Journal Information
Vol. 7. Num. 3.July - September 2018
Pages 203-402
Download PDF
More article options
Vol. 7. Num. 3.July - September 2018
Pages 203-402
Original Article
DOI: 10.1016/j.jmrt.2018.05.014
Open Access
Characterization of the aging state of modified HP steels by ultrasonic signal processing and the effect of creep voids in the interdendritic region
Natalie C. de Siqueiraa,b,
Corresponding author

Corresponding author.
, Monica P. Arenasa,b, Priscila D. de Almeidab, Leonardo S. Araújoa, Carlos B. Ecksteinc, Laudemiro Nogueirac, Luiz H. de Almeidaa, Gabriela R. Pereiraa,b
a Department of Metallurgical and Materials Engineering, Federal University of Rio de Janeiro, COPPE-UFRJ, Rio de Janeiro, RJ 21941-972, Brazil
b Laboratory of Non-Destructive Testing. Corrosion and Welding (LNDC), UFRJ, Rio de Janeiro, RJ 21941-596, Brazil
c Petrobras, Rio de Janeiro, Brazil
This item has received

Under a Creative Commons license
Article information
Full Text
Download PDF
Figures (4)
Show moreShow less
Tables (2)
Table 1. Aging state and operating temperature.
Table 2. Ultrasonic longitudinal velocity, Young's modulus, and acoustic impedance of analyzed samples.
Show moreShow less

HP austenitic stainless steels are extensively used in the oil and gas industry as radiant tubes in reformer and pyrolysis furnaces. These steels exhibit high resistance to oxidation, thermal stability, and creep resistance at elevated temperatures.

Their microstructure changes as a function of temperature, leading to different aging states. In addition, creep voids can be formed due to furnace operating conditions such as time and stress. Therefore, the characterization of the microstructural conditions is necessary for service life assessment. This study aims at characterizing two modified HP steel samples with different aging states and volumetric fraction of creep voids via scanning electron microscopy and ultrasonic testing. Fast Fourier transforms were applied to the second backwall echo of the acquired ultrasonic signal. Spectral analysis established a correlation between the microstructural conditions and the signal response. According to the methodology developed, the ultrasonic parameters and spectral analysis results were sensitive to microstructural changes caused by aging and the presence of creep voids.

HP austenitic steels
Creep voids
Ultrasonic testing
Spectral analysis
Ultrasonic parameters
Full Text

Centrifugally cast modified HP stainless steels are often used as radiant tubes in steam reformer and pyrolysis furnaces. These tubes are designed to operate under extreme conditions, such as high internal pressure at elevated temperatures for more than 100,000h [1].

HP steels have certain advantages over nickel-based super steels, such as competitive pricing, thermal stability, resistance to oxidation at elevated temperatures, and high creep strength [2]. Steam reformer tubes are subjected to different temperatures along their length, causing microstructural changes. A more detailed study of the temperature profile along reformer furnace tubes has been conducted by Le May et al. [1]. Each operating temperature range has been classified [3] based on its microstructural features. An as-cast HP-Nb-modified steel shows an austenitic matrix decorated with complex primary interdendritic carbide precipitates. Below 650°C, the microstructure is stable. However, when exposed to temperatures between 650°C and 1000°C, the material ages. This phenomenon is characterized by the coalescence of primary carbides and extensive secondary interdendritic precipitation of chromium-rich carbides [4]. Furthermore, at temperatures above 700°C, niobium-rich carbides (NbC) initiate an in situ transformation to the G phase, a Ni-Nb-based silicide with stoichiometry of Ni16Nb6Si7[4,5]. Previous studies [1,2,6] state that creep voids are preferentially formed at the G phase/matrix interface. When these voids are interconnected, microcracks are formed, which may lead to creep failure. This is one of the most common types of breakdown observed in steam reformer tubes.

Ultrasonic testing (UT) is commonly used for the detection of defects and microstructural evaluation, as wave propagation is significantly affected by the presence of different phases. Some ultrasonic parameters can be measured by UT, such as yield strength, fracture toughness, grain size, ultrasonic velocity, and attenuation [7–9]. Ultrasonic data are usually presented in the time domain; however, for complex analyses, such as the evaluation of microstructural changes, the extraction of information in the frequency domain may be necessary [7]. Spectral analysis and Fast Fourier Transform (FFT) have been widely applied to ultrasonic signals for materials characterization [7,10].

Birring et al. [11–13] evaluated the influence of creep damage degradation in samples exposed to hydrogen attack through ultrasonic velocity measurements. All cases presented a decrease in the ultrasonic velocity: longitudinal, shear, surface, and creeping waves showed sensitivity to this type of damage. Other studies [14–16] also demonstrated that creep damage can be characterized by a decrease in ultrasonic velocity and increase in signal attenuation. Kruger et al. [17] used spectral analysis to evaluate cast-iron samples and found that materials showing higher attenuation had lower amplitude spectra. In another study [18], hydrogen damage was determined through a decrease in the spectral amplitude.

These studies showed that UT, relying on spectral analysis, is sensitive to small microstructural variations. Thus, the present study aims at characterizing two different aging states of a modified HP steel with creep voids using scanning electron microscopy (SEM) and non-destructive UT. Ultrasonic signals were analyzed in the time domain using the cross-correlation function and in the frequency domain using FFT. Longitudinal wave velocities, Young's moduli, and acoustic impedance values were also measured.

2Experimental method2.1Microstructure

Two samples of a modified HP-Nb steel were extracted from a reformer tube operated for over 77,000h. The working pressure in the reformer furnace was kept constant at approximately 5MPa [19]. The samples were exposed to different operating temperatures; therefore, they presented different aging states, as described in Table 1. The inner diameter and wall thickness of the tube were 112mm and 11.50mm, respectively.

Table 1.

Aging state and operating temperature.

Aging state  Operating temperature 
IV  800–900°C 

Crosssectional metallographic samples were extracted from the tube for further SEM analyses in backscattered electron mode (BSE). The samples were ground up to 1500 mesh using emery paper, polished with 1μm diamond paste, followed by ultrasonic cleaning for 5min in alcohol. The microstructural characterization was performed with a VEGA3 TESCAN system at an accelerating voltage of 20kV. The public-domain ImageJ® digital image processing software was used to quantify the volumetric fraction of chromium and niobium carbides, G phase, and creep voids [20].

2.2Ultrasonic testing

UT was performed via the pulse echo technique (Sonotron Isonic 2005) using a single non-focused probe (Imasonic®) with a nominal frequency of 1.6 MHz. Testing was performed in contact mode with honey as the coupling material. A Tecktronix MSO 4034 oscilloscope set at 250 MA/s was used for signal acquisition. For each aging state, 30 signals were obtained over an area of 40mm×120mm. This procedure was repeated six times to ensure technical validation. Matlab® was then used to post-process the collected data. Ultrasonic signals were cross-correlated [21] to correct the phase shift between the data acquired for each sample. Longitudinal ultrasonic velocities, Young's moduli, and acoustic impedance values were quantified via the cross-correlation function.

The average signal for each sample was used as its single ultrasonic signal in the time domain. The DC level (zero frequency) was removed from each sample's signal. Signal analyses were performed on the second backwall echo using 512-point windowing. FFT was applied to evaluate the signals in the frequency domain [21]. Some ultrasonic parameters, such as velocity and acoustic impedance, were calculated based on the ASTM E494-10 standard [22]. Young's modulus and acoustic impedance were calculated using indexed values of density and Poisson's modulus [23].


The microstructure of the aging state IV sample (Fig. 1a) is composed of blocky primary Cr-rich carbides, secondary intradendritic precipitates of Cr-rich carbides (dark), and white and gray regions identified as the NbC and G phases, respectively. In this case, a small volume fraction of creep voids can also be observed in the interdendritic regions. On the other hand, the samples exposed to temperatures of the order of 900°C presented a microstructure typical to aging state V and a higher volume fraction of creep voids. The NbC phase is almost completely transformed to the G phase [4] and aligned voids are located in the interdendritic region (Fig. 1b).

Fig. 1.

SEM micrographs in BSE mode of samples with different aging states: (a) IV and (b) V.


Thirty micrographs were taken for each aging state to obtain the volumetric fraction of the phases present. The procedure to evaluate these images was based on the ASTM E1382 standard [20]. Volumetric fractions of the NbC and G phases were counted as one due to their close gray levels. Volumetric fractions of the Cr-rich carbides and voids were determined separately due to their significantly different gray levels. It was observed that for aging state V, the volumetric fraction of chromium carbides, G phase, and voids increased, as shown in Fig. 2. In contrast, the volumetric fraction of the niobium carbides decreased due to the in situ transformation of the NbC phase to the G phase [4].

Fig. 2.

Bar diagram showing the volumetric fraction of phases present in each sample.


Signal processing was performed for the second backwall echo of the analyzed samples. Fig. 3 shows the superposition of the data in the time domain for aging states IV and V. The aging state IV sample signal showed a higher amplitude than the aging state V sample signal. The signal attenuation may be explained by the presence of creep voids and phases, which contributed to the loss of energy due to scattering [24]. In this study, Rayleigh scattering (λD) was predominant as the incident wavelengths (λ) in the material were much greater than the average scattered sizes (D) [25]. Fig. 2 shows that the volumetric fraction of voids in the aging state V sample was higher, justifying the increase in attenuation of the ultrasonic waves [15,16,26].

Fig. 3.

Time domain signal of second backwall echo in HP steel samples. A higher amplitude is observed for state IV and a larger attenuation for state V.


Calculated ultrasonic velocities, Young's moduli, and acoustic impedance values are shown in Table 2. The samples with aging states IV and V presented ultrasonic velocities of 6011.54±19.37 and 5922.73±13.52m/s, respectively. The lower longitudinal velocity in the aging state V sample was due to the presence of creep voids [27]. These contributed to a decrease in the Young's modulus and acoustic impedance.

Table 2.

Ultrasonic longitudinal velocity, Young's modulus, and acoustic impedance of analyzed samples.

Aging state  Ultrasonic longitudinal velocity[m/s]  Young's modulus [MPa×103Acoustic impedance [kg/m2s×106
IV  6011.54±19.37  211.01±1.57  47.25±0.15 
5922.73±13.52  204.88±1.07  46.55±0.11 

Thus, the ultrasonic parameters presented higher values in the aging state IV sample. In contrast, a decrease in the values of these parameters was observed in the aging state V sample due to the phases present and the higher volumetric fraction of creep voids. These results agree with previous studies, which show that the ultrasonic velocity is lower in the presence of creep damage [6,28,29].

In accordance with Papadakis et al. [30], another method to assess a material's microstructure is spectral analysis. This process enables the observation of changes in the frequency domain according to variations in the microstructural conditions [12,31–33].

Spectral analysis was performed on the second backwall echo. The central frequency was found to be 1.465MHz for both samples, in accordance with the transducer's bandwidth. The spectral amplitude for each sample was influenced by its microstructural conditions. The mean frequency spectra for the two samples and their standard deviations can be observed in Fig. 4 The samples in state V had lower amplitude than those in state IV.

Fig. 4.

Second backwall echo spectra of state IV and V samples (mean±standard deviation (σ)).


Thus, UT could clearly differentiate between aging states IV and V. The state IV sample exhibited a higher amplitude in the time domain and spectral analysis, as shown in Figs. 3 and 4, respectively. In contrast, the state V sample presented lower amplitudes due to the attenuation of the ultrasonic signal in the presence of creep voids and phases. In the present study, a variation in the phase morphology was observed (Fig. 1). On the other hand, the volumetric fraction of voids, the most significant factor affecting the propagation of ultrasonic waves, increased, thus reducing the longitudinal velocity and signal amplitude. The presence of phases and voids differently affected the ultrasonic velocity, since velocity tends to be higher on the initial aging states [14] but decreases in the presence of voids [12,14–16]. Previous studies have widely reported that signal attenuation increases and longitudinal velocity decreases in the presence of voids and microcracks [4–16].


Non-destructive UT of HP stainless steel samples established a correlation between signal response and microstructural changes. The ultrasonic parameters and FFT applied to the second backwall echo were strongly affected by microstructural changes caused by aging and the presence of creep voids. The signal amplitudes were attenuated due to changes in the phases present and increase in the volumetric fraction of creep voids.

Conflicts of interest

The authors declare no conflicts of interest.


The authors acknowledge CNPq, FAPERJ, and FINEP for financial support; PETROBRAS for supplying the samples; and the team of Non-Destructive Testing LNDC/COPPE/UFRJ for scientific support.

I. Le May, T.L. Silveira da, C.H. Vianna
Criteria for the evaluation of damage and remaining life in reformer furnace tubes
Int J Press Vessel Pip, 66 (1996), pp. 233-241
L.H. de Almeida, A.F. Ribeiro, I. Le May
Microstructural characterization of modified 25Cr–35Ni centrifugally cast steel furnace tubes
Mater Charact, 49 (2003), pp. 219-229
G.D. Barbabela, L.H. de Almeida, T.L. da Silveira, I. Le May
Role of Nb in modifying the microstructure of heat-resistant cast HP steel
Mater Charact, 26 (1991), pp. 193-197
G.D. Barbabela, L.H. de Almeida, T.L. da Silveira, I. Le May
Phase characterization in two centrifugally cast HK stainless steel tubes
Mater Charact, 26 (1991), pp. 1-7
R.A. Pedro Ibañez, G.D. de Almeida Soares, L.H. de Almeida, I. Le May
Effects of Si content on the microstructure of modified-HP austenitic steels
Mater Charact, 30 (1993), pp. 243-249
G. Sposito, C. Ward, P. Cawley, P.B. Nagy, C. Scruby
A review of non-destructive techniques for the detection of creep damage in power plant steels
NDT E Int, 43 (2010), pp. 555-567
S.E. Krüguer, J.M.A. Rebello, P.C. De Camargo
Hydrogen damage detection by ultrasonic spectral analysis
NDT E Int, 32 (1999), pp. 275-281
A. Vary
3rd ed., ASNT Handbook, (2007)pp. 588
P. Palanichamy, A. Joseph, T. Jayakumar, B. Raj
Ultrasonic velocity measurements for estimation of grain size in austenitic stainless steel
NDT Int, 28 (1995), pp. 179-185
A. Kumar, T. Jayakumar, B. Raj
Ultrasonic spectral analysis for microstructural characterization of austenitic and ferritic steels
Philos Mag A, 80 (2000), pp. 2469-2487
A.S. Birring, D.G. Alcazar, J.J. Hanley, S. Gehl
Detection of creep damage by ultrasonics
Rev Prog Quant Nondestruct Eval, (1989), pp. 1833-1840
A.S. Birring
Nondestructive detection of material degradation caused by creep and hydrogen attack
Mater Res Soc Symp Proc, 142 (1989), pp. 119-130
Materials Research Society
A.S. Birring, M.L. Bartlett, K. Kawano
Ultrasonic detection of hydrogen attack in steels
Corrosion, 45 (1989), pp. 259-263
H. Willems
Investigation of creep damage in alloy 800H using ultrasonic velocity measurements, Nondestruct. Charact. Mater. II
M. Hirao, T. Morishita, H. Fukuoka
Ultrasonic velocity change with creep damage in copper
Metall Trans A, 21 (1990), pp. 1725-1732
C.M. Sayers, M. Hirao, T. Morishita
Grain boundary compliance and ultrasonic velocities in pure copper undergoing intergranular creep
Int J Solids Struct, 297 (2016), pp. 130-131
S.E. Kruger, W. Gao, C. Glorieux, J. Charlier, J.M.A. Rebello
(1998)pp. 312-314
S.E. Kruger, J.M.A. Rebello, P.C. De Camargo
Hydrogen damage detection by ultrasonic spectral analysis
NDT E Int, 32 (1999), pp. 275-281
T.L. da Silveira, I. Le. May
Reformer furnaces: materials, damage mechanisms, and assessment
Arab J Sci Eng, 31 (2006), pp. 99-119
ASTM E1382
Standard test methods for determining average grain size using semiautomatic and automatic image analysis
A.V. Oppenheim, R.W. Schafer, J.R. Buck
Discrete-time signal processing
2nd ed., Prentice Hall, (1999)
ASTM E494-10
Standard practice for measuring ultrasonic velocity in materials
ASTM Stand, (2011), pp. 1-15
S.F.S. America
Steel casting handbook, high alloy data sheets heat series, Supplement 9
N.K. Batra
Ultrasonic spectral analysis for characterization of composites
Rev Prog Quant Nondestruct Eval, (1989), pp. 1527-1533
E.P. Papadakis
Ultrasonic attenuation caused by scattering in polycrystalline media
J Acoust Soc Am, 37 (1965), pp. 711-717
H. Willems, G. Dobmann
Early detection of creep damage by ultrasonic and electromagnetic techniques
Nucl Eng Des, 128 (1991), pp. 139-149
J.S.C. Filho, D.B. Baroni, C.A. Lamy, M.S.Q. Bittencourt, M.S. Motta, J. Su
Neural network to measure volumetric fraction of a bubbly column using ultrasonic wave. COBEM 2009
20th International Congress of Mechanical Engineering,
B. Raj, B.K. Choudhary, R.K. Singh Raman
Mechanical properties and non-destructive evaluation of chromium molybdenum ferritic steels for steam generator application
Int J Press Vessel Pip, 81 (2004), pp. 521-534
T. Ohtani, H. Ogi, M. Hirao
Acoustic damping characterization and microstructure evolution in nickel-based superalloy during creep
Int J Solids Struct, 42 (2005), pp. 2911-2928
E.P. Papadakis, K.A. Fowler, L.C. Lynnwort
Ultrasonic attenuation by spectrum analysis of pulses in buffer rods: method and diffraction corrections
J Acoust Soc Am, 53 (1973), pp. 1336-1343
F.M. Sears, B.P. Bonner
Ultrasonic attenuation measurement by spectral ratios utilizing signal processing techniques
IEEE Trans Geosci Remote Sens, GE-19 (1981), pp. 95-99
N. Lunn, S. Dixon, M.D.G. Potter
High temperature EMAT design for scanning or fixed point operation on magnetite coated steel
NDT E Int, 89 (2017), pp. 74-80
W. Sachse
Ultrasonic pulse spectroscopy of a solid inclusion in an elastic solid
J Acoust Soc Am, 56 (1974), pp. 113-116
NOTICE Undefined index: identificador (includes_ws_v2/librerias/html/item.php[1194])

Paper was part of technical contributions presented in the events part of the ABM Week 2017, October 2nd to 6th, 2017, São Paulo, SP, Brazil.

Copyright © 2018. Brazilian Metallurgical, Materials and Mining Association
Journal of Materials Research and Technology

Subscribe to our newsletter

Article options
Cookies policy
To improve our services and products, we use cookies (own or third parties authorized) to show advertising related to client preferences through the analyses of navigation customer behavior. Continuing navigation will be considered as acceptance of this use. You can change the settings or obtain more information by clicking here.