Study of a drawing-quality sheet steel. I: Stress/strain behaviors and Lankford coefficients by experiments and micromechanical simulations

Abstract

The sheet-metal industry uses Lankford coefficients and the forming-limit curve, FLC, as standards for characterizing a sheet’s ability to be stretched and deep drawn. Investigators have recently made significant advances in computer codes that predict these measures of formability. However, complete experimental data sets that provide input properties and verification data for the simulations rarely exist for a single material. The current investigation focused on obtaining such data for a single drawing-quality steel sheet. Measurements intended for the calibration and initial verification of the simulation code include uniaxial-tension tests, through-thickness and plane-strain compression experiments, and quantitative texture – orientation distribution function – evaluations, while a comparison between measured and simulated Lankford coefficients, Part I, and an FLC, Part II, provide a rigorous verification of the computer simulations. In order to initially verify the simulations, we performed through-thickness and plane-strain compression measurements. A key experimental result was that the flow curve in free, through-thickness compression – an experiment that corresponds to biaxial stretching – lies 18% above the uniaxial tensile data. The plane-strain compression curve is another 11% above the free-compression stress/strain data. We measured the Lankford coefficients, as a function of angle to the rolling direction, for the same steel sheet, finding the maximum values in and at 90° to the rolling direction, 1.59 and 1.89 respectively. A minimum Lankford coefficient of 1.19 was measured at 45° to the rolling direction. For calibrating a rate-dependent visco-plastic self-consistent polycrystal model we needed only to measure the material’s initial texture and to fit power-law and saturation-hardening laws to our tensile data. This kept the set of adjustable parameters to a minimum. Without other adjustments to the model, we predicted the correct stress levels in the free and channel-die compression experiments as well as values of Lankford coefficients. These successes indicate that the polycrystal model should be capable of simulating the entire FLC, Part II.