Testing the impact of creep on the X12CrMoWVNbN steel magnetic properties

Transcription

1 11th European Conference on Non-Destructive Testing (ECNDT 214), October 6-1, 214, Prague, Czech Republic More Info at Open Access Database Testing the impact of creep on the X12CrMoWVNbN1-1-1 steel magnetic properties Maciej ROSKOSZ, Andrzej RUSIN, Krzysztof FRYCZOWSKI, Michał BIENIEK Institute of Power Engineering and Turbomachinery, Silesian University of Technology, Gliwice, Poland s: Abstract The Barkhausen effect and the hysteresis loop were measured on 15 mm long cylindrical X12CrMoWVNbN1-1-1 steel specimens with a diameter of 4 mm. An analysis was conducted of the changes that describe the effects under analysis quantitatively. New specimens (made of a forging in the as-delivered state) and post-creep specimens (destroyed due to creep) were compared. The aim was to find magnetic quantities for which the impact of creep-related changes in the mechanical properties and microstructure is the most visible. For the Barkhausen effect, the steel X12CrMoWVNbN1-1-1 creep processes result in a rise in the maximum value of the amplitude of the components of the Fast Fourier Transform (FFT) spectrum and the Barkhausen noise energy. Analyzing the hysteresis loop, drops in coercivity were found after the creep process was completed. The results of the testing are the basis for further research on the impact of creep processes on changes in magnetic properties. The ultimate object of the research is to determine the correlations between changes in values of magnetic parameters and the progress of creep. Keywords: Barkhausen noise, coercivity, creep, X12CrMoWVNbN1-1-1 steel 1. Introduction The X12CrWMoVNbN1-1-1 steel is an advanced high-chromium ferritic steel used as the rotor material in ultra-supercritical power plants. The steel is characterized by a tempered martensite matrix and it is strengthened by means of solution hardening of such elements as Mo and W and precipitation hardening of such phases as M 23 C 6 (M: Cr, Fe, Mo, W, etc.) type carbides and MX (M: V, Nb, etc. and X: C, N) type carbonitrides. In typical applications, the basic process contributing to the loss of durability of this steel, apart from variable loads, is creep. The use of the Barkhausen effect in the testing of wear processes is the subject of numerous studies [1-3]. Tests were performed to analyze the possibility of using the Barkhausen effect to assess the progress of creep in specimens made of the X12CrMoWVNbN1-1-1 steel. The testing results comprise the extreme states of the material: as-received state (new material) and post-creep state the state of the specimen destroyed during the creep test. If the results of analyses that describe the Barkhausen noise quantitatively are different for the two cases mentioned above, a basis will arise to develop a wider programme of testing the impact of creep on the Barkhausen effect. 2. Experimental details Three as-received specimens (marked as N1, N2 and N3) and three post-creep test specimens (marked as, and ) were analyzed. Details of the conditions prevailing during the creep tests for individual specimens are presented in Tab. 1. The specimens had the form of a 15 mm long cylinder with a 4 mm diameter. In the as-received state, the material was subjected to heat treatment according to the following specification: 165 o C / 4,75 h / polymer 68 o C / 4,5 h / air 69 o C / 8,75 h / air.

2 Tab. 1. Creep testing details Specimen No Creep test temperature [ C] Stress [MPa] Time to destruction [h] The Barkhausen noise was measured with the MEB4-C device, supplied by the company Mag-Lab s.c. from Gdańsk. The tests were performed using a coil in two configurations to take account of all basic measurement parameters listed in Tab. 2. The applied test stand is shown in Fig. 1. Tab. 2. Configurations of basic parameters of measurements Configuration C1 Configuration C2 sampling frequency f p [khz] 8 8 number of measuring points [thousand] magnetizing current amplitude [A] 2 2 magnetizing current change rate [A/s] max. value of discrimination voltage [V] 1 1 pre-amplifier gain 1-time 1-time main amplifier gain 2dB 35dB Fig. 1. Test stand (1 measuring coil, 2 magnetizing coil, 3 tested specimen, 4 core closing the magnetic circuit I-shaped, 5 core closing the magnetic circuit C-shaped, 6 support) In order to analyze the possibility of the creep process advancement assessment, the Barkhausen noise was subjected to quantitative analyses. The following were analyzed: the Barkhausen noise envelope, the impulse number of counts, the signal FFT spectrum, the Barkhausen noise energy and the magnetic hysteresis loop characteristic points.

3 3. Creep process impact on changes in microstucture The tested as-received steel has the structure of a high-tempered martensite with carbide formations (sub-grains) inside martensite blocks and on their boundaries (Fig. 2) with a small number of non-metallic oxide inclusions. Chemical composition analyses of the formations show that the carbides are characterized by a higher and lower content of coal and mutually dissolved chemical elements resulting from the steel chemical composition, which indicates that these carbides are probably of the M 23 C 6 type (where M Cr, Fe, Mo, W, V) and of the MC type (where M W, V, Nb), or the MX type (where X C, N). This is confirmed by the reference literature data [4, 5]. For post-creep specimens tested in the temperature of, the occurrence of a high-tempered martensite structure and a small number of non-metallic, mainly oxide, inclusions are found. The steel structure reveals small creep pores and voids with an irregular distribution. No significant changes in the high-tempered martensite structure morphology are found compared to the as-received state except an additional process of formation of tiny phases rich in W (about 15%) and Mo (about 14%), corresponding to the composition of the Laves phase of the (Fe,Cr) 2 (W,Mo) type. The Laves phase is formed predominantly on the grain and block boundaries, and locally only inside tempered martensite blocks. The phase is formed independently on the boundaries, or it is directly adjacent and bound to carbide M 23 C 6, on which its nucleation proceeds. For post-creep specimens tested in the temperature of 68 o C (Fig. 3) i 7 o C, the occurrence of a high-tempered martensite structure with carbide formations with chemical compositions corresponding to the M 23 C 6 and MC carbides, as well as the presence of creep pores and voids are found. The process of soaking in these temperatures caused a barely noticeable coagulation of carbides inside the blocks and a creation of a more continuous formation network on the boundaries. The appearance of carbides with a bigger size resulting from their partial growth is also found. Compared to the post-creep specimen tested in the temperature of 68 o C, the post-creep specimen tested in the temperature of 7 o C shows a more advanced process of carbide coagulation, where also the carbides on the block (sub-grain) boundaries became clearly coagulated. Fig. 2. Steel structure in as-received state; etched specimen, SEM image; within tempered martensite blocks and on the boundaries numerous tiny formations of carbides Fig. 3. Specimen 2 destroyed in a creep test: 68 o C / 12 MPa / 239 h; etched microsection; BSE image; high-tempered martensite with carbide formations inside blocks and on boundaries; locally bigger carbide formations and creep pores

4 4. Analysis of testing results 4.1. The Barkhausen noise envelope analysis Five measuring series were performed for each specimen. For each magnetization period, 2 RMS values were determined. For the RMS values, resultant magnetizing current values were found. The differences between the current values obtained in this manner for individual magnetization periods were of the order 1-4 A. Therefore, it is assumed that the timedependent distribution of the magnetizing current values is identical for all magnetization periods. This made it possible to find the arithmetic mean of the RMS values from 1 magnetization periods. For data obtained in this way, individual envelopes of the Barkhausen noise were plotted for each measuring series. Fig. 4 shows the Barkhausen noise envelopes made for the material in the as-received and post-creep states for Configuration C1. The Barkhausen noise envelopes for post-creep specimens differ from each other clearly, which may prove the impact of the creep process parameters on the Barkhausen noise. The envelopes of specimens in the as-received and postcreep states of the material practically coincide with each other and it is impossible to use them as diagnostic signals. Similar results were obtained for Configuration C2. 7 U BN [V] I m [A] Fig. 4. Barkhausen noise envelopes from 5 measuring series specimens (black); specimens,, (red) Configuration C Analysis of the number of counts The single impulse method allows an analysis of the Barkhausen effect with respect to the distribution of the amplitude of impulses for a certain range of changes in the strength of the external magnetic field H or the entire hysteresis loop. The simplest variant of this type of analysis is counting the impulses with an amplitude exceeding the set level of voltage. This requires a single-threshold discriminator and a counter controlled by a time gate. The simplest synthetic information obtained in this type of analysis is a single quantity, i.e. the number of impulses counted in one magnetization period. This quantity gives an idea of the number of domain walls that cross the pinning sites, and after the discrimination level is selected also an assessment of these pinning sites. For each specimen 3 measuring series with 1 magnetization periods each were performed and from them the mean value of the number of counts was determined for the set

5 discrimination voltage. Fig. 5 presents the distribution of the number of counts depending on discrimination voltage U d for specimens in the as-received and post-creep states. A detailed analysis was conducted of the value of the number of counts for individual discrimination voltage values. An example of the most visible differences is shown in Fig. 6. No voltage value was found for which clear differences that could be used as a diagnostic signal occurred in the number of counts for the as-received and post-creep material states. Similar results were obtained for Configuration C2. N BN U d [V] 32 N BN Fig. 5. Distribution of the mean number of the Barkhausen noise N BN depending on discrimination voltage U d Configuration C1 specimens N1, N2, N3 (black), specimens,, (red) Fig. 6. The number of counts of impulses for discrimination voltage U d =6V, Configuration C Analysis of changes in quantities describing the magnetic hysteresis loop The analysis comprised changes in two parameters characterizing the hysteresis loop: remanence induction B r and coercivity H C. Remanence induction B r is defined as the value of magnetic induction B at the external magnetic field H=. Coercivity H C is the value of the external magnetic field H at which magnetic induction assumes the value B=. Fig. 7 presents example magnetic hysteresis loops for the steel under analysis. Fig. 8a and Fig. 8b present the values of coercivity H C of the as-received and post-creep specimens for Configuration C1 and Configuration C2, respectively. The values differ clearly, which indicates a possibility of using them to identify the state of the creep process. However, no differences are observed between the values of magnetic remanence B r.

6 2 B [T] H [A/m] Fig. 7. Hysteresis loops for specimens and Configuration C1 1 H C [A/m] H C [A/m] a) Configuration C1 b) Configuration C2 Fig. 8. Values of coercivity H C of specimens made of as-received and post-creep state material 4.4. Analysis of changes in the Barkhausen noise energy The Barkhausen noise energy, which is also referred to as the Barkhausen noise intensity, is defined as the time integral of squared impulse voltage V i, and it is a parameter describing the impulse distribution and voltage in the duration time T of one magnetization period by means of a single value. E = V dt (1) Determined values of the Barkhausen noise energy are presented in Fig. 9a for Configuration C1 and in Fig. 9b for Configuration C2. Differences between the energy values for the material in the as-received and post-creep states occur for Configuration C1. In the case of Configuration C2, a clear difference can only be seen for Specimen.

7 3 E BN E BN a) Configuration C1 b) Configuration C2 Rys. 9. Values of the Barkhausen noise energy EBN of specimens made of as-received and post-creep state material 4.5. FFT spectrum analysis of the signal The Barkhausen noise was analyzed using the FFT. The FFT spectra obtained for the material in the as-received and post-creep states are compared in Fig. 1a and Fig. 1b. Clear differences can be seen between the values of the FFT spectrum amplitude for the specimens in the as-received state and for the specimens after the creep test in the range of 3 to 1 khz. 8 A FFT [mv] f [khz] 8 A FFT [mv] f [khz] a) frequency range -1 khz b) frequency range khz Fig. 1. Distribution of the FFT amplitude depending on frequency Configuration C1 specimens (black), specimens,, (red)

8 5. Conclusions The tests and the quantitative analysis of the Barkhausen effect and of the magnetic hysteresis loop characteristic points prove the possibility of assessing the progress of creep in specimens made of the X12CrMoWVNbN1-1-1 steel. It is possible to distinguish quantities for which the differences in the Barkhausen effect components between specimens in the as-received and in the post-creep test states may be used in future to develop a wider programme of testing the impact of the creep process on the Barkhausen effect. The biggest and unequivocal changes occurred for the maximum amplitude of the FFT spectrum components, coercivity H c and the Barkhausen noise energy. The analysis of the number of counts and of the Barkhausen noise envelope did not show an unequivocal possibility of differentiating between the two states of specimens. No differences were observed between the values of magnetic remanence B r. It was also shown that differences in the Barkhausen noise were more visible for a lower magnetization frequency. Acknowledgements The results presented in this paper were obtained from research work co-financed by the Polish National Centre for Research and Development within the framework of Contract SP/E/1/ 67484/1 Strategic Research Programme Advanced technologies for obtaining energy. Development of a technology for highly efficient zero-emission coal-fired power units integrated with CO 2 capture. References 1. J N Mohapatra, A K Ray, J Swaminathan, A Mitra, 'Creep behaviour study of virgin and service exposed 5Cr.5Mo steel using magnetic Barkhausen emissions technique', Journal of Magnetism and Magnetic Materials, Volume 32, Issue 18, September 28, Pages M J Sablik, B Augustyniak, L Piotrowski, 'Modeling incipient creep damage effects on Barkhausen noise and magnetoacoustic emission', Journal of Magnetism and Magnetic Materials, Volumes , Supplement, May 24, Pages E523 E525, Proceedings of the International Conference on Magnetism (ICM 23). 3. Dong-Won Kim, Dongil Kwon 'Quantification of the Barkhausen noise method for the evaluation of time-dependent degradation', Journal of Magnetism and Magnetic Materials, Volume 257, Issues 2 3, February 23, Pages Huiran Cui, Feng Sun, Ke Chen, Lanting Zhang, Rongchun Wan, Aidang Shan, Jiansheng Wu 'Precipitation behavior of Laves phase in 1%Cr steel X12CrMoWVNbN1-1-1 during short-term creep exposure', Materials Science and Engineering A 527 (21) G Götz, W Blum 'Influence of thermal history on precipitation of hardening phases in tempered martensite steel of type X12CrMoWVNbN1-1-1', Mater. Sci. Eng. A, 348/1-2:21 27, 23.