神经网络与响应面法相结合分析既有混凝土桥梁的可靠性

The Reliability Assessment of Existing Bridge Based on Neural Network and Response Surface Method

响应面有限元的可靠度计算是采用有限元数值模拟来解决功能函数不能明确表示的结构可靠度计算问题的一类方法,对于大型复杂结构的可靠度分析有重要的意义.提出了基于BP神经网络与响应面法相结合的结构可靠度计算的几何分析方法,通过得出的可靠指标对一座既有桥梁进行了可靠性分析.数值试验表明,该方法建立的神经网络模型可以很好地拟合真实的极限状态函数,在既有桥梁可靠性分析中具有广阔的应用前景.

Reliability calculation based on response surface FEM method utilizes the FEM numerical simulation to solve the reliability problems in which the analytical performance function can not be given. It is important for calculating the reliability of the large and complicated structures. A geometric analysis method of structural reliability calculation is proposed based on BP neural network and response surface method. A reliability assessment of the simply-supported RC bridge beam is also conducted through the reliability index. The numerical experiments show that the model of neural network by this method can simulate the true ultimate state function, and it has a prospect of wide application in the reliability assessment of existing bridge.