For construction control of a cantilever-erected bridge, parametric sensitivity analysis (SA) is especially essential by providing quantitative indices on how important a structural parameter may influence structural response. Different from the conventional first-order derivative method dealing with a small variation of parameters, in this paper, a global SA method considering random distributions of parameters is presented. Ratio of response's coefficient of variation (CV) to parameter's CV is defined as the global sensitivity index, which can be obtained by carrying out uncertainty assessment on structural responses with respect to given uncertainties of parameters. An artificial neural networks embedded Monte Carlo method is applied for the uncertainty analysis. Finally, application of the proposed global SA method in cantilever erection practice of a cable-stayed bridge is presented, and importance of material parameters and load parameters for structural responses obtained.
[1]
Mohammad Bagher Menhaj,et al.
Training feedforward networks with the Marquardt algorithm
,
1994,
IEEE Trans. Neural Networks.
[2]
C. Fortuin,et al.
Study of the sensitivity of coupled reaction systems to uncertainties in rate coefficients. I Theory
,
1973
.
[3]
N J Gimsing.
CABLE SUPPORTED BRIDGES: CONCEPT AND DESIGN. SECOND EDITION
,
1997
.
[4]
Francis T.K. Au,et al.
Acceleration of Levenberg-Marquardt training of neural networks with variable decay rate
,
2003,
Proceedings of the International Joint Conference on Neural Networks, 2003..
[5]
Joseph A. C. Delaney.
Sensitivity analysis
,
2018,
The African Continental Free Trade Area: Economic and Distributional Effects.
[6]
Erik H. Vanmarcke,et al.
Random Fields: Analysis and Synthesis.
,
1985
.