on-line monitoring technique.

An advanced Wigner–Ville time–frequency analysis of Lamb wave signals based upon an autoregressive model for efficient damage inspection

It is quite straightforward to generate and acquire ultrasonic guided waves in metallic or composite structures to investigate structural defects; however, the interpretation and evaluation of the reflected/transmitted signal to extract useful information is a challenging task. This is primarily due to the dispersion and multimodal behavior of Lamb waves, which are dependent on the frequency of the excitation wave and the thickness of the material under investigation. These multimodal and dispersive behaviours lead to a complex waveform structure, and therefore require an advanced signal processing technique to decipher useful information in the time and/or frequency domains. For this purpose, due to its desirable mathematical properties, the Wigner–Ville distribution (WVD) is considered to be a powerful tool for generating time–frequency (TF) spectra and estimating the temporal and spectral features of this type of complex signal. However, due to its quadratic nature, undesirable cross-terms and spurious energies are also generated, which limit the readability and the interpretation of the spectrum. To suppress this effect, this paper describes the use of an autoregressive model based upon Burg’s maximum entropy method to modify the kernel of the discrete WVD. This technique was applied to ultrasonic Lamb wave signals obtained using various numerical and experimental configurations to extract the useful discriminatory spectral and temporal information required for the mode identification, localization, and quantification of structural damage. When used for damage localization, the proposed method precisely estimated the distance between two closely spaced notches in a metallic plate from different simulated noisy signals with a maximum uncertainty of 5%, based upon an excellent TF energy distribution. Moreover, the energy concentration (EC) of the TF energy distribution, in combination with the variation of its instantaneous frequency (IF) curve, was also effective in identifying the overlapping modes of the Lamb wave signal. Finally, for damage quantification, three TF-based damage indices, namely EC, TF flux, and IF, were extracted from five sets of specimens using the proposed TF scheme and used to train the regression model. Model testing proved that the damage indices have the potential to predict crack sizes precisely and reliably.

Causal dilated convolutional neural networks for automatic inspection of ultrasonic signals in non-destructive evaluation and structural health monitoring

This paper presents a deep learning network that performs automatic detection of defects by inspecting full ultrasonic guided wave signals excited in plate structures. The findings show that the algorithm, which is an adaptation of WaveNet, and hence is based on causal dilated convolutional neural networks, is effectively able to learn features and/or patterns related to the presence of waves scattered from damage, thus eliminating the need for any feature engineering to be performed by human operators. The network outperformed the widely used conventional approach that combines the optimal baseline selection and baseline signal stretch compensation methods when tested on two different datasets. The first dataset consisted of finite element simulated Lamb wave signals acquired in a pitch-catch configuration on a steel plate across a 50 °C range of temperature variations, and the second was a publicly available experimental dataset of Lamb wave signals also acquired in pitch-catch mode on a composite plate with a 40 °C range of variations. The improvements over the conventional approach are particularly encouraging when analyzing signals at temperatures well outside the temperature range available in the set of baseline signals, hence suggesting that this class of algorithms can complement or substitute existing methods, especially when testing occurs at unseen environmental and operational conditions, or when the effects of sensor drift make conventional methods less effective.

Lamb wave based damage detection in metallic plates using multi-headed 1-dimensional convolutional neural network

Lamb wave based damage diagnosis holds potential for real-time structural health monitoring; however, analysing the Lamb wave response possess challenge due to its complex physics. Data-driven machine learning (ML) algorithms are often more effective in identifying the damage-related features from these complex responses. However, in analysing such complex responses the ML algorithms requires extensive data pre-processing and are often not suitable for real-time damage detection. This paper presents a deep learning multi-headed 1-dimensional convolutional neural network (1D-CNN) architecture capable to operate directly on raw discrete time-domain Lamb wave signals recorded from a thin metallic plate. The multi-headed configuration consisting of two parallel 1D-CNN layers is capable to learn higher order damage-related features and enhances robustness of overall classification performance. To train the adopted 1D-CNN algorithm a diverse database is also constructed consisting 216 numerically and 24 experimentally generated responses of a thin 1.6 mm Al-5052 plate structure. The diversification of training database is achieved by varying parameters like scanning length, scanning frequency and adding different levels of white noises to the captured responses. Later, the trained 1D-CNN architecture is tested against two separated unseen test-databases. The first test database consist of experimentally generated 12 samples with notch-like damage and 12 samples of pristine condition. The proposed 1D-CNN classifier generalizes well on the unseen samples and decisively predicts the outcome for 23 out of 24 samples of first test database. The second test database consists of 108 unseen FE simulated samples capturing additional damage scenarios. In the second test phase, the model has correctly predicted the condition of all the 108 samples.

Damage Classification of Composites Based on Analysis of Lamb Wave Signals Using Machine Learning

Abstract Composite materials have a myriad of applications in complex engineering systems, and multiple structural health monitoring (SHM) strategies have been developed. However, these methods are challenging due to signal attenuation and excessive noise interference in composite materials. Signal processing can capture a small difference between the input–output signals associated with the severity of the damage in composites. Thus, the research question is “can signal processing techniques reduce the required number of features and assess the randomness of fatigue damage classification in composite materials using machine learning (ML) algorithms?” To answer this question, piezo-electric signals for carbon fiber reinforced polymer (CFRP) test specimens were taken from NASA Ames prognostics data repository. A framework based on a comparative analysis of signals was developed. For the first specific aim, the effectiveness of features based on statistical condition indicators of the sensor signals were evaluated. For the second specific aim, actuator-sensor signal pair were analyzed using cross-correlation to extract two features. These features were used to train and test four supervised ML algorithms for damage classification and their performance was discussed. For the third specific aim, randomness in the dataset of fatigue damage of the specimens was assessed. Results showed that by signal processing, the requirement of features for training ML was reduced with the improvement in the performance of ML. The randomness was captured by the utilization of two specimens from the same material. This work contributes to the improvement of intelligent damage classification of composite materials, operating under complex working conditions.

A novel time–frequency transform for broadband Lamb waves dispersion characteristics analysis

Owing to carrying rich information about structure flaws, broadband Lamb waves are considered as a promising tool for non-destructive testing. However, since every Lamb wave mode has its own dispersion characteristics, the feature extraction among broadband multimodal Lamb wave is challenging. Time–frequency representation is significantly effective to analyze dispersive signals. In this article, taking advantages of the idea of dispersion compensation, two kinds of time–frequency domain dispersion analysis methods for broadband Lamb wave were proposed. The first one is based on the concept of the general parameterized time–frequency transform. A kernel function related to group delay was designed and the time–frequency compensation transform was proposed. The other one combines the segment linear mapping technique and the short-frequency Fourier transform, called the time–frequency de-dispersion transform. Both these two methods work well in representing multimodal Lamb wave signals with high resolution. However, time–frequency de-dispersion transform outperforms in representing multipath Lamb waves than time–frequency compensation transform. Moreover, a mode purification strategy was also proposed for distinguishing the interested mode from interferences. According to verification in synthetic and experimental data, not only the multimodal components but also multipath echoes are represented in time–frequency plane with high resolution. Finally, the proposed method shows a great robustness to inaccuracies in the dispersion data.

A Damage Detection Method Using Neural Network Optimized by Multiple Particle Collision Algorithm

A critical task of structural health monitoring is damage detection and localization. Lamb wave propagation methods have been successfully applied for damage identification in plate‐like structures. However, Lamb wave processing is still a challenging task due to its multimodal and dispersive characteristics. To address this issue, data‐driven machine learning approaches as artificial neural network (ANN) have been proposed. However, the effectiveness of ANN can be improved based on its architecture and the learning strategy employed to train it. The present paper proposes a Multiple Particle Collision Algorithm (MPCA) to design an optimum ANN architecture to detect and locate damages in plate‐like structures. For the first time in the literature, the MPCA is applied to find damages in plate‐like structures. The present work uses one piezoelectric transducer to generate Lamb wave signals on an aluminum plate structure and a linear array of four transducers to capture the scattered signals. The continuous wavelet transform (CWT) processes the captured signals to estimate the time‐of‐flight (ToF) that is the ANN inputs. The ANN output is the damage spatial coordinates. In addition to MPCA optimization, this paper uses a quantitative entropy‐based criterion to find the best mother wavelet and the scale values. The presented experimental results show that MPCA is capable of finding a simple ANN architecture with good generalization performance in the proposed damage localization application. The proposed method is compared with the 1‐dimensional convolutional neural network (1D‐CNN). A discussion about the advantages and limitations of the proposed method is presented.

The Effect of Piezoelectric Fiber Rosette Configurations on Lamb Wave Direction Detection for Damage Localization

Piezoelectric fiber rosettes respond to the directivity characteristics of Lamb waves, and therefore, are useful in detecting the Lamb wave propagation direction. Considering material damage as a secondary wave source, two piezoelectric fiber rosettes are arranged to measure the scattered wave propagation directions for damage localization. The influences of various rosette configurations, i.e., 45°‐rectangular, 135°‐rectangular, 60°‐delta, and 120°‐delta, on the estimation accuracy of the propagation direction are investigated in this paper. The response of the piezoelectric fiber to the A0 mode Lamb wave under narrowband tone‐burst excitation is theoretically derived. Experimental tests and piezoelectric coupling simulations are performed to obtain the Lamb wave signal of each fiber. The matching pursuit (MP) algorithm is applied to extract the weak damage‐related wave packet by using Hann‐windowed narrowband excitation as an atom. The Lamb wave propagation directions are estimated based on the error function. The accuracies of the directions with 4 types of rosette configurations are compared, and their error sources are discussed. The results show that the accuracy of the 135°‐rectangular configuration is relatively satisfactory, and the errors depend on the size and location of each fiber in the rosette. The proposed damage localization method is validated by experimental tests. The predicted locations are close to the actual damage location. The research results are significant for piezoelectric fiber rosette design and optimization and damage location without wave speed or time‐of‐flight information in complex or irregular structures.

A technique based on nonlinear Hanning-windowed chirplet model and genetic algorithm for parameter estimation of Lamb wave signals

Parameter estimation techniques based on chirplet models and intelligent algorithms can realize the simultaneous multi-information extraction of signals. They have attracted considerable attention for the processing of Lamb wave signals to detect defects and evaluate the material properties. Influenced by their dispersive nature, Lamb wave signals possess nonlinear instantaneous frequencies and asymmetric envelopes. However, the classical chirplet models are established with either Gaussian windows or linear chirps. They are inadequate for characterizing the dispersion features of ultrasonic signals whose excitations are modulated by Hanning windows. In our previous work, a nonlinear Hanning-windowed chirplet (NHWC) model with nine parameters was proposed to realize the full characterization of waveforms. However, the large number of parameters limits its application. A simple NHWC model with seven parameters was designed by submitting the same nonlinear phase modulation term into the Hanning-windowed sine function in this paper. Furthermore, a real-coded multi-objective genetic algorithm was developed to realize the parameter estimation of signals by combining a clustering algorithm and the NHWC model. Different strategies were adopted to ensure the convergence of the algorithm. The maximum extreme values were adopted to realize adaptive discretization of the search space and the updating of parameters. The parameters in the NHWC models were divided into implicit and explicit parts, and different strategies were applied to update them. The clustering algorithm and a sorting combination method were employed to generate a Pareto set. Experimental results showed that the parameter estimation with the simplified NHWC model exhibited a more robust performance than that of the model that contained nine parameters for characterizing the Lamb wave signals with or without the dispersion features. The arrival time, amplitude, and instantaneous frequencies of wave packets were identified with the parameter estimation technique.

Damage Detection Method Based on Continuous Wavelet Transformation of Lamb Wave Signals

A damage estimation method based on continuous wavelet transformation (CWT) of the normalized Lamb wave signals is proposed here. Lamb waves are actuated and sensed using piezoelectric (lead zirconate titanate, PZT) transducers arranged in the form of square detection cells on a plate-like structure. Excitation sequences based on pitch–catch and pulse–echo configurations are tested for the same arrangement of the transducers. The possibilities of the existence of damage for each actuator–senor pair are formed by using the normalized coefficient of CWT. The size of the possible damage region is directly controlled through envelopes defined by the coefficients of CWT, and no additional parameter is required to define its size. The aggregate damage image is constructed by the fusion of damage possibilities from all actuator–sensor pairs using damage indices based on conjunctive and compromised fusion schemes. The results indicate that the proposed method can estimate the location and severity of multiple damage with signals directly from the damaged plate, without the need of baseline signals from the undamaged plate, and the time-compensated signals provide better damage imaging than the raw signals. The most accurate and computationally inexpensive combination is the pulse–echo configuration with damage index based on conjunctive image fusion scheme. The method is computationally inexpensive and can be applied for multiple damage estimations in large structures to reduce the evaluation cost and inspection time during on-line structural health monitoring.

Unsupervised damage clustering in complex aeronautical composite structures monitored by Lamb waves: An inductive approach

Structural Health Monitoring (SHM), i.e. the action of monitoring structures in real-time and in an automated manner, is a major challenge in several industrial fields such as aeronautic. SHM is by nature a very high dimensional data-driven problem that possesses several specificities when addressed as a machine learning problem. First of all data in damaged cases are rare and very costly as the generation of damaged data is not always possible and simulations are not reliable especially when dealing with complex structures. SHM is thus by nature an unsupervised problem. Furthermore, any incoming sample should be instantaneously clustered and handcrafted damage indexes are commonly used as a first dimension reduction step due to large datasets to be processed. As a consequence, unsupervised dimensionality reduction (DR) techniques that project very high dimensional data into a two or three-dimensional space (such as t-SNE or UMAP) are very appealing in such a context. However, these methods suffer from one major drawback which is that they are unable to cluster any unknown incoming sample. To solve this we propose to add inductive abilities to these well know methods by associating their projection bases with Deep Neural Networks (DNNs). The resulting DNNs are then able to cluster any incoming unknown samples. Based on those tools, a SHM methodology allowing for unsupervised damage clustering with dimensionality reduction is presented here. To demonstrate the effectiveness of the method, results of damage classification on large experimental data sets coming from complex aeronautical composite structures monitored through Lamb waves are shown. Furthermore, several DR techniques have been benchmarked and recommendations are derived. It is demonstrated that the use of raw Lamb wave signals instead of the associated damage indexes is more effective. This non-intuitive result helps to reduce the gap between laboratory research and the actual start-up of SHM activities in industrial applications.

Delamination area quantification in composite structures using Gaussian process regression and auto-regressive models

After detecting initial delamination damage in a hotspot region of a composite structure monitored through a data-driven approach, the user needs to decide if there is an imminent structural failure or if the system can be kept in operation under monitoring to track the damage progression and its impact on the structural safety condition. Therefore, this study proposes delamination area quantification by stochastically interpolating global damage indices based on Gaussian process regression and taking into account uncertainty. Auto-regressive models are applied to extract damage-sensitive features from Lamb wave signals, and the Mahalanobis squared distance is used to compute damage indices. Two sets of laboratory tests are used to demonstrate the effectiveness of this methodology—one in carbon–epoxy laminate with simulated damage under temperature changes to show the general steps of the procedure, and a second test involving a set of carbon fiber–reinforced polymer coupons with actual delamination caused by repeated fatigue loads. Various levels of progression damage, measured by the covered area of delamination, are monitored using piezoelectric lead zirconate titanate patches bonded to the structural surfaces of these setups. The Gaussian process regression proved to be capable of accommodating the uncertainties to relate the damage indices versus the damaged area. The results exhibit a smooth and adequate prediction of the damaged area for both simulated damage and actual delamination.

Frequency Aliasing-Based Spatial-Wavenumber Filter for Online Damage Monitoring

The spatial-wavenumber filter method can extract the specific mode of the Lamb wave, thereby distinguishing the incident wave and the damage reflection wave. This method has been widely studied for damage imaging. However, the diameter of piezoelectric transducer (PZT) sensor limits the spatial sampling wavenumber of the linear PZT sensor array, which limits the application of this method because of the Nyquist–Shannon sampling theorem. Therefore, the wavenumber filtering range of spatial-wavenumber filter should be less than half of the spatial sampling wavenumber. In this paper, a frequency aliasing based spatial-wavenumber filter for online damage monitoring is proposed. In this method, the wavenumber filtering range is extended to the spatial sampling wavenumber, and two wavenumber results will be calculated as for the frequency aliasing. Subsequently, the wavenumber of the received Lamb wave signal can be obtained according to the average arrival time difference between the two adjacent sensors in the linear PZT sensor array. Finally, the damage is localized using the spatial-wavenumber filter and cruciform PZT sensor array. This method was validated on an epoxy laminate plate. The maximum damage localization errors are less than 2 cm. It is indicated that this method can extend the spatial-wavenumber filtering range to the spatial sampling wavenumber and the application of spatial-wavenumber filter-based online damage monitoring.

Lamb Wave-Based Damage Localization Feature Enhancement and Extraction Method for Stator Insulation of Large Generators Using VMD and Wavelet Transform.

Lamb waves are used to locate any damage in the stator insulation structure of large generators. However, it is difficult to extract the features of Lamb wave signals in a strong background noise environment, thus significantly reducing the accuracy with which the damage is located. This paper proposes a method based on variational mode decomposition (VMD) and wavelet transform to enhance and extract the location features of stator insulation damage signals of large motors. First, considering that the characteristics of VMD are sensitive to noise, the Lamb wave detection signal is decomposed, denoised, and reconstructed; the reconstructed signal is then wavelet-transformed to extract the time of flight (TOF) of the damage-scattered wave as the damage location feature; finally, the damage location is determined using the TOF features. The proposed method is experimentally tested and verified under various noise environments. The results show that the VMD and wavelet transform methods can significantly improve the signal-to-noise ratio of Lamb wave detection signals and the accuracy with which the damage is located under strong background noise. This study extends the applicability of Lamb wave-based non-destructive detection of stator insulation damage in complex environments.

Efficient Lamb-wave based damage imaging using multiple sparse Bayesian learning in composite laminates

Lamb wave techniques have been widely used for structural health monitoring (SHM) and nondestructive testing (NDT). To deal with dispersive and multimodal problems of Lamb wave signals, many signal processing methods have been developed. A spatially distributed array of piezoelectric transducers is generally adopted for both transmission and reception of Lamb waves. When imaging the damage in composite laminates, it is necessary to meet the need of processing array signals with high efficiency. In this paper, the multiple sparse Bayesian learning (M-SBL) strategy is employed for damage imaging. Multiple residual signals including damage-reflection waves are decomposed into a sparse matrix of location-based components simultaneously. An appropriate dictionary is designed to match the damage-reflection waves instead of interference waves. The key to success is to obtain the sparse matrix of weighting coefficients through the M-SBL algorithm. Damage imaging can be achieved efficiently using the delay-and-sum (DAS) method with sparse coefficients in time-domain. Results from the experiment in composite laminates demonstrate the effectiveness of the proposed method.

A Damage Localization Method With Multimodal Lamb Wave Based on Adaptive Polynomial Chirplet Transform

Ultrasonic Lamb wave testing allows health monitoring for plate structures and has been extensively applied in the modern industry. To obtain the location and further form the damage imaging, the time of flight (TOF) is a pivotal parameter. However, the dispersive and multimodal nature of the Lamb wave obstructs the accurate measurement of TOF. The objective of this article is to develop efficient techniques to separate multimodal Lamb wave signals and accomplish accurate damage localization. The narrowband Lamb wave is generated by the excitation signal with a center frequency of 150 kHz. The fundamental symmetric and antisymmetric modes are exploited. The adaptive polynomial chirplet transform (CT) is proposed to decompose the Lamb wave signals and identify the modes. Then, the instantaneous frequency characteristics are utilized to extract time information. Thus, the TOF is obtained and damage location can be calculated. The reconstruction error is defined to quantitate the difference between the reconstructed signal and the received signal. The obtained reconstruction errors from the simulation and experiment are 0.1631 and 0.2875, respectively. The varied defect location is adopted to obtain the TOF estimation. The results show that the proposed method outperforms the CT and cross correlation, especially for the longer propagation distances where the dispersion effect is distinct. For experiments considering the defect and boundary reflection, the localization errors of the collinear defect are 2.51% and 3.04% for the two Lamb wave modes. In comparison with smoothed-pseudo-Wigner–Ville distribution and CT, the proposed method owns superior performance on mode identification and location accuracy.