Property Measurements and Moisture Diffusion Simulations During Potash Caking and Cake Strength Predictions

Researchers have demonstrated that the hygroscopicity of bulk potash exposed to humid air causes the adsorption and condensation of water vapour on potash particle surfaces. The accumulated water forms liquid bridges near contact points between potash particles and dissolves salt impurities on potash particle surfaces until the brine bridges are saturated. When the ambient air becomes dry, the accumulated water evaporates, the dissolved salt recrystallizes, and the crystalline bridges form between potash particles. A sufficient number of crystal bridges per unit volume in bulk potash can cause strong potash cakes. This can lead to costly handling problems. It is the purpose of this research to investigate the effects of external compressive load and ambient air humidity on caking properties and cake strengths of bulk potash during long duration storage. The research is separated into the four parts: discrete element modelling, potash sample preparation and fracture testing, simulation of water vapour diffusion into test cells, and cake strength correlations. The discrete element method (DEM) was used to predict the tensile force distributions on particles and cake strength of a compressed potash ring specimen when the specimen is fractured in the centrifuge. In the planar DEM model, a tensile force concentrating on an inter-particle contact region can be transformed to the local-averaged tensile stress distributed on the equivalent cross section corresponding to the particle. The local-averaged tensile stresses are consistent with the microscopic tensile stresses in the solid continuum model when the Poisson’s ratio is small. The DEM model indicates that the tangential tensile forces are the main factor causing fractures of potash specimens while the radial tensile forces are insignificant. The uniform tangential tensile forces imply that the fracture interface of a specimen should be parallel to the radial direction of the specimen (i.e. normal to the main tensile force). During sample preparation, potash ring specimens were subjected to four nominal sample depths (0, 4, 8, and 16 m) and three values of moisture adsorption humidity (69, 75, and 84%). Two values of moisture desorption humidity (52 and 11%) were used to make crystal bridges in the potash samples. Test sample preparation periods from weeks to months were used. Springs were used to generate the external compressive loads equivalent to the nominal sample depths. The caking properties of potash specimens, such as mass, height, porosity, and moisture content, were measured during sample preparation. After caking, specimens were ruptured in the centrifuge to test their fracture rotary speeds. Data for fracture rotary speed were input into the DEM model to compute cake strengths of potash specimens. The data analyses showed that the higher external equivalent loads during sample preparation cause the bigger variations of sample height and porosity. The moisture content of potash specimens was determined by the moisture adsorption and desorption humidity. When the moisture adsorption humidity was not higher than 75%, cake strengths of potash samples were enhanced with an increase of the external equivalent load. If the moisture adsorption humidity reached 84%, the external load hardly affected the potash cake strength. A numerical model of water vapour diffusion with the boundary condition of free mass convection was developed for two sub-groups of uncompressed potash specimens corresponding to those with the complete and reduced top sample surfaces exposed to the ambient air. The numerical model predicted the spatial distributions of local relative humidity and moisture content in transient states inside a potash specimen and the variations of free mass convection coefficient and macroscopic sample moisture content over time. The predictions of sample moisture content for the two sample sub-groups were in agreement with the corresponding experimental data. The big free mass convection coefficient during moisture desorption meant a high mass transfer rate near the potash sample top surface. The almost uniform spatial distributions on the maximum and residual local moisture contents in a potash specimen showed that the variations of local moisture content were similar inside the specimen during moisture adsorption and desorption. It was also expected that the crystal bridges forming in the specimen after caking were analogous. A correlation was developed to predict cake strengths of potash specimens by using the ambient air humidity for moisture adsorption and desorption and the external equivalent load. The upper limitations of relative humidity and external load were 75% and 16 m. The correlation showed that cake strength of potash specimens was enhanced with an increase of moisture adsorption relative humidity conditions and external load, and with a decrease in moisture desorption relative humidity conditions