基于声发射和深度置信网络的钢筋混凝土梁损伤识别方法研究
|
摘要
为准确识别混凝土结构的损伤,采用声发射技术对钢筋混凝土梁四点弯破坏试验进行损伤监测。首先利用参数关联法分析了钢筋混凝土梁的损伤演化过程;随后通过峰度值指标进一步评价试件梁裂缝损伤演化,划分钢筋混凝土梁损伤阶段,并建立深度置信网络(DBN)训练声发射信号样本集;最后通过Matlab构建DBN模型并进行训练,比较DBN与BP神经网络的识别效果。结果表明,将DBN提取的声发射特征参数作为输入向量,可以减小噪声和环境因素等对损伤识别结果的影响,提高损伤识别的精度。
In order to accurately identify the damage of concrete structures, an experiment on a reinforced concrete beams was carried out under four-point bending test. The destruction of the beam was monitored using the acoustic emission(AE) test system. The damage evolution process of the reinforced concrete beam was revealed by correlation analysis method of AE parameters. Through the Kurtosis, index specific evaluation of specimen beam cracks damage, typical failure stages of reinforced concrete beams and the AE signal parameters training database of DBN were established. Finally, the DBN network model was designed and trained with Matlab. The results obtained from the above-mentioned method were compared with those based on BP neural networks.Results show that, the influence of noise and false mode information on damage identification results can be reduced. The accuracy of damage identification is improved by taking deep belief nets to extract features of AE parameters as damage signature.
In order to accurately identify the damage of concrete structures, an experiment on a reinforced concrete beams was carried out under four-point bending test. The destruction of the beam was monitored using the acoustic emission(AE) test system. The damage evolution process of the reinforced concrete beam was revealed by correlation analysis method of AE parameters. Through the Kurtosis, index specific evaluation of specimen beam cracks damage, typical failure stages of reinforced concrete beams and the AE signal parameters training database of DBN were established. Finally, the DBN network model was designed and trained with Matlab. The results obtained from the above-mentioned method were compared with those based on BP neural networks.Results show that, the influence of noise and false mode information on damage identification results can be reduced. The accuracy of damage identification is improved by taking deep belief nets to extract features of AE parameters as damage signature.
引文
[1] 杨智春,于哲峰. 结构健康监测中的损伤检测技术研究进展[J]. 力学进展, 2004, 34(2): 215-223. (YANG Zhichun, YU Zhefeng. Progress of damage detection for structural health monitoring [J]. Advances in Mechanics, 2004, 34 (2): 215-223. (in Chinese))
[2] 韩建德,刘金龙,王曙光,等. 声发射技术在混凝土材料及其耐久性中的应用研究进展[J]. 材料导报, 2014, 28(1): 110-115. (HAN Jiande, LIU Jinlong, WANG Shuguang, et al. Recent situation in research on technique of acoustic emission for concrete material and its durability [J]. Materials Review, 2014, 28(1): 110-115. (in Chinese))
[3] NAIR A, CAI C S. Acoustic emission monitoring of bridges: review and case studies [J]. Engineering Structures, 2010, 32(6): 1704-1714.
[4] 丁幼亮,邓扬,李爱群. 声发射技术在桥梁结构健康监测中的应用研究进展[J]. 防灾减灾工程学报, 2010, 30(3):341-351. (DING Youliang, DENG Yang,LI Aiqun. Application of acoustic emission technology in bridge structure health monitoring is reviewed[J]. Journal of Disaster Prevention and Mitigation Engineering, 2010, 30(3):341-351. (in Chinese))
[5] 胡少伟,米正祥. 标准钢筋混凝土三点弯曲梁双K断裂特性试验研究[J].建筑结构学报, 2013, 34(3): 152-157. (HU Shaowei, MI Zhengxiang. Experimental study on double-K fracture characteristic of standard reinforced concrete three-point beam [J]. Journal of Building Structures, 2013, 34 (3):152-157. (in Chinese))
[6] 岳健广,镇东. 受弯钢筋混凝土柱宏/微观损伤演化声发射监测试验与评估研究[J]. 建筑结构学报, 2017, 38(8): 156-166.(YUE Jianguang, ZHEN Dong. Acoustic emission monitoring experiments and assessment method for macro-micro-scale seismic damage evolution of bending-type reinforced concrete columns [J]. Journal of Building Structures, 2017, 38(8): 156-166. (in Chinese))
[7] 王柏生. 结构损伤检测与识别技术[M]. 杭州:浙江大学出版社, 2000. (WANG Baisheng. Structural damage detection and identification technology[M]. Hangzhou: Zhejiang University Press, 2000. (in Chinese))
[8] 冉志红,屈俊童,和飞. 桥梁结构损伤诊断的模式识别理论及其工程应用[M]. 北京:科学出版社,2011:243-249. (RAN Zhihong, QU Juntong, HE Fei. Pattern identification theory and its application of bridge structural damage diagnosis [M]. Beijing: Science Press, 2011:243-249. (in Chinese))
[9] 王英. 基于人工智能方法的预应力混凝土梁式桥损伤识别研究[D].成都:西南交通大学, 2009.(WANG Ying. Damage identification of prestressed concrete girder bridge based on artificial intelligence [D]. Chengdu: Southwest Jiaotong University, 2009. (in Chinese))
[10] CHIEMENTIN X, MBA D, CHARNLEY B, et al. Effect of the denoising on acoustic emission signals [J]. Journal of Vibration and Acoustics, 2010, 132(3): 031009.
[11] 陈翠平. 基于深度信念网络的文本分类算法[J]. 计算机系统应用, 2015, 24(2):121-126.(CHEN Cuiping. Text categorization based on deep belief network [J]. Computer Systems & Applications, 2015, 24(2):121-126. (in Chinese))
[12] HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks [J]. Science, 2006, 313(5786): 504-507.
[13] 施徐敢,张石清,赵小明. 融合深度信念网络和多层感知器的人脸表情识别[J]. 小型微型计算机系统, 2015, 36(7): 1629-1632. (SHI Xugan, ZHANG Shiqing, ZHAO Xiaoming. Facial expression recognition by integrating deep belief networks with multi-layer perceptron[J]. Journal of Chinese Computer Systems, 2015, 36(7):1629-1632. (in Chinese))
[14] ANTONI J. The spectral kurtosis: a useful tool for characterising non-stationary signals[J]. Mechanical Systems and Signal Processing, 2006, 20(2): 282-307.
[15] HINTON G E. Training products of experts by minimizing contrastive divergence[J]. Neural Computation, 2002, 14(8): 1771-1800.
[16] BENGIO Y, LAMBLIN P, DAN P, et al. Greedy layer-wise training of deep networks[J]. Advances in Neural Information Processing Systems,2007,19: 153-160.
[17] 吴胜兴,王岩,李佳,等. 混凝土静态轴拉声发射试验相关参数研究[J]. 振动与冲击, 2011, 30(5): 196-204. (WU Shengxing, WANG Yan, LI Jia, et al. Parameters of acoustic emission test of concrete under static uniaxial tension [J]. Journal of Vibration and Shock, 2011, 30(5): 196-204. (in Chinese))
[18] 柳小桐. BP神经网络输入层数据归一化研究[J]. 机械工程与自动化, 2010(3):122-123.(LIU Xiaotong. The BP neural network input layer data nor-malization research [J]. Mechanical Engineering and Automation, 2010(3):122-123. (in Chinese))
[2] 韩建德,刘金龙,王曙光,等. 声发射技术在混凝土材料及其耐久性中的应用研究进展[J]. 材料导报, 2014, 28(1): 110-115. (HAN Jiande, LIU Jinlong, WANG Shuguang, et al. Recent situation in research on technique of acoustic emission for concrete material and its durability [J]. Materials Review, 2014, 28(1): 110-115. (in Chinese))
[3] NAIR A, CAI C S. Acoustic emission monitoring of bridges: review and case studies [J]. Engineering Structures, 2010, 32(6): 1704-1714.
[4] 丁幼亮,邓扬,李爱群. 声发射技术在桥梁结构健康监测中的应用研究进展[J]. 防灾减灾工程学报, 2010, 30(3):341-351. (DING Youliang, DENG Yang,LI Aiqun. Application of acoustic emission technology in bridge structure health monitoring is reviewed[J]. Journal of Disaster Prevention and Mitigation Engineering, 2010, 30(3):341-351. (in Chinese))
[5] 胡少伟,米正祥. 标准钢筋混凝土三点弯曲梁双K断裂特性试验研究[J].建筑结构学报, 2013, 34(3): 152-157. (HU Shaowei, MI Zhengxiang. Experimental study on double-K fracture characteristic of standard reinforced concrete three-point beam [J]. Journal of Building Structures, 2013, 34 (3):152-157. (in Chinese))
[6] 岳健广,镇东. 受弯钢筋混凝土柱宏/微观损伤演化声发射监测试验与评估研究[J]. 建筑结构学报, 2017, 38(8): 156-166.(YUE Jianguang, ZHEN Dong. Acoustic emission monitoring experiments and assessment method for macro-micro-scale seismic damage evolution of bending-type reinforced concrete columns [J]. Journal of Building Structures, 2017, 38(8): 156-166. (in Chinese))
[7] 王柏生. 结构损伤检测与识别技术[M]. 杭州:浙江大学出版社, 2000. (WANG Baisheng. Structural damage detection and identification technology[M]. Hangzhou: Zhejiang University Press, 2000. (in Chinese))
[8] 冉志红,屈俊童,和飞. 桥梁结构损伤诊断的模式识别理论及其工程应用[M]. 北京:科学出版社,2011:243-249. (RAN Zhihong, QU Juntong, HE Fei. Pattern identification theory and its application of bridge structural damage diagnosis [M]. Beijing: Science Press, 2011:243-249. (in Chinese))
[9] 王英. 基于人工智能方法的预应力混凝土梁式桥损伤识别研究[D].成都:西南交通大学, 2009.(WANG Ying. Damage identification of prestressed concrete girder bridge based on artificial intelligence [D]. Chengdu: Southwest Jiaotong University, 2009. (in Chinese))
[10] CHIEMENTIN X, MBA D, CHARNLEY B, et al. Effect of the denoising on acoustic emission signals [J]. Journal of Vibration and Acoustics, 2010, 132(3): 031009.
[11] 陈翠平. 基于深度信念网络的文本分类算法[J]. 计算机系统应用, 2015, 24(2):121-126.(CHEN Cuiping. Text categorization based on deep belief network [J]. Computer Systems & Applications, 2015, 24(2):121-126. (in Chinese))
[12] HINTON G E, SALAKHUTDINOV R R. Reducing the dimensionality of data with neural networks [J]. Science, 2006, 313(5786): 504-507.
[13] 施徐敢,张石清,赵小明. 融合深度信念网络和多层感知器的人脸表情识别[J]. 小型微型计算机系统, 2015, 36(7): 1629-1632. (SHI Xugan, ZHANG Shiqing, ZHAO Xiaoming. Facial expression recognition by integrating deep belief networks with multi-layer perceptron[J]. Journal of Chinese Computer Systems, 2015, 36(7):1629-1632. (in Chinese))
[14] ANTONI J. The spectral kurtosis: a useful tool for characterising non-stationary signals[J]. Mechanical Systems and Signal Processing, 2006, 20(2): 282-307.
[15] HINTON G E. Training products of experts by minimizing contrastive divergence[J]. Neural Computation, 2002, 14(8): 1771-1800.
[16] BENGIO Y, LAMBLIN P, DAN P, et al. Greedy layer-wise training of deep networks[J]. Advances in Neural Information Processing Systems,2007,19: 153-160.
[17] 吴胜兴,王岩,李佳,等. 混凝土静态轴拉声发射试验相关参数研究[J]. 振动与冲击, 2011, 30(5): 196-204. (WU Shengxing, WANG Yan, LI Jia, et al. Parameters of acoustic emission test of concrete under static uniaxial tension [J]. Journal of Vibration and Shock, 2011, 30(5): 196-204. (in Chinese))
[18] 柳小桐. BP神经网络输入层数据归一化研究[J]. 机械工程与自动化, 2010(3):122-123.(LIU Xiaotong. The BP neural network input layer data nor-malization research [J]. Mechanical Engineering and Automation, 2010(3):122-123. (in Chinese))