WELDING RESEARCH cooling and outside surface heating by hot combustion flue gas originating from the direct fire furnace immediately below. A heat transfer analysis was conducted using the analysis procedures discussed in the Finite Element Model section, where it was assumed that the cool air flows quickly inside the roll and could keep a temperature of 204C (400F) without being heated by the roll inner surface. The furnace temperature was cyclic from 1066 to 204C. Figure 15 shows the predicted temperature at three points on the roll shell for ten cycles. The predicted temperatures on the roll outer surface (N147), on the roll inner surface (N148), and on the inner surface of the end bell (N3284) are 668, 651, and 646C, respectively. The temperature on the roll outer surfaces reduces from 1066 to 668C. Creep-fatigue analysis was conducted using the analysis procedures discussed in the Finite Element Model section by inputting the predicted temperature history from the heat transfer analysis and applying mechanical loads. Figure 16 shows the predicted distribution of effective creep strain with inside cooling. Effective creep strain was significantly reduced by inner surface cooling of the roll by air, as illustrated by comparisons of Figs. 16 and 9. Figure 17 shows the distribution of the predicted maximum principal stress with inside roll cooling. Stress at the weld root was reduced significantly with inside cooling, as illustrated by comparisons of Fig. 10A (without inside cooling) to Fig. 17 (with inside cooling). Figure 18 shows the evolution of temperature and maximum principal stress with inside cooling. The stress range with inside cooling is about 105 MPa, while the stress range without inside cooling is about 140 MPa, resulting in a 25% reduction. The numerical evaluation of the inside cooling concept shows that the inside cooling method can lower the roll shell temperature, reduce the creep strain at the weld toe, and reduce the stress range at the weld root. Therefore, this method could be an effective way to improve the creep-fatigue life of the furnace roll in a production line. Feasibility Study of Inside Cooling A fully coupled heat transfer and CFD analysis was conducted using the one-fourth model (Fig. 8) to evaluate the effect of the cooling system design inside the roll on the temperature of the roll shell. In the design, cool air flowed into the roll through a center tube and out the roll via the four holes around the center hole. The outer surfaces of the roll were heated from the hot air in the furnace. Before turning on the cool air, the roll was heated to the operating temperature according to the furnaceheating process — Fig. 5. Figure 19 shows the roll temperature at the start of cooling utilizing the cool air. To apply the cooling process smoothly, the cool-air pressure was linearly ramped up from atmospheric pressure to 0.62 MPa at the inlet in one minute. Figure 20 shows the predicted air temperature inside the roll at 3, 15, and 27 s. At 3 s, the inlet pressure is 0.130 MPa. The air temperature near the end of the center tube is heated up by the hot roll shell. The air temperature is about 525 K. As time increases, the air temperature increases. At 27 s, the air temperature inside the roll reaches 800 K. Therefore, instead of cooling the roll shell using the flowing air, the air is heated by the roll shell. The design is not efficient to cool the roll shell. Conclusion Multiphysics analyses, including a heat transfer analysis, a creep-fatigue analysis, and a computational fluid dynamics analysis, were conducted to understand the failure of a welded roll in the vertical furnace of a continuous hot-dip coating line and propose solutions to improve the creep-fatigue lifetime. The heat transfer analysis was conducted to predict the temperature history of the welded roll by modeling heat convection heating from hot air inside the furnace using ABAQUS and a user-developed subroutine. The creep-fatigue analysis was performed by inputting the predicted temperature history, applying a mechanical load, and defining boundary conditions. The CFD analysis was used to evaluate a cooling system 440-s WELDING JOURNAL / NOVEMBER 2016, VOL. 95 design that intends to cool the roll by flowing high-pressure cool air inside the roll. Based on the analysis results, the following conclusions could be drawn: • Three factors: difference of material properties between the filler metal and base metal, furnace temperature variation during service, and high operating temperature contributed to the roll failure. • Reducing the furnace temperature variation during roll service can lower the stress in the welded joint to improve the creep-fatigue life of the roll. • Using EBW without filler metal to replace FCAW to weld the roll, the creep strain and stress in the interface between the filler metal and the base metal can be eliminated so the creepfatigue life of the roll can be improved. • Applying cool air inside the roll can lower the roll temperature to reduce the creep strain and stress to increase the roll lifetime, but it requires an effective cooling system design. • The predicted high tensile hoop stress can be used to explain the observed cracks near the weld along the direction perpendicular to the roll hoop direction, and the predicted highly localized maximum principal stress at the weld root can be used to explain the observed crack at the weld root, which validates the models indirectly. References 1. Ewald, J., Sheng, S., Klenk, A., and Schellenberg, G. 2001. Engineering guide to assessment of creep crack initiation on components by two criteria diagram. International Journal of Pressure Vessels and Piping 78: 937–949. 2. Nikbin, K. 2013. Creep/fatigue crack growth testing, modeling and component life assessment of welds. Procedia Engineering 55: 380-393. 3. Marie, S., and Delaval, C. 2001. Fatigue and creep-fatigue growth in 316 stainless steel cracked plates at 650°C. International Journal of Pressure Vessels and Piping 78: 847-857. 4. Vacchieri, E., Holdsworth, S. R., Cost, A., Poggio, E., Riva, A., and Villari, P. 2014. Creep-fatigue interaction in two gas turbine Ni based superalloys subjected to service-like conditions. Materials at High Temperatures 31(4): 348–356. 5. Ji, D. M., Zhang, L. C., Ren, J., and Wang, D. 2015. Creep-fatigue interaction and cyclic strain analysis in P92 steel based
Welding Journal | November 2016
To see the actual publication please follow the link above