cg5052 - agitation and mixing - 2 - 2015.pdf
TRANSCRIPT
-
1
The surface area for cooling to reactor volume ratio decreases upon scale-up
Hazard!!
-
2
53
SSDNP3
SSDNQ
23
SST DN ST
NV
Q
Torque:
52
2SS
S
q DNN
PT
222 tipspeedDNV
TSS
T
q
Tip speed:
SSSS DNDN
Average impeller
shear:
SN SSDN
Maximum impeller
shear:
Agitated tank hydrodynamics
For turbulent conditions and geometric similarity
-
3
53
SSDNP3
SSDNQ
23
SST DNS
T
NV
Q
Torque:
52
2SS
S
q DNN
PT
222 tipspeedDNV
TSS
T
q
Tip speed:
SSSS DNDN
Average impeller
shear:
SN SSDN
Maximum impeller
shear:
Agitated tank hydrodynamics
For turbulent conditions and geometric similarity
Basics of fluid flow and shear
conditions scale differently!
-
4
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
-
5
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
Different phenomena scale
differently!
-
6
-
7
Unclear how to scale a single
phenomena!
-
8
The overall scaling-up problem
chemical reaction
yield selectivity safety
liquid-liquid extraction
yield efficiency
crystallization
purity size and shape polymorph
blending and homogenisation heat transfer to or from wall mass and heat transfer between phases
generation of contact surface area between phases
FLOW SHEAR
1. Agitation creates:
2. Which leads to:
3. On which the process goal
depends :
phenomena that scales
differently often in a complex manner, in
combination and changing
over time
-
Scaling-up - the overall problem
Difficult to characterise how the process goal depends on what we directly can achieve by the agitation - processes are
complex
Difficult to characterise in detail and scale-up how agitation influences on surface area generation, homogenisation, and
heat and mass transfer - hydrodynamics of agitated tanks are
complex
Process experiments have to be performed and
semi-empirical scaling-up procedures used
-
10
SCALING-UP PROCEDURE
1) define the process need
2) carefully identify all mixing parameters
that may have an influence on the process
3) select the most critical mixing parameters
4) scale on the most critical parameters
5) carefully review the behavior on each scale
-
11
Change the agitation rate to examine the role of power input -plot your process result versus N3 in a log-log plot
Use the small scale experiments to try to identify
the role of mixing
-
12
Use the small scale experiments to try to identify the role of mixing
Change the agitation rate to examine the role of power input -plot your process result versus N3
Change the impeller diameter to examine whether flow or shear is governing
Change impeller blade width to examine whether macroscale or microscale mixing is important
3
SSDNQ
SSDNMaximum shear:
Flow:
Use the small scale experiments to try to identify
the role of mixing
-
13
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
Assumes turbulent conditions and geometric similarity
-
14
Non-geometric scale-up
Why geometric similarity?
Geometry do not control no mixing in itself!
Geometric similarity occupies one degree of freedom that can be used to control mixing!
Allow different geometry to enable an additional mixing related scale-up criterion
Ex liquid-liquid contacting:
Droplet generation occurs in the impeller region
Droplet coalescence occurs in the bulk region
1,2, TT
2
1,
2,
1,
2,
S
S
T
T
D
D
-
15
Mixing requirements
Require rapid attainment of uniformity:
fast chemical reactions
competitive chemical reaction
reaction crystallizations
liquid-liquid and gas-liquid processes influenced by mass transfer
Uniformity less important:
Heat transfer
Blending of miscible liquids
Slow chemical reactions
Suspending of solids
More easy performed in small scale and are more difficult to scale up
Can usually be scaled with few difficulties
-
16
Some final remarks
It is impossible to have all conditions equal in different scales Preoccupation with scale-up rules truncate thinking Geometric similarity do not capture an important mixing
property
Use intelligent experiments to identify what aspects of the mixing that is governing your process goal
Be aware of that the governing mechanism can depend on the scale
-
17
-
A reactor in the pilot plant has a volume of 1000 litres and a diameter of
1 m. It can be equipped with a propeller with a diameter of 0.3 m, or a
45o pitched blade turbine with a diameter of 0.5 m. It has been decided
to purchase reactors to the process lab to simulate 1000 litre in a
smaller scale.
Determine the dimensions of the reactors if the volumes are 20 or 50
litres, respectively, and the reactors are geometrically equivalent with
the pilot plant reactor.
-
A 1000 L baffled tank has the diameter 1 m and can be equipped with a
Rushton turbine, a propeller or a 45 pitched blade turbine. Determine the flow condition, power input and pumping capacity for 50, 100, 200 and
400 rpm and agitation diameters: 0,33 and 0,45.
Data:
Density 1000 kg/m3
Viscosity: 0.001 Ns/m2
Rushton turbine Ne=5.2 Nq=0.72
Propeller Ne=0.5 Nq=0.8
45 pitched blade turbine Ne=1.3 Nq=0.79
-
2Re SS
DN 53
SS
pDN
PN
3
SS
qDN
QN
Re P Q
-
You have done experiments with water in laboratory experiments and
are now considering what will happen with the power consumption
when you the same experiments with the process liquor:
Water Process liquor
=1000 kg/m3 =800 kg/m3
0.001 Ns/m2 0.005 Ns/m2
How will the power consumption in the process liquor compare
to that in the aqueous phase at
a.equal agitation rate and turbulent conditions?
b.equal Reynolds number and turbulent conditions
1
-
A catalytic powder is charged into a solution in an agitated baffled tank.
In 1000 litre scale this process is carried out with rotation speed 100 rpm
and stirrer diameter 0.3 m. To study the process you would like to scale
down to 6 litre scale. Determine appropriate stirrer speed and impeller
diameter in the 6 litre scale. Assume geometric similarity and that the
liquid is water.
22
1) Criterion: keep particles suspended from the bottom:
2) Suspending charactersitic: 85.0 SJS DN
3/1VDS mD labS 055.0,
85.0
arg,
,
arg,
,
elS
labS
elJS
labJS
D
D
N
N rpmN labS 425,
3/1
argarg,
,
el
lab
elS
labS
V
V
D
D
85.0 SJS DN
2
-
A catalytic powder is charged into a solution in an agitated baffled
tank. In 1000 litre scale this process is carried out with rotation
speed 100 rpm and stirrer diameter 0.3 m. To study the process
you would like to scale down to 6 litre scale. Determine
appropriate stirrer speed and impeller diameter in the 6 litre scale.
Assume geometric similarity and that the liquid is water.
1) Criterion: keep particles suspended from the bottom:
2) Suspending characteristic: 85.0 SJS DN
mD labS 055.0,
rpmN labS 425,
Turbulent?
2Re SS
DN
150000Re arg el
21100Re lab
Turbulent?
2
-
A catalytic powder is charged into a solution in an agitated baffled
tank. In 1000 litre scale this process is carried out with rotation speed
100 rpm and stirrer diameter 0.3 m. To study the process you would
like to scale down to 6 litre scale. Determine appropriate stirrer speed
and impeller diameter in the 6 litre scale. Assume geometric similarity
and that the liquid is water.
Determine also the specific power input required assuming that the
power number is 4
24
1) Criterion: keep particles suspended from the bottom:
2) Suspending charactersitic: 85.0 SJS DN
3) Geometric similarity and turbulent conditions: 23
ssT DN
55.0
arg,
,
arg,
,
elS
labS
elT
labT
D
D
53
ssp DNNP
V
DNN sspT
53
3/45arg, mWelT
3/115, mWlabT
2
-
A solid-liquid and liquid-liquid reaction was scaled-up to 3500 gal. The
scaling-up was performed based on the just suspended state of the
solid particles. However, upon analyzing the yield of the full scale
production plant, it soon became clear that the productivity had
decreased, giving only about 10% of the expected productivity. What
had gone wrong?
3
-
26
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
3
-
A solid-liquid and liquid-liquid reaction was scaled-up to 3500 gal. The
scaling-up was performed based on the just suspended state of the
solid particles. However, upon analyzing the yield of the full scale
production plant, it soon became clear that the productivity had
decreased, giving only about 10% of the expected productivity. What
had gone wrong?
Would need more agitation to generate the same droplet
surface area per unit volume
There is more coalescence in the larger tank because of
longer circulation times
The liquid-liquid contacting forgotten
3
-
28
You have investigated a chemical reaction by labexperiements. In the
process two liquid solutions are mixed in the presence of a solid
catalyst. The catalyst particles have a diameter of 2 mm and a density
of 1900 kg/m3. The liquid phase has a density of 900 kg/m3 and the
viscosity is 0.002 Ns/m2. The experiments have been performed in a 1
L tank with four baffles and 0.11 m diameter, agitated by a pitched blade
turbine having a diameter 40 % of the tank diameter. In the experiments
it has been observed that the yield increases with increasing agitation
rate upto 700 rpm beyond which there is no improvement.
Measurements show that the reaction has gone to completion within 2
hours and that the temperature increases at most by 10 . The process is now to be transferred to the plant and be performed in a 4 m3 tank.
What should be the agitation in the plant and are there any particular scaleup issues to be aware of
4
-
29
catalyst particles: 2 mm and 1900 kg/m3.
liquid density: 900 kg/m3.
Viscosity: 0.002 Ns/m2
1 L tank, four baffles, pitched blade turbine, diameter 40 % of the
tank diameter.
700 rpm is the optimum
reaction time: 2 hours
temperature increases at most by 10 Plant: 4 m3 tank.
Scaling-up is normally based on geometric similarity do we have that?
Do we have turbulent conditions? The chemical reaction is obviously influenced by the agitation Liquid-liquid blending Solids suspending Evolution of heat
4
-
30
catalyst particles: 2 mm and 1900 kg/m3.
liquid density: 900 kg/m3.
1 L tank, four baffles, pitched blade turbine, diameter 1/3 of the tank
diameter.
600 rpm is the optimum
reaction time: 2 hours
temperature increases at most by 10 Plant: 4 m3 tank.
Scaling-up is normally based on geometric similarity do we have that?
Do we have turbulent conditions The chemical reaction is obviously influence by the agitation Liquid-liquid blending Solids suspending Evolution of heat 10 is not insignificant since the plant will have 15 times less
area/unit volume Normally the agitation rate cannot be increased sufficiently to
compensate Longer time is required for the process
4
-
31
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
4
-
Scaling-up of heat transfer 5
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of blending
2
1,
2,
1,
2,
S
S
T
T
D
D
Scaling-up of liquid-
liquid dispersion 1,2, TT
Scaling-up of solid-
liquid suspending
55,0
1,
2,
1,
2,
S
S
T
T
D
D
3/2
1
2
V
V
3/5
1
2
V
V
3/55.0
1
2
V
V
23
SST DN 1,2, SS NN
SS DN
85.0 SJS DN
3/2 SS DN
4