复合材料固化相关黏弹性性能演化及残余应力分析

Analysis on the process dependent viscoelastic properties and residual stresses of composites during cure

通过对复合材料固化度和温度相关黏弹性本构方程的分析,定义一个能综合反映固化度和温度等对复合材料黏弹性性能影响的无量纲参数De_m。当参数De_m都大于10~2时,复合材料基体处于流动状态;当参数De_m都小于10^(-2)时,复合材料为弹性状态;仅当部分参数De_m小于10~2而大于10^(-2)时,复合材料处于黏弹性状态。以AS4纤维/3501-6树脂复合材料为例,基于对其参数De_m在典型固化工艺过程中的演化,研究该复合材料黏弹性性能的发展过程,发现基于参数De_m分析得到的凝胶点时间与实验结果一致。根据复合材料黏弹性性能对残余应力发展的影响,将复合材料残余应力计算分为流动阶段和黏弹性阶段,并建立了相应的状态相关黏弹性本构模型。最后通过与原始模型预测结果的比较验证了提出的本构模型,表明本文提出的计算方法与原始黏弹性本构模型计算结果一致,但大大降低了计算所需的时间和存储空间。

Based on the analysis of the degree of cure and temperature dependent viscoelastic constitutive model of composite, the dimensionless parameters defined as Den were proposed to characterize the viscoelasticity of composite in this paper. The parameters Dem can comprehensively reflect the effects of the degree of cure and temperature on the viscoelasticity of a composite during cure. When all of the values of Dem are higher than 10^2 , the composite matrix is in a flow state. When all of the values of Dem are less than 10^-2 , the composite is in an elastic state. Only when one or more Dem lie between 10^2 and 10^-2 , the composite is in viscoelastic state. The behavior of the AS4 fiber/3501-6 polymer matrix composite was used as example, the development of the viscoelasticity of the composite was analyzed according to the evolution of the parameters Dem during the typical curing cycle. The result shows that the gelation time determined by the palameters Dem is consistent with that obtained by experimental tests. According to the effect of viscoelasticity of composite on the accumulation of residual stress, the analysis procedure on the residual stress was divided into the flow stage and the viscoelasticity stage, and the cure state dependent viscoelastic constitutive models were put forward. The proposed models were verified by comparing their predictions with those of the original constitutive model. The results show that the residual stresses predicted by the proposed models are consistent with those predicted by the original model, but the computing time and storage space required by the proposed models are greatly reduced.