 ·Home
·Table of Contents
·Materials Characterization and testing | Ultrasonic Examination of Thermal Sprayed Coatings with Frequency Analysis
H. HATANAKA, T. ARAKAWA,
Advanced Production Engineering Center, Ishikawajima-Harima Heavy Industries Co., Ltd.
1 Shin-Nakahara-cho, Isogo-ku, Yokohama 235-8501, Japan
TEL: 81-45-759-2193, FAX: 81-45-759-2125,
E-mail: hiroaki_hatanaka@ihi.co.jp
K. NAMBA, I. KAJIGAYA,
Power Plant Division, Ishikawajima-Harima Heavy Industries Co., Ltd.
3-2-16 Toyosu, Koto-ku, Tokyo 135-8733, Japan
Contact
|
Abstract
Thermal sprayed coating is presently being applied to surfaces of structures such as boiler components to prevent erosion and/or corrosion. Methods for the nondestructive evaluation of degradation and damage, such as decline in adhesive strength and decrease in coating thickness should thus be established. This paper proposes nondestructive methods for evaluating adhesive strength of thermal sprayed coating and measuring coating thickness by ultrasonic testing. Adhesive strength was evaluated based on spectral intensity at a certain frequency band of ultrasonic waveform. Spectral intensity at a certain frequency band increases with adhesive strength correspondingly. In coating thickness measurement, to enhance signal-to-noise ratio of ultrasonic waveforms, the wavelet transform was used to analyze the ultrasonic waveform. The echo from the boundary between the coating and base metal could be readily determined.
Keywords: thermal sprayed coating, nondestructive evaluation, frequency, adhesive strength, wavelet transform, ultrasonic testing
Introduction
Because of diversification of boiler and fuel, thermal sprayed coatings are widely used to prevent erosion and/or high-temperature corrosion. Methods for nondestructive evaluation of degradation or damage such as surface flaw, delamination, reduction in adhesive strength and decrease in coating thickness should be made available for maintenance and quality control. The evaluation of coatings has so far been carried out by methods attended with destruction.
Test pieces with 50Cr-50Ni coatings were prepared for use in preventing boiler components from high-temperature corrosion, and nondestructive evaluations were established. This paper proposes a means to evaluate adhesive strength and confirms the possibility of measuring thicknesses of thermal sprayed coatings by ultrasonic testing. Evaluation of adhesive strength method uses ultrasonic waves which pass through the boundary between the coating and base metal. After transforming the ultrasonic waveform into a frequency domain, adhesive strength is evaluated from its higher frequency components. Measurement of coating thickness method uses wavelet analysis to the ultrasonic waveform. Wavelet analysis is applied to enhance the signal-to-noise ratio of waveform and the echo from the boundary between coating and base metal is determined.
Evaluation of Adhesive Strength of Thermal Sprayed Coatings
Attenuation of higher frequency components of ultrasonic waves
A plate with 50Cr-50Ni coating, having an artificial delaminated part was prepared and normal beam method was examined from the coating surface. On comparing bottom echo of base metal in non-delaminated and delaminated parts, the echo height in delaminated part was found small. There may be discontinuity at the boundary between the coating and base metal in the delaminated part. Thermal sprayed coating is thin, echo from the boundary is buried in the transmitter pulse and noise in the neighborhood, and therefore it is difficult to find the echo from the boundary. However, delaminated and non-delaminated parts can be found from difference between echo heights. Echo height in the delaminated part may be small due to attenuation of higher frequency components of ultrasonic wave and it may be possible to evaluate adhesive strength on the basis of higher frequency components. The Fourier transform is defined as,
where F(w) and f(t) denote the Fourier transformed function and original function, respectively. Observed echo height h(t), whose unit [V], is proportional to sound pressure p. The unit of Fourier transform H(w) of h(t) is [Vt] and that of S(w), the area formed by H(w) and the frequency axis, is [V] ,the same as echo height h(t). This process transforms the sound pressure into frequency components. Sound pressure p is expressed as,
where x, p0, anda denote propagation distance of ultrasonic wave, initial sound pressure and attenuation constant, respectively. The attenuation constant a is proportional to the power of frequency w. With increase in frequency, sound pressure becomes smaller. Because of surface roughness and porosity of coating material, higher frequency components of ultrasonic wave attenuate. With discontinuity as in the case of delamination at the boundary, higher frequency components attenuate more. Frequency data of the first bottom echo and its multiple reflections for delaminated and non-delaminated parts, using probe with frequency broadband of 10MHz, are shown in Figure 1. Higher frequency components of ultrasonic waves, which pass through the sprayed coatings, attenuate and frequency components become approximately 0 at 10MHz. The spectrum in the delaminated part shows lower frequency components compared to the non-delaminated part. Adhesive strength of coating may thus be evaluated based on the spectral intensity S(w) at a certain frequency band.
Fig 1: Example of frequency spectrum of ultrasonic waves which pass through the boundary between the coating and base metal |
Experimental procedure and delamination tests
To use the present method to evaluate adhesive strength, test pieces with 50Cr-50Ni coatings were prepared by plasma spraying. Carbon steel plates 6mm in thickness were used as base metals. To vary adhesive strength of coatings, sprayed distances and blasting materials were varied when the test pieces were prepared. Nominal coating thickness was 450mm. To obtain ultrasonic test data, a normal longitudinal wave probe with frequency broadband of 10MHz was used. Delamination tests were then conducted.
The present method is mainly influenced by bottom surface conditions of base metal, since this method uses bottom echoes of the base metal. If the bottom surface of the base metal is rough due to corrosion and/or erosion, the echo becomes small. When evaluating only by bottom echo of the base metal, there is the misconception, "there are no-good parts". Spectral intensity becomes small, when the bottom echo of base metal is small. However, the ratio of spectral intensity at a certain frequency band to the all may be the same whether the bottom surface is rough or not. It may thus be possible to evaluate adhesive strength of coating from ratio of spectral intensity at a certain frequency band to the all spectral intensity. The ratio of intensity from 6MHz to 10MHz to that from 1MHz to 15MHz is shown in Figure 2. This ratio increased with adhesive strength and adhesive strength may be found from this ratio.
Fig 2: Ratio of intensity from 6MHz to 10MHz to that from 1MHz to 15MHz vs. adhesive strength |
Thickness Measurement of Thermal Sprayed Coatings
Experimental procedure and results
Test pieces whose coating thicknesses are 200mm, 300mm, 450mm, 600mm, 800mm and 1000mm were prepared by plasma spraying. To obtain ultrasonic waveform, a twin crystal, normal longitudinal wave probe with frequency broadband of 10MHz was used and ultrasonic wave was put perpendicular into the test piece from the coating surface. With a single crystal probe, it is hard to determine the echo from the boundary between the base metal and coating, since coating is thin and the boundary echo is buried in the transmitter pulse and noise in the neighborhood. To analyze ultrasonic waveforms easily, the twin crystal probe was used. Discrete wavelet transform was applied for analysis of ultrasonic waveforms and Daubechies wavelet of order 10 (db10) was used as the mother wavelet.
|  |
| Fig 3: Discrete wavelet transform of ultrasonic waveforms at coating thickness of 450mm |
Ultrasonic waveform and decomposed waveforms by db10 wavelet of the test piece with 450mm coating thickness are shown Figure 3. The waveform on the upper left is the original waveform S0. The bottom echo of the base metal appears later. S0 is decomposed to A1 and D1, A1 is decomposed to A2 and D2, and so on. S0 is thus written as,
Fig 4: Original waveforms S0 (left side) and A3 (right side) with change in coating thickness |
D1 is the highest frequency component from D1 to D6, and D2 is the half frequency component of D1. D3 is the quarter frequency component of D1. The peak at 4.5ms of A3 is a echo from the boundary between the coating and base metal. Test piece data are shown in Figure 4. Waveforms on the left side are the original waveforms and on the right side, A3 after decomposition. At 1000mm, a peak appears around 5ms and shifts to the time axis as thickness decreases. The correlation between peak time of the echo from the boundary and coating thickness and fitted by linear function are shown in Figure 5. The ultrasonic wave velocity of this coating material was found to be 2800m/s.
Fig 5: Correlation between echo peak time and coating thickness |
Applications of this method
The thickness of conventional material is determined based on time differences between the first bottom echo and second bottom echo when the ultrasonic wave velocity of the material is known. For coating thickness measurement, the coating is thin and correct ultrasonic wave velocity of the coating material is unclear. Moreover, the second boundary echo can hardly be observed. Thus, sprayed coating test pieces with known correct thickness must be used for measurement. By wavelet analysis, it is possible to measure the echo from the boundary between coatings and base metal easily since the signal-to-noise ratio of waveform can be enhanced.
An electromagnetic thickness meter can be used for the coating thickness measurement of non-magnetism coating material when the base metal has magnetism. With the present method, it is possible to measure coating thickness regardless of magnetism of coating materials and base metals.
Conclusion
Methods for evaluating coating adhesive strength and for measuring coating thickness by ultrasonic testing were established. Adhesive strength was evaluated based on spectral intensity at a certain frequency band of ultrasonic waveform. Spectral intensity at a certain frequency band increases with adhesive strength correspondingly. By this method, a coating can be shown to have weak adhesive strength and/or be thick. In coating thickness measurement, the twin crystal probe was used to lessen the influence of the transmitter pulse. To enhance signal-to-noise ratio of ultrasonic waveforms, the wavelet transform was used to analyze the ultrasonic waveform. The echo from the boundary between the coating and base metal could be readily determined. By combining these two methods, it will be possible to measure the thickness and adhesive strength simultaneously and evaluate the damage of thermal sprayed coatings.
References
- G. Chen, A. Yamaguchi and K. Miya,"Signal processing for detection and characterization of defects from composite eddy current signals", insight, Vol.39 No.11 (1997) 785-789.
- K. Namba, Y. Mizo and I. Kajigaya, "Boiler Application of Thermal Sprayed Coatings for High-Temperature Corrosion and Erosion Protection", Ishikawajima-Harima Engineering Review, Vol.38 No.3 (1998) 181-188.
- Y. J. Chen, Y. W. Shi and Y. P. Lei, "Use of a wavelet analysis technique for the enhancement of signal-to-noise ratio in ultrasonic NDE", Insight, Vol.38, No.11 (1996) 800-803.