Page 100 - Engineered Interfaces in Fiber Reinforced Composites
P. 100

Chapter 3.  Measurements of  interfacelinterlaminar properties   83





                                 1                               Y
                                             t            h,l

                            Fig. 3.33. CLS specimen for mixed mode interlaminar fracture tests.



                                                                                  (3.35)

                where the subscripts 1 and 2 refer to the sections indicated in Fig 3.33. Because CLS
                gives total energy consisting of mode I and mode 11, the individual components may
                be determined from beam theory using finite element analysis, For  unidirectional
                specimens with the delamination placed at the mid-plane, beam theory gives a value
                GI/GI-II = 0.205 (Brussat et  al.,  1977). Details of  the  expressions for  the  strain
                energy release rate of other mixed-mode tests are not treated here as the stress states
                are much more complicated than in the pure mode I1 ENF test (Whitney, 1989).
                  It should be noted that a mode I1 crack-resistance R-curve may also be obtained
                for some fiber-matrix  systems (Vu-Khanh, 1987). In conjunction with the mode I
                R-curve the additional information of a mode I1 R-curve will be of great use to the
                composite design engineers. This is increasingly the view of the ESfS task group on
                delamination crack growth resistance.
                3.4.4. Mode I  edge delamination  tests


                  The problem of  delamination along the  straight  free edge of  laminates, which
                takes place under an in-plane uniaxial load, has attracted  significant investigation
                because  the  presence  and  growth  of  edge  delamination  may  cause  progressive
                reduction  in  the  laminate's  stiffness and  residual strength.  In  severe cases, this
                fracture phenomenon acts as a precursor to final failure of the laminates. The free
                edge delamination is attributed to the existence of interlaminar stresses, which are
                highly localized in the neighborhood of a free edge. The magnitude and distribution
                of  these interlaminar stress components vary widely and depend on the laminate
                layup, stacking sequence, properties of the composite constituents and the nature of
                loading.  Comprehensive  reviews  of  the  experimental  observations  have  been
                presented  (Kim,  1989) and  a  micromechanics analysis  of  the  edge delamination
                test (EDT) has been performed (Whitney, 1989).
                  Due  to  the complex mixed-mode nature  of  composite delamination, no  closed
                form  solutions  have  been  developed  yet  to  express  the  influence  of  governing
                parameters  that  control  the  edge delamination  behavior.  Under  tensile loading,
                delamination is normally preceded by a number of transverse cracks, particularly in
                the 90" plies. Because of the presence of these cracks, the location of delamination is
                not unique as in the case of compressive loading, which invariably results in gross
                buckling of the laminate. The path of delamination along the axial direction varies
   95   96   97   98   99   100   101   102   103   104   105