ABSTRACT
Thus all terms in eqn. (2.47) which relate to F and the hypothetical curve are replaced by Fe = Fme/mo, where Fe is the critical (final) load after crack growth. In the limit this would be the point of unstable growth. This substitution givesS
- { 2nk [FJ"-1}(mo)2 F; d ( I ) G = I + n + I me me 2B da mo (2.49)
The expression is admittedly approximate but is questioned here on two counts. Firstly, no attempt seems to have been made to formulate G for the growing crack at the onset of growth. It seems to be tacitly assumed that G as already evaluated in eqn. (2.47) (identified here as ] mJ is indeed the same as G at initiation for the growing crack (identified in the foregoing as lin)' There may indeed be little difference, but the point does not seem to be clearly established, although both energy available and remote dissipation rate may differ for the two cases. Also the modification itself appears fallacious in that the load F and the hypothetical load F are, for extensive yielding, related by plasticity effects as well as elasticity. In the limit of near collapse behaviour the relationship is of the form derived in connection with Fig. 2.24, where the plastic change is related to change in collapse load. Presumably for small-scale yielding this term is small in relation to the elastic compliance effect, but in that case it is accepted that G = G and LEFM gives an acceptable estimate of behaviour although there may be an intermediate range where the plastic component of the change in load is neither negligible nor estimable from the collapse load.
