Multi-Objective Particle Swarm Optimization of Sensor Distribution Scheme with Consideration of the Accuracy and the Robustness for Deformation Reconstruction

Feifei Zhao; Hong Bao; Song Xue; Qian Xu

doi:10.3390/s19061306

Multi-Objective Particle Swarm Optimization of Sensor Distribution Scheme with Consideration of the Accuracy and the Robustness for Deformation Reconstruction

Sensors (Basel). 2019 Mar 15;19(6):1306. doi: 10.3390/s19061306.

Authors

Feifei Zhao¹, Hong Bao², Song Xue³, Qian Xu⁴

Affiliations

¹ Key Laboratory of Electronic Equipment Structure Design of Ministry of Education, Xidian University, Xi'an 710071, China. ffzhao@stu.xidian.edu.cn.
² Key Laboratory of Electronic Equipment Structure Design of Ministry of Education, Xidian University, Xi'an 710071, China. hbao@xidian.edu.cn.
³ Key Laboratory of Electronic Equipment Structure Design of Ministry of Education, Xidian University, Xi'an 710071, China. sxue@xidian.edu.cn.
⁴ Xinjiang Observatory, National Astronomical Observatories, Chinese Academy of Science, Urumqi 830011, China. xuqian@xao.ac.cn.

Abstract

For the inverse finite element method (iFEM), an inappropriate scheme of strain senor distribution would cause severe degradation of the deformation reconstruction accuracy. The robustness of the strain⁻displacement transfer relationship and the accuracy of reconstruction displacement are the two key factors of reconstruction accuracy. Previous research studies have been focused on single-objective optimization for the robustness of the strain⁻displacement transfer relationship. However, researchers found that it was difficult to reach a mutual balance between robustness and accuracy using single-objective optimization. In order to solve this problem, a bi-objective optimal model for the scheme of sensor distribution was proposed for this paper, where multi-objective particle swarm optimization (MOPSO) was employed to optimize the robustness and the accuracy. Initially, a hollow circular beam subjected to various loads was used as a case to perform the static analysis. Next, the optimization model was established and two different schemes of strain sensor were obtained correspondingly. Finally, the proposed schemes were successfully implemented in both the simulation calculation and the experiment test. It was found that the results from the proposed optimization model in this paper proved to be a promising tool for the selection of the scheme of strain sensor distribution.

Keywords: deformation reconstruction; inverse finite element method; multi-objective particle swarm optimization; optimal model; transfer relationship.