Multiaxial sensor placement optimization in structural health monitoring using distributed wolf algorithm. Issue 4 (1st October 2015)

• Record Type: Journal Article
• Authors: Yi, Ting‐Hua; Li, Hong‐Nan; Wang, Chuan‐Wei
• Is Part Of: Structural control and health monitoring. Volume 23:Issue 4(2016)
• Page Start: 719
• Page End: 734
• Publication Date: 2015-10-01
• Subjects: Structural health monitoring -- optimal sensor placement -- distributed wolf algorithm -- modal assurance criterion -- benchmark structure; Structural engineering -- Periodicals; Structural control (Engineering) -- Periodicals; Automatic data collection systems -- Periodicals; Detectors -- Periodicals; 624.17
• DOI: 10.1002/stc.1806
• Languages: English
• ISSNs: 1545-2255

Abstract

Summary: Optimal sensor placement technique plays a key role in the design of an effective structural health monitoring system. Recent advances in sensing technologies have also promoted using multiaxial sensors to perform efficiently and economically monitoring for civil engineering structures. However, the available evaluation criteria for the optimal sensor placement can only guarantee that the optimization is conducted in a single structural direction but not in multi‐dimension space, which may result in the non‐optimal placement of multiaxial sensors. To tackle this issue thoroughly, a new multiaxial optimal criterion termed as the triaxial modal assurance criterion is developed by taking account into three translational degrees of freedom as a single unit in the Fisher information matrix. Afterwards, a novel distributed wolf algorithm is proposed to improve the optimization performance in identifying the best sensor locations. The dual‐structure coding method is improved and adopted to represent the solution. The shuffling strategy is proposed to enhance the searching capability and convergence performance. The attacking process is also modified to prevent the algorithm from being trapped in a local minimum. The effectiveness of the proposed scheme is investigated by the benchmark structure developed by the University of Central Florida, USA. The results clearly demonstrate that the proposed distributed wolf algorithm outperforms the existing algorithm in its global optimization capability. Copyright © 2015 John Wiley & Sons, Ltd.