摘要
由于纤维缠绕聚合物基复合材料（FWRP）与结构同时形成，因此缠绕成型工艺过程控制（如纤维缠绕张力）对FWRP制品结构抗力具有重大影响。人工神经网络具有超强非线性映射能力，可在特定条件下将缠绕张力与缠绕结构抗力两者由相关关系转化为确定性关系（函数关系）。本文分别以碳纤维和玻璃纤维缠绕聚合物基复合材料制成NOL环试件。在实验基础上采用人工神经网络结合遗传算法对纤维缠绕张力进行优化，取代传统的“试凑法”，优化结果与实验吻合。
Because the filament wound reinforced plastics （FWRP） and structures form synchronously, the technical parameters of FWRP such as the fiber winding tension have significant influence on the structural resistance of FWRP. And the artificial neural network has powerful non-linear mapping ability so that under the given condition the relation of the winding tension and structural resistance can be transformed into a certain relation （functional relation）. In this paper the glass fiber and carbon fiber are wound into NOL ring respectively. According to the experiment, the optimization for the fiber winding tension based on genetic algorithm-neural network, which substitutes the traditional cut-and-try method, has achieved a better result.
出处
《玻璃钢／复合材料》
CAS
CSCD
北大核心
2007年第4期3-5,17,共4页
Fiber Reinforced Plastics/Composites
基金
国家自然科学基金（50475034）
关键词
纤维缠绕
聚合物基复合材料
缠绕工艺
纤维缠绕张力
优化
filament winding
reinforced plastics
winding technology
fiber winding tension
optimization
