Optimizing FBG sensor layout of tunnel monitoring using improved multi-objective snow ablation optimizer based on radial basis function

• Published in: Measurement : journal of the International Measurement Confederation Vol. 242; p. 116289
• Main Authors: Xing, Rongjun, Zhao, Zhongchao, He, Chuan, Xu, Pai, Zhu, Daiqiang, Li, Yufu, Li, Yujun, Yang, Zewen
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 01.01.2025
• Subjects: Chaos mapping; FBG sensor layout; Multi-objective optimization; Pareto frontier; Radial Basis Function; Snow ablation optimizer; FBG sensor layout; Snow ablation optimizer; Multi-objective optimization; Pareto frontier; Chaos mapping; Radial Basis Function
• ISSN: 0263-2241
• DOI: 10.1016/j.measurement.2024.116289

•We proposed an improved algorithm named IMOSAO for FBG sensor layout.•The objective function of monitoring performance was constructed by RBF.•We employed Tent mapping to generate uniformly distributed initial solutions.•Dual-population mechanism was applied for balancing exploration and exploitation.•Results achieved average accuracy above 95.72 % and minimum accuracy 84.61 %. In order to improve the optimization criterion and optimization algorithm, this study introduced the objective function of monitoring performance combining the reconstruction error of strain based on Radial Basis Function and the sensor deployment cost, and proposed the improved multi-objective snow ablation optimizer (IMOSAO) with external archive and Tent mapping. Firstly, initial solutions with uniform distribution were generated by Tent mapping and decimal coding. Then, they were updated by the dual-population mechanism balancing exploration and exploitation. Secondly, the Pareto frontiers of sensor layouts in the external archive were obtained after stop updating. The effectiveness of the improved parts of IMOSAO were validated in ablation studies, numerical and physical experiments. The optimized layouts achieved promising results with the cost reduction of 39 %, average accuracy above 95.72 %, minimum accuracy 84.61 % and least R2 0.9961. Finally, the algorithm reduced operation time and offered superior convergence compared with others. These findings underscore its significant potential in optimizing sensor layout of tunnel monitoring and provide a new view for other measurement sensors.