在研究支持向量机模拟受载混凝土五个超声参数（波速、首波幅值、主频幅值、非线性系数和超声波谱面积）与应力的关系时，为提高其运算效率和寻优结果。通过数值模拟，对比了不同的归一化方式和核函数对寻找超声参数与应力相关性效果的影响；并采用试验探究了经验参数、遍历算法优化参数和粒子群算法优化参数三种支持向量机的模拟效果。结果表明：提出的首项归一化方式更适用于超声参数与应力这一特殊问题，核函数选择径向基核函数效果更优；不同的支持向量机模拟效果差异明显。简单支持向量机模拟效果较差，在低应力阶段尤为明显；遍历算法优化的模型得到的结果效果较好，但是计算时间过长，在低应力阶段判断误差较大，预测的结果高于真实值；粒子群算法优化支持向量机相比于遍历算法大幅减少了计算时间，模拟的结果效果也最好。
Abstract
In the study of the relationship between five ultrasonic parameters (wave velocity, first wave amplitude, main frequency amplitude, non-linear coefficient and ultrasonic spectral area) and stress in loaded concrete simulated by support vector machines, in order to improve their operational efficiency and optimisation results. The effect of different normalization methods and kernel functions on the effectiveness of finding the correlation between ultrasonic parameters and stress is compared through numerical simulations, and the simulation results of three types of support vector machines, namely empirical parameters, traversal algorithm optimized parameters and particle swarm algorithm optimized parameters, are investigated experimentally. The results show that the proposed first normalization method is more suitable for the special problem of ultrasonic parameters and stress, and the kernel function is better than the radial basis kernel function; the simulation results of different support vector machines differ significantly, and the simulation results of the simple support vector machine are poor, especially in the low stress stage; the results obtained by the model optimized by the traversal algorithm are better, but the computation time is too long, and the judgment error in the low stress stage is The particle swarm algorithm optimises the support vector machine to reduce the computation time significantly compared to the traversal algorithm and gives the best simulation results.
关键词
支持向量机 /
粒子群算法 /
首项归一化 /
受载混凝土应力 /
超声参数
{{custom_keyword}} /
Key words
support vector machine /
particle swarm algorithms /
first term normalization /
loaded concrete stresses /
ultrasonic parameters
{{custom_keyword}} /
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] Hughes D S, Kelly J L. Second-order Elastic Deformation of Solids[J]. Physical Review, 1953, 92(5): 1145.
[2] Demirboa. Relationship between Ultrasonic Velocity and Compressive Strength for High-volume Mineral-admixtured Concrete[J]. Cement & Concrete Research, 2004, 34(12): 2329-2336.
[3] 赵 岩，徐 晓，郝文秀. 超声波检测混凝土梁工作应力状态研究[J]. 河北农业大学学报，2019，42(01): 122-125.
ZHAO Yan, XU Xiao, HAO Wen-xiu. Ultrasonic Detection of the Working Stress State of Concrete Beams[J]. Journal of Hebei Agricultural University, 2019, 42(01): 122-125.
[4] LIU Z, MA C, WEI X. P-Wave Velocity and Energy Evolution Process of Concrete in Uniaxial Loading–Unloading Tests[J]. Journal of Materials in Civil Engineering, 2021, 33(6): 04021110.
[5] 朱金颖，陈龙珠，严细水. 混凝土受力状态下超声波传播特性研究[J]. 工程力学，1998(03): 111-117.
ZHU Jin-ying, CHEN Long-zhu, YAN Xi-shui. Research on Ultrasonic Wave Propagation Characteristics of Concrete under Force[J]. Engineering Mechanics, 1998(03): 111-117.
[6] Cortes C, Vapnik V. Support-vector Networks[J]. Machine Learning, 1995, 20(3): 273-297.
[7] Drucker H, Burges C J C, Kaufman L, et al. Support Vector Regression Machines[J]. Advances in Neural Information Processing Systems, 1997, 9: 155-161.
[8] Chapelle O, Vapnik V, Bousquet O, et al. Choosing Multiple Parameters for Support Vector Machines[J]. Machine Learning, 2002, 46(1): 131-159.
[9] Samui P. Slope Stability Analysis: A Support Vector Machine Approach[J]. Environmental Geology, 2008, 56(2): 255-267.
[10] 黄 昕. 高分辨率遥感影像多尺度纹理、形状特征提取与面向对象分类研究[D]. 武汉: 武汉大学，2009.
HUANG Xin. Research on Multi-scale Texture and Shape Feature Extraction and Object-oriented Classification of High-Resolution Remote Sensing Images[D]. Wuhan: Wuhan University, 2009.
[11] 程 灿，李建勇，徐文胜，等. 基于支持向量机与粒子滤波的刀具磨损状态识别[J]. 振动与冲击，2018，37(17): 48-55+71.
CHENG Can, LI Jianyong, XU Wensheng, et al. Tool Wear State Identification Based on Support Vector Machine and Particle Filtering[J]. Journal of Vibration and Shock, 2018, 37(17): 48-55+71.
[12] 胡晓依，荆云建，宋志坤，等. 基于CNN-SVM的深度卷积神经网络轴承故障识别研究[J]. 振动与冲击，2019, 38(18): 173-178.
HU Xiaoyi, JING Yunjian, SONG Zhikun, et al. Bearing Fault Identification Of Deep Convolutional Neural Network Based on CNN-SVM[J]. Journal of Vibration and Shock, 2019, 38(18): 173-178.
[13] Nieto P J G, García-Gonzalo E, Lasheras F S, et al. Hybrid PSO–SVM-based Method for Forecasting of the Remaining Useful life for Aircraft Engines and Evaluation of its Reliability[J]. Reliability Engineering & System Safety, 2015, 138: 219-231.
[14] 贺玲凤，刘 军. 声弹性技术[M]. 北京: 科学出版社，2002.
HE Ling-feng, LIU Jun. Acoustic Elasticity Technology[M]. Beijing: Science Press, 2002.
[15] 朱自强，喻 波，密士文，等. 超声波在混凝土中的衰减特征[J]. 中南大学学报(自然科学版)，2014，45(11): 3900-3907.
ZHU Zi-qiang, Yu Bo, MI Shi-wen, et al. Characteristics of Ultrasonic Waves in Concrete[J]. Journal of Central South University (Natural Science Edition), 2014, 45(11): 3900-3907.
[16] 王怀亮，宋玉普. 不同尺寸混凝土试件受压状态下超声波传播特性研究[J]. 大连理工大学学报，2007(01): 90-94.
WANG Huai-liang, SONG Yu-pu. Research on Ultrasonic Wave Propagation Characteristics of Concrete Specimen of Different Sizes under Compression[J]. Journal of Dalian University of Technology, 2007(01): 90-94.
[17] 王 菲. 低应力条件下混凝土超声传播特性的三维数值模拟研究[D]. 重庆: 重庆交通大学，2019.
WANG Fei. Three-dimensional Numerical Simulation of Concrete Ultrasonic Propagation Characteristics under Low Stress Conditions[D]. Chongqing: Chongqing Jiaotong University, 2019.
[18] 刘晓宙，朱金林，尹 昌，等. 岩石等非线性介观弹性固体材料的谐波特性的超声研究[J]. 物理学进展，2006(Z1): 386-390.
LIU Xiao-zhou, ZHU Jin-lin, YIN Chang, et al. Ultrasonic Research on the Harmonic Characteristics of Non-linear Mesoscopic Elastic Solid Materials Such as Rocks[J]. Progress in Physics, 2006(Z1): 386-390.
[19] 许 颖，张何勇，王青原. 基于非线性超声技术的热损伤混凝土微裂纹检测实验研究[J]. 振动与冲击，2021，40(05): 126-135.
XU Ying, ZHANG Heyong, WANG Qingyuan. Experimental Study on Microcrack Detection of Thermally Damaged Concrete Based on Nonlinear Ultrasonic Technology[J]. Journal of Vibration and Shock, 2021, 40(05): 126-135.
[20] 赵明阶，吴德伦. 单轴加载条件下岩石声学参数与应力的关系研究[J]. 岩石力学与工程学报，1999(01): 51-55.
ZHAO Ming-jie, WU De-lun. Relationship between Rock Acoustic Parameters and Stress under Uniaxial Loading[J]. Chinese Journal of Rock Mechanics and Engineering, 1999(01): 51-55.
[21] 林军志，赵明阶，杨洪武. 改性混凝土声学参数与应力相关性试验研究[J]. 岩土力学，2009，30(S1): 69-74.
LIN Jun-zhi, ZHAO Ming-jie, YANG Hong-wu. Experimental Study on the Correlation between Acoustic Parameters and Stress of Modified Concrete[J]. Rock and Soil Mechanics, 2009, 30(S1): 69-74.
[22] 奉国和. SVM分类核函数及参数选择比较[J]. 计算机工程与应用，2011，47(03): 123-124+128.
FENG Guo-he. SVM Classification Kernel Function and Parameter Selection Comparison[J]. Computer Engineering and Applications, 2011, 47(03): 123-124+128.
[23] 汤荣志，段会川，孙海涛. SVM训练数据归一化研究[J]. 山东师范大学学报(自然科学版)，2016，31(04): 60-65.
TANG Rong-zhi, DUAN Hui-chuan, SUN Hai-tao. SVM Training Data Normalization Research[J]. Journal of Shandong Normal University (Natural Science Edition), 2016, 31(04): 60-65.
[24] 郑小霞，钱 锋. 高斯核支持向量机分类和模型参数选择研究[J]. 计算机工程与应用，2006(01): 77-79.
ZHENG Xiao-xia, QIAN Feng. Research on Gaussian Kernel Support Vector Machine Classification and Model Parameter Selection[J]. Computer Engineering and Applications, 2006(01): 77-79.
[25] Eberhart R, Kennedy J. Particle Swarm optimization[C]// Proceedings of the IEEE International Conference on Neural Networks. 1995, 4: 1942-1948.
[26] 张战廷，刘宇锋. ABAQUS中的混凝土塑性损伤模型[J]. 建筑结构，2011，41(S2): 229-231.
ZHANG Zhan-ting, LIU Yu-feng. The Plastic Damage Model of Concrete in ABAQUS[J]. Building Structure, 2011, 41(S2): 229-231.
{{custom_fnGroup.title_cn}}
脚注
{{custom_fn.content}}
PDF(1689 KB)
