Abstract
Low-cycle fatigue failure has been widely accepted as the key mechanism causing damages of coke drums during cyclic thermal-mechanical loadings. Common damages of coke drums known as bulging and cracking are associated with accumulative plasticity caused by thermal and mechanical strains. External repairs using temper-bead welding techniques are implemented to repair welds in the damaged areas of coke drums, which provide structural support to the vessels. Compared with matching filler metals, Ni-base fillers including alloy 625 and alloy 182 are compatible with both low-alloy steel base metal and internal clads in terms of weldability and thermal expansion. However, the differences of yield strengths and cyclic hardening behaviors of nickel-base alloys from base metals compromise the fatigue resistances of weld joints. In this study, alloy 182 and alloy 625 repair coupons were evaluated and compared based on isothermal low-cycle fatigue tests. Low-cycle fatigue behaviors of both weld metals and 1.25Cr-0.5Mo base metal were measured at 1.0%, 1.5% and 2.0% strain amplitudes. Test results indicate both nickel-base filler metals exhibit overmatching strength over the base metal due to cyclic hardening. Low-cycle fatigue tests of Heat Affected zone (HAZ) samples show the failures of alloy 625 weld joints occur in the base metal, while the failures of alloy 182 weld joints occur along the fusion boundary. The observations show strength mismatch and fatigue resistance are the key factors to determine failure locations of the joints. In addition, cyclic hardening coefficients based on kinematic hardening model were extracted from experimental data to simulate the cyclic behaviors of the weld joints. Finite element simulation results were shown to be consistent with experimental data at stabilized cycles. Cyclic behaviors of weld metal and base metal within a weld transition sample were calculated based on the numerical model.
Low Cycle Fatigue Properties of Alloy 617 base Metal and Weld Joint at Room Temperature
