基于灰色-神经网络的大跨度斜拉桥参数识别

Parameter Identification of Long-Span Cable-Stayed Bridge Based on Grey-Neural Network

为解决大跨度斜拉桥施工过程中观测噪声对结构参数识别的影响,以苏通大桥为工程背景,提出了基于灰色-神经网络的施工全过程参数识别方法.灰色系统理论与人工神经网络相融合,在小样本和数据不完备的情况下可以进行结构参数识别,并具有预测功能.苏通大桥参数识别的结果表明,参数识别后的计算值与实际结构挠度响应间的最大误差减小77%,与传统识别方法相比,灰色-神经网络方法的识别精度提高50%.

To solve the structural parameter identification construction process of long-span cable-stayed bridges, problem with measurement noises during the an entire-processed parameter identification method based on grey-neural network was proposed. The fusion of the grey system theory and the artificial neural network makes it possible to identify and forecast structural parameters even under the condition of insufficient samples or incomplete data. The identification results for the Su-Tong bridge show that with the proposed method the maximum error between post-identification result and real structural response decreases by 77% and the estimation accuracy is raised by 50% than with traditional ones.