Can statistical methods optimize complex multicomponent mixtures for sintering ceramic granular materials? A case of success with synthetic aggregates

Abstract

The relationship between the proportions of multicomponent mixtures with the technological properties of ceramic granular materials (synthetic aggregates) has been studied using statistical methods. The four phases involved in the formulations have been: kaolin (K) as aluminosilicate source; cork powder (C) as organic carbon source; sodium carbonate (N) as flux and pyrite (P) as source of iron and sulfur. The Mixture Experiments - Design of Experiments (ME-DOE) has been the statistical methodology applied from the initial configuration of the 36 starting formulations to the final validation of the models and optimums. After granulation, artificial aggregates have been obtained by sintering in a rotary kiln, and their main technological properties have been determined. Bloating index (BI), particle density (ρrd), water absorption (WA24) and crushing strength (S) were selected as the four key characteristics to be modeled and optimized, using response surface and effect plots to assess the effect of K, C, N and P on such properties. 32 out of 36 starting varieties met the density criteria for lightweight aggregates. In the optimum formulations obtained, the minimum percentage of K was 83 wt%, so that the variations in the percentages of P, C and N were the critical variables for determining the final properties of the aggregate. The contrast between experimental and estimated data has shown that the models fit adequately, indicating that this type of approach may have enormous potential for future research on artificial aggregates and other ceramic materials.