基于LSTM深度学习网络的分布动载荷识别

PDF(2978 KB)

振动与冲击 ›› 2024, Vol. 43 ›› Issue (11) : 126-134.

论文

郭安丰1，吴邵庆1,2

作者信息 +

Distributed dynamic load identification based on LSTM deep learning network

GUO Anfeng1, WU Shaoqing1,2

Author information +

文章历史 +

摘要

提出一种基于LSTM深度学习网络的分布动载荷识别新方法。首先，建立结构有限元模型并对载荷作用区域进行平面化和子区域网格划分，构建子区域上以形函数形式分布的动载荷和有限元模型节点动响应之间的传递关系，建立节点处应变动响应与对应子区域上分布动载荷的样本库；其次，利用Meyer小波对样本库中的时域样本进行特征提取，并基于LSTM深度学习网络训练子区域上分布动载荷与有限元模型节点应变动响应的传递关系；最后，开展了数值仿真研究，利用有限元模型仿真应变动响应识别了三维壁板结构表面的分布动载荷，验证了所提出方法的有效性。研究旨在为服役状态下壁板结构上动载荷环境预示提供技术支撑。

Abstract

A novel distributed dynamic load identification method based on LSTM deep learning network is proposed. Firstly, finite element model of structure is built and flatting as well as sub-regions partitioning on the area where load acts is conducted, the transitive relationship between dynamic load on sub-regions with the form of shape functions and nodal dynamic response on finite element model is constructed，then the sample database containing nodal strain response and distributed dynamic load on the corresponding sub-regions can be established; Secondly, Meyer wavelet is adopted to extract the features of the time-domain samples in the database and then the LSTM deep learning network is used to train the transitive relationship between the nodal strain response of finite element model and distributed dynamic load on the corresponding sub-regions; Finally, numerical simulations are conducted in which the simulated strain response is adopted to identify the distributed dynamic load on a three-dimensional panel structure, the effectiveness of the proposed method is verified. This research work aims to provide technical support on dynamic load environment prediction on the panel structure in service.

导出引用

郭安丰1，吴邵庆1,2. 基于LSTM深度学习网络的分布动载荷识别[J]. 振动与冲击, 2024, 43(11): 126-134

GUO Anfeng1, WU Shaoqing1,2. Distributed dynamic load identification based on LSTM deep learning network[J]. Journal of Vibration and Shock, 2024, 43(11): 126-134

参考文献

[1] 周盼, 张权, 率志君, 等. 动载荷识别时域方法的研究现状与发展趋势 [J]. 噪声与振动控制, 2014, 34(01): 6-11. Zhou Pan，Zhang Quan，Shuai Zhijun，et al. Research status and development trend of time domain method for dynamic load identification [J]. Noise and Vibration Control, 2014, 34(1): 6-11. [2] Wada D, Sugimoto Y, Murayama H, et al. Investigation of inverse analysis and neural network approaches for identifying distributed load using distributed strains [J]. Transactions of the Japan Society for Aeronautical and Space Sciences, 2019, 62(3): 151-61. [3] Liu H, Liu Q, Liu B, et al. An efficient and robust method for structural distributed load identification based on mesh superposition approach [J]. Mechanical Systems and Signal Processing, 2021, 151:107383. [4] Lei Y, Zhang F, Zhang Y, et al. Identifying Distributed Dynamic Loading in One Spatial Dimension Based on Combing Wavelet Decomposition and Kalman Filter with Unknown Input [J]. Journal of Aerospace Engineering, 2021, 34(4):04021025. [5] Zheng Y, Wu S, Li Y, et al. Identify the spatially-correlated random fluctuating pressure on structure from strain data [J]. Aerospace Science and Technology, 2021, 119:107182. [6] Jiang J, Luo S, Zhang F. One novel dynamical calibration method to identify two-dimensional distributed load [J]. Journal of Sound and Vibration, 2021, 515:116465. [7] Liu J, Li K. Sparse identification of time-space coupled distributed dynamic load [J]. Mechanical Systems and Signal Processing, 2021, 148:107177. [8] Liu Y, Wang L, Li M, et al. A distributed dynamic load identification method based on the hierarchical-clustering-oriented radial basis function framework using acceleration signals under convex-fuzzy hybrid uncertainties [J]. Mechanical Systems and Signal Processing, 2022, 172:116465. [9] Zheng Y, Wu S, Fei Q. Distributed Dynamic Load Identification on Irregular Planar Structures Using Subregion Interpolation [J]. Journal of Aircraft, 2021, 58(2): 288-299. [10] 高文静, 周焕林, 陶然. 基于布谷鸟搜索算法的动载荷识别[J]. 重庆大学学报, 2020, 43(06):30-39. [Gao W, Zhou H, Tao R. Dynamic load identification based on the cuckoo search algorithm [J]. Journal of Chongqing University, 2020, 43(6): 30-39.] [11] Yu B, Wu Y, Hu P, et al. A non-iterative identification method of dynamic loads for different structures [J]. Journal of Sound and Vibration, 2020, 483:115508. [12] 夏鹏, 杨特, 徐江, 等. 利用时延神经网络的动载荷倒序识别[J]. 航空学报, 2021, 42(7):224452. [Xia Peng, Yang Te, Xu Jiang, et al. Reversed time sequence dynamic load identification method using time delay neural network [J]. Acta Aeronautica et Astronautica Sinica, 2021, 42(7):224452.] [13] Yang H, Jiang J, Chen G, et al. A Recurrent Neural Network-Based Method for Dynamic Load Identification of Beam Structures [J]. Materials, 2021, 14(24):7846. [14] Yang H, Jiang J, Chen G, et al. Dynamic load identification based on deep convolution neural network [J]. Mechanical Systems and Signal Processing, 2023, 185:109757. [15] 曾俊玮, 季元进, 任利惠, 等.卡尔曼滤波器与神经网络串行的轮胎载荷识别模型[J]. 振动与冲击,2023,42(11):262-270. [Zeng J, Ji Y, Ren L, et al. A serial tire load identification model based on Kalman filter and neural network. Journal of Vibration and Shock, 2023, 42(11): 262-270.] [16] 张志宏, 张宏, 陈有, 等.基于遗传神经网络的履带行驶系统载荷识别方法[J]. 振动与冲击,2022,41(03):54-61. [Zhang Z, Zhang H, Chen Y, et al. Load identification method of track driving system based on genetic neural network. Journal of Vibration and Shock, 2022, 41(3): 54-61.] [17] Zhang Q, Luo X. Finite element method for developing arbitrary surfaces to flattened forms [J]. Finite Elements in Analysis and Design, 2003, 39(10): 977-984. [18] McCartney J, Hinds BK, Seow BL, et al. An energy based model for the flattening of woven fabrics [J]. Journal of Materials Processing Technology，2000, 107(1-3):312-318.