中文摘要：

根据质量影响吊杆自振频率的特性,用附加质量法增加识别参数,实现对短吊杆张力的识别。随机生成吊杆的张力和弯曲刚度,并运用有限元方法计算出对应的附加质量前后吊杆的自振频率,构成神经网络的教师数据,训练神经网络,拟合吊杆的自振频率与弯曲刚度和张力之间的非线性关系,进而建立基于附加质量法和神经网络的短吊杆张力和弯曲刚度识别系统。以1组长度为2～10m的短吊杆和在建拱桥的2根试验短吊杆为例,通过数值模拟检验识别方法及系统的有效性,识别误差分别为0.5‰和6%左右,表明所提出的识别方法和系统可行和有效,且识别精度比弦振动理论计算方法有显著提高。

英文摘要：

In order to determine the unknown parameters like flexural rigidity in tension identification of short hangers,a new method called additional mass method was proposed based on the significant effect of mass on the natural frequency of the hangers. Artificial Neural Network （ANN） was employed to solve the nonlinear relationship among frequencies,tension and flexural rigidity. The tension and flexural rigidity of hanger were randomly generated,and the natural frequencies before and after the attachment of the mass were calculated by finite element method,these parameters were used as trainning data in ANN,which was founded to identify the tension and flexural rigidity of hangers. A group of short hangers length from 2～10 m were studied,and field experiments were taken in an arch bridge under construction to validate the feasibility and effectivity of this approach. The identification errors are about 0.5‰ and 6% resepectivily,which were better than the string theory method obviously.