[论文解读] Fully convolutional networks for structural health monitoring through multivariate time series classification

Luca Rosafalco, Andrea Manzoni+2 more|arXiv (Cornell University)|Feb 12, 2020

Structural Health Monitoring Techniques参考文献 41被引用 54

一句话总结

本论文提出一种基于仿真的分类方法，使用全卷积网络（Fully Convolutional Networks）来检测并定位结构健康监测（SHM）中的多变量时间序列的损伤，离线在物理基础的数字孪生上训练，达到高达95%的正确损伤分类。

ABSTRACT

We propose a novel approach to Structural Health Monitoring (SHM), aiming at the automatic identification of damage-sensitive features from data acquired through pervasive sensor systems. Damage detection and localization are formulated as classification problems, and tackled through Fully Convolutional Networks (FCNs). A supervised training of the proposed network architecture is performed on data extracted from numerical simulations of a physics-based model (playing the role of digital twin of the structure to be monitored) accounting for different damage scenarios. By relying on this simplified model of the structure, several load conditions are considered during the training phase of the FCN, whose architecture has been designed to deal with time series of different length. The training of the neural network is done before the monitoring system starts operating, thus enabling a real time damage classification. The numerical performances of the proposed strategy are assessed on a numerical benchmark case consisting of an eight-story shear building subjected to two load types, one of which modeling random vibrations due to low-energy seismicity. Measurement noise has been added to the responses of the structure to mimic the outputs of a real monitoring system. Extremely good classification capacities are shown: among the nine possible alternatives (represented by the healthy state and by a damage at any floor), damage is correctly classified in up to 95% of cases, thus showing the strong potential of the proposed approach in view of the application to real-life cases.

研究动机与目标

• 将损伤检测与定位在结构健康监测中表述为分类问题。
• 开发能够处理不同长度的多变量时间序列的基于FCN的架构。
• 使用物理基础的数字孪生在离线对损伤情景进行训练，以应对不同负荷下。
• 在使用训练好的FCNs进行在线监测时实现实时损伤分类。

提出的方法

• 使用一个三卷积块、批归一化、ReLU，以及全局池化然后再通过softmax进行分类的全卷积网络。
• 通过使用独立的卷积分支来处理多数据源，并通过拼接层融合特征。
• 训练两个分类器：G_d用于损伤检测（健康与损伤）以及G_l用于损伤定位（楼层上的损伤，加上未损状态）。
• 从一个简化的物理结构模型构建训练、验证和测试数据集，加入传感器噪声以模拟真实测量。
• 通过创建W个多变量序列并在最终分类前进行数据融合，允许变长的时间序列。
• 使用Adam优化的交叉熵损失，结合学习率调度和早停准则进行训练。

实验结果

研究问题

• RQ1 FCN是否能够在不同加载条件下从多变量时间序列中可靠地区分健康状态与损坏状态？
• RQ2 基于FCN的方法是否能够将损伤定位到结构的具体楼层？
• RQ3 该方法对测量噪声和不同时间序列长度的鲁棒性如何？
• RQ4 离线训练的模型是否能够在在线监测中对未见过的加载情景进行泛化？

主要发现

• 损伤分类在九种可能状态（健康或任意楼层的损伤）下的准确率达到最高95%。
• FCN架构通过独立分支和融合有效处理不同长度的多变量时间序列。
• 在数值基准测试（一个八层剪切建筑）中，噪声扩增数据下显示出极高的分类性能。
• 离线训练使用数字孪生生成多样的损伤和加载情景，使在部署后能够进行实时在线分类。
• 该方法利用FCN自动提取判别特征，减少工程特征设计的需要。