基于微结构识别的单向复合材料导热系数预估

以T300碳纤维/环氧树脂基单向复合材料为例，考虑纤维周围间隙缺陷的影响，建立了基于微观图像识别的等效导热系数预估方法.首先利用图像识别技术处理材料微观电镜照片，然后依据纤维体积分数稳定性判据应用几何重构技术建立了代表性单元，并通过在代表性单元（RVE）内部交界面处添加接触热阻的方法引入间隙缺陷的影响，最终利用有限元方法模拟得到等效导热系数（ETC）.研究发现:间隙的位置对等效导热系数影响微弱；随着间隙缺陷占比和厚度的增加，等效导热系数显著降低；间隙缺陷占比大于0.8，无量纲间隙缺陷厚度小于0.15时，单向纤维增韧复合材料的等效导热系数受间隙影响最突出；相对于纤维和基体理想接触的情况，考虑间隙缺陷后，等效导热系数最大降幅可达52.1%. 

Thermal conductivity prediction of unidirectional composites based on microstructure identification

Taking the T300 carbon fiber/epoxy resin matrix unidirectional composites as an example, an equivalent thermal conductivity prediction method based on microscopic image recognition was established, in consideration of the influence of the interface defect. The microstructure scanning electron microscope photos were processed first through image recognition. Then according to the fiber volume fraction stability criterion, the representative volume element(RVE) were built by geometry reconstruction technique. Meanwhile, the thermal contact resistance was added at the interfaces in the RVEs to represent the influence of the interface defects. Finally, the prediction results of equivalent thermal conductivity (ETC) were obtained by the finite element method. Results show that the randomness of the interface defect''s position affects ETC weakly, while ETC decreases significantly with the increasing proportion or the thickness of the interface defect. When the the interface defect proportion is greater than 0.8 and the dimensionless thickness of the interface defect is less than 0.15, the most significant influence on ETC appears. Compared with the simulations of ignoring the interface defect between fiber and matrix, the biggest drop of ETC can reach 52.1% considering the interface defect.