基于统计学习理论的运营桥梁结构的响应辨识

Identification of operational bridge structure responses using statistical learn theory

基于统计学习理论中的结构风险最小原理,在多重不确定性因素下,提出了对运营桥梁结构的动力响应进行辨识的方法.计算了路面随机谱激励下的运营结构的响应,并引入了统计学习理论,对其辨识过程中存在的结构风险进行了分析,提出了使其风险最小的方法.通过数值算例,对路面随机谱激励以及噪声二重不确定性因素下的结构响应,进行了重构,并与欠拟合与过拟合的结果进行了比较.计算表明,基于结构风险最小的方法,能有效地对二重不确定性因素下的土木结构响应进行辨识,并进一步讨论了算法的改进.

A method to indentify the responses of an operational bridge structure to dynamitic loads under multiuncertain factors is proposed by using the principle of structural risk minimum from statistical learn theory. The bridge responses with a moving vehicle are numerical simulated based on a simplified model under road excitations and Gaussian white noises, and the structural risk of the identifications are investigated. The under-fitting and the over-fitting results are compared to the identification results with statistical learn theory concerning the double uncertainties, which to illustrate the method validity. The further work of the method is discussed.