JOURNAL OF YANGTZE RIVER SCIENTIFIC RESEARCH INSTI ›› 2018, Vol. 35 ›› Issue (4): 67-71.DOI: 10.11988/ckyyb.20160857

• ENGINEERING SAFETY AND DISASTER PREVENTION • Previous Articles     Next Articles

Application of Optimized Support Vector Machines and V/S Analysis to Tunnel Deformation Prediction and Trend Judgment

ZHANG Bi   

  1. Shaanxi Railway Institute, Weinan 714000, China
  • Received:2016-08-20 Online:2018-04-01 Published:2018-04-16

Abstract: Tunnel deformation is of obvious non-linear characteristics. In the present research, models based on the Kalman filter and a variety of optimization support vector machines are built for accurate prediction. The applicability of various models is discussed, and further combinatorial prediction is conducted. Meanwhile, V/S analysis is adopted to calculate the Hurst index of deformation series for deformation trend judgment. The judgment result iscompared with prediction result in the aim of obtaining the comprehensive deformation rules of tunnel. Results suggest that the least squares support vector machine has the optimum effect, and the deformation in the next four cycles would keep increasing. Moreover, the Hurst index of deformation series and deformation rate series is 0.845 and 0.602, respectively, both larger than 0.5, indicating that the deformation in late stage would experience a sustained growth, which is consistent with the prediction result.

Key words: tunnel, support vector machine, coefficient of variation, V/S analysis, deformation prediction, trend judgment

CLC Number: