- 上一篇：ECO测试系统英文文献和中文翻译
- 下一篇：涡轮机叶片的铣削英文文献和翻译
of both aerated and unaerated stirred vessels. The model was
validatedwithmeasured pulse response curves, using either a fluo-
rescent or a hotwater tracer. Jaworski, Bujalski,Otomo, andNienow
(2000) reported the simulation results of a dual Rushton turbine
the using CFD code Fluent. The predicted mixing time was 2–3
times higher than the experimental data. Deshpande and Ranade
(2003) used a modified computational snapshot approach for pre-
dicting the interaction between the flows generated by two Rushton
impellers. In each case the CFD complexity implies that the final
results depend on a considerable number ofmodelling options and
assumptions.In most CFD simulations the baffles, impeller disc, and impeller
blades are treated as zero thickness walls. This assumption is
unrealistic since studies have shown that impeller blade thick-
nesses influence mixing properties (Bujalski, Nienow, Chatwin,
& Cooke, 1987; Rutherford, Mahmoudi, Lee, & Yianneskis, 1996;Yapici, Karasozen, Schäfer, & Uludag, 2008). Their studies indi-
cated that the power number decreases with increasing disc
thickness, while the mixing time increased with increasing disc
thickness.
In this work the actual dimensions of a stirred tank reactor
were modelled and the thickness of baffles and impeller blades
were not neglected. The flow field was simulated using LES with
a Smagorinsky–Lilly subgrid scalemodel, and the flownumber and
mixing time were simulated in a fully baffled tank reactor stirred
with two standard six-blade Rushton turbines. To account for the
impeller revolution the sliding mesh (SM) approach was used. The
paper presents a comparison between the experimental and simu-
lation results for the radial profiles of axial and radial components
of the velocity at different impeller rotational speeds.
A series of experiments were promoted for the validation of the
simulation results. The simulated velocity and stirring power input
were comparedwith the PIV results and themixing time data eval-
uated by the PLIF technique. The results indicate the usefulness of
this approach for furtherwork on devising a general purposemixer
design tool.
2. Experiments
The design and dimensions of the stirred tankwith two Rushton
impellers of standard geometry used in this work are shown in
Fig. 1. The stirred tank consists of a flat-bottomed glass cylinder
(refractive index 1.47 at a wavelength of 587 nm) having an inner
diameter T of 0.30m. Four vertical baffles are symmetrically placed
around the tank wall, each with a width one-tenth of the tankdiameter (l =0.1T). The Rushton stirrer with two impellers, which
is of standard design with a diameter (D) equal to one-third of the
tank diameter (D/T = 1/3), is located with a clearance from the tank
bottom of about half the tank diameter (C1 = 0.55T). The upper
impeller is placed C = 0.7T above the lower one. A motor drives
the turbines and the stirring speed is measured using a calibrated
digital oscilloscope. In order to reduce the optical refractive index
effects at the cylindrical surface of the tank, it is placed in a square
glass vessel. The stirred tank is filled with tap water as the main
continuous phase fluid, the surface of which is C2 = 0.55T above the
upper impeller. The square vessel is also filledwithwater to reduce
light refraction at the interface. For more details on the geometry
of the stirred tank used in this study see Guillard, Trägårdh, and
Fuchs (1999) and Moghaddas (2004).
A double-cavity 2mJ×25mJNd:Yag (continuum) pulsed laser is
used to produce a beamat a wavelength of 532 nm. The laser beam
passes through a plano-concave lens to produce a two-dimensional