Multi-influencing Factor Weighted WPSO–SVM Prediction of Subway Tunnel Settlement under GRA Supports

• DOI: 10.25103/jestr.172.24
• Publication Date: 2024-05-12T09:24:31Z
• Authors: Caiyun Gao
• Keywords: Settlement (finance); Tunnel Construction

With the rapid economic development and urbanization in China, subway systems have become primary mode of urban rail transit.However, during operation, tunnels may experience settlement deformation due to various influencing factors.To guarantee safe operation eliminate potential safety hazards, tunnel prediction has important significance.However, existing studies seldom discussed effects weighting factors on prediction.In addition, optimization support vector machine (SVM) using particle swarm (PSO) often suffers from issues such as local premature convergence.To address these problems, grey relational analysis (GRA) weighted (WPSO) SVM were combined, a GRA-WPSO-SVM model was constructed.This applied predict Sanyao Section Xi'an Exhibition Center China.Prediction results compared with those PSO-SVM root mean square error (RMSE), relative (MRE), correlation coefficient evaluation metrics.Results demonstrate that, RMSE MRE are 0.0008 m 1.9707%, which better than SVM.Moreover, exhibit strong measured data tunnels, 0.93.Obviously, is effective.The proposed method provides an evidence for trends.