Multi-scale Entropy Feature Extraction of Magnetic Signal and Observation of Magnetic Domain in Tension of Carbon Steel Specimen
-
摘要: 为揭示碳钢试样拉伸损伤过程磁信号多尺度特征, 选取45钢试样开展准静态拉伸实验, 提取磁场强度法向分量Hp(y)和磁场强度梯度K的多尺度熵特征, 结合上述特征, 选取支持向量机方法构建了损伤评估模型。利用原子力显微法观测各拉伸损伤阶段的磁畴形貌, 提取了磁畴相位角特征。结果表明: 各拉伸损伤阶段Hp(y)均存在对应于隐性损伤区域的零值点, 随着损伤程度增加, Hp(y)绝对值增大, K值曲线在Hp(y)零值点处出现峰值。各阶段Hp(y)多尺度熵值随尺度因子的增加呈上升趋势。随着损伤程度的增加, Hp(y)多尺度熵值降低, 而磁畴相位角呈逐渐上升趋势, 颈缩和断裂阶段Hp(y)多尺度熵值再次升高, 磁畴相位角则开始下降, 所构建的损伤评估模型具有83.3%的损伤识别精度。本文成果可为铁磁构件在役检测及损伤评估提供方法模型支撑。Abstract: In order to reveal the multi-scale characteristics of the magnetic signal in tension of carbon steel specimens, 45 steel specimens were selected to carry out quasi-static tensile experiments, and the multi-scale entropy characteristics of the normal component Hp(y) and the magnetic field intensity gradient K in the magnetic field were extracted. Combining the above features, the support vector machine method was selected to construct a damage assessment model. The magnetic domain morphology at each tensile damage stage was observed by atomic force microscopy, and the phase angle characteristics of the magnetic domain were extracted. Results show that there exists a zero point corresponding to the hidden damage area of Hp(y) at each tensile damage stage. As the damage degree increases, the absolute value of Hp(y) increases, and the K value curve has a peak at the zero point of Hp(y). The multi-scale entropy of Hp(y) at each stage increases with the increasing of scale factor. As the damage degree increases, the Hp(y) multi-scale entropy value decreases, while the phase value of the magnetic domain increases. The Hp(y) multi-scale entropy value increases again in the necking and fracture stage, and the phase value of the magnetic domain begins to decrease. The damage recognition accuracy of the damage assessment model is 83.3%. The results can provide method and model support for the in-service inspection and damage assessment of ferromagnetic components.
-
表 1 45钢试样元素质量分数
ω(Mn) ω(Cr) ω(Co) ω(P) ω(S) ω(C) ω(Si) ω(Cu) ω(Ni) ω(Fe) 0.66% 0.21% 0.081% 0.019% 0.025% 0.45% 0.22% 0.18% 0.21% Bal. 表 2 部分磁特征样本数据
特征参数 初始阶段 弹性阶段 屈服阶段 强化阶段 颈缩阶段 断裂阶段 |Hp(y)|max 43 134 163 194 339 394 Kmax 0.387 88 0.703 03 0.903 03 1.018 18 1.521 21 13.787 88 MSE(Hp(y)) MSE(τ=1) 0.030 4 0.013 5 0.008 3 0.007 4 0.005 4 0.027 8 MSE(τ=2) 0.045 2 0.014 7 0.009 4 0.008 7 0.009 0.054 9 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ MSE(τ=10) 0.135 1 0.029 2 0.013 2 0.02 0.042 4 0.064 5 MSE(K)) MSE(τ=1) 0.860 3 0.971 8 0.717 8 0.584 2 0.461 9 0.057 1 MSE(τ=2) 1.092 5 1.145 2 0.897 3 0.751 3 0.660 6 0.094 1 ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ ⋮ MSE(τ=10) 0.851 7 0.752 7 0.727 4 0.557 7 0.676 0.191 2 -
[1] SHI P P, SU S Q, CHEN Z M. Overview of researches on the nondestructive testing method of metal magnetic memory: status and challenges[J]. Journal of Nondestructive Evaluation, 2020, 39(2): 43 doi: 10.1007/s10921-020-00688-z [2] WANG H P, DONG L H, WANG H D, et al. Effect of tensile stress on metal magnetic memory signals during on-line measurement in ferromagnetic steel[J]. NDT & E International, 2021, 117: 102378 [3] 冷建成, 李政达, 王玉洁, 等. 循环应力对磁记忆效应影响的试验研究[J]. 材料导报, 2019, 33(10): 1723-1727, 1733 doi: 10.11896/cldb.18020090LENG J C, LI Z D, WANG Y J, et al. Experimental study on the effect of cyclic stress on magnetic memory effect[J]. Materials Reports, 2019, 33(10): 1723-1727, 1733 (in Chinese) doi: 10.11896/cldb.18020090 [4] 邸新杰, 李午申, 白世武, 等. 金属磁记忆检测信号的二维谱熵特征[J]. 焊接学报, 2006, 27(11): 69-72 doi: 10.3321/j.issn:0253-360X.2006.11.018DI X J, LI W S, BAI S W, et al. Two-dimension spectrum entropy feature for metal magnetic memory signal[J]. Transactions of the China Welding Institution, 2006, 27(11): 69-72 (in Chinese) doi: 10.3321/j.issn:0253-360X.2006.11.018 [5] 白鹭, 任吉林, 范振中. 基于小波包能量谱的磁记忆检测方法[J]. 无损检测, 2008, 30(10): 744-746 https://www.cnki.com.cn/Article/CJFDTOTAL-WSJC200810024.htmBAI L, REN J L, FAN Z Z. The metal magnetic memory testing based on wavelet packet energy spectrum[J]. Nondestructive Testing, 2008, 30(10): 744-746 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-WSJC200810024.htm [6] MOONESAN M, KASHEFI M. Effect of sample initial magnetic field on the metal magnetic memory NDT result[J]. Journal of Magnetism and Magnetic Materials, 2018, 460: 285-291 doi: 10.1016/j.jmmm.2018.04.006 [7] 樊清泉, 任尚坤, 任仙芝, 等. 外加磁场对Q235钢力磁效应影响试验研究[J]. 中国测试, 2019, 45(6): 46-53 https://www.cnki.com.cn/Article/CJFDTOTAL-SYCS201906009.htmFAN Q Q, REN S K, REN X Z, et al. Experimental study on the influence of external magnetic field on magneto-mechanical effect for Q235 steel[J]. China Measurement & Test, 2019, 45(6): 46-53 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-SYCS201906009.htm [8] HUANG H H, QIAN Z C. Effect of temperature and stress on residual magnetic signals in ferromagnetic structural steel[J]. IEEE Transactions on Magnetics, 2017, 53(1): 6200108 [9] SHI P P, ZHANG P C, JIN K, et al. Thermo-magneto-elastoplastic coupling model of metal magnetic memory testing method for ferromagnetic materials[J]. Journal of Applied Physics, 2018, 123(14): 145102 doi: 10.1063/1.5022534 [10] 任文坚, 孙金立, 陈曦, 等. 地磁场中应力对磁畴组织结构影响的试验研究[J]. 机械工程学报, 2013, 49(2): 8-13 https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201302003.htmREN W J, SUN J L, CHEN X, et al. Experimental study on effect of stress on magnetic domain structure under geomagnetic condition[J]. Journal of Mechanical Engineering, 2013, 49(2): 8-13 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-JXXB201302003.htm [11] ZHU W, MA Z W, YAN J, et al. Magnetic domain structures and their evolution in quasi-two-dimensional ferromagnet Cr5Te8[J]. Journal of Magnetism and Magnetic Materials, 2020, 512: 167019 doi: 10.1016/j.jmmm.2020.167019 [12] SALAZAR W, MORENO-ALDANA L C, DEL BUSTO J W S. Magnetic domain structures of Sr3Co2Z hexaferrite by TEM[J]. Journal of Magnetism and Magnetic Materials, 2020, 501: 166423 doi: 10.1016/j.jmmm.2020.166423 [13] 刘涛, 鲍宏, 朱达荣, 等. 基于磁记忆和表面纹理特征融合的再制造毛坯疲劳损伤评估[J]. 中国机械工程, 2018, 29(13): 1615-1621 doi: 10.3969/j.issn.1004-132X.2018.13.016LIU T, BAO H, ZHU D R, et al. Fatigue damage evaluation of remanufacturing cores using feature fusion of magnetic memory and surface texture[J]. China Mechanical Engineering, 2018, 29(13): 1615-1621 (in Chinese) doi: 10.3969/j.issn.1004-132X.2018.13.016 [14] 李思岐, 俞洋, 党永斌, 等. 基于改进的支持向量回归机算法的磁记忆定量化缺陷反演[J]. 工程科学学报, 2018, 40(9): 1123-1130 https://www.cnki.com.cn/Article/CJFDTOTAL-BJKD201809014.htmLI S Q, YU Y, DANG Y B, et al. Metal magnetic memory quantitative inversion of defects based on optimized support vector machine regression[J]. Chinese Journal of Engineering, 2018, 40(9): 1123-1130 (in Chinese) https://www.cnki.com.cn/Article/CJFDTOTAL-BJKD201809014.htm [15] CHEN L F, CUI X L, LI Z H, et al. A new deep learning algorithm for SAR scene classification based on spatial statistical modeling and features re-calibration[J]. Sensors, 2019, 19(11): 2479 doi: 10.3390/s19112479