Content area
Abstract
A flexible boundary cubical triaxial tester was built to study the three-dimensional load response of dry, cohesive powders. The tester was used to load a powder sample along isotropic, deviatoric, and mean effective stress paths. The isotropic, or hydrostatic triaxial compression (HTC), tests were used to evaluate the critical state theory based modified Cam-clay constitutive parameters, $\lambda$ (loading slope) and $\kappa$ (unloading-reloading slope). Conventional triaxial compression (CTC) and mean effective stress (MES) test results were used to evaluate the slope of the critical state line, M. Test results yielded $\lambda$ = 0.045, $\kappa$ = 0.011, and M = 0.98. The modified Cam-clay model was verified by comparing the model predicted stress-strain behavior of powder with cubical triaxial tester experimental data. The mean average relative difference (ARD) and mean absolute difference between measured and predicted curves for: (1) voids ratio during HTC tests were 0.07% and 0.001, respectively, (2) deviatoric stress during MES tests were 15.7% and 5.6 kPa, respectively, (3) volumetric strain during MES tests were 22.4% and 0.4%, respectively, (4) deviatoric stress during CTC tests were 20.7% and 6.1 kPa, respectively, and (5) volumetric strain during CTC tests were 50.3% and 1.2%, respectively. A finite element model (FEM), using the modified Cam-clay equations, was used to predict incipient flow behavior of wheat flour in a hopper bin. The incipient flow regime within the powder mass was defined as comprising nodes in the FEM mesh that exceeded 7% axial strain. The FEM predictions were validated using a transparent plastic laboratory size hopper bin. The hopper bin was designed to ensure a predominantly two-dimensional flow so that visual observation and video recording of powder mass during discharge could be made. The bin dimensions were: bin height = 305 mm, width = 152 mm, and depth = 51 mm; hopper height = 190 mm, outlet width = 203 mm, depth = 51 mm, and angle = 15$\sp\circ$ from vertical. The first experimentally observed dynamic arch location and profile was used to validate the finite element model predictions. The finite element model predicted values were within the 95% CI of the measured values.
