Experimental Studies of Heat and Mass Transfer in Fixed Beds - 2

he traditional approach to the problem of transport of heat and mass in a fixed bed has been to introduce "effective" parameters, which lump together all the physical phenomena contributing to the transport. These parameters are then estimated from experimental data by regression analysis of appropriate models, which usually assume the fluid flowing through the tube to be in unidirectional axial "plug" flow. There exists a large degree of disagreement between the data from various workers, especially for the dimensionless heat transfer coefficient or wall Nusselt number.

Our earliest experimental work was aimed at getting the effective radial thermal conductivity k_r and the wall heat transfer coefficient h_w in low-N (5-10) fixed beds of spheres. We looked at the "length effect" on the parameters [Dixon, A.G., 1985] and identified places where the experimental rig could be improved to eliminate this effect. We then extended our heat transfer measurements to include packings of cylinders and rings [Dixon, A.G., 1988]. The behaviour of the parameters at the lowest values of N motivated us to start studying bed voidage, structure and heat transfer at very low N (< 4). We looked at angular temperature variations [Dixon, A.G., 1993], statistical features [Dixon, A.G., 1994] and most recently the effects of N, d_p and d_t.

The large variability that is seen in the results is linked to the strong changes in bed packing structure that are observed as N decreases from 10 downwards, and these are especially influential for N < 4. Our research here has focused on studying the structural features of very low-N fixed beds, and relating these features to the observed heat transfer behavior [Dixon, A.G., 1997; Dixon, A.G. and van Dongeren, J.H., 1998].