An efficient method for computational flow-based simulation of heat transfer in a rotary kiln with pilot scale validation

Abstract

• Counter-current rotary kiln 5.27 m long, operated with two different feeds of gravel. • Temperature measurements taken along the kiln under various operating conditions. • Novel fast efficient computational model developed to predict bed temperature. • Predicted temperature profile in reasonable agreement with the measurements. • Model allows rapid, but reliable assessment of kiln design and operating conditions. Rotary kilns are used to carry out heating and chemical transformation in several mineral processing operations. A series of experimental runs was carried out using the realistic but controlled arrangement of a pilot rotary kiln to obtain temperature data for the validation of a novel model that couples a computational fluid dynamics model of the flame to a granular bed model through heat transfer terms. The rotary kiln was 5.27 m long, with an internal diameter of 410 mm, and two different feeds of inert standard commercial gravel were used in the runs, with the kiln operating in counter-current operation. Each of the feeds was run through the kiln under various operating conditions (feed rate, kiln rotation speed, fuel rate and weir height), and kiln temperature profiles were measured during each run. The effect of various operating parameters on heat transfer within the kiln was then determined. A fast and efficient computational model of the pilot kiln was developed to predict the burner flame and heat transfer in the kiln; the granular bed was modelled using a one-dimensional model which was coupled in a pragmatic way to the computational fluid dynamics gas phase model. Dependence of the predicted temperature profile on operating parameters is in reasonable agreement with the measurements, and the model can be run rapidly for efficient assessment of a large number of design and operating condition options. While there have been an increasing number of papers published in recent years reporting very detailed models of the motion in granular beds in rotary drums that are extremely valuable for a fundamental understanding of issues such as segregation, such models are very computationally intensive, and are still not feasible for industrial-scale kilns in which the particles are small. This paper indicates that good predictions of process parameters such as bed temperature are possible without such very time-consuming computations. Moreover, the experimental temperature measurements will be useful for future rotary kiln studies.