涡激响应尾流双振子模型的参数反演方法

尾流双振子模型是研究圆柱结构涡激振动响应的重要模型,模型参数的准确确定对悬浮隧道设计理论具有重要意义.首先通过降阶法将多变量二阶非线性常微分方程组的尾流双振子模型变换为一阶方程组.然后给出一种新型的圆柱结构水槽试验设计方案,其中试验模型的刚度能够较好反映悬浮隧道等实际工程结构的刚度,基于相机动态捕捉和视频识别计算机程序,获取圆柱结构在水平和竖直方向的位移试验数据.基于试验结果和龙格-库塔方法求解一阶方程组,采用BP神经网络智能算法对模型参数进行反演,同时利用遗传算法对神经元的初始权值和阈值进行优化,所得结果平均误差仅为5.50％,优于遗传算法和未优化的BP神经网络模型.结果表明,基于遗传算法优化的BP神经网络智能算法能够精确实现尾流双振子模型的参数确定,为圆柱结构涡激振动响应分析提供理论基础.

Parameter inversion method of double wake oscillators model with vortex induced response

The double-wake oscillator model is an important model to study the vortex induced vibration response of a cylindrical structure,and the accuracy of model parameters is of great significance to the theoretical design of a submerged floating tunnel.Firstly,the double-wake oscillator model with the form of multivariable second order nonlinear ordinary differential equations is transformed into the first order equations by the method of order reduction.Then,a new test design scheme of the cylindrical structure is given,in which the stiffness of the experimental model can better reflect the stiffness of the practical engineering structure such as the submerged floating tunnel,and through the camera dynamic capture and computer program of video recognition,the displacements of the cylindrical structure in horizontal and vertical directions are obtained.Based on the test results and Runge-Kutta method,BP neural network intelligent algorithm is used to the inversion of model parameters,and a genetic algorithm is used to optimize the initial weights and thresholds of neurons.The average error of the results is only 5.50%,which is superior to the genetic algorithm and the unoptimized BP neural network model.The results show that the BP neural network intelligent algorithm based on the genetic algorithm optimization can determine the parameters of the double-wake oscillator model accurately,which will provide theoretical basis for the vortex induced vibration response analysis of cylindrical structures.