Natural Convection Effects on Magnesium Solution Mining

Abstract

Raw magnesium chloride can be recovered using solution mining at a depth of a 1500 to 2000 meters. Underground caverns are formed in stacked layers of bischofite , carnallite and halite. The salt layers consist for a large part of NaCl and but contain a minimum of at least 35% magnesium chloride. Recovery is implemented by injection of fresh water in the salt layer, which becomes saturated with magnesium. The carnallite and bischofite layers largely consist of the less soluble NaCl, with interdispersed magnesium salts. A cylindrical model of magnesium recovery is presented that consists of a central open space, an annular space filled with a porous salt (NaCl) layer with an outer boundary that consists of a bischofite layer or carnallite layer. The central cavity is an open cylinder filled with a solution of NaCl and MgCl2. The presence of KCl can be disregarded. Fresh water is injected into the centre of the central cavity. Brine is extracted at a distance below the injection point. The natural convection flows of the fresh injection water in the cavity are investigated. Calculations show that the central cavity contains a solution of more or less constant composition except near the central axis of the cavity. Fresh injection water is lighter than the brine and therefore it migrates to the top of the cavity while mixing with the brine. The concentration of the brine near the injection point at the axis of the central cavity increases rapidly so that the roof will not be exposed to brines that are able to dissolve significant amounts of NaCl. Adjacent to the central cavity there is a concentric annular space, which consists of a skeleton of crystalline NaCl. This acts as a porous medium. The outer radius of the porous medium is adjacent to the undisturbed bischofite layer. At the outer radius the concentration of bischofite and sodium chloride are given by the saturated equilibrium conditions. Also at the outer radius there is a no flow boundary. At the inner radius the concentrations are constant due to the mixing conditions in the central cavity. The values of these concentrations are given by the cavity growth model. The Elder model is used to simulate natural convection flows through a porous annular cylinder of low permeability. The enhancement in the transfer rate due to natural convection flows with respect to mass diffusion is expressed in the Sherwood number. Simulations were conducted on porous media with a permeability of 2e-12 m^2 or less. The maximum enhancement factor resulting from the simulations is two and a half. For higher permeabilities the Brinkman model is used. The cavity growth model describes the rate of change of the inner and outer radius of the porous medium. The rate of change of the outer radius is determined by the diffusive mass flux of MgCl2. The rate of change of the inner radius is determined by the dissolution of the NaCl skeleton. However, the equilibrium concentration of NaCl in the presence of dissolved MgCl2 is rather small leading to a low dissolution rate of NaCl. Results of the cavity growth model are presented in terms of the inner and outer radius of the annular cylinder as a function of time. The cavity growth model requires enhancement factors in the order of 1e5 to get magnesium concentrations that are comparable to field data. Such high enhancement factors can possibly only be obtained using the Brinkman model. ?