基于GA-LSTM的高层建筑结构地震响应的分散控制研究

摘要
图/表
参考文献(29)
相关文章 (15)

全文: PDF (1385 KB) HTML (1 KB)
输出: BibTeX | EndNote (RIS)

摘要

高层建筑结构复杂，自由度数庞大，采用主动抗震控制方法时结构的精确模型无法建立，整体控制目标也难以获取。为此，将长短时记忆(LSTM)网络理论和大系统分散控制理论结合，提出一种基于LSTM的智能分散控制方法；利用LSTM深度学习框架构建不同类型的分散控制器，并根据Lyapunov稳定性理论推导分散控制器稳定的充分条件，运用遗传算法(GA)对LSTM框架的初始学习率进行优化，提高分散控制器的收敛速度和预测精度。以20层Benchmark模型作为被控对象，研究GA-LSTM分散控制方法的控制性能，并与集中控制效果进行对比分析。结果表明，基于GA-LSTM的智能分散控制方法可以简化控制器结构，对比集中控制可能产生的整体失效现象，具有更高的可靠性和更好的控制效果。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	高经纬
	涂建维
	刘康生
	李召

关键词 ：分散控制, 长短时记忆(LSTM)网络, Lyapunov稳定性理论, 遗传算法(GA), 结构振动控制

Abstract：

High-rise buildings have complex structures and huge degrees of freedom.When active seismic control method is used, it is difficult to establish a precise structure model, and the overall control target is hard to achieve.Thus, an intelligent decentralized control method based on long short-term memory (LSTM) networks was proposed, based on the LSTM theory combined with the large-scale system decentralized control theory.Different decentralized controller types were constructed using the LSTM deep learning framework, and the sufficient conditions for the stability of decentralized controllers were derived according to the Lyapunov stability theory.A genetic algorithm (GA) was used to optimize the initial learning rate of the LSTM framework to improve the convergence speed and prediction accuracy of the decentralized controller.Taking a 20-layer benchmark model as a controlled object, the control performance of the GA-LSTM decentralized control method was studied and its effect was compared with the centralized control effect.The results show that the intelligent decentralized control method based on GA-LSTM simplifies the controller structure.Compared with the overall failure phenomena that may occur in the centralized control, it has higher reliability and better control effect.

Key words： decentralized control long short-term memory (LSTM) networks Lyapunov stability theory genetic algorithm (GA) structural vibration control

收稿日期: 2019-10-17 出版日期: 2021-05-28

引用本文:

高经纬，涂建维，刘康生，李召. 基于GA-LSTM的高层建筑结构地震响应的分散控制研究[J]. 振动与冲击, 2021, 40(10): 114-122.
GAO Jingwei，TU Jianwei，LIU Kangsheng，LI Zhao. Decentralized control for the seismic response of high-rise building structures based on GA-LSTM. JOURNAL OF VIBRATION AND SHOCK, 2021, 40(10): 114-122.

链接本文:

http://jvs.sjtu.edu.cn/CN/ 或 http://jvs.sjtu.edu.cn/CN/Y2021/V40/I10/114

［1］徐怀兵,欧进萍.设置混合调谐质量阻尼器的高层建筑风振控制实用设计方法［J］.建筑结构学报, 2017,38(6): 144-154.
XU Huaibing, OU Jinping.Design method for wind-induced vibration control of high-rise buildings with hybrid tuned mass dampers ［J］.Journal of Building Structures, 2017,38(6): 144-154.

［2］雷永勤，杜永峰.基于神经网络的结构振动智能主动容错控制算法研究［J］.振动与冲击, 2014,33(13): 117-122.
LEI Yongqin, DU Yongfeng.Smart active fault tolerant control algorithm based on neural network for structural vibration control［J］.Journal of Vibration and Shock, 2014,33(13): 117-122.

［3］李宏男,李瀛,李钢.地震作用下建筑结构的分散控制研究［J］.土木工程学报, 2008,41(9): 27-33.
LI Hongnan, LI Ying, LI Gang.Decentralized control of structures under earthquakes［J］.China Civil Engineering Journal, 2008,41(9): 27-33.

［4］潘兆东,谭平,周福霖.土木工程结构保性能PID协调分散控制研究［J］.振动与冲击, 2018,37(12): 89-95.
PAN Zhaodong, TAN Ping, ZHOU Fulin.A study on guaranteed cost PID coordinated decentralized control for civil engineering structures［J］.Journal of Vibration and Shock, 2018,37(12): 89-95.

［5］潘兆东,谭平,周福霖.土木工程结构模糊滑模分散控制(DFSMC)研究［J］.振动与冲击, 2017,36(20): 107-111.
PAN Zhaodong, TAN Ping, ZHOU Fulin.A decentralized fuzzy sliding mode control (DFSMC) for civil engineering structures［J］.Journal of Vibration and Shock, 2017,36(20): 107-111.

［6］AMINI F, HASSANAL M, JAVANBAKHT M.Optimized and decentralized pulse control of seismically excited steel structures［J］.International Journal of Steel Structures, 2017,17(2): 631-642.

［7］LOPEZ-FRANCO M, SANCHEZ E N, ALANIS A Y, et al.Decentralized control for stabilization of nonlinear multi-agent systems using neural inverse optimal control［J］.Neurocomputing, 2015,168: 81-91.

［8］郭丽丽,丁世飞.深度学习研究进展［J］.计算机科学, 2015,42(5): 28-33.
GUO Lili, DING Shifei.Research progress on deep learning［J］.Computer Science, 2015,42(5): 28-33.

［9］HINTON G, SALAKHUTDINOV R.Reducing the dimensionality of data with neural networks［J］.Science, 2006,313(5786): 504-507.

［10］ZHANG H, TAN J W, ZHAO C Y.A fast detection and grasping method for mobile manipulator based on improved faster R-CNN［J］.Industrial Robot-The International Journal of Robotics Research and Application, 2020,47(2): 167-175.

［11］LI X, CAO L, TIONG A M H, et al.Distal-end force prediction of tendon-sheath mechanisms for flexible endoscopic surgical robots using deep learning［J］.Mechanism and Machine Theory, 2019,134: 323-337.

［12］HOCHREITER S, SCHMIDHUBER J.Long short-term memory［J］.Neural Computation, 1997,9(8): 1735-1780.

［13］SUNDERMEYER M, NEY H, SCHLUTER R.From feedforward to recurrent LSTM neural networks for language modeling［J］.IEEE/ACM Transactions on Audio, Speech, and Language Processing, 2015,23(3): 517-529.

［14］WANG Y, GAN D H, SUN M Y, et al.Probabilistic individual load forecasting using pinball loss guided LSTM［J］.Applied Energy, 2019,235: 10-20.

［15］涂建维,高经纬,李召,等.基于长短时记忆网络的结构智能控制算法研究［J］.华中科技大学学报(自然科学版), 2019,47(12): 110-115.
TU Jianwei, GAO Jingwei, LI Zhao, et al.Research on structural intelligent control algorithms based on long short-term memory networks［J］.Journal of Huazhong University of Science and Technology (Natural Science Edition), 2019,47(12): 110-115.

［16］袁磊,甘庆鹏,李开成,等.基于深度学习与遗传算法的动车组与ATP车载设备接口试验测试序列优化生成［J］.铁道学报, 2018,40(3): 88-94.
YUAN Lei, GAN Qingpeng, LI Kaicheng, et al.Optimized generation of test sequences for interface type test between train and ATP onboard equipment using deep learning and genetic algorithm［J］.Journal of the China Railway Society, 2018,40(3): 88-94.

［17］KANADA Y.Optimizing neural-network learning rate by using a genetic algorithm with per-epoch mutations［C］// International Joint Conference on Neural Networks (IJCNN).Vancouver: IEEE, 2016.

［18］汪权,庄嘉雷,张俊,等.地震作用下高层建筑结构的重叠分散控制研究［J］.计算力学学报, 2015,32(1): 48-52.
WANG Quan, ZHUANG Jialei, ZHANG Jun, et al.Overlapping decentralized control of tall buildings under earthquakes［J］.Chinese Journal of Computational Mechanics, 2015,32(1): 48-52.

［19］谭平,潘兆东,周福霖.复杂结构重叠分散控制理论研究与数值分析［J］.建筑结构学报, 2018,39(1): 69-77.
TAN Ping, PAN Zhaodong, ZHOU Fulin.Theoretical investigation and numerical analysis of overlapping decentralized control for complex structure［J］.Journal of Building Structures, 2018,39(1): 69-77.

［20］CHEN X B, XU W B, HUANG T Y, et al.Pair-wise decomposition and coordinated control of complex systems［J］.Information Sciences, 2012,185(1): 78-99.

［21］许庆虎.地震激励下参数不确定建筑结构的重叠分散控制方法研究［D］.合肥: 合肥工业大学, 2018.

［22］PALACIOS-QUINONERO F, ROSSELL J M, KARIMI H R.Semi-decentralized strategies in structural vibration control［J］.Modeling, Identification and Control, 2011,32(2): 57-77.

［23］LEI Y, WU D T, LIN Y.A decentralized control algorithm for large-scale building structures［J］.Computer Aided Civil & Infrastructure Engineering, 2012,27(1): 2-13.

［24］GERS F A, SCHMIDHUBER J, CUMMINS F.Learning to forget: continual prediction with LSTM［J］.Neural Computation, 2000,12(10): 2451-2471.

［25］张荣,李伟平,莫同.深度学习研究综述［J］.信息与控制, 2018,47(4): 385-397.
ZHANG Rong, LI Weiping, MO Tong.Review of deep learning［J］.Information and Control, 2018,47(4): 385-397.

［26］宋哲, 陈文卿, 徐志伟.基于神经网络的悬臂梁在线辨识与振动主动控制［J］.振动与冲击, 2013,32(21): 204-208.
SONG Zhe, CHEN Wenqing, XU Zhiwei.Active vibration control of a cantilever beam based on neural network online identification ［J］.Journal of Vibration and Shock, 2013,32(21): 204-208.

［27］KU C C, LEE K Y.Diagonal recurrent neural networks for dynamic systems control［J］.IEEE Transactions on Neural Networks, 1995,6(1): 144-156.

［28］欧进萍.结构振动控制:主动、半主动和智能控制［M］.北京: 科学出版社, 2003.

［29］OHTORI Y, CHRISTENSON R E, SPENCER B F, et al.Benchmark control problems for seismically excited nonlinear buildings［J］.Journal of Engineering Mechanics, 2004,130(4): 366-385.