模糊神经网络在钢筋混凝土桥梁结构可靠度分析中的应用

Application of fuzzy neural network in reliability analysis of reinforced concrete bridge

钢筋混凝土桥梁结构可靠度分析中存在着模糊性问题,计算结构可靠度的传统方法没有考虑结构的模糊性问题.将模糊神经网络用于钢筋混凝土桥梁结构的可靠度计算,用神经网络来构造模糊系统,实现输入、输出隶属函数的自动调整,并给出了模糊神经网络计算结构可靠度的步骤.最后通过两个算例来计算结构可靠度,计算结果表明,模糊神经网络法计算结果与JC法计算结果比较接近,该方法可以用于桥梁结构的可靠度计算,特别是可用于功能函数不能显示表达的可靠度计算.该方法的缺点是计算用时稍长以及训练、检验样本的来源存在问题.

Many fuzzy problems exist in reliability analysis of reinforced concrete bridges, traditional methods can＇ t solve this fuzzy problem. Fuzzy neural network is used for reliability calculation of reinforced concrete bridges. Fuzzy system is constituted by neural network to make automatic adjustment of input and output membership functions be carried out. Every step of calculating reliability is given. Finally, the result of two cases is close to the result of JC method. Experimental example demonstrates that the method is used for calculating bridge reliability, especially calculating the reliability when the function can not be determined. Disadvantages of the method are long-time calculation and the lack of training and inspection samples.