Physics-Informed Multi-LSTM Networks for Metamodeling of Nonlinear Structures
Physics-Informed Multi-LSTM Networks for Metamodeling of Nonlinear Structures
复制标题
DOI：
10.1016/j.cma.2020.113226
复制
发表时间：
2020-02
期刊：
影响因子：
--
通讯作者：
Ruiyang Zhang;Yang Liu;Hao Sun-
中科院分区：
文献类型：
--
作者：
Ruiyang Zhang;Yang Liu;Hao Sun-
本文介绍了一种创新的物理深度学习框架，用于在数据稀缺的情况下对非线性结构系统进行元建模。基本概念是将可用但不完整的物理知识（例如，物理定律、科学原理）转化为深度的长短期记忆（LSTM）网络，从而在可行的解空间内约束和促进学习。物理约束嵌入在损失函数中，以加强模型训练，即使可用的训练数据集非常有限，也可以准确地捕获潜在的系统非线性。对于动力结构，考虑了运动方程、状态相关性和滞回本构关系等物理规律来构造物理损失。特别地，提出了两种物理信息的多LSTM网络架构用于结构元建模。通过两个示例（例如，受到地面运动激励的非线性结构）。事实证明，嵌入式物理可以缓解过拟合问题，减少对大训练数据集的需求，并提高训练模型的鲁棒性，以便通过外推能力进行更可靠的预测。因此，基于物理学的深度学习范式优于经典的非物理学引导的数据驱动神经网络。
This paper introduces an innovative physics-informed deep learning framework for metamodeling of nonlinear structural systems with scarce data. The basic concept is to incorporate available, yet incomplete, physics knowledge (e.g., laws of physics, scientific principles) into deep long short-term memory (LSTM) networks, which constrains and boosts the learning within a feasible solution space. The physics constraints are embedded in the loss function to enforce the model training which can accurately capture latent system nonlinearity even with very limited available training datasets. Specifically for dynamic structures, physical laws of equation of motion, state dependency and hysteretic constitutive relationship are considered to construct the physics loss. In particular, two physics-informed multi-LSTM network architectures are proposed for structural metamodeling. The satisfactory performance of the proposed framework is successfully demonstrated through two illustrative examples (e.g., nonlinear structures subjected to ground motion excitation). It turns out that the embedded physics can alleviate overfitting issues, reduce the need of big training datasets, and improve the robustness of the trained model for more reliable prediction with extrapolation ability. As a result, the physics-informed deep learning paradigm outperforms classical non-physics-guided data-driven neural networks.