Relationship Between Heat Transfer and Residual Stresses In Three Carburizing Steels

Monday, September 10, 2012: 5:00 PM
Atlantic D (Radisson Blu Aqua)
Mr. Daniel S. Baker , Severstal NA, Dearborn, MI
Dr. John G. Speer , Advanced Steel Processing & Products Research Center Colorado School of Mines, Golden, CO
Dr. David K. Matlock , Advanced Steel Processing & Products Research Center Colorado School of Mines, Golden, CO
Dr. Chester J. Van Tyne , Advanced Steel Processing & Products Research Center Colorado School of Mines, Golden, CO
Heat treatment simulations have grown in importance since they can help to develop processes and reduce the need for trial and error experiments. Simulations can also provide insight into the metallurgical phenomena which occur during heat treatment that are not easily observed using standard laboratory methods. In the present study DANTE® subroutines in conjunction with ABAQUS® were used to simulate heat treatments of three carburizing grade steels, 4120, 4320, and 8620, for a bar with various diameters and a length of 254 mm. The results of these simulations provide relationships between quenching method, microstructure development, and residual hoop stresses. In a non-carburized condition after oil quenching with a low heat transfer rate, a tensile hoop stress component was predicted at and near the surface of the bar. After high intensity quench with a high heat transfer rate, a compressive hoop stress component was predicted at the surface. The Biot number is commonly used to characterize the importance of heat flow at the surface relative to heat flow via conduction through the solid component during heating or cooling for a component of given volume and surface area. The Biot number where the residual hoop stress component on the surface switches from tension to compression was defined in this study as the “critical” Biot number. For a Biot number greater than approximately 2.5 the hoop stress component on the surface of the part was compressive. The value of the critical Biot number was not dependent on alloy content over the range examined.