运营隧道病害多因素综合等级评价方法研究

Research on multi-factor comprehensive grade evaluation method of operational tunnel diseases

在传统神经网络结构的基础上,文章将模糊理论与人工前向神经网络相结合,并采用神经网络学习算法对隶属函数的中心点、宽度向量以及输出层的连接权值进行了优化,提出了隧道结构健康状态评价的模糊神经网络预测模型。同时,将收集整理已形成的隧道病害分级样本库作为训练数据集和测试数据集,文章对模糊神经网络模型进行了训练和测试,分别实现了单一结构病害健康状态评价和多种病害耦合作用健康状态评价。训练结果证明,模糊神经网络模型可以迅速、方便、准确地实现隧道结构健康状态评价,从而为隧道的全生命周期运营养护提供依据。

On the basis of traditional neural network structures,this article combines fuzzy theory with artificial feedforward neural networks,and uses neural network learning algorithms to optimize the center point,width vector,and connection weight of the output layer of the membership function.A fuzzy neural network prediction model for evaluating the health status of tunnel structures is proposed.At the same time,the collected and organized tunnel disease classification sample library will be used as the training and testing datasets.The article trains and tests the fuzzy neural network model,achieving the health status evaluation of a single structural disease and the health status evaluation of multiple disease coupling effects,respectively.The training results have shown that the fuzzy neural network model can quickly,conveniently,and accurately evaluate the health status of tunnel structures,thereby providing a basis for the full lifecycle operation and maintenance of tunnels.