.png){kind=icon}
返回
Sequential sensor placement for damage detection under frequency-domain dynamics
.png){kind=icon}
PRE
AI
DOI：10.1016/j.finel.2025.104315.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Identification and monitoring of damage have a growing importance in the maintenance of structures. A robust active sensing framework that integrates model-based inference and optimal sensor placement is proposed. By tightly coupling measured data and data acquisition scenarios, a simultaneous approach of damage estimation and sensor placement can be used to continuously and accurately assess a structure. In this work, a partial differential equation-constrained formulation for damage estimation is first developed using a conventional model-updating approach with a penalization damage parameter. Then, this formulation is linearized around the damage estimator to produce an Optimal Experimental Design (OED) problem for desirable sensor locations. Hence, the simultaneous sensing framework is postulated using a Fisher Information Matrix (FIM)-based approach as follows: given a current candidate damage state associated with the most up-to-date sensor information, find the next sensor location that minimizes some metric of the FIM and update the damage estimator. The sensing framework is also enhanced by introducing a Modified Error in Constitutive Equations (MECE) functional in the damage estimator. Adding MECE makes the framework more robust by limiting the damage estimator from being trapped in local minima. Through numerical examples, we show that our approach produces accurate damage estimators using a small number of sensor locations. In addition, we compare our results to those obtained using random sensor selections and expertly selected locations.
Keyword:
Optimal experimental design
Damage detection
Inverse problems
期刊
IF:
3.5
论文数:
2.6K
被引数:
5.1K
机构
引用论文
.png){kind=icon}
分享收藏EFFICIENT D-OPTIMAL DESIGN OF EXPERIMENTS FOR INFINITE-DIMENSIONAL BAYESIAN LINEAR INVERSE PROBLEMS
OAAI
OAAI分享收藏
查看更多内容