纤维丝束带复合材料的有效性能预测与实验测试

Effective property prediction and experiment test for fiber tow composites

  • 摘要: 依据纤维丝束带复合材料的相关几何结构参数值和所确定的纤维丝束带特征体积单元(RVE)模型几何结构尺寸,以有限元软件MSC.Patran/Nastran为平台建立纤维丝束带复合材料RVE有限元模型并在模型中置入相应的制备缺陷。各类制备缺陷的置入均采用删除网格单元的方法,置入裂纹型制备缺陷时偏移裂纹两侧单元相对面以获得当前裂纹宽度,置入孔洞型制备缺陷时尽量模拟其真实形貌。根据复合材料力学关于材料各性能参数的定义和细观力学基本理论推导了有限元计算细观力学(FECM)方法预测复合材料有效弹性性能和有效热膨胀性能的过程。根据FECM方法预测了不含制备缺陷、含单一制备缺陷和含各类制备缺陷时的弹性常数和有效热膨胀系数。结果表明:各类制备缺陷的存在均会使弹性模量和剪切模量减小,泊松比和热膨胀系数可能增大也可能减小。通过与实验测试结果对比分析可知,数值预测结果普遍比实验测试结果偏大,但总体效果较为理想,最大相对误差为6.04%。

     

    Abstract: According to the value of geometrical structure parameters and determined geometrical structure size of representative volume element (RVE) model of the fiber tow composites, finite element software MSC.Patran/Nastran as a platform to establish RVE finite element models of fiber tow composite and corresponding manufacturing defects were also implemented into the models. Grid element deletion method was utilized to implement all types of manufacturing defects. When crack-like manufacturing defects implemented into the RVE models, the crack width was obtained by moving the two opposite surfaces of the elements at the both sides of offset crack. The real morphologies of void-like manufacturing defects should be simulated as much as possible. According to material property parameter definitions of mechanics of composites and basic theories of micro-mechanics, the finite element computational micro-mechanics (FECM) methods to predict effective elastic properties and effective thermal expansion properties of composites were derived. Based on FECM methods, elastic constants and effective thermal expansion coefficients without manufacturing defects, containing a single manufacturing defect and containing all types of manufacturing defects were respectively predicted. The results show that the presence of all manufacturing defects will reduce elastic modulus and shear modulus, while Poisson's ratios and coefficients of thermal expansions may increase or decrease. Through comparison with the experimental test results, numerical prediction results are generally a little larger than experiment test results, but the overall effect is ideal. Wherein, the maximum relative error is 6.04%.

     

/

返回文章
返回