BP神经网络预测织物拉伸性能

Predicting Fabric Tensile Property with BP Neural Network

采用BP神经网络技术建立反映织物结构参数、纱线参数与织物拉伸性能间关系的三层神经网络模型,根据影响织物拉伸性能的各种参变量,用动量-学习率自适应调整的BP算法训练模型。通过预测值和实验值的比较,表明用神经网络方法预测织物拉伸性能有相当的准确性,从而在一定程度上实现用神经网络预测织物的拉伸性能。

In the paper, artificial neural network (ANN) is adopted to analyze fabric tensile property. The BP neural network of three-layers is constructed to obtain the relationship between fabric structural parameters, yarn parameters and fabric tensile property . The important factors that affect tensile property are analyzed. The predicted values and test values are compared and are in good agreement.