基于神经网络算法的救生筏橡胶基布质量预测方法

Prediction Method of Rubber Base Fabric Quality of Life Raft Based on Neural Network Algorithm

为了降低救生筏橡胶基布检测成本,提高预测速度及准确性,将橡胶基布的坯布抗拉强度和扯断伸长率作为输入变量,采用多层感知器神经网络建立橡胶基布质量预测模型,实现对橡胶基布抗拉强度和撕裂强度的预测。试验结果表明:多层感知器神经网络预测速度快捷、精度高,在样本较少的情况下,对经纬向抗拉强度和撕裂强度的预测值能较好地适应生产质量控制中的精度要求;运用该质量预测模型,为合理确定救生筏产品质量安全区提供了高效准确的依据。

In order to reduce the detection cost of rubber base fabric for life raft and improve the prediction speed and accuracy,the tensile strength and elongation at break of rubber base fabric are taken as input variables,and the quality prediction model of rubber base fabric is established by multi-layer perceptron neural networks to predict the tensile strength and tear strength of rubber base fabric.The experimental results show that the prediction speed of multi-layer perceptron neural networks is fast and the accuracy is high.Under the condition of few samples,the prediction value of longitudinal and latitudinal tensile strength and tear strength can better adapt to the precision requirements of production quality control.Using this quality prediction model,it provides an efficient and accurate basis for reasonably determining the safety zone of life raft product quality.