- Email: [email protected]
Infrared thermography as a new method for quality control of sheet metal parts in the press shop
archives of civil and mechanical engineering 12 (2012) 148–155
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/acme
Infrared thermography as a new method for quality control of sheet metal parts in the press shop Q. Ana,n, D. Hortiga, M. Merkleinb a
Daimler AG, Werk Sindelfingen, Press Shop F4 Sindelfingen, Ka¨sbru¨nnler Str., 71063 Sindelfingen, Germany Friedrich-Alexander-University Erlangen-Nuremberg, Chair of Manufacturing Technology, Egerland Str. 13, Erlangen, Germany
b
ar t ic l e in f o
abs tra ct
Article history:
Active dynamic thermography was proved to be a new alternative for crack detection. To
Received 27 April 2012
achieve zero failure sheet metal production in the press shop, this method was investigated
Accepted 27 April 2012
for different steel components, which were directly collected from series production. And the
Available online 4 May 2012
cracks in the presented specimens differ in length and in depth, and a series study on these
Keywords:
surface cracks is executed. The crack detection capabilities of active thermography with two
NDT
different excitation sources were compared: optical and inductive excited thermography.
Active dynamic thermography
Results of experiments obtained by these two techniques are showed and compared with
Crack detect
each other. Opportunities of applying these methods in press shop are discussed.
Sheet metal parts
& 2012 Politechnika Wrocławska. Published by Elsevier Urban & Partner Sp. z.o.o. All rights reserved.
Quality control
1.
Introduction
Infrared thermography (IRT) was at first an invention for military applications. However, due to the rapid development of thermal cameras since the 1970s, this technology has been increasingly used in civilian areas. Because of its several advantages, such as non-contacts with the inspected object and detection capability of subsurface failure [7], IRT is widely applied for nondestructive testing (NDT). Furthermore, materials with either low or high thermal conductivity are able to be detected using IRT with proper mathematical evaluations [7,11]. For example, cracks or impact of lamination. Furthermore, IRT system requires very short testing time, which enable its application on automatic quantitative material defects detection [1,2,7]. These properties of IRT also match the requirements of quality control systems for sheet metal parts, and it can be a solution for an automated defects detection system in the press shop. In this paper, two of the most common IRT techniques will be introduced, which are the optical and inductive excited pulsed-phase thermography (PPT). Specimens for the experiments are collected and cut off from the series production of
luggage trunk door. Detection results of defects with these two IR techniques were compared, in order to approach a better suitable method for quality controlling of specific sheet metal parts.
2.
Material and methods
Under the definition of IRT, a distinction is made between passive and active thermography. In passive thermography, measurements of temperature decreases provide temperature profiles, and an abnormal temperature profile indicates a problem in the specimen. In the active thermography, extra energy is brought onto the specimen, so that there would be a significant temperature difference [1,7].
2.1.
Active thermography
The fundamental of NDT procedures nowadays is mainly based on the active thermography, because of its many advantages, such as more reliable information in its results,
n
Corresponding author. Tel.: þ49 7031 90 76255; fax: þ49 7113 05 214. E-mail address: [email protected] (Q. An).
1644-9665/$ - see front matter & 2012 Politechnika Wrocławska. Published by Elsevier Urban & Partner Sp. z.o.o. All rights reserved. http://dx.doi.org/10.1016/j.acme.2012.04.008
archives of civil and mechanical engineering 12 (2012) 148–155
and insensibility of influence from environments [1]. Fig. 1 exhibits a typical classification of all common IRT methods. Among all these active thermography techniques, pulsedphase thermography (PPT) and lock-in thermography (LIT) are mostly applied [4,5]. Thus, the methods for detects defection in sheet metal parts were focused on PPT and LIT techniques. According to [3,8,9,12], main advantages and disadvantages of these two techniques are listed in Table 1. As described in [6,7], PPT combines promising features from two older thermographic techniques. It is as rapid and easy as pulsed thermography to develop and requires a shorter evaluation time. Moreover PPT provides phase delay images as Lock-in thermography. Accordingly, PPT is safe and easy to deploy NDT technique, giving the possibility to rapidly inspect large and complex surfaces. So it is more applicable for defect detection in sheet metal parts, and was therefore discussed in this paper, covering two exitation techniques.
2.2.
Pulsed-phase-thermography
In PPT, a short burst of excitation is applied to the specimen, and the heating pulse occurs in milliseconds with its shape
149
being approximately rectangular. This is very important, because this pulse includes in the frequency speckle a high lobe near zero frequency and smaller lobes in low frequencies, which improve entirely the signal–noise-ratio [9,10]. Temperature history of one pixel is normally pattered into two sections: heating and cooling. A strong contribution is however included in the cooling period, so the result evaluation is applied to this area with a Fast Fourier Transform (FFT) algorithm. As it is known, FFT can calculate a time domain into a frequency domain, which exhibits enhanced thermographic results in phase and amplitude images [10,11]. Fig. 2 shows a graphical explanation of data acquisition and processing by PPT, and it is based on [7,9,13]. Only the phase image processing is showed, and the amplitude image processing follows the same principle. The temperature of pixel (i,j) is Tij, which decreases after impulse excitation till Tij(N). N is the sum of sampling images, Dt is the sampling interval, so NDt is the whole measure time. When the themogram sequence is processed using FFT, the real and imaginary transform will be [7] Fn ¼ Dt
N 1 X
ðkDtÞej2mk=N ¼ Ren þ Imn
ð1Þ
k¼0
where n designates the frequency increment (n¼ 0.1y.N). Re and Im are the real and imaginary parts of transform. The FFT algorithm is used in each pixel for the whole image sequence, and then thermograms will be transformed into phase and amplitude images, by using ([7]) the following equation: Imn ð2Þ |n ¼ tan1 Ren An ¼
Fig. 1 – Classification of active thermography methods.
Table 1 – Advantages and disadvantages comparison of the pulsed-phase and lock-in thermography. Method Pulsed-phase thermography (PPT)
Advantages
Lock-in thermography (LIT)
Disadvantages
Rapid testing time Suitable for unknown defects Relative easy excitation’s control
Advanced detection depth Relative low heating volume
Low detectable depth Relative high heating volume
Necessity of finding optical lock-in by unknown defects Complex testing equipments
ð3Þ
The calculation with FFT is finished with software packages such as MatLabs. When the pulsed heating reaches the specimen, thermoenergy expands into its surrounding area. And the thermoenergy intensity increases sharply where the crack is located, because cracks present a barrier for the heat transmission. Correspondingly, the crack temperature increases more rapidly than the sound area, this change can be measured with an IR camera. Even more, after the evaluation with FFT, cracks present a more characterized pattern in the phase or amplitude image. Also, phase is less affected than thermal date by problems such as non-uniform heating, surface emissivity variations and non-planar surfaces [7,10,12,13].
2.3.
Longer testing time
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ren 2 þ Imn 2
Specimens
In order to prove the PPT’s availability for defects on sheet metal parts, two specimens are selected from the press shop. They represent the very typical defect of this sheet metal part. However cracks on the two specimens have different lengths and depths. Like in practice, it is difficult to characterize defects that existed in the serial production. As described in Section 2, experiments in this paper are based on the PPT technique. There are two most commonly used excitation resources in the PPT technique: the optical
150
archives of civil and mechanical engineering 12 (2012) 148–155
T
(t)
(f)
test duration test duration
FFT
N
1 2 3 1 2 3 T
(1)
N T
f
t
(N)
(N)
(1)
Fig. 2 – Date acquisition and processing by PPT.
Fig. 5 – Form and size of specimen II.
Fig. 3 – The cutting locations of specimen I and specimen II.
Specimen IR camera crack
Flash lamp Fig. 4 – Form and size of specimen I.
excitation is normally the flash lamp, and the inductive excitation is a coil connected with an induction generator. The material of these two specimens for the experiments is mild steel, DC05 with electro-zinc coating and a thickness of 0.8 mm. Both were from the same sheet metal part, namely the luggage trunk door, but they were cut from different locations, as shown in Fig. 3. Specimen I contains a 10 mm long and o1 mm wide crack on the radius area, as shown in Fig. 4.The crack’s depth is 50.8 mm, it can be observed, that the sheet metal of this crack is not totally thinned out. Fig. 5 shows the specimen II, it has a 30 mm long and about 1 mm wide crack also on the radius area. This crack is no more a surface defect, while the sheet metal along the crack is totally drawn apart.
PC Synchronisation Controlling Image Processing Evaluation
Fig. 6 – Basic configuration of optical excited PPT system.
2.4.
Experimental procedures
The experimental configuration for optical excited PPT is shown in Fig. 6. Pulsed heating generated by a flash lamp is absorbed and generates a thermal wave propagating from the surface into the specimen where it is reflected at thermal
archives of civil and mechanical engineering 12 (2012) 148–155
boundaries, and the thermal wave reflection is measured with an IR camera. A test computer is used for all relevant controlling and evaluation processes. The flash lamp for the conducted experiments is part of a flash system IPL-System from CryLas GmbHs, and it emits a maximal heating of 25 J excitation within 30 ms, such configuration is directly applied in the experiment. Accordingly, the excitation capacity was 25 J and the excitation duration was less than 30 ms. It should be noted that, the flash lamp must lighten exactly the crack area, otherwise, reflected disruptions would be too strong for the evaluation later [11]. The IR-Camera is the ImageIR 8300 with cooled InSb Detector from InfraTec GmbHs, which has a resolution from 640 512 pixel and NEDT from o25 mK. The sampling frequency in the experiments with flash lamp as excitation was 12.5 Hz and sampling time was 2 s. Fig. 7 shows the basic configuration of the inductive excited PPT system. A coil powered by an induction generator introduces eddy current into the specimen, while an IR camera records the temperature changes in the specimen over the sampling time. The induction generator’s model is a TruHeat HF ¨ TTINGER Electronic of TRUMPF Groups, and it has a 3010 HU maximal capacity of 10 kW and excitation frequency of 10% capacity at about 200 kHz, this configuration was also the
Fig. 7 – Basic configuration of inductive excited PPT system.
151
excitation capacity in experiments with inductive excitation, and the excitation duration was 50 ms. The coils are specially made from 4 mm 4 mm rectangular copper pipe for each specimen, because of their complex geometry. Unlike optical excited PPT, there is a direct interaction between the heating mechanism and the defect in the inductive excited PPT, which leads to a strong contrast in the crack area, especially when the crack is oriented vertical to the coil and on the surface of the specimen. The ImageIR 8300 is also used for experiments with inductive excitation. Because the crack in specimen I is relatively shorter than the crack in specimen II, the thermoconduction in the specimen happened so rapidly, the sampling frequency is raised to 20 Hz, in order to observe significant results. The sampling frequency for specimen II was still 12.5 Hz. The sampling time for both specimens was 2 s. In order to prove the PPT’s availability for defects on sheet metal parts, two specimens are selected from the press shop. They represent the very typical defect of this sheet metal part. However cracks on the two specimens have different
Fig. 9 – Specimen I’s phase and amplitude images by the flash lamp and induction excited PPT.
Fig. 8 – Comparison of specimen I’s thermographic data from flash lamp and induction excited PPT.
152
archives of civil and mechanical engineering 12 (2012) 148–155
lengths and depths. As like in the practice, it is difficult to characterize defects existed in the serial production. As described in Section 2, experiments in this paper are based on the PPT technique. There are two most commonly excitation’s resources in the PPT technique: the optical excitation is normally the flash lamp, and the inductive excitation is a coil connected with an induction generator.
3.
Results
Three types of figures can be obtained by the experiments. To approach a suitable technique for defect detection in sheet metal parts, thermograms are compared at first and secondly the phase and amplitude figures. The phase-difference-diagrams provide a precise phase value comparison in conclusion.
3.1.
Specimen I
Fig. 8 shows that the temperature raised to maximal 306 K. With a room temperature of 295 K, the temperature increase is 9 K according to Fig. 8. This amount of heating would not influence the sheet metal properties. Comparing with flash lamp, the inductive excitation reached a higher heating treatment. If one only compares the phase images in Fig. 9, both flash lamp and induction excited PPT illustrated remarkablly a typical crack pattern, which is either light or dark line. However this pattern can hardly be seen in the amplitude image of flash lamp excited PPT, as in Fig. 9. A more precise phase value can be observed in Fig. 10. The crack phase value is the mean of the whole phase values along the crack’s line pattern, and the value of sound area is a
Fig. 10 – Phase values comparison between crack in specimen I and sound area.
archives of civil and mechanical engineering 12 (2012) 148–155
153
Fig. 11 – Comparison of specimen II’s thermographic data from flash lamp and induction excited PPT.
Phase value differences in Figs. 10 and 13 indicate that, if the crack is longer, the phase value difference between the crack and the sound area is correspondingly more significant. This phenomenon is more remarkable in experiments with inductive excitation.
4.
Fig. 12 – Specimen II’s phase and amplitude images by the flash lamp and induction excited PPT.
phase value of points beside the crack. By the flash lamp excitation, there is only a slight phase value difference of 50.5 rad between the 10 mm long crack and the sound area. While the phase value difference is remarkable by the inductive excitation.
3.2.
Specimen II
Fig. 11 presents the thermographic data of specimen II. By flash lamp excitation, the crack line can be observed. Both the phase and amplitude image from these two experiments show a clear typical crack line pattern, as shown in Fig. 12. Also Fig. 13 indicates a greater phase value difference between crack and the sound area. The peak of value difference by induction excitation was raised by about 0.5 rad more than by the flash lamp excitation.
Discussion
Results of experiments in this paper prove the possibility of using PPT technique to detect cracks in sheet metal parts. In the phase images, a characterized pattern from cracks can be observed, which simplifies the procedure of locating the crack. Primarily the problems of applying thermographic technique on metal parts, such as inhomogeneous emission coefficient on the metal surface and sensibility of light disruption, are solved through the image processing operation in the PPT. Results of the experiments with inductive excitation show enhanced phase values differences between the crack and the sound area. Consequently, inductive excited PPT is evaluated as a more robust procedure for defect detection in sheet metal parts. It is therefore suggested to apply inductive excitation in this case. Furthermore, according to the phase value comparisons, the crack length has a mathematical connection with the phase value difference. If this connection could be established in follow up experiments, then PPT can also provide a measurement of the crack size. More experiments with various cracks lengths should be carried out to build this mathematical model.
5.
Conclusion
This paper aims to present a new method for quality controlling of sheet metal parts. Experiments have also proved the feasibility of using inductive excited PPT as a better solution, alternatively to only visual inspection conducted by human
154
archives of civil and mechanical engineering 12 (2012) 148–155
Fig. 13 – Phase values comparison between crack in specimen II and sound area.
beings in press shop. Further investigations are suggested to complete this quality control procedure. [4]
references
[1] Aderhold J., Meinlschmidt P, Active and passive inlinethermography, Guidelines for Thermography Technique— Nondestructive Testing Using Image Processing, 2011, pp. 12–14 (in German). [2] J. Baumenn, U. Netzelmann, R. Bilgram, T. Hierl, V. Carl, H.R. Schubach, G. Zenzinger, Investigations of the active thermography testing for validation of the high performance manufacturing process—results of a BMBF-Project, in: Proceedings of the DGZfP Conference 2005, Presentation no. 57, Rostock (in German). [3] K. Chatterjee, S. Tuli, G. Pickering, D. Almond, A comparison of the pulsed, lock-in and frequency modulated thermography
[5]
[6]
[7]
[8]
nondestructive evaluation techniques, NDT&E International 44 (2011) 655–667. A. Dillenz, T. Zweschper, G. Riegert, G. Busse, Progress in phase angle thermography, Review of Scientific Instruments 74 (1) (2003) 417–419. C. Ibarra-Castanedo, D. Gonza´lez, X. Maldague, Automatic algorithm for quantitative pulsed phase thermography calculations, in: Proceedings of the 16th WCNDT 2004, vol. 9, no. 11, Montreal, Canada, 2004, pp. 921–931. C. Ibarra-Castanedo, X. Maldague, Defect depth retrieval from pulsed phase thermographic data on plexiglas and aluminium samples, Thermosense XXVI, vol. 5405, 2004 (pp. 348–356). X. Maldague, Theory and Practice of Infrared Technology for Nondestructive Testing, Electrical and Computing Engineering Department, University Laval, pp. 1–11. U. Netzelmann, The thermography procedures: pulsed and lockin thermography, Guidelines for Thermography Technique— Nondestructive Testing Using Image Processing, 2011, pp. 15–18 (in German).
archives of civil and mechanical engineering 12 (2012) 148–155
[9] U. Netzelmann, G. Walle, S. Lugin. Active Algorithms for the thermograms evaluation in the lock-in and pulsed thermography, Guidelines for Thermography Technique–Non-Destructive Testing Using Image Processing, 2011, pp. 46–49 (in German). [10] S. Pickering, D. Almond, Pulsed matched excitation energy comparison of the pulsed and lock-in thermography NDE techniques, NDT&E International 41 (2008) 501–513. [11] U. Siemer, Use of Thermography as a Nondestructive Test Method in the Automotive Industry, Dissertation, University Saarbru¨cken, 2010.
155
[12] G.Y. Tian, J. Wilson, L. Cheng, D.P. Almond, Pulsed eddy current thermography and applications, Lecture Notes in Electrical Engineering 96 (1) (2011) 205–231. [13] A. Ziadi, F. Galmiche, X.. Maldague, Pulsed shaping in infrared thermography for nondestructive evaluation, Review of Scientific Instruments 74 (1) (2003) 411–413.
Available online at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/acme
Infrared thermography as a new method for quality control of sheet metal parts in the press shop Q. Ana,n, D. Hortiga, M. Merkleinb a
Daimler AG, Werk Sindelfingen, Press Shop F4 Sindelfingen, Ka¨sbru¨nnler Str., 71063 Sindelfingen, Germany Friedrich-Alexander-University Erlangen-Nuremberg, Chair of Manufacturing Technology, Egerland Str. 13, Erlangen, Germany
b
ar t ic l e in f o
abs tra ct
Article history:
Active dynamic thermography was proved to be a new alternative for crack detection. To
Received 27 April 2012
achieve zero failure sheet metal production in the press shop, this method was investigated
Accepted 27 April 2012
for different steel components, which were directly collected from series production. And the
Available online 4 May 2012
cracks in the presented specimens differ in length and in depth, and a series study on these
Keywords:
surface cracks is executed. The crack detection capabilities of active thermography with two
NDT
different excitation sources were compared: optical and inductive excited thermography.
Active dynamic thermography
Results of experiments obtained by these two techniques are showed and compared with
Crack detect
each other. Opportunities of applying these methods in press shop are discussed.
Sheet metal parts
& 2012 Politechnika Wrocławska. Published by Elsevier Urban & Partner Sp. z.o.o. All rights reserved.
Quality control
1.
Introduction
Infrared thermography (IRT) was at first an invention for military applications. However, due to the rapid development of thermal cameras since the 1970s, this technology has been increasingly used in civilian areas. Because of its several advantages, such as non-contacts with the inspected object and detection capability of subsurface failure [7], IRT is widely applied for nondestructive testing (NDT). Furthermore, materials with either low or high thermal conductivity are able to be detected using IRT with proper mathematical evaluations [7,11]. For example, cracks or impact of lamination. Furthermore, IRT system requires very short testing time, which enable its application on automatic quantitative material defects detection [1,2,7]. These properties of IRT also match the requirements of quality control systems for sheet metal parts, and it can be a solution for an automated defects detection system in the press shop. In this paper, two of the most common IRT techniques will be introduced, which are the optical and inductive excited pulsed-phase thermography (PPT). Specimens for the experiments are collected and cut off from the series production of
luggage trunk door. Detection results of defects with these two IR techniques were compared, in order to approach a better suitable method for quality controlling of specific sheet metal parts.
2.
Material and methods
Under the definition of IRT, a distinction is made between passive and active thermography. In passive thermography, measurements of temperature decreases provide temperature profiles, and an abnormal temperature profile indicates a problem in the specimen. In the active thermography, extra energy is brought onto the specimen, so that there would be a significant temperature difference [1,7].
2.1.
Active thermography
The fundamental of NDT procedures nowadays is mainly based on the active thermography, because of its many advantages, such as more reliable information in its results,
n
Corresponding author. Tel.: þ49 7031 90 76255; fax: þ49 7113 05 214. E-mail address: [email protected] (Q. An).
1644-9665/$ - see front matter & 2012 Politechnika Wrocławska. Published by Elsevier Urban & Partner Sp. z.o.o. All rights reserved. http://dx.doi.org/10.1016/j.acme.2012.04.008
archives of civil and mechanical engineering 12 (2012) 148–155
and insensibility of influence from environments [1]. Fig. 1 exhibits a typical classification of all common IRT methods. Among all these active thermography techniques, pulsedphase thermography (PPT) and lock-in thermography (LIT) are mostly applied [4,5]. Thus, the methods for detects defection in sheet metal parts were focused on PPT and LIT techniques. According to [3,8,9,12], main advantages and disadvantages of these two techniques are listed in Table 1. As described in [6,7], PPT combines promising features from two older thermographic techniques. It is as rapid and easy as pulsed thermography to develop and requires a shorter evaluation time. Moreover PPT provides phase delay images as Lock-in thermography. Accordingly, PPT is safe and easy to deploy NDT technique, giving the possibility to rapidly inspect large and complex surfaces. So it is more applicable for defect detection in sheet metal parts, and was therefore discussed in this paper, covering two exitation techniques.
2.2.
Pulsed-phase-thermography
In PPT, a short burst of excitation is applied to the specimen, and the heating pulse occurs in milliseconds with its shape
149
being approximately rectangular. This is very important, because this pulse includes in the frequency speckle a high lobe near zero frequency and smaller lobes in low frequencies, which improve entirely the signal–noise-ratio [9,10]. Temperature history of one pixel is normally pattered into two sections: heating and cooling. A strong contribution is however included in the cooling period, so the result evaluation is applied to this area with a Fast Fourier Transform (FFT) algorithm. As it is known, FFT can calculate a time domain into a frequency domain, which exhibits enhanced thermographic results in phase and amplitude images [10,11]. Fig. 2 shows a graphical explanation of data acquisition and processing by PPT, and it is based on [7,9,13]. Only the phase image processing is showed, and the amplitude image processing follows the same principle. The temperature of pixel (i,j) is Tij, which decreases after impulse excitation till Tij(N). N is the sum of sampling images, Dt is the sampling interval, so NDt is the whole measure time. When the themogram sequence is processed using FFT, the real and imaginary transform will be [7] Fn ¼ Dt
N 1 X
ðkDtÞej2mk=N ¼ Ren þ Imn
ð1Þ
k¼0
where n designates the frequency increment (n¼ 0.1y.N). Re and Im are the real and imaginary parts of transform. The FFT algorithm is used in each pixel for the whole image sequence, and then thermograms will be transformed into phase and amplitude images, by using ([7]) the following equation: Imn ð2Þ |n ¼ tan1 Ren An ¼
Fig. 1 – Classification of active thermography methods.
Table 1 – Advantages and disadvantages comparison of the pulsed-phase and lock-in thermography. Method Pulsed-phase thermography (PPT)
Advantages
Lock-in thermography (LIT)
Disadvantages
Rapid testing time Suitable for unknown defects Relative easy excitation’s control
Advanced detection depth Relative low heating volume
Low detectable depth Relative high heating volume
Necessity of finding optical lock-in by unknown defects Complex testing equipments
ð3Þ
The calculation with FFT is finished with software packages such as MatLabs. When the pulsed heating reaches the specimen, thermoenergy expands into its surrounding area. And the thermoenergy intensity increases sharply where the crack is located, because cracks present a barrier for the heat transmission. Correspondingly, the crack temperature increases more rapidly than the sound area, this change can be measured with an IR camera. Even more, after the evaluation with FFT, cracks present a more characterized pattern in the phase or amplitude image. Also, phase is less affected than thermal date by problems such as non-uniform heating, surface emissivity variations and non-planar surfaces [7,10,12,13].
2.3.
Longer testing time
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Ren 2 þ Imn 2
Specimens
In order to prove the PPT’s availability for defects on sheet metal parts, two specimens are selected from the press shop. They represent the very typical defect of this sheet metal part. However cracks on the two specimens have different lengths and depths. Like in practice, it is difficult to characterize defects that existed in the serial production. As described in Section 2, experiments in this paper are based on the PPT technique. There are two most commonly used excitation resources in the PPT technique: the optical
150
archives of civil and mechanical engineering 12 (2012) 148–155
T
(t)
(f)
test duration test duration
FFT
N
1 2 3 1 2 3 T
(1)
N T
f
t
(N)
(N)
(1)
Fig. 2 – Date acquisition and processing by PPT.
Fig. 5 – Form and size of specimen II.
Fig. 3 – The cutting locations of specimen I and specimen II.
Specimen IR camera crack
Flash lamp Fig. 4 – Form and size of specimen I.
excitation is normally the flash lamp, and the inductive excitation is a coil connected with an induction generator. The material of these two specimens for the experiments is mild steel, DC05 with electro-zinc coating and a thickness of 0.8 mm. Both were from the same sheet metal part, namely the luggage trunk door, but they were cut from different locations, as shown in Fig. 3. Specimen I contains a 10 mm long and o1 mm wide crack on the radius area, as shown in Fig. 4.The crack’s depth is 50.8 mm, it can be observed, that the sheet metal of this crack is not totally thinned out. Fig. 5 shows the specimen II, it has a 30 mm long and about 1 mm wide crack also on the radius area. This crack is no more a surface defect, while the sheet metal along the crack is totally drawn apart.
PC Synchronisation Controlling Image Processing Evaluation
Fig. 6 – Basic configuration of optical excited PPT system.
2.4.
Experimental procedures
The experimental configuration for optical excited PPT is shown in Fig. 6. Pulsed heating generated by a flash lamp is absorbed and generates a thermal wave propagating from the surface into the specimen where it is reflected at thermal
archives of civil and mechanical engineering 12 (2012) 148–155
boundaries, and the thermal wave reflection is measured with an IR camera. A test computer is used for all relevant controlling and evaluation processes. The flash lamp for the conducted experiments is part of a flash system IPL-System from CryLas GmbHs, and it emits a maximal heating of 25 J excitation within 30 ms, such configuration is directly applied in the experiment. Accordingly, the excitation capacity was 25 J and the excitation duration was less than 30 ms. It should be noted that, the flash lamp must lighten exactly the crack area, otherwise, reflected disruptions would be too strong for the evaluation later [11]. The IR-Camera is the ImageIR 8300 with cooled InSb Detector from InfraTec GmbHs, which has a resolution from 640 512 pixel and NEDT from o25 mK. The sampling frequency in the experiments with flash lamp as excitation was 12.5 Hz and sampling time was 2 s. Fig. 7 shows the basic configuration of the inductive excited PPT system. A coil powered by an induction generator introduces eddy current into the specimen, while an IR camera records the temperature changes in the specimen over the sampling time. The induction generator’s model is a TruHeat HF ¨ TTINGER Electronic of TRUMPF Groups, and it has a 3010 HU maximal capacity of 10 kW and excitation frequency of 10% capacity at about 200 kHz, this configuration was also the
Fig. 7 – Basic configuration of inductive excited PPT system.
151
excitation capacity in experiments with inductive excitation, and the excitation duration was 50 ms. The coils are specially made from 4 mm 4 mm rectangular copper pipe for each specimen, because of their complex geometry. Unlike optical excited PPT, there is a direct interaction between the heating mechanism and the defect in the inductive excited PPT, which leads to a strong contrast in the crack area, especially when the crack is oriented vertical to the coil and on the surface of the specimen. The ImageIR 8300 is also used for experiments with inductive excitation. Because the crack in specimen I is relatively shorter than the crack in specimen II, the thermoconduction in the specimen happened so rapidly, the sampling frequency is raised to 20 Hz, in order to observe significant results. The sampling frequency for specimen II was still 12.5 Hz. The sampling time for both specimens was 2 s. In order to prove the PPT’s availability for defects on sheet metal parts, two specimens are selected from the press shop. They represent the very typical defect of this sheet metal part. However cracks on the two specimens have different
Fig. 9 – Specimen I’s phase and amplitude images by the flash lamp and induction excited PPT.
Fig. 8 – Comparison of specimen I’s thermographic data from flash lamp and induction excited PPT.
152
archives of civil and mechanical engineering 12 (2012) 148–155
lengths and depths. As like in the practice, it is difficult to characterize defects existed in the serial production. As described in Section 2, experiments in this paper are based on the PPT technique. There are two most commonly excitation’s resources in the PPT technique: the optical excitation is normally the flash lamp, and the inductive excitation is a coil connected with an induction generator.
3.
Results
Three types of figures can be obtained by the experiments. To approach a suitable technique for defect detection in sheet metal parts, thermograms are compared at first and secondly the phase and amplitude figures. The phase-difference-diagrams provide a precise phase value comparison in conclusion.
3.1.
Specimen I
Fig. 8 shows that the temperature raised to maximal 306 K. With a room temperature of 295 K, the temperature increase is 9 K according to Fig. 8. This amount of heating would not influence the sheet metal properties. Comparing with flash lamp, the inductive excitation reached a higher heating treatment. If one only compares the phase images in Fig. 9, both flash lamp and induction excited PPT illustrated remarkablly a typical crack pattern, which is either light or dark line. However this pattern can hardly be seen in the amplitude image of flash lamp excited PPT, as in Fig. 9. A more precise phase value can be observed in Fig. 10. The crack phase value is the mean of the whole phase values along the crack’s line pattern, and the value of sound area is a
Fig. 10 – Phase values comparison between crack in specimen I and sound area.
archives of civil and mechanical engineering 12 (2012) 148–155
153
Fig. 11 – Comparison of specimen II’s thermographic data from flash lamp and induction excited PPT.
Phase value differences in Figs. 10 and 13 indicate that, if the crack is longer, the phase value difference between the crack and the sound area is correspondingly more significant. This phenomenon is more remarkable in experiments with inductive excitation.
4.
Fig. 12 – Specimen II’s phase and amplitude images by the flash lamp and induction excited PPT.
phase value of points beside the crack. By the flash lamp excitation, there is only a slight phase value difference of 50.5 rad between the 10 mm long crack and the sound area. While the phase value difference is remarkable by the inductive excitation.
3.2.
Specimen II
Fig. 11 presents the thermographic data of specimen II. By flash lamp excitation, the crack line can be observed. Both the phase and amplitude image from these two experiments show a clear typical crack line pattern, as shown in Fig. 12. Also Fig. 13 indicates a greater phase value difference between crack and the sound area. The peak of value difference by induction excitation was raised by about 0.5 rad more than by the flash lamp excitation.
Discussion
Results of experiments in this paper prove the possibility of using PPT technique to detect cracks in sheet metal parts. In the phase images, a characterized pattern from cracks can be observed, which simplifies the procedure of locating the crack. Primarily the problems of applying thermographic technique on metal parts, such as inhomogeneous emission coefficient on the metal surface and sensibility of light disruption, are solved through the image processing operation in the PPT. Results of the experiments with inductive excitation show enhanced phase values differences between the crack and the sound area. Consequently, inductive excited PPT is evaluated as a more robust procedure for defect detection in sheet metal parts. It is therefore suggested to apply inductive excitation in this case. Furthermore, according to the phase value comparisons, the crack length has a mathematical connection with the phase value difference. If this connection could be established in follow up experiments, then PPT can also provide a measurement of the crack size. More experiments with various cracks lengths should be carried out to build this mathematical model.
5.
Conclusion
This paper aims to present a new method for quality controlling of sheet metal parts. Experiments have also proved the feasibility of using inductive excited PPT as a better solution, alternatively to only visual inspection conducted by human
154
archives of civil and mechanical engineering 12 (2012) 148–155
Fig. 13 – Phase values comparison between crack in specimen II and sound area.
beings in press shop. Further investigations are suggested to complete this quality control procedure. [4]
references
[1] Aderhold J., Meinlschmidt P, Active and passive inlinethermography, Guidelines for Thermography Technique— Nondestructive Testing Using Image Processing, 2011, pp. 12–14 (in German). [2] J. Baumenn, U. Netzelmann, R. Bilgram, T. Hierl, V. Carl, H.R. Schubach, G. Zenzinger, Investigations of the active thermography testing for validation of the high performance manufacturing process—results of a BMBF-Project, in: Proceedings of the DGZfP Conference 2005, Presentation no. 57, Rostock (in German). [3] K. Chatterjee, S. Tuli, G. Pickering, D. Almond, A comparison of the pulsed, lock-in and frequency modulated thermography
[5]
[6]
[7]
[8]
nondestructive evaluation techniques, NDT&E International 44 (2011) 655–667. A. Dillenz, T. Zweschper, G. Riegert, G. Busse, Progress in phase angle thermography, Review of Scientific Instruments 74 (1) (2003) 417–419. C. Ibarra-Castanedo, D. Gonza´lez, X. Maldague, Automatic algorithm for quantitative pulsed phase thermography calculations, in: Proceedings of the 16th WCNDT 2004, vol. 9, no. 11, Montreal, Canada, 2004, pp. 921–931. C. Ibarra-Castanedo, X. Maldague, Defect depth retrieval from pulsed phase thermographic data on plexiglas and aluminium samples, Thermosense XXVI, vol. 5405, 2004 (pp. 348–356). X. Maldague, Theory and Practice of Infrared Technology for Nondestructive Testing, Electrical and Computing Engineering Department, University Laval, pp. 1–11. U. Netzelmann, The thermography procedures: pulsed and lockin thermography, Guidelines for Thermography Technique— Nondestructive Testing Using Image Processing, 2011, pp. 15–18 (in German).
archives of civil and mechanical engineering 12 (2012) 148–155
[9] U. Netzelmann, G. Walle, S. Lugin. Active Algorithms for the thermograms evaluation in the lock-in and pulsed thermography, Guidelines for Thermography Technique–Non-Destructive Testing Using Image Processing, 2011, pp. 46–49 (in German). [10] S. Pickering, D. Almond, Pulsed matched excitation energy comparison of the pulsed and lock-in thermography NDE techniques, NDT&E International 41 (2008) 501–513. [11] U. Siemer, Use of Thermography as a Nondestructive Test Method in the Automotive Industry, Dissertation, University Saarbru¨cken, 2010.
155
[12] G.Y. Tian, J. Wilson, L. Cheng, D.P. Almond, Pulsed eddy current thermography and applications, Lecture Notes in Electrical Engineering 96 (1) (2011) 205–231. [13] A. Ziadi, F. Galmiche, X.. Maldague, Pulsed shaping in infrared thermography for nondestructive evaluation, Review of Scientific Instruments 74 (1) (2003) 411–413.