摘要
本文提出了依据神经网络方法完成大型空间网架结构节点固结系数的识别,实现空间网架结构有限元模型修正的方法。文中,通过有效独立法优化配置有限数量传感器的安放位置。在此基础上,利用从优化位置处的结构时程响应中获取的结构自振频率和局部位移模态信息,本文建立了基于径向基神经网络的空间网架结构有限元模型修正的计算方法。数值仿真的计算结果表明,该方法能有效地用于空间网架结构的有限元模型修正。
The finite element model correction of large space truss is done by identifying its joint fixity factor based on neural net. The finite sensors are placed in an optimal fashion by effective independence method. The structure frequency and local modal shape are acquired from structural time-history responses at optimal sensor placement location, then the RBF neural net is established to correct the finite element model of space truss. The numerical example shows that the method can be effectively used to correct the finite element model of space truss structures.
出处
《地震工程与工程振动》
CSCD
北大核心
2003年第4期83-89,共7页
Earthquake Engineering and Engineering Dynamics
基金
国家自然科学基金会主任基金(50148020)
国家自然科学基金(50278057)
