Evaluation of deformation behavior in LPSO-magnesium alloys by AE clusteringand inverse analysis.
Long-period Stacking ordered (LPSO) Mg alloys were attracted attention due to its high properties. The materials used in this study were unidirectional solidification (DS) [Mg.sub.85][Y.sub.6][Zn.sub.9], as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2] and as-cast [Mg.sub.85][Y.sub.6][Zn.sub.9]. Compression test and fracture toughness test were conducted with AE measurement. In compression test of DS [Mg.sub.85][Y.sub.6][Zn.sub.9], kink deformation was generated from the early deformation stage. The number of microcrack was increased from about half of yield stress and it was rapidly increased around the yield point. The clustering analysis was applied to the extracted AE events. The AE events were divided into two clusters. The observation results showed that cluster one was considered as the AE events generated by kink deformation and cluster two was considered as the AE events generated by microcrack. The inverse analysis was applied to the extracted AE events. The displacement of height generated by kink deformation was estimated in compression test. It is considered that the dynamic behavior of deformation and fracture in LPSO-Mg alloys was evaluated by AE method.
Keywords: Kink deformation, LPSO-Mg alloys, Clustering, Inverse analysis
A unique deformation mechanism in synchronizes long-period stacking ordered (LPSO) structures contributes to generate an excellent macroscopic mechanical property [1-6]. Analysis of dynamic deformation mechanism in nanometer scale is expected to develop new materials with better performance. Observation by measurement and analysis equipment such as transmission electron microscope is very useful to understand the deformation mechanism. However, these analyses are usually applicable to only surface of materials and basically give static images. Combination with other techniques will be effective to obtain dynamic mechanism in detail. Acoustic emission (AE) has been used to obtain dynamic information in mechanical deformation which is difficult to be analyzed using other methods.
AE is well known as an elastic wave in materials due to deformation, fracture and so on. Macroscopic mechanical properties of materials such as strength, elongation, fracture toughness, fatigue strength and corrosion characteristics are related to dynamic microscopic behaviors in materials which are mostly sources of AE. One of the advantages of using AE method is to globally monitor the deformation and fracture behaviors in samples with high sensitivity. Many researches have been done to understand the deformation or fracture behaviors in various materials including metals, ceramics, polymers and composites by analysis of AE behaviors. However, electrical or mechanical noises sometime interfere in the effective AE signals related fracture behaviors. Recently our research group has developed a novel AE analysis system with continuous recording function of AE waveforms in order to avoid noises and extract useful information related to deformation behaviors in materials . Our system also includes advanced waveform analysis methods such as wavelet transform with high frequency resolution and accurate location algorithm by Akaike information criteria (AIC) picker.
AE parameters such as amplitude, energy, duration etc. are used to identify AE sources and explain the dynamic evolution of deformation and fracture behaviors in materials. Also clustering by AE parameters gives better understanding of aspect of deformation in material testing. However, it is usually very difficult to identify dislocation movement with continuous type AE signals. Recently novel cluster analysis of AE signals using frequency components has been successfully applied to understand the evolution of plastic deformation in Mg alloys [8-9]. In this paper, mechanical properties of LPSO-Mg alloys with different microstructures were investigated using tensile test, compression test and fracture toughness test. Two materials with different composition, [Mg.sub.97][Zn.sub.1][Y.sub.2] and [Mg.sub.85][Zn.sub.6][Y.sub.9], were used for experiments in this paper. [Mg.sub.97][Zn.sub.1][Y.sub.2] has a volume fraction of [alpha]-Mg phase with 75% and 18R LPSO-Mg phase with 25%, and [Mg.sub.85][Zn.sub.6][Y.sub.9] demonstrates a fully18R LPSO-Ma phase. AE signals during tensile, compression and fracture toughness tests were recorded and analyzed. Frequency analysis for each AE event suggested that deformation in [alpha]-Mg phase generated a higher frequency signal and deformation in LPSO-Mg phase a generated a lower frequency one. Evolution of deformation mechanism during material tests was also analyzed by clustering techniques using frequency components.
2. Experimental and Analytical Procedures
The materials used in this study were unidirectional solidification (DS) [Mg.sub.85][Y.sub.6][Zn.sub.9], as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2] and as-cast [Mg.sub.85][Y.sub.6][Zn.sub.9]. The surfaces of the materials were observed by a scanning electron microscope (SEM, JSM-6510LA, JEOL). The parallel section surface of specimens of unidirectional solidification [Mg.sub.85][Y.sub.6][Zn.sub.9] was shown in Fig. 1(a), vertical section was shown in Fig. 1(b), as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2] was shown in Fig. 1(c) and as-cast [Mg.sub.85][Y.sub.6][Zn.sub.9] was shown in Fig. 1(d). The LPSO phase in DS [Mg.sub.85][Y.sub.6][Zn.sub.9] was grown to the parallel direction to solidification direction. DS [Mg.sub.85][Y.sub.6][Zn.sub.9] and as-cast [Mg.sub.85][Y.sub.6][Zn.sub.9] had LPSO single phase, as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2] had LPSO phase and [alpha]-Mg phase. The volume fraction of LPSO phase was about 25% and [alpha]-Mg phase was about 75% in as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2]. The DS [Mg.sub.85][Y.sub.6][Zn.sub.9] specimens of compression test were prepared. The size of compression test specimen was 5x5x5 mm with cube shape. The as-cast [Mg.sub.97][Y.sub.1][Zn.sub.2] and as-cast [Mg.sub.85][Y.sub.6][Zn.sub.9] specimens of fracture toughness test were prepared.
[FIGURE 1 OMITTED]
Compression test was conducted at constant crosshead speed of 5x[10.sup.-4] mm/s by a universal testing machine (AG-5000C, Shimazu Co.) (Fig. 2(a)). The load direction was parallel to solidification direction. The surface observation was conducted by a video microscope (VHX-600, KEYENCE Co.) or high-speed camera (HPV-X2, Shimazu Co., Kirana, Specialised Imaging Ltd., Phantom Miro M110, vision Research Co.) during compression test. The maximum frame rate of high-speed camera was 5 Mfps. The surface observation was performed by SEM and a laser microscope (LEXT, OLYMPUS Co.) after the compression test. The height and size of kink deformation were measured by a laser microscope. Four-point-bending tests were carried out at a constant displacement rate of 8.33x[10.sup.-3] mm/s with AE measurement. During the tests, the crack path was observed by microscope. The positions of AE sensors were sides of specimens (Fig. 2(b)). AE sensors were AE254SMH177 (Fuji ceramics, Japan). Continuous Waveform Memory (CWM), which was developed by our group, was used to perform AE measurement. The high-pass filter was set at 100 kHz. The RMS voltage was measured with 200 ms time constant.
AE signal was measured by continuous wave memory (CWM) during both tests. The sampling rate was 10 MHz, and the range of measurement was 5 V. Two AE sensors (M304A, Fuji Ceramics Co.) were used in compression test. The AE signal was inputted for the trigger of the high-speed camera in order to observe the generation of kink deformation. The high pass filter of 100 kHz was applied to the AE signal for removing mechanical noise. AE events were extracted from filtered AE wave measured in each tests. The energy spectrum of the AE events was calculated by fast Fourier transform (FFT). The FFT length was 512 samples near the rising point of each AE events. The energy values of each 19.55 kHz were extracted and Savizky-Golay filter was applied to these energy values for smoothing the energy spectrum. The all energy values in the filtered spectrum were divided by the maximum energy values of filtered spectrum in each AE events to reduce the effect of AE amplitude. Normalizing was applied to the all energy values in the spectrum.
Principal component analysis (PCA) was conducted to the all energy values in the spectrum. Clustering analysis was applied by using calculated several principal components of each AE events. K-means method was used for clustering analysis. The value of k in compression test was decided by average silhouette coefficient. Average silhouette coefficient was calculated. High average silhouette coefficient means that the data of same cluster was gathered at short distance and different clusters were far away with each other. The k value at maximum average silhouette coefficient was used for analysis.
[FIGURE 2 OMITTED]
3. Results and Discussion
The specimens were compressed to eleven different stress levels in compression tests of DS specimens. The part of generation of kink deformation showed low brightness (Fig. 3). Kink deformation fraction at each stress levels was calculated by using this difference of brightness. Kink deformation was started to generate from low stress level and this fraction rapidly increased near the yield point. The specimen surface was divided roughly, and the number of grains introduced kink deformation was calculated. The number of grain introduced kink deformation rapidly increased in low stress levels and stabilized after rapid rise. First kink deformation generated at each grain in early deformation stage represented by white-colored, and further kink deformation generated at each grain already introduced kink deformation in later deformation stage represented by grey-colored.
In the compression tests, the cumulative contribution rate was over 0.7 where the number of principal components was 3. Average silhouette coefficient was calculated at k value from two to seven in this principal components, and it showed maximum value at k = 2. The clustering analysis was applied to the AE events in compression test at k = 2. Fig. 4(a) shows the average energy spectrum of each cluster. The energy of cluster one showed high value at low frequency, and that of cluster two showed high value at high frequency. The cumulative number of AE events in each cluster was calculated (Fig. 4(b)). AE events of cluster one were measured from low stress level and dominant under 100 MPa. On the other hand, AE events of cluster two were increased around 100 MPa, and the specimen was yielded at the time of rapid increase of cluster two.
[FIGURE 3 OMITTED]
[FIGURE 4 OMITTED]
The discussion about the classification of AE sources by clustering analysis is effective for understanding the deformation behavior. The kink fraction behavior coincided well with the AE events behavior of cluster one in early deformation stage, and it did not corresponded with the AE events behavior of cluster one at all in later deformation stage. The kink fraction was rapidly increased and few AE events of cluster one were measured in later deformation stage. On the other hand, the number of grains introduced kink deformation accorded with the AE events behavior of cluster one through the compression test. Our previous papers demonstrated that the first twin deformations in each grains emitted high AE energy, and the twin deformations after the first ones emitted low AE energy in the tensile test of rolled AZ31 [10-15]. As well as twin deformation, it was suggested that the kink deformations in early deformation stage emitted high AE energy and the kink deformations in later deformation stage emitted low AE energy. Therefore, the AE source of cluster one was considered as the kink deformation generated in early deformation stage. The microcrack behavior coincided with the AE events behavior of cluster two. The AE source of cluster two was considered as the microcrack. AE event generated by kink deformation tended to have high energy at high frequency and AE event generated by microcrack tended to have high energy at low frequency.
In this study, compression test and fracture toughness test were conducted for the understanding of deformation behavior in DS LPSO-Mg alloys. The following conclusions were drawn.
(1) Kink deformation was generated from the early deformation stage. The number of microcrack was increased from about half of yield stress and the rapid increase of microcrack caused the yield in compression test of DS [Mg.sub.85][Y.sub.6][Zn.sub.9].
(2) The classification of AE sources by clustering analysis is effective for the evaluation of the dynamic behavior of kink deformation and microcrack.
This work was supported by a Grant-in-Aid for Scientific Research on Innovative Areas, "Synchronized Long-Period Stacking Ordered Structure", from the Ministry of Education, Culture, Sports, Science and Technology, Japan (No.23109001). The Mg alloy samples used in the present study were supplied by Kumamoto University and Osaka University.
[1.] Y. Kawamura, K. Hayashi, A. Inoue and T. Masumoto: Mater. Trans., 42 (2001) 1172-1176.
[2.] Inoue, M. Matsushita, Y. Kawamura, K. Amiya, K. Hayashi and J. Koike: Mater. Trans. 43 (2002) 580-584.
[3.] T. Itoi, T. Semiya, Y. Kawamura and M. Hirohashi: Scr. Mater. 51 (2004) 107-111.
[4.] E. Abe, Y. Kawamura, K. Hayashi and A. Inoue: Acta Mater. 50 (2002) 3845-3857.
[5.] M. Matsuda, S. Ii, Y. Kawamura, Y. Ikuhara and M. Nishida: Mater. Sci. Eng. A, 393 (2005) 269-274.
[6.] X. H. Shao, Z. Q. Yang and X. L. Ma: Acta Mater. 58 (2010) 4760-4771.
[7.] K. Ito and M. Enoki: Mater. Trans. 48 (2007) 1221-1226.
[8.] E. Pomponi and A. Vinodradov: Mechan. Systems Signal Process. 40 (2013) 791-804.
[9.] A. Vinogradov, D. Orlov, A. Danyuk and Y. Estrin: Mater. Sci. Eng. A 621 (2015) 243-251.
[10.] Y.P. Li and M. Enoki: Mater. Trans. 48 (2007) 1215-1220.
[11.] Y.P. Li and M. Enoki: Mater. Trans. 48 (2007) 2343-2348.
[12.] Y.P. Li and M. Enoki: Mater. Trans. 49 (2008) 1800-1805.
[13.] Y.P. Li and M. Enoki: Mater. Sci. Eng. A 536 (2012) 8-13.
[14.] T. Yasutomi and M. Enoki: Mater. Trans. 53(2012) 1611-1616.
[15.] Y. Muto, K. Matsumoto, T. Shiraiwa and M. Enoki: J. Japan Inst. Metals, 78 (2014) 381-387.
Manabu ENOKI (1), Yuki MUTO (1), Takayuki SHIRAIWA (1)
(1) Department of Materials Engineering, The University of Tokyo; Tokyo, Japan Phone: +81 3 5841 7126, Fax: +81 3 5841 7181;
|Printer friendly Cite/link Email Feedback|
|Author:||Enoki, Manabu; Muto, Yuki; Shiraiwa, Takayuki|
|Publication:||Journal of Acoustic Emission|
|Date:||Jan 1, 2016|
|Previous Article:||Monitoring of hydrogen assisted SCC on martensitic stainless steel by acoustic emission technique.|
|Next Article:||Measurement of electromagnetic emission as a tool to study fracture processes of carbon fibre reinforced polymers.|