A comparison of the pulsed, lock-in and frequency modulated thermography nondestructive evaluation techniques

doi:10.1016/j.ndteint.2011.06.008

NDT & E International

Volume 44, Issue 7, November 2011, Pages 655-667

https://doi.org/10.1016/j.ndteint.2011.06.008 Get rights and content

Abstract

Pulsed, lock-in and frequency modulated thermography are three alternative nondestructive evaluation techniques. The defect imaging performance of these techniques are compared using: matched excitation energy; the same carbon fiber composite test piece and infrared camera system. The lock-in technique suffers from “blind frequencies” at which phase images for some defects disappear. It is shown that this problem can be overcome by using frequency modulated (chirp) excitation and an image fusion algorithm is presented that enhance phase imaging of defects. The signal-to-noise ratios (SNRs) of defect images obtained by the three techniques are presented. For the shallowest defects (depths 0.25 and 0.5 mm, 6 mm diameter), the pulsed technique exhibits the highest SNRs. For deeper defects the SNRs of the three techniques are similar in magnitude under matched excitation energy condition.

Highlights

► Three thermographic techniques are compared under matched excitation energy. ► For shallow defects, pulsed technique produces the best result, based on SNR. ► For deeper defects, lock-in and FMTWI performances are comparable to that of pulsed. ► Blind frequency effect in lock-in thermography is overcome using FMTWI. ► Image fusion algorithm of multiple phase images is proposed and implemented.

Introduction

Pulsed [1], [2] and lock-in [3] thermography are the most commonly used thermographic nondestructive evaluation techniques. The two techniques are distinctly different but are deployed in the inspection of similar components/structures. In general these techniques are suitable for the detection of shallow planer defects, e.g. delamination in composites or adhesion defect in surface coatings.

In pulsed thermography, the sample surface is instantaneously heated using an optical flash. Over time the surface heat penetrates into the material producing asymptotic cooling. If sub-surface faults are present, heat-flow is obstructed. These result in relatively slower cooling of the defective regions than their sound counterparts. Consequently relative hot spots appear on the surface above the defects. The hot-spots thus obtained in pulsed thermography eventually fade away to achieve thermal equilibrium.

By contrast, in lock-in thermography the sample surface is heated by periodically modulated lamps and thermograms are captured under the periodic sinusoidal heating. The analysis of these thermograms mainly considers the phase shift of the thermal response of the defective regions with respect to the sound regions of the material [3]. Fourier transformation is extensively used for this purpose. The phase and magnitude information, thus obtained for each of the pixels in the thermogram, are stored in the form of 2D-matrices and subsequently converted into a gray scale images known as the phase image and magnitude image. Since in this technique, the images are extracted at the same frequency as used in the excitation, it is called lock-in thermography with an analogy to the working principle of lock-in amplifier. The excitation frequency is chosen based on the diffusion length of the thermal signal. The thermal diffusion length is the distance over which the signal strength falls to 1/e of that at the surface. Since the material acts as a low-pass filter, lower frequency signals propagate further into the material. However, if the defect lies beyond the diffusion length, it is not likely to be seen.

Since, pulsed and lock-in thermography are intrinsically different, their comparison is problematic. An objective comparison between them, based on the estimation of the signal-to-noise ratio between the defective regions and the background noise under matched excitation energy condition on a carbon fiber composite, was reported earlier [4]. It showed the superiority of the pulsed technique over lock-in for shallower defects. However, for deeper defects the performances of the techniques were similar. In the comparison, lock-in phase images were chosen over magnitude images as they are least affected by the variation of surface emissivity and illumination. However, the presence of blind frequencies [4], [5], [6] poses a major problem in the utilization of phase images in lock-in thermography. These are the frequencies at which the defect does not exhibit any phase shift with respect to the sound region although the thermal diffusion length is longer than the depth of the defect. This effect is primarily a 3-dimensional heat-flow phenomenon which cannot be accounted for using a traditional 1-dimensional heat-flow model. Thus if lock-in thermography is performed at only one frequency, there is always a chance of overlooking some of the defects in the phase image. Of course the inspection can be repeated over multiple frequencies to avoid the short-coming but certainly this increases its duration.

This paper proposes an application of a relatively new thermographic technique, named frequency modulated thermal wave imaging (FMTWI) [7], [8], as a means of overcoming the problem of blind frequencies. The excitation, presently used in FMTWI, comprises a linear up-chirp whose bandwidth spans over the frequencies necessary to produce thermal diffusion lengths to cover the range of defect depths in question. As per Fourier theorem, such a signal may be viewed as a superposition of multiple sinusoidal signals having frequencies which are integral multiples of the fundamental. Thus FMTWI, in essence, can be termed as superposed lock-in thermography which facilitates extraction of multiple phase and magnitude images from a single run without lengthening its duration.

In contrast to pulsed excitation, where a large number of frequencies constructively interfere at an instance in time, in FM excitation, the constructive interference is absent due to the nature of phase distribution over the constituent frequencies. However, using signal processing, it is possible to make them interfere to generate a pulse-like response. This technique is known as pulse compression, and was mentioned as a major advantage of FMTWI over lock-in technique [7]. Matched filtering is used to achieve this. In the matched filter, individual constituent frequencies are delayed in such a way that they produce constructive interference at the output of the filter.

This paper compares the image quality obtained from FMTWI to that obtained from pulsed and lock-in techniques. Finally, as a measure to overcome the problem of blind frequencies, an image fusion algorithm is proposed and demonstrated.

Section snippets

Test piece

Fig. 1 is a drawing of the test piece employed in all tests described in this paper. It is an approximately 7 mm thick carbon fiber composite board containing artificial defects formed by sets of 2, 4 and 6 mm diameter flat-bottomed holes at depths from the surface ranging from 0.25 to 2.5 mm in 0.25 mm increments. The four 12 mm diameter holes at the edge of the sample are at depths of 2, 2.5, 3 and 3.5 mm. The test piece was painted with acetone soluble black acrylic paint to provide a greater

Choice of frequencies in lock-in and FMTWI

Both in lock-in thermography and FMTWI, the target is periodically heated using external heat sources, which produce highly attenuated thermal waves inside the material. The possibility that a defect can be detected by such a wave depends on its depth being less or similar to the thermal diffusion length. The thermal diffusion length ( $μ$ ), the distance over which the wave's amplitude falls to 1/e of that at the surface, is a measure of its attenuation. $μ$ is related to the excitation frequency

Energy matching

In order to make an objective comparison of pulsed, lock-in and FMTWI techniques, the effective excitation energies for each of the techniques have to be made identical. In pulsed thermography, the energy is fixed by the flash lamp and its driver electronics while in lock-in and FMTWI, the test piece is periodically heated using tungsten–halogen flood lamps whose power and time of exposure can be readily adjusted. Since these lamps cannot take away any energy from the test piece, their periodic

Calculation of phase and amplitude images in lock-in and FMTWI

There are two steps in the process—(a) offset trend removal and (b) Fourier transformation. The former is essential because of the presence of DC heating which gradually ramps up the average temperature of the test piece. The Fourier transformation is erroneous under such a condition. To remove the trend, straight lines are fitted to the time varying temperature data pixel by pixel using a least square fit algorithm and then the fitted trend is subtracted from the original data. Once the offset

Experimental setup

For all the pulsed, lock-in and FMTWI testing a TWI Thermoscope system [9] with an Indigo Merlin camera was used. The ThermoScope system is an integrated pulsed thermographic system employing a medium wavelength infrared camera and an integrated flash heating system outputting a heating pulse of approximately 2 kJ for 2 ms. The Indigo Merlin is an electrically cooled InSb focal plane array (FPA) camera with a 12-bit digital output and a resolution of 320×256 (width×height). The camera has a

Pulsed thermography

The raw frames from the pulsed thermography experiment are shown in Fig. 5. It is found that the signal-to-noise ratio exhibits time variation as both the noise floor and the signal magnitude change with frames. The plots of the SNRs vs time for the 6 mm defects are shown in Fig. 6. It clearly shows that for each defect, there exists a frame where its signal-to-noise ratio becomes maximum. These best SNRs, which are listed in Table 4, are chosen for comparison with the data obtained from lock-in

Comparison with TSR

It may be considered unfair to compare the raw pulsed images with mathematically processed amplitude and phase images obtained from the lock-in and FMTWI experiments. Thus, to make a fair comparison, the pulsed images were enhanced with the time series reconstruction (TSR) algorithm [12]. Steps followed are summarized below:

•
The average of pre-flash frames was subtracted from the rest of the pulsed experiment video.
•
Polynomials of the form $y = a_{0} + a_{1} x + a_{2} x^{2} + a_{3} x^{3} + a_{4} x^{4}$ were fitted to the logarithmic

FMTWI image fusion

The presence of blind frequencies imposes a risk of overlooking defects by the lock-in technique, if the frequency is not chosen correctly. Thus in lock-in tests, a set of frequencies have to be tried out to confirm that no such anomalies have occurred. This obviously would increase the duration of the test. However in FMTWI, it is possible to extract all the phase images from a single run. Hence it is considerably faster.

In principle, in FMTWI, it should be possible to obtain a fused image

Conclusion

In this paper, three thermographic techniques (viz pulsed, lock-in and FMTWI) are compared based on effective excitation energy matching in time domain. It is shown that while for shallow defects in CFRP sample, the pulsed technique provides the best signal-to-noise ratio, its performance decreases for deeper defects. However, first derivative images obtained from TSR processed pulsed data show significant improvement in the detectability of the deeper features. Also, for these defects, lock-in

References (14)

S.K. Lau et al.
A quantitative analysis of pulsed video thermography
NDT& E Int.
(1991)
S.G. Pickering et al.
Matched excitation energy comparison of the pulse and lock-in thermography NDE techniques
NDT &E Int.
(2008)
X.V. Maldague et al.
Advances in pulsed phase thermography
Infrared Phys. Technol.
(2002)
J.M. Milne et al.
The non-destructive evaluation of composites and other materials by thermal pulse video thermography
Proc. SPIE
(1985)
G. Busse et al.
Thermal wave imaging with phase sensitive modulated thermography
J. Appl. Phys.
(1992)
W. Bai et al.
Evaluation of defects in composite plates under convective environments using lock-in thermography
Meas. Sci. Technol.
(2001)
C. Wallbrink et al.
The effect of size on the quantitative estimation of defect depth in steel structures using lock-in thermography
J. Appl. Phys.
(2007)

There are more references available in the full text version of this article.

Cited by (156)

Fluorescence thermography based on smartphone and its application in defect inspection
2024, Sensors and Actuators A: Physical
Thermography can be used for nondestructive detection of surface and subsurface defects in objects. Generally, thermal images are recorded with an infrared imager. However, infrared imager cannot work accurately when the inspected surface has a high reflectivity. This paper proposes a fluorescence thermography method in visible light range based on a smartphone, with a principle of fluorescence integral intensity ratio temperature sensing. Considering that the R, G, and B components of each pixel in a fluorescence image are the integral intensity of partial fluorescence spectrum, the intensity ratio of any two of components R, G and B can also reflect temperature according to the spectroscopic principle. Therefore, temperature-dependent fluorescence images can be calibrated for thermal imaging. A YAG:Ce phosphor-based composite membrane was used as the temperature probe, and a smartphone was used to take fluorescence photos and efficiently convert the photos into thermal images through a self-developed APP. The smartphone-based fluorescence thermal imager was verified by applications of defect inspection. A shielded hot soldering iron and dummy defected samples were successfully inspected with the imaging system. Compared with the technique of infrared thermography, the smartphone-based fluorescence thermography has the advantages of applicability to argenteous metals, low cost, and high spatial resolution.
An active infrared system for identifying the flaking disease in Qingyang North Grotto Temple
2024, Journal of Cultural Heritage
The grotto temples, which are a type of stone cultural heritage, have immense historical, cultural, and scientific value. Owing to long-term natural weathering and human activities, grotto temples are highly susceptible to the flaking damage. This damage pattern arises owing to the varying cohesion between the inner and outer layers of sandstone. However, due to the intrinsic constraints imposed by the internal environment of grotto temples, current methods still fail to produce accurate evaluation results in a nondestructive manner. In this study, an active thermal stimulation device was specially designed to identify flaking damages in grotto temples. Using this device, field tests were performed in Cave 32 of the Northern Grotto Temple in Qingyang. Infrared detection data obtained from field tests revealed that the different air layer thicknesses significantly affect the corresponding thermal behavior. A cooling index (CI) is proposed as an indicator to express the air layer thickness. To avoid damaging the structures and to further reveal the functional relationship, a finite element model (FEM) of a flake was built according to the actual size and measured parameters of the surrounding rock mass. By combining the above methods, the relationship between the CI and air layer thickness was established, where a linear tendency was obtained. Finally, the effectiveness of the developed device in reliably quantifying the air layer thickness associated with the configuration of flaking was determined, revealing the broad application prospects of the device in heritage conservation.
Transmissive laser lock-in thermography for highly sensitive and online imaging of real interfacial bubbles in wafer bonding
2023, Infrared Physics and Technology
Online and non-destructive detection of interfacial bubbles in wafer bonding is a critical approach to enhance the yield of advanced packaging. However, the commonly used scanning acoustic microscopy (SAM) is time-consuming and may cause additional risk due to soaking in water. The conventional thermography is suffering from low sensitivity and resolution. In this work, we demonstrate for the first time an effective method for highly sensitive and online imaging of real interfacial bubbles in wafer bonding using transmissive laser lock-in thermography (LLT). The transmissive LLT system projects a modulated laser beam onto target wafer surface and records the thermal image sequence on the other side. The bubbles distinctly emerge with clear edges in the extracted lock-in phase images. To enhance the detectivity, the relationship between phase image contrast and LLT parameters was explored. As a result, we can visualize and detect the bubbles larger than 500 μm at depth of 725 μm even under the interference of metal pattern with a speed of 130 mm²/10 s, which is 38% faster than that of SAM. Besides, the LLT phase images contain both bubbles and metal patterns information. The ratio between bubble depth and width reaches 1.45, which is about 2-fold of the latest infrared researches. Above all, this work paves the way for high-sensitive imaging of various tiny internal defects in advanced packaging using infrared thermography.
Comparison in the transient regime of four lock-in thermography algorithms by means of synthetic and experimental data on CFRP
2023, NDT and E International
Lock-in thermography is considered a well-established technique for investigating fiber-reinforced polymer composite (FRP) materials. These materials are characterized by a low thermal diffusivity and for this motivation, lock-in tests on composite materials generally are time consuming and then, for some applications, not suitable for industrial in-situ inspections. A way for reducing the testing time is to carry out the lock-in tests in the transient regime, although the most well-known lock-in algorithms were originally developed for the post-processing analysis in the steady-state regime. In this regard, the aim of this paper is to present a comprehensive comparison, in the transient regime, of four well-known lock-in thermography algorithms: the Signal Reconstruction (SR), Fast Fourier Transform (FFT), Dual Integration (DI), and Four Points method (4PM). In particular, to evaluate and compare how the transient regime can affect the accuracy and precision of the algorithms in determining the phase contrast of defects, several experimental tests were carried out on a carbon fibre-reinforced polymer (CFRP) test piece having different artificial defects by modulating the sine excitation periods by means of halogen lamps. Furthermore, synthetic data were used for simulating and investigating the four lock-in algorithms in a steady-state regime, as well. In addition, in order to evaluate the robustness of each algorithm, some offline processing analyses were carried out on the experimental data to evaluate the performance of each algorithm as a function of two user-controllable parameters, the number of frames and the number of cycles, and one not directly controlled by the operators, the dropped frame. For each analysis, qualitative and quantitative results in terms of phase and normalized phase contrast are presented and discussed.
Prediction of defect depth in GFRP composite by square-heating thermography
2023, Infrared Physics and Technology
Square-heating thermography was used to predict the depth of glass fiber reinforced plastic (GFRP) flat-bottom holes during its cooling period. Four methods were proposed for depth prediction using a specified characteristic peak time. Square-heating thermography cooling theory was analyzed using numerical methods, the linear relationship between the logarithmic characteristic time and logarithmic depth of the GFRP’s flat-bottom holes was obtained, and the slope and intercept of the linear curve were mainly related to the thermal diffusivity and the heating time. Finally, the flat-bottom hole depth was predicted using experimental cooling data from GFRP samples.
Size determination of interior defects by reconstruction of subsurface virtual heat flux for step heating thermography
2023, NDT and E International
Step heating thermography is a non-destructive testing (NDT) technique that inspects interior defects by observing surface temperature rises during a long pulse heat stimulation. Due to the heat diffusion induced blurring of defect shapes, it is a challenging task to accurately determine the sizes of deep and small defects. This work focuses on estimating the size of a subsurface defect by an inversion of temperature images for the step heating thermography, to overcome the information loss during the propagation of thermal waves. We assume a space- and time-dependent virtual heat flux on the defect surface, which is reconstructed from surface temperature measurements by solving a two-dimensional inverse heat conduction problem. The local future time concept is combined with the Tikhonov's regularization to stabilize the inverse solution. A prudent choice of the regularization parameter is carried out through the L-curve. The feasibility of the proposed approach is assessed by numerical simulations and experiments. It is shown that the inversion brings an improvement in defect sizing accuracy for defects with low width-to-depth ratios and an efficient suppression of measurement noises.

View all citing articles on Scopus

^☆: This work was supported by British Council under its UKIERI programme. It also formed part of a Naval Research Board, DRDO India supported project and the core research programme of the UK Research Centre in NDE funded by the Engineering and Physical Science Research Council.

View full text

A comparison of the pulsed, lock-in and frequency modulated thermography nondestructive evaluation techniques☆

Abstract

Highlights

Introduction

Section snippets

Test piece

Choice of frequencies in lock-in and FMTWI

Energy matching

Calculation of phase and amplitude images in lock-in and FMTWI

Experimental setup

Pulsed thermography

Comparison with TSR

FMTWI image fusion

Conclusion

NDT& E Int.

NDT &E Int.

Infrared Phys. Technol.

The non-destructive evaluation of composites and other materials by thermal pulse video thermography

Proc. SPIE

Thermal wave imaging with phase sensitive modulated thermography

J. Appl. Phys.

Evaluation of defects in composite plates under convective environments using lock-in thermography

Meas. Sci. Technol.

The effect of size on the quantitative estimation of defect depth in steel structures using lock-in thermography

J. Appl. Phys.