地震研究

为了实现钢筋混凝土(Reinforced Concrete,RC)框架结构地震响应的快速预测,提出了一种基于人工神经网络的RC框架结构地震响应预测方法,设计低层、多层和小高层共3个典型RC框架结构作为研究对象,以四川雅安地区为目标场地,基于条件均值谱选取地震动记录作为输入并进行弹塑性时程分析,所得样本数据用于训练人工神经网络。以地震动强度信息和结构信息为输入预测结构响应,同时对模型进行参数敏感性分析。结果表明:建立的人工神经网络模型具有较好的泛化性能,平均谱加速度具有最高的平均影响值,提出的方法为快速预测RC框架结构地震响应提供了方法借鉴。

To realize the rapid prediction of seismic response of reinforced concrete(RC)frame structure,a seismic response prediction method of RC frame structure based on the artificial neural network(ANN)is proposed.Three typical RC frame structures of low-rise,multi-storey and small high-rise were designed as research examples.Ya'an area in Sichuan was taken as the target site.Based on the conditional mean spectrum(CMS),the ground motion records were selected as input and the elastic-plastic time history analysis was carried out.The obtained sample data were used to train the artificial neural network.The information of the ground motion intensity and the structures were used as input to predict the structural response,and the parameters' sensitivity of the model was analyzed.The results show that the established artificial neural network model has good generalization performance,and the average spectral acceleration(AvgSa)has the highest average impact value(MIV).The proposed method provides a reference for the rapid prediction of the seismic response of the RC frame structures,and has a good application prospect.

引言
1 RC框架结构模型设计
2 地震动记录选取及增量动力分析
3 基于人工神经网络的结构响应预测
4 神经网络模型的参数敏感性分析
5 结论

图1 结构的平面(a)及立面(b)图<br/>Fig.1 Plane(a)and elevation(b)of the structures

图1 结构的平面(a)及立面(b)图
Fig.1 Plane(a)and elevation(b)of the structures

图2 3层(a)、8层(b)、14层(c)RC框架结构梁、柱截面配筋图<br/>Fig.2 Reinforcement of the beams and columns of the 3-storey(a),8-storey(b),14-storey(c)RC frame structures

图2 3层(a)、8层(b)、14层(c)RC框架结构梁、柱截面配筋图
Fig.2 Reinforcement of the beams and columns of the 3-storey(a),8-storey(b),14-storey(c)RC frame structures

图3 纤维单元材料Concrete02(a)和Steel02(b)本构模型<br/>Fig.3 Constitutive model of the fiber element materials Concrete02(a)and Steel02(b)

图3 纤维单元材料Concrete02(a)和Steel02(b)本构模型
Fig.3 Constitutive model of the fiber element materials Concrete02(a)and Steel02(b)

图4 试件FW1(a)和FW2(b)滞回曲线标定结果<br/>Fig.4 Calibration results of hysteresis curves of specimens FW1(a)and FW2(b)

图4 试件FW1(a)和FW2(b)滞回曲线标定结果
Fig.4 Calibration results of hysteresis curves of specimens FW1(a)and FW2(b)

图5 14层RC框架结构包络条件均值谱<br/>Fig.5 Envelope CMS of a 14-storey RC frame structure

图5 14层RC框架结构包络条件均值谱
Fig.5 Envelope CMS of a 14-storey RC frame structure

图6 3层(a)、8层(b)、14层(c)RC框架结构谱型匹配结果<br/>Fig.6 Spectral matching results of the 3-storey(a),8-storey(b),14-storey(c)RC frame structures

图6 3层(a)、8层(b)、14层(c)RC框架结构谱型匹配结果
Fig.6 Spectral matching results of the 3-storey(a),8-storey(b),14-storey(c)RC frame structures

图7 最大层间位移角(a)和最大水平位移(b)分布数据统计结果<br/>Fig.7 Statistics of the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

图7 最大层间位移角(a)和最大水平位移(b)分布数据统计结果
Fig.7 Statistics of the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

表1 神经网络模型参数选取<br/>Tab.1 Selection of the parameters of the neural network model

表1 神经网络模型参数选取
Tab.1 Selection of the parameters of the neural network model

图8 基于人工神经网络的结构响应预测流程图<br/>Fig.8 Flow chart of the structural response predictionbased on the artificial neural network

图8 基于人工神经网络的结构响应预测流程图
Fig.8 Flow chart of the structural response predictionbased on the artificial neural network

图9 隐含层神经元数目对应的平均误差指标<br/>Fig.9 Average error indexes corresponding to thenumber of hidden layer neurons

图9 隐含层神经元数目对应的平均误差指标
Fig.9 Average error indexes corresponding to thenumber of hidden layer neurons

图10 人工神经网络模型结构<br/>Fig.10 Structure of the artificial neural network model

图10 人工神经网络模型结构
Fig.10 Structure of the artificial neural network model

图11 最大层间位移角(a)和最大水平位移(b)的神经网络模型回归结果<br/>Fig.11 Regression results of the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

图11 最大层间位移角(a)和最大水平位移(b)的神经网络模型回归结果
Fig.11 Regression results of the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

表2 最大层间位移角评价结果<br/>Tab.2 Evaluation of the maximum inter-storey displacement angle

表2 最大层间位移角评价结果
Tab.2 Evaluation of the maximum inter-storey displacement angle

表3 最大水平位移评价结果<br/>Tab.3 Evaluation of the maximum horizontal displacement

表3 最大水平位移评价结果
Tab.3 Evaluation of the maximum horizontal displacement

表4 各个特征参数MIV值及排序<br/>Tab.4 MIV-values of the characteristic parametersand their sorting

表4 各个特征参数MIV值及排序
Tab.4 MIV-values of the characteristic parametersand their sorting

图12 特征参数对最大层间位移角(a)和最大水平位移(b)的MIV值<br/>Fig.12 MIV values of the characteristic parameters to the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

图12 特征参数对最大层间位移角(a)和最大水平位移(b)的MIV值
Fig.12 MIV values of the characteristic parameters to the maximum inter-storey displacement angle(a)and the maximum horizontal displacement(b)

崔智全,付旭云,钟诗胜,等.2013.小波网络平均影响值的航空发动机自变量筛选[J].计算机集成制造系统,19(12):3062-3067.
Cui Z Q,Fu X Y,Zhong S S,et al.2013.Independent variable selection of aero-engine based on average influence value of wavelet network[J].Computer Integrated Manufacturing Systems,19(12):3062-3067.(in Chinese)
霍俊荣.1989.近场强地面运动衰减规律的研究[D].哈尔滨:中国地震局工程力学研究所.
Huo J R.1989.Study on attenuation law of near field strong ground motion[D].Harbin:Institute of Engineering Mechanics,China Earthquake Administration.(in Chinese)
冀昆,温瑞智,任叶飞.2016.中国地震安全性评价中天然强震记录选取[J].哈尔滨工业大学学报,48(12):183-188.
Ji K,Wen R Z,Ren Y F.2016.Selection of natural strong earthquake records in seismic safety evaluation of China[J].Journal of Harbin Institute of Technology,48(12):183-188.(in Chinese)
温瑞智,冀昆,任叶飞.2021.工程地震动输入—从传统抗震设防到韧性提升[M].北京:地震出版社.
Wen R Z,Ji K,Ren Y F.2021.Engineering ground motion input:From traditional seismic fortification to toughness improvement[M].Beijing:Seismological Press.(in Chinese)
许卫晓,程扬,杨伟松,等.2021.RC框架—抗震墙并联结构体系拟静力试验[J].吉林大学学报(工学版),51(1):268-277.
Xu W X,Cheng Y,Yang W S,et al.2021.Quasi-static test of RC frame-shear wall dual structure system[J].Journal of Jilin University(Engineering and Technology Edition),51(1):268-277.(in Chinese)
GB 50010—2010,混凝土结构设计规范[S].
GB 50010—2010,Code for design of concrete structures[S].(in Chinese)
GB 50011—2010,建筑抗震设计规范[S].
GB 50011—2010,Code for seismic design of buildings[S].(in Chinese)
JGJ 3—2010,高层建筑混凝土结构技术规程[S].
JGJ 3—2010,Technical specification for concrete structures of high-rise building[S].(in Chinese)
Abadi M,Agarwal A,Barham P,et al.2016a.TensorFlow:Large-scale machine learning on heterogeneous distributed systems[J/OL].e-prints arXiv,(2016-03-14)[2023-02-22].https://arxiv.org/pdf/1603.04467.pdf.
Abadi M,Agarwal A,Barham P,et al.2016b.Tensorflow:large-scale machine learning on heterogeneous distributed systems[R].CoRR,doi:10.48550/arXiv.1603.04467.
Arias A.1970.A measure of earthquake intensity in seismic design for nuclear power plants[M].Cambridge:MA MIT Press.
Baker J W,Cornell C A.2005.A vector-valued ground motion intensity measure consisting of spectral acceleration and epsilon[J].Earthquake Engineering and Structural Dynamic,34(10):1193-1217.
Baker J W.2011.Conditional mean spectrum:tool for ground-motion selection[J].Journal of Structural Engineering,137(3):322-331.
Bishop C M.2006.Pattern recognition and machine learning[M].New York:Springer.
Boger Z,Guterman H.1997.Knowledge extraction from artificial neural network models[J].Computational Cybernetics and Simulation,62(4):3030-3035.
Cornell C A,Krawinkler H.2000.Progress and challenges in seismic performance assessment[J].PEER Center News,3(2):1-3.
Dolce M,Prota A,Borzi B,et al.2021.Seismic risk assessment of residential buildings in Italy[J].Bulletin of Earthquake Engineering,(19):2999-3032.
Harirchian E,Lahmer T,Rasulzade S.2020.Earthquake hazard safety assessment of existing buildings using optimized multi-layer perceptron neural network[J].Energies,13(8):2060.
Harrington P.2018.Multiple versus single set validation of multivariate models to avoid mistakes[J].Critical Reviews in Analytical Chemistry,48(1):33-46.
Haykin S.2009.Neural networks and learning machines[M].New Jersey,United States:Pearson.
Jalayer F,Cornell C A.2009.Alternative non-linear demand estimation methods for probability-based seismic assessments[J].Earthquake Engineering and Structural Dynamic,38(8):951-972.
Jalayer F,De R R,Manfredi G.2015.Bayesian Cloud analysis:efficient structural fragility assessment using linear regression[J].Bulletin of Earthquake Engineering,13(4):1183-1203.
Karsoliya S.2012.Approximating number of hidden layer neurons in multiple hidden layer BPNN architecture[J].International Journal of Engineering Trends and Technology,3(6):714-717.
Kim T,Song J,Kwon O S.2020.Probabilistic evaluation of seismic responses using deep learning method[J].Structural Safety,84(4):131-139.
Kohrangi M,Bazzurro P,Vamvatsikos D.2016a.Vector and scalar IMs in structural response estimation,part I:Hazard analysis[J].Earthquake Spectra,32(3):1507-1524.
Kohrangi M,Vamvatsikos D,Bazzurro P.2016b.Implications of intensity measure selection for seismic loss assessment of 3-D buildings[J].Earthquake Spectra,32(4):2167-2189.
Lu P Z,Chen S Y,Zheng Y J.2012.Artificial intelligence in civil engineering[J].Mathematical Problems in Engineering,(11):1-22.
Morfidis K,Kostinakis K.2018.Approaches to the rapid seismic damage prediction of RC buildings using artificial neural networks[J].Engineering Structures,165(15):120-141.
Oh B W,Glisic B,Park S W,et al.2020.Neural network-based seismic response prediction model for building structures using artificial earthquakes[J].Journal of Sound and Vibration,468:109-115.
Picard R R,Cook R D.1984.Cross-validation of regression models[J].Journal of American Statistical Association,79(387):575-583.
Ruder S.2017.An overview of gradient descent optimization algorithms[J/OL].e-prints arXiv,(2017-06-15)[2023-02-22].https://arxiv.org/pdf/1609.04747.pdf.
Sheela K,Deepa S N.2013.Review on methods to fix number of hidden neurons in neural networks[J].Mathematical Problems in Engineering,(6):1-11.
Vamvatsikos D,Cornell C A.2002.Incremental dynamic analysis[J].Earthquake Engineering and Structural Dynamic,31(3):491-514.
Wang Z Y,Pedroni N,Zentner I,et al.2018.Seismic fragility analysis with artificial neural networks:application to nuclear power plant equipment[J].Engineering Structure,162:213-225.
Wang Z.2020.Abstract universal approximation for neural networks[D].New York:Cornell University.
Xu Y,Goodacre R.2018.On splitting training and validation set:a comparative study of cross-validation,bootstrap and systematic sampling for estimating the generalization performance of supervised learning[J].Journal of Analysis and Testing,2(3):249-262.
Yepes E C,Silva V,Rossetto T,et al.2016.The global earthquake model physical vulnerability database[J].Earthquake Spectral,32(4):2567-2585.
Yoshikaw A H,Goda K.2014.Financial seismic risk analysis of building portfolios[J].Natural Hazards Review,15(2):112-120.

备注

引言

1 RC框架结构模型设计

2 地震动记录选取及增量动力分析

3 基于人工神经网络的结构响应预测

4 神经网络模型的参数敏感性分析

5 结论

期刊信息