Convective Heat Transfer in a Moving Bed with Special Reference to Melting

Abstract

Previous workers have been unable to correlate data on forced convection melting of metals or organic solids in packed beds in the accepted manner of heat transfer factor or Nusselt number against Reynolds number. it was recognised that melting in a tower furnace was restricted to a short zone at the bottom so, to simplify the problem, Akers (2) simulated the upper zone under moving bed conditions. His heat transfer coefficients and those of Norton 26) were as much as an order of magnitude less than those obtained by other workers using fixed beds. It was considered that this discrepancy had to be explained before proceeding to the analysis of melting beds. The present work has shown that it stemmed from two main causes: dislocations of the bed near the walls permitted bypassing of the fluid in the regions of high voidage; with long beds the terminal temperature differences tended to zero so measurement errors became unusually significant. By operation with vessel to particle diameter ratios not less than 13 and by selection of conditions to prevent small terminal differences, data have been obtained for 6, 9 and 12 mm diameter glass ballotini heated by air which are in close agreement with published fixed bed results. They are correlated by: Nu = 0.183 x Re0.834 for 95<Re<1662. A previous model of adiabatic moving bed heat transfer with intra-particle temperature distribution has been extended to the non-adiabatic case. <A new digital computer approach to solution of the model has showm that a previous hybrid-analogue computer method was unstable. Digital numerical integration has been used for analysis of experimental results and for prediction of general parameters which will greatly simplify the design of moving beds.