基于神经网络的部分预应力混凝土梁荷载-裂缝模型研究

The Load and Crack Width Model for Partially Prestressed Concrete Beams Based on Neural Network

混凝土结构裂缝形成和发展的过程十分复杂 ,具有一定的随机性 ,难于用常规的方法进行建模。应用神经网络较强的函数映射能力和联想、记忆功能 ,对部分预应力混凝土梁的荷载 裂缝关系进行建模 ,通过试验数据进行验证 ,效果良好 ,证明应用神经网络来对部分预应力混凝土梁的荷载 裂缝关系进行建模是可行的。

The initiation and propagation of cracks in concrete structure are very complex and to some extent, and it is difficult to model it in a regular way. A model based on neural network, which is good at association and generalization, is proposed to study the relationship between the crack width and the applied load in partially prestressed concrete beams. The model is verified by some test results with satisfaction.