B. Brickstad, NUMERICAL-ANALYSIS OF IPIRG CRACKED PIPE EXPERIMENTS SUBJECTED TO DYNAMIC AND CYCLIC LOADING, International journal of pressure vessels and piping, 55(3), 1993, pp. 395-422
This report contains results of a finite element study aiming to ident
ify the influence of loading history and geometry for cracked pipes su
bjected to complex loading. The experiments have been performed within
the International Piping Integrity Research Group (IPIRG) Program. Th
e majority of the numerically analyzed experiments were conducted on s
traight pipes with an outside diameter of 168 mm and containing a larg
e circumferential through-wall crack. The considered pipes were loaded
in four-point bending under displacement control and at a temperature
of 288-degrees-C. The types of loading were combinations of either qu
asi-static or dynamic and also monotonic or cyclic loading with differ
ent loading ratios R. Some analyses were also performed on surface-cra
cked pipes subjected to slow, monotonic loading. In the finite element
study, 20-node solid elements were used for the through-wall cracked
pipes and a combination of shell and non-linear line spring elements f
or the surface-cracked pipes. Stable crack growth was simulated by gra
dual node relaxation and crack closure is accounted for by using simpl
e contact elements. The J-integral for a remote contour is calculated
and used as a characterizing fracture parameter although the cyclic lo
ading violates the theoretical basis for this procedure. The near-tip
J can not be used for growing cracks because of the weak energy singul
arity. The results of the numerical study confirm the trends from the
experiments in that a high loading rate has a negative influence on th
e fracture properties of the studied carbon steel and that large cycli
c loading, especially at R = - 1, lowers the apparent JR-curve for bot
h carbon and stainless steels. To some extent geometry effects appear
to be present when comparing the results from pipes containing surface
cracks and through-wall cracks with results from CT specimens. These
effects are more pronounced for large amounts of stable crack growth t
han at initiation.