基于BP神经网络的股线力学性能数值预报

Numerical Prediction of Mechanical Properties of Plied Yarn Using BP Neural Network

基于股线的参数,建立了S形函数连接的BP神经网络模型。采用Levenberg-Marquardt优化算法,用MATLAB6.1进行仿真,对股线的拉伸力学性能进行了预报。预报结果与实验结果吻合良好。

Based on the parameters of plied yarn, a modeling of Neural Network is determined. Using Levenberg-Marquardt method, the tensile properties of plied yarn is predicted with Matlab 6.1 software. The predicted results are compared with those of experiments, and it is showed that they are in good agreement with each other.