Simulation of Mixing Processes 1
Simulation of Mixing Processes
1
The Influence of the Mixing in the Process.
Moshe Bentolila
09/06/2015 2
Contents
• Motivation
• VisiMix Model
• VisiMix Applications in Industry
• Conclusions
09/06/2015 3
Motivation
09/06/2015 4
Common Questions
Did we cover the main parameters during the process
development?
Will our facilities will be appropriate for the developed
process?
Does the equipment offer is good for the process?
What about safety and runaway scenario?
Do our process is robust?
Does the operational range parameters are large enough for
the manufacture facilities?
09/06/2015 5
The Goal Once the Science of the process (Chemistry, Biology or physics) is known well, a common situation during the process transfer from lab to production or from site to site is the gap between the old and new results. Our first goal is to develop a process that will run properly in
the first trial on a new scale or site, similar to our successful results in the lab or in the old facility.
In order to achieve this, we need to evaluate the process with the same conditions we will have in the production phase. The main parameters we change are the hydrodynamics of
the system. If we are able to identify and control these parameters we will be able to achieve to the available and optimal solution.
09/06/2015 6
Mixing Simulation Software
R&D
Production
Design
QbD
Data and Results Management
09/06/2015 7
VisiMix Model
09/06/2015 8
Technology - Simulation of Mixing Processes
09/06/2015 9
Typical Mixing Parameters
Application * Key process and scale up parameters
Newtonian/ Non Newtonian Hydrodynamics and scale up
•Circulation flow rate •Local turbulence values •Shear rates
Blending- Single Phase mixing •Macro and micro mixing times •Max./ Min. concentration difference C
Suspension, Crystallization, Dissolution •Max. local conc’s •Max. shear rate •Crystal collision energy
Liquid liquid mixing Emulsification, Heterogeneous org. synth.
•Drop size dist. •Surface specific mixing time
Gas injection, Absorption, Gas liquid reactors •Gas hold up •Specific surface •KL a- spec. mass trans. rate
Biotechnology •Oxygen mass transfer rate
Heat transfer in tanks with different heat/cooling devices
•Media temp’ •Heat transfer coeff. •Specific heat/cool rate
09/06/2015 10
Flow chart of the mathematical model and calculations INITIAL DATA
EQUIPMENT SUBSTANCES REGIME [type, design, size] [phases, composition, properties] [flow rates,
process parameters]
HYDRODYNAMICS
power consumption, circulation rate, forces, flow pattern, local
flow velocities
TURBULENCE
Macro-Scale Turbulent Mixing Micro-Scale Local Turbulence
Distribution of Turbulent Dissipation
MODELING OF MACRO-SCALE AND MICRO-SCALE
MIXING-DEPENDENT PHENOMENA
single-phase mixing, pick-up of solids, solid distribution, drop
breaking, coalescence, heat transfer, heating/cooling dynamics,
mass transfer,etc.
DYNAMIC CHARACTERISTICS OF MIXING-
DEPENDENT PROCESSES AT THE TRANSIENT STAGE
09/06/2015 11
A simplified scheme of mixing in the turbulent regime
1 - Central zone;
2 - peripheral zone
3 - upper level of liquid
4 - shaft
5 - torque
6 - wall
7 - agitator’s blade
q - circulation flow rate;
D - eddy diffusivity
Wax - average axial
circulation velocity
09/06/2015 12
Typical VisiMix Reactor
Mbot
ω
Mbaf Mbl
Mwall
09/06/2015 13
VisiMix Model
Main Purpose of VisiMix Modeling : Analysis of processes based on simulation of mixing-dependent phenomena
09/06/2015 14
Basic Characteristics of VisiMix Physical Models
• All the initial assumptions are based on fundamental scientific data.
• All parameters of the models are functions of basic flow characteristics, and not of the equipment specific features.
• All experimental coefficients in the equations have a clear physical meaning and are defined by independent measurements.
• The results of modeling are always verified by experiments with agitators and tanks of different types and sizes.
• All the calculation methods have passed a stage of industrial applications.
09/06/2015 15
General Equilibrium of Momentum:
T
R
(r)dr;tgvRtgV;
VρwHfπR
wall M
T
tgT
0
12
2: wallof Resistance2
2
;
r
r
(r)dr/tg
vρi
hi
Ziς
ior M
)i
(rtgv
ib
ih
iZ
iς
i M r
2
1
22
2
2
:devices Internal
0i
ΣMbot
Mwall
Mimp
ΣM
;2
2 :k Agitator rtg r-voωrdr; U
R(r)U
ρbl,k
hbl,k
Zimp,k
Zbl,kς
imp,k M
imp
inr
09/06/2015 16
The mathematical description of the tangential flow is based on the
momentum balance. For steady state conditions, the general
equilibrium is presented as the balance of the agitator torque and
flow resistance moments of the tank wall, its bottom and baffles;
these moments are expressed in terms of flow resistance and
calculated using empirical functions for the resistance factors (fw, fbl,
etc.,)
Coefficient Evaluation
09/06/2015 17
Coefficient Evaluation
09/06/2015 18
Local equilibrium of momentum
;0 shearresimp dMdMdM
;2
)(2
rdrrU
hZZdMimp
in
R
rblblimpblimp
;22
)()( 2
2
drrrv
rfdMtg
bot
);(2 2rHddM shear
;dr
dvturb
drdvLturb
2
Agitator:
Resistance:
Shear:
09/06/2015 19
The system includes also an equation of the turbulent
transfer of shear momentum expressed in terms of the
"mixing length of Prantl" hypothesis :
Prandtl’s Mixing Length
•Analogous to the kinetic theory of gases
•Used because ‘it works’
Suppose ‘lumps’ of fluid move
randomly from one shear layer
to another, a distance l apart.
This carries momentum and the
velocity difference must
therefore be related to the
turbulence
y
y1
y2 l
(y)u
Kolmogorov Representation
• The large eddies absorb the kinetic energy from the main flow provided their characteristic frequency u/L is tuned to the frequency of the main flow. L is the scale of large eddies
• The energy supplied at the highest hierarchical level, corresponding to the largest length scale L, is expanded to induce motions at the lower levels characterized by smaller length λk. Within a wide inertial interval (Kolmogorov scale) both dissipation and external supply energy are negligible, and thus energy is only transferred from one mode to another.
Visimix.Ltd 21
Numerical Solution
09/06/2015 22
Axial circulation
;axtgbl PPPP
;
2
sin)(3
drvr
hZP
imp
in
R
r
tgimp
blblblbl
;2
)(3
3
tgw
bftg
bfbfbftg VHRfRv
SZP
,22
0
2
zz
ubltgax HrQV
PPPP
;
0rr
zturbz
dr
dv
dr
dvLturb
2
Total consumption of power:
Impeller blades :
Tangential flow :
Axial circulation :
09/06/2015 23
The description of the meridional circulation is based on the analysis
of energy distribution in the tank volume, and the calculations are
performed using the results of modeling of the tangential flow.
Axial circulation
1
2 3
4
5
09/06/2015 24
09/06/2015 25
VisiMix Application in Industry
09/06/2015 26
09/06/2015 28
09/06/2015 29
09/06/2015 30
31 Visimix.Ltd
QbD Methodology (J.M. Berty, CEP, 1979)
LABORATORY (R&D)
BENCH SCALE (RC1,Mini Pilot)
PILOT
Demo – Simulation (Visimix, Dynochem,CFD)
PLANT (Production)
LABORATORY (R&D)
BENCH SCALE (RC1,HEL)
PILOT (Mini Pilot)
PLANT (Pilot, Production)
Scale Down
Final Design
Build
Design Analyze
New Process with Mixing
Lab and Prod Calculations
Moshe Bentolila, Roberto Novoa, and Wayne Genck, Michal Hasson, Efrat Manoff, "Computer Aided Process Engineering at Chemagis" , PHARMACEUTICAL ENGINEERING July/August 2011. 30-38
09/06/2015 32
Non ideal stirring – non homogeneity • Before performance of scale up experiments VisiMix
simulation was used to check suspension at different Mini Pilot Reactors:
Reactor 7603 7605 7605 7607
Volume, L 10 25 25 50
RPM 500 (Max) 400 500 (Max) 150 (Max)
Main Characteristic
Liquid – Solid Mixing
Solid suspension quality
Complete suspension is questionable.
Partial settling of solid phase may
occur.
Complete suspension is
expected.
Complete suspension is
expected.
Complete suspension is questionable.
Partial settling of solid phase may
occur. Max. degree of non uniformity of solid
distribution
AXIAL, % 22.3 10.3 29.1 132
RADIAL, % 65.7 34.3 76.3 90.8
Not all Mini Pilot reactor are capable of full suspension of POCA.
33
Methodology
In the three years since the commencement of this process, their engineers achieved a high level of proficiency in the use of the simulation models in order to analyze the results as a function of the process operational parameters. After three years of working with this integrated plan – they have summarized their knowledge and experiences up to this point, as follows: 1. VisiMix products, when integrated in the validation process (up until they
achieved a stable process and confirmed the production) – helped to reduce the number of lost production batches (each batch valued in millions of dollars) - from 100 to just under 10 batches – (review slide 25)
2. VisiMix program used in conjunction with another simulation tool - as
reported in this presentation - contributed to the improvement of the teamwork style and professionalism. Net results were observed throughout the implementation of the new solution in better project development: EOR (end of reaction) time reductions, projects development time and cost reduction, and in addition - increasing the expertise and qualification level of the professional staff.
The presentation can be review on the Visimix Website in the References - Users Publications page. (Scale up optimization using simulation experiments-Chemagis
presentation) 09/06/2015 34
Methodology
09/06/2015
# produced lots needed until a stable process is achieved
35
VisiMix Application
Homogeneous Reaction
09/06/2015 36
Chemical Reaction
Lab experiment
without stirring HClFeeding
upper from the top HClFeeding the
close to the impeller low HClFeeding the agitator velocity
close to the impeller high HClFeeding the velocity
Results VisiMix
Visimix.Ltd 39
Optimax VisiMix Optimax Model
Results VisiMix
Visimix.Ltd 40
Reaction -Reported Examples
Bromination of 1,3,5-trimethoxybenzene 1,3,5-trimethoxybenzene (A) offers three equally relative
sites where halogen (B) can substitude one hydrogen atom. The rate constants k1, k2 and k3 for successive substitution are
not 3:2:1 because of the strong deactivation of subsequent electrophilic substitutions by halogens.
Because k1>>k2>>k3 much more R (monobromo) and S (dibromo) should finally be present starting from approximately equimolar quantities of A and B (b~1).
The methoxy group strongly activates bromination and measured product distributions show substantially more S than would be expected knowing that k1>>k2.
(J.R. Bourne & F. Kozicki “ Mixing effects during the bromination of 1,3,5, -trimethoxybenzene” Chem. Eng. Sci. 32 (1977) 1538)
Reaction -Reported Examples
Results Reported
The trend was clear. (stopped flow apparatus whose mixing time is ~ 1 ms).
N(rpm) 0 213 425 1063 Stopped flow
A(%) 22.2 19.9 18.3 13.5 4
R(%) 57.9 61.3 64.5 73.4 87
S(%) 19.9 18.8 17.2 13.1 9
Reaction -Reported Examples
Hydrolysis of ethyl monochloroethanoate and neutralisation of HCl. The following reactions compete for NaOH (the limiting reagent) and have been widely used
NaOH + HCl -> NaCl + H2O k1
NaOH + CH2ClCOOC2H5 -> C2H5OH + CH2ClCOONa k2 In such experiments alkali is added to a stirred, acidic ester solution. Rate constants at 298K are k1 = 1.3 x 108 m3/mol.s k2 = 0.030 m3/mol.s the product distribution can berepresented by the yield of alcohol relative to the limiting reagent (B) and denoted by XQ.
XQ = Q/Bo
J. Baldyga and J.R. Bourne, “Turbulent Mixing and Chemical Reactions” Wiley (1999)
Reaction -Reported Examples
S - just below the liquid surface
XQ = 0.266
D - in discharge stream of turbine
XQ = 0.153
I - in suction stream of turbine
XQ = 0.138
VisiMix Application
Crystallization
09/06/2015 45
ProcessesMixing Parameters for Crystallization
For calculation of the Mass transfer coefficient, it is
necessary to enter a number of additional initial data,
including the Diffusivity of the solute. In our case the
problem consists not in prediction, but in reproduction
of the same value of the Mass transfer coefficient.
Mixing Parameters for Crystallization Processes
Cooling crystallization of API in 6000 liter reactor.
After investigation it was found that the tip diameter of the Agitator was damaged and the real diameter is a 80 % of the reported one.
new.vsm-3107VisiMix_R .vsm3107VisiMix_R
X (v,90) < 250 micron Campaign
195 First
325 Second
Visimix.Ltd 48
VisiMix Application
Gas Liquid Reaction
09/06/2015 49
Description
• Gas – Liquid reaction in pilot scale ~ 1000 liter is finished after 4 Hour
• Same in Production 4 Days.
Visimix.Ltd 50
VisiMix Application
Safe Process Assess
09/06/2015 51
Developing an innovative way to dramatically improve the safety of the chemical processes. By: Mr.Nekhamkin-
17.06.2015 -10:30 am- Hall 9.1, Room: Logos / Genius
High Shear Rate at Chemical Fast Reactions
09/06/2015 52
An innovative new technology to better utilize processes for both RSD (Rotor Stator Dispersers) and/or Emulsifiers. By: Dr. Kokotov- 18.06.2015 -16:40 pm- CMF, Room: Harmonie 5
Application VisiMix
Process and Quality Problem
Process R-6826
Feed R-Cl R-NH2 + R’-Cl t-D-R-R’
Impurity
t-L-R-R’
09/06/2015 53
Impurity results at laboratory and in production
System volume Impeller type RPM impurity]%[
Laboratory reactor 0.63 lit
rotor stator 15,000 rpm 0%
3-blade
1,500 rpm 0.3%
800 rpm 0.6%
100 rpm 1.5%
Production
R-6826 2,978 lit
bottom – flat blade
up - turbofoil 140 rpm 0.3% - 0.6%
Correlation between shear rates and the impurity concentration
54
R’- Cl (liquid)
R- NH2
Working with rotor stator at laboratory scale
Problem
How to scale up ?
Potential Saving :
MORE than 250 K$
55
Rotor Stator Technology
09/06/2015 56
Calculating shear forces with VisiMix
The required shear rate can not be achieved in the production reactor
Lab impeller Rotor stator R-6826
system Impeller type RPM impurity]%[ Turbulent shear rate
[1/s]
Laboratory reactor
rotor stator 15,000 rpm 0% 780,000
3-blade
1,500 rpm 0.3% 32,900
800 rpm 0.6% 12,900
100 rpm 1.5% 580
R-6826 bottom – flat blade
up - turbofoil 140 rpm 0.3% - 0.6% 15,200
09/06/2015 57
Conclusion
Using VisiMix Products support you can
understand better your processes
Reduce dramatically your Scaling up processes and Scaling down
Save a huge amount of Time & Money ($1,000,000 +)
The VisiMix Products are friendly and easy to use with very quick results.
The VisiMix results are based on a systematic and seriously experimental checking – and found very reliable.
VisiMix Projects Parameters and Data Base allows you to share and transfer the data with colleagues in the company.
Thank you for your attention
09/06/2015
59