基于模糊神经网络的既有钢筋混凝土桥梁构件可靠性评估

介绍了模糊神经网络的基本理论知识,根据其知识体系和既有桥梁可靠性的研究特点,结合相关文献的规定和研究成果,引入与所检测的截面裂缝宽度、应变值和应变值相关的评估参数Z,利用既有钢筋混凝土桥梁养护等级标准给出了构件技术状况等级与可靠指标的对应关系,然后根据模糊神经网络的理论优势,对输入变量(实测数据:裂缝宽度,应变值和应变值)进行训练,计算出构件在不同荷载作用下的可靠指标,其可用于既有桥梁结构构件可靠性评估中。并结合一实桥拆除构件的试验结果,得到了一些有益的结论,验证了本方法应用于既有钢筋混凝土桥梁构件可靠性评估的可行性。

Reliability Evaluation of Existing Reinforced Concrete Bridge Members Based on Neuro-fuzzy Network

Theory knowledge of neuro-fuzzy network is introduced. Incorporating with rules and results of relevant literatur, evaluation parameter Z of a section in relation to the width of crack,strain and deflection is applied,based on the knowledgeable system and research characteristics of the existing bridge reliability. Using the maintenance condition levels of the existing reinforced concrete bridge, the relationship between the technical state grade of bridge-members and the reliability index are given .Based on the theoretic advantage of neurofuzzy network , the reliability index is calculated after training input variable(actual testing data e.g. the width of crack, strain and deflection). It would be used to evaluate reliability index with consideration of different loads of the existing reinforced concrete member. Some useful conclusions are obtained and the feasibility of this calculating method is testified,by destructive test of the structure member removed from the existing bridge.