Creep crack initiation prediction considering constraint effect for pressurized pipelines with circumferential surface cracks

Abstract

AbstractTheoretical 2‐parameter C*‐Q* approaches and numerical simulation were conducted to investigate the creep crack initiation (CCI) time and the effect of constraints induced by the geometrical sizes of pipelines with circumferential surface cracks. The theoretical enhancement model of the C*‐Q* approaches under the transient creep condition, which considered the load‐independent constraint parameter Q*, was proposed to predict the CCI time around the crack front. The results revealed that the distribution regulation of Q* along the crack front for circumferential internal surface cracks and external surface cracks was similar. The maximum constraint level occurred near the deepest crack front part for cracks with small a/c (a/c < 0.4), while it occurred near the free surface for cracks with large a/c (a/c > 0.4). The constraint values at the same position (2Φ/π) increased with the increasing of the crack depth when a/c kept constant. In addition, the circumferential internal surface cracks of pipelines were proved more dangerous than the external surface cracks with the same geometrical size. Furthermore, the CCI times were decided by the peak values of constraint, or the CCI firstly occurred at the position where the constraint level was maximum. Additionally, the variation of hydrostatic stresses, triaxiality, and multiaxial strain factor considering the constraint parameter Q* was discussed. The suitability of the analytical C*‐Q* approaches was verified to predict CCI. The comparison of CCI times between the analytical approach and the BS 7910 as well as the finite element results demonstrated that the solutions under stress intensity factor–Riedel‐Rice control (initially by stress intensity factor and then by transient creep stress or Riedel‐Rice conditions)—were more accurate when internal pressure P < 18 MPa, but the solutions under Hutchinson‐Rice‐Rosengren–Riedel‐Rice control (initially by plastic Hutchinson‐Rice‐Rosengren control and then by Riedel‐Rice conditions) were more appropriate when P > 18 MPa. Finally, the accuracy of C*‐Q* 2‐parameter approaches and the suitability of finite element simulation were validated.