论文部分内容阅读
以玻纤增强聚丙烯(FRPP)复合管材缠绕成型性能及其关联的成型过程参数为对象,基于留一法交叉检验的手段,采用径向基神经网络(RBFNN)建模的方法研究了FRPP复合管材缠绕成型性能预测模型。结果表明,预测层间剪切强度、环向拉伸强度时,分别存在的一种RBFNN模型的检验最大相对误差≤0.001%。对学习好的RBFNN模型进行敏感度分析表明,玻纤增强聚丙烯(FRPP)复合管材缠绕成型性能及其关联的成型过程参数之间的关系为高度非线性关系。性能最好的工艺参数分析说明所建立的模型是可靠的。
Based on the method of leaving one-way crossover test and using the method of radial basis-nerve network (RBFNN) modeling to study the characteristics of the FRPP composite pipe winding and forming process and its associated molding process parameters, Pipe Wound Molding Performance Prediction Model. The results show that the maximum relative error (≤0.001%) of a RBFNN model exists when predicting interlaminar shear strength and circumferential tensile strength respectively. Sensitivity analysis of the well-studied RBFNN model shows that there is a highly non-linear relationship between the winding-forming properties of glass-fiber reinforced polypropylene (FRPP) composite pipe and its associated molding process parameters. The best performing process parameter analysis shows that the model is reliable.