The detailed compilation of measured data is given in Ref.
[ 18], together with evaluation and model simulation results.
Dynamic viscosities and densities for water related to actual
temperatures during experiments. Pipe roughnesses were
interpreted from clear-water data generated before and after
each series of tests and relative roughnesses were close to
smooth pipe conditions. The sensitivity of relative pipe
roughness on model results was found to be small for the data
analysed here. Mean values of solids concentration were used
in model calculations. This simplification allows for a devi-
ation in calculated losses of about 5% compared to individual
measured values at velocities of practical interest.
Velocities at limit of stationary deposition, V~, for the prod-
ucts were calculated for investigated concentrations accord-
ing to procedures suggested in Wilson et al. [1]. The
maximum velocity at limit of stationary deposition is then
evaluated from particle size and pipe diameter using a nom-
ographic chart. The concentration at which this maximum
velocity occurs is calculated and V~ at the actual concentration
can then be estimated. In these calculations, it is here assumed
that the particle size is represented by dso and that Cb is 0.60.
Since some of the products contain the most critical size
fraction for deposition (d= 0.5 mm), a comparison with Vs
for this particle size was also made.
6. Results and discussion
Measured data and model predictions are compared in
terms of hydraulic gradient (i m) and excess pressure gradient
(q~) versus mean velocity (V). Figs. 4-7 feature results for
iron ores, having mean volumetric concentrations of about
23% (hematite) and 14-29% (magnetite). Results for
crushed rock ( C = 6-13%) are shown in Fig. 8. Values of w,
Vso and M were calculated from Eqs. (A2)-(A5) (Appen-
dix) and from particle size data in Fig. 2 and Table 1. Cal-
culated values of V~ (based on dso), M and Vso are given for
each product.
Figs. 4 and 5 show that observed losses for the hematite
ore agree with the model at a concentration of 23% for veloc-
ities of 3-4 m/s. However, Figs. 6 and 7 show that the model
tends to underestimate friction losses for the slightly heavier
and coarser magnetite ore at concentrations of 14-29%. The
scatter of the experimental results was rather large because
the aim of the test was to make quick measurements with