基于概率神经网络的桩-承台的损伤识别

On Damage Localization of Thick Pile Cap Based on Probabilistic Neural Network

通过有限元计算证明了桩厚承台的空间桁架传力机理,对其桁架传力中的拉、压杆进行损伤模拟,分别得到不同损伤程度和未损伤工况下的模态参数,并用概率神经网络进行了损伤位置的识别.结果表明,用概率神经网络对基于空间桁架传力机理的厚承台进行损伤识别是可行的.

The load transfer mechanism of space truss of the thick pile cap was verified by the finite element method and then the different levels of damaged cases and the undamaged case were simulated at the tension rods and the compression bars of the load transfer mechanism. The mode parameters were obtained respectively. The probabilistic neural network was employed to locate the damage of the pile cap. The results showed that to locate the damage in thick pile cap based on load transfer mechanism of space truss model was feasible.