论文部分内容阅读
将层次分析等理论更快速方便地应用于结构评估工程实践,研究基于现有理论的自动化、智能评估方法.借助Matlab模糊数学、神经网络工具箱,采用线性映射的层次分析方法完成桥梁耐久性等级评估.FIS结构和BP模型评估结果的对比表明:受隶属函数限制,FIS结构评估结果较粗糙,但由于模糊关系的特性,处理复杂输入输出关系时效果较好;而当样本数据集在超空间中线性可分时,神经网络对已有数据的拟合精度更高,但只能处理与训练样本同构的评估数据.由此证明:利用Matlab软计算模型实现桥梁工作性能的自动化智能评估是可行的.
The theory of analytic hierarchy process (AHP) is more quickly and conveniently applied to the engineering practice of structural assessment, and the method of automation and intelligence assessment based on the existing theory is studied.Using the Matlab fuzzy mathematics and neural network toolbox, the level of durability of the bridge The comparison between the FIS structure and the BP model shows that the results of the FIS structure assessment are rough due to the membership functions and the results of the FIS structure evaluation are rather rough. However, due to the characteristics of the fuzzy relation, the effect is better when dealing with complex input- The neural network can fit the existing data more accurately but can only deal with the isomorphic evaluation data of the training samples.This proves that the automatic intelligent evaluation of the working performance of bridge using Matlab soft computing model is feasible.