为了进一步研究体外预应力混凝土结构体外索极限应力的确定方法,在分析影响体外预应力混凝土结构体外索极限应力因素的基础上,运用神经网络预测的原理,采用误差反向传播神经网络即BP神经网络对体外索的索力进行了预测。预测结果表明,应用BP神经网络模型能准确的预测体外索的索力。预测结果和试验结果相比误差在10%以内,可以满足实际工程的需要。
In order to determine the maximum stress of the externally prestressed concrete structures,various factors that influence the stress were analyzed,and the theory of BP(Back Propagation) neural network was used to predict the external tendon force.The results show that the BP neural network model can predict the limit of the external tendon force and the error of the results is less than 10% compared to the experiment results,which is good enough to satisfy practical engineering needs.