基于频响函数提升小波总能量的桁架结构随机模型修正

为了解决模型修正问题中的随机性,构建了一种基于提升小波总能量的随机模型修正方法.首先,将结 构的加速度频响函数进行提升小波变换,并提取提升小波总能量来代替加速度频响函数;然后,以待修正参数作为输入,提升小波总能量为输出构建响应面代理模型代替原来的有限元模型;接着,运用蒙特卡洛抽样抽取响应样本,并设定阈值筛选响应样本;最后,以代理模型预测得到的响应和抽样所得真实响应之间的差值最小为 目标函数,通过布谷鸟优化算法寻优求解待修正参数的均值.算例表明,所提方法修正后参数的最大误差小于3.3％,相应的频响函数曲线重合度高.

Stochastic model updating of truss structures based on total energy of lifting wavelet transform of frequency response function

In order to deal with the randomness in model updating problems,a stochastic model updating method based on the total energy of the lifting wavelet transform is constructed.Firstly,the acceleration frequency response function of the structure is subjected to lifting wavelet transformation and the total energy of the lifting wavelet transform is extracted to replace the acceleration frequency response function.Then,the parameters to be updated are taken as input and the total energy of the lifting wavelet transform as the output to construct a response surface model instead of the original finite element model.And then,Monte Carlo sampling is used to extract response samples,and optimize the response samples by setting a threshold.Finally,the minimum difference between the response predicted by the surrogate model and the real response obtained by sampling is defined as the objective function,the mean of each parameter to be updated is solved by iterative optimization through the cuckoo optimization algorithm.The calculation example shows that the maximum error of the parameters from the proposed method is less than 3.3%,and the corresponding frequency response function curve has a high degree of agreement.