439s

Welding Journal | November 2016

WELDING RESEARCH Fig. 19 — Cooling system starting temperature. Fig. 20 — Fluid temperature evolution during inside cooling. NOVEMBER 2016 / WELDING JOURNAL 439-s near-weld cracks found in the failed roll, as shown in Fig. 13B. The stress magnitude is low because the roll is at a temperature of 427C. If the roll cools to room temperature, the stress magnitude would be higher. Although the modeled weld has a different orientation than the failed roll, the hoop stress pattern will be the same between the modeled weld and the failed weld. The tensile stress near the weld could induce the cracks near the weld as shown in Fig. 13C, D. This study suggests that the cool-to temperature has a significant effect on creep strain and stress during temperature cycles. Therefore, controlling the cool-to temperature between cycles will reduce the creep-fatigue strain and stress to extend the creep-fatigue lifetime. In addition, the analyses also shows that the difference in material properties between the weld (N117) and the base metal (MO-RE®1) is the root cause of roll failure. If the same material is used for both the weld and base metal, the high creep-fatigue strain and stress could be reduced. Improvement Study 1: Electron Beam Welding The study of the effect of cool-to temperature between cycles on the creep-fatigue process shows that the different material properties between the base material and the filler metal are the root cause of high creepfatigue strain and stress at the weld toes and the weld root. One option to reduce the high creep-fatigue strain and stress is to use the same material for both the filler metal and the base metal during the welding of the roll using FCAW. Another option is to use a newer welding process such as EBW without any filler metals. Since using the base metal as filler metal could result in weldability problems during FCAW of the roll, EBW was studied to explore the possibility of improving the creep-fatigue life. After welding using EBW, PWHT will be conducted to eliminate the weld residual stresses and restore material properties. Therefore, weld residual stress does not need to be included in the creep-fatigue analysis. The weld material properties could be assumed to be the same as the base material properties. A ten-cycle creep-fatigue analysis was conducted by replacing the material properties in the weld with the base material properties while keeping the other conditions the same as Case 1 to study the effect of an electronbeam welded joint on creep-fatigue strain and stress. Figure 14 shows the predicted effective plastic strain and maximum principal stress with EBW without the use of filler metal. The effective creep strain plot uses the same scale as in Fig. 9. The striking difference in Figs. 14A and 9 illustrates the elimination of the localized concentration of an effective creep strain in Fig. 14A. Similarly, maximum principal stress is plotted as the same scale used in Fig. 10. It is found that high stress at the weld root disappears, as shown in Fig. 14B. Analysis results show that EBW could be an effective method to improve the creep-fatigue life of the furnace roll in the continuous hot-dip coating line. A preliminary welding test shows that EBW could be used to produce a defect-free weld of MORE ®1. A further study was conducted to predict the creep strain and stress in the electric beam welded roll by including the weld size, shape, and residual stress resulting from EBW (Ref. 30). Analysis results confirmed that EBW is an effective method to significantly reduce creep strain and stress. Improvement Study 2: Inside Cooling A material’s resistance to creep greatly depends upon the temperature to which the material is exposed. If the roll operating temperature can be lowered, the creep life of the roll could be extended. A method was numerically examined by applying cool air inside of the roll to lower the roll shell wall temperature. The temperature on the roll shell is the result of competing effects of inside surface


Welding Journal | November 2016
To see the actual publication please follow the link above