Abstract
Micromechanical models for the prediction of the longitudinal strength of metal-matrix composites are often based on one constituent reaching its elastic limit. This procedure can grossly underestimate the useful or effective strength of the composite. At elevated temperatures, the yield strength of the matrix material can become very low relative to the strength of the fiber (in extreme cases as low as 5-10 % of the fiber value). Even though the matrix has yielded, further load can be carried by the fibers. The problem can be further complicated by the formation of a third phase or a reaction zone between the matrix and the fiber. The fact that this reaction zone has distinct material properties must be accounted for in any analysis. To get a valid picture of the composite's true load carrying ability the complete elastic-plastic stress-strain behavior must be studied. This paper presents results of such analyses for a system with a monotonically increasing load up to failure of the first constituent. The three constituents are modeled as strain-hardening materials. Results are first obtained for a simple bilinear model of the stress-strain be havior in each material. A more detailed study was also conducted in which the three materials were modeled with curvilinear post-yield behavior. Results are presented for some representative constituent material properties. These results indicate the methods of analyses outlined in this paper are valid and easy to use. The results also clearly show that the entire elastic-plastic stress-strain behavior of the metal-matrix composite is more im portant than simply determining the initial yielding of one component. Strength predic tions based on the first component yielding are seen to be often far too conservative.
Get full access to this article
View all access options for this article.