论文部分内容阅读
桥墩计算长度系数与其墩高、墩身截面尺寸以及边界条件等均相关,该值的确定在工程界一直未能得到妥善解决。基于MATLAB径向基神经网络原理,首先通过有限元与欧拉公式计算采集训练数据样本,进而利用MATLAB对数据进行训练创建径向基神经网络;然后通过随机选取若干组测试数据,对比分析有限元与欧拉公式、径向基网络拟合以及线性拟合3种方法下的计算长度系数。对比结果显示,基于MATLAB径向基神经网络拟合所得的墩柱计算长度系数,与有限元结合欧拉公式所得结果吻合良好,同时具备快捷方便的特点,可以在今后的工程设计中加以应用。
The calculation of the length factor of pier is related to the height of pier, the section size of pier body and the boundary conditions. The determination of this value has not been properly solved in engineering field. Based on MATLAB Radial Basis Function (RBFNN) theory, the training data samples were acquired by finite element and Euler equations firstly, and then the data were trained by MATLAB to create radial basis neural networks. Then, by randomly selecting several sets of test data, And Euler’s formula, radial basis network fitting and linear fitting three kinds of calculation of length coefficient. The comparison results show that the calculated length coefficient of piers based on MATLAB Radial Basis Function Neural Network fitting well with the results obtained by the Eulerian finite element method combined with the fast and convenient features can be applied in the future engineering design.