【摘要】 结合5种混凝土延性柱耗能器在低周期反复荷载作用下的试验数据研究,利用神经网络的工作原理,通过建立神经网络的输入层、隐含层、输出层,确定输入单元、输出单元和隐含层节点数,从而建立了BP神经网络的模型,并根据已有的部分试验数据数据.对网络进行训练,对各种混凝土延性柱耗能器骨架曲线进行了预测拟合,实现混凝土延性柱耗能器骨架曲线的数字化,使其成为具有分析和判断的拟合曲线功能,完整的描绘混凝土延性柱耗能器的骨架曲线,为后续混凝土延性柱耗能器性能研究的仿真模拟提供了可靠的数据模型.结果表明,这种方法是可行的.更多还原

【Abstract】 Combined with 5 kinds of concrete ductility column energy dissipator at low cycle load test data research,the working principle of neural network,and by establishing a neural network’s input layer,hidden and output layer,determine inputs unit,output unit and hidden node number,and to establish the BP neural network model,and part of the test data according to the existing data.Networks are trained to of all kinds of concrete ductility column energy dissipator skeleton curve fitting,forecast the realization concrete ductility column energy dissipator skeleton curve digital,make it become with analysis and judgment of the fitting curve function,complete description of concrete column energy dissipator ductility of skeleton curves,for the subsequent concrete ductility column energy dissipator performance simulation study provides the reliable data model.The results show that the method is feasible.更多还原