A Design Rule for Prediction of the Just Suspended SPEED of Mixed Slurries

A DESIGN RULE FOR PREDICTION OF THE JUST SUSPENDED SPEED OF MIXED SLURRIES

Inci Ayranci, Theodore Ng, Maria Garcia, Arthur W. Etchells, Suzanne M. Kresta

Design for solids suspension in stirred tanks is based on an empirical correlation known as the Zwietering correlation (Zwietering, 1958). This correlation has many limitations, one of which limits the industrial use significantly – the correlation's applicability only for unimodal slurries. With the current design rule the just suspended speed, N_js, of the mixed slurry is determined assuming that the slurry consists of only the particle with the maximum N_js. Our previous work on binary slurries (Ayranci and Kresta, 2011) showed that this design rule fails depending on the particle size, particle density and the concentration of the two solid phases. The limitation of the design correlation for mixed slurries and the failure of the current design rule clearly show that there is need for a model to predict the N_js of mixed slurries. The objective of this study is to propose a model to predict the N_js of mixed slurries at relatively low concentrations where particle-particle interactions do not play a significant role. We propose two models: the power model and the momentum model. The power model states that the power required for the complete off bottom suspension of the mixture is a summation of the power of the unimodal slurries of the solid phases present in the mixture. Similarly, the momentum model is based on the sum of momentum of the unimodal slurries. The two models were tested with binary mixtures of seven type of solids in water: nickel, bronze, glass of two particle sizes, sand, urea formaldehyde and ion exchange resin. The total solids loadings were increased up to 35 wt% (weight percent). First, the unimodal slurry data was used to test the concentration dependence of the Zwietering correlation. Then the two models were tested. The unimodal data showed that the Zwietering correlation cannot predict the concentration dependence accurately. The power model was successful in predicting the mixed slurry N_js up to 20 wt% solids loading.