中文摘要：

为提高混凝土桥墩塑性铰区箍筋用量的计算精度，针对已有文献所提出的延性桥墩塑性铰区箍筋用量的近似计算公式的不足，在其基础上，BI入径向基网络理论，改进其计算方法．以轴压比、纵筋配筋率、混凝士轴心抗压强度、实测延性系数为输入参数，力学含箍率为输出参数，建立精确径向基网络模型，以多组试验数据为训练样本，对网络进行训练，得到具有高度非线性映射关系的网络模型，通过新的试验数据对网络进行了检验，并与采用非线性最小二乘法拟合公式的计算结果进行了比较．结果表明：用训练成熟的径向基网络进行仿真，避免了诸多人为因素的影响，大大提高了计算结果的精度，使得计算更加准确、高效、因此，径向基网络应用于延性桥墩塑性铰区箍筋用量计算是可行并实用的．

英文摘要：

In order to improve the accuracy of calculation for stirrup ratio in plastic hinge regions of concrete piers, an improved method is presented based on Radial Basis Function Neural Network （RBFNN） and present existing approximate calculation methods. A RBFNN was built up with the axial compressive ratio, the longitudinal rebar ratio, the concrete axial strength and the ductility coefficient considered as input factors and the ratio of mechanical stirrup adopted as an output factor. Trained by some experimental samples, RBFNN with highly non-linear reflection relationship was founded and checked by other samples. It is proved to be more ef- fective and accurate by comparing with the results gained by the method of non-linear least square. At the same time, it also avoids many man-induced factors. So based on RBFNN, the improved method is applicable and practical in calculating the stirrup ratios in plastic hinge regions of concrete piers.