EDECAD

1
EDECAD
Efficient DEsign and Control of Agglomeration in
Spray Drying Processes
Project team Full time Dr. Geraldo Nhumaio (the
presenter) Contributors Dominic Kirkman, Anna
Carruthers, Dr. Yusuf Al-Suleimani Dr. Hassan
Abduljalil Supervisors Dr. A. P. Watkins
Prof. A. J. Yule Atomisation and Sprays Research
Group
2
Contents

• Introduction
• Definition of spray drying
• Definition of agglomerates
• Advantages of agglomeration by spray drying
• EDECAD project participants
• UMIST role in the project
• Test cases on collisions and agglomerations
• UMIST cold laboratory case
• Bremen cold industrial case
• Anhydro (APV) hot industrial case
• Summary of achievements

considerable time to be spent here
3
Definition of spray drying
4
Definition of agglomerates
5
Why the need for agglomerates

• Porous agglomerates (e.g. instant coffee/milk,
  laundry detergents, agrochemicals)
• Have improved solubility in water
• Disperse better
• Have low propensity to formation of lumps
• Have dust-free flow (e.g., in vending machines)

6
Advantages of agglomeration by spray drying over
other drying technologies

• Quick drying -vs- minimum heat shock to
  processed material (friendly for heat sensitive
  products)
• Processed sticky material does not contact solid
  surfaces until dry gt equipment experiences
  minimum corrosion problems

7
EDECAD project participants
8
UMIST role in the project

• Provision of experimental spray data (initial
  conditions and validation data) for two
  interacting sprays in a wide range of boundary
  and nozzle operating conditions
• To coordinate the development and implementation
  of the collision sub-model, i.e., the
  consortium Work Package 2 (WP2)
• To implement and test novel sub-models on
  collision, agglomeration and drying using the
  (in-house) Spray3D code

9
The UMIST twin-fluid atomizers cold case
10
The UMIST twin-fluid atomizers cold case (Nozzles
used in the study)
External mix AutoJet Air Atomizing Nozzles
SUV113A from Spraying Systems Co.
Sonic flow
Liquid velocity at orifice 0 20 m/s Air
velocity at orifice 0 340 m/s
11
The UMIST twin-fluid atomizers cold case
(Measurement levels for initial spray data)
12
The UMIST twin-fluid atomizers cold case
(Measurement levels for numerical validation)
?45
13
The UMIST twin-fluid atomizers cold case
(Different test cases and conditions)
14
The UMIST twin-fluid atomizers cold case (CFD
simulations)
15
The UBremen (industrial, cold HP) case
16
The UBremen (industrial, cold HP) case
17
The APV/Anhydro/Invensys (industrial, hot HP)
case
18
The APV/Anhydro/Invensys (industrial, hot HP)
case
19
The APV/Anhydro/Invensys (industrial, hot HP)
case
20
The APV/Anhydro/Invensys (industrial, hot HP)
case
21
Summary of achievements (mainly for reporting at
EU HQ in Brussels)

• Novel validation data to aid the prediction of
  impacting water sprays were produced and made
  available to all the EDECAD consortium partners
• A 3rd yr project (rated excellent) was achieved
  in the spray drying topic
• A M.Sc. and a Ph.D. graduates were trained in
  the field of spray drying technology
• A research associate gained experience in
  developing FTN routines for collision
  agglomeration modelling as well as for
  post-processing of PDA data
• Higher collision and agglomeration rates occur
  where larger particles negotiate their paths with
  massive clouds of smaller momentum droplets

22
Appendix
Would you like to see more information? If so,
please see next slides after the meeting.
23
The UMIST Spray3D code(General Features)

• Staggered grid (at each location 3 different
  cells are used to store different vectors and
  scalar quantities)
• Transient solution for both continuum and
  discrete phases
• Non-iterative (pressure) predictor corrector
  method is used
• All parcels are tracked regardless of being
  collectors or collected
• Basic features of this code were used as
  reference for the improved (i.e., stochastic)
  collision model

24
Original UMIST Spray3D collision
sub-model(Simultaneous Particle Tracking)

• The colliding droplets are assumed to be in the
  same cell
• Droplets within the cell are assumed uniformly
  distributed
• The probability that the collector parcels
  undergo n collisions with the smaller ones within
  a time step is obtained from the linearized form
  of Poisson distribution (Sommerfeld, 2000)
• A collision occurs if a random number X,
  generated between 0 and 1, is less than Pcoll

25
Original UMIST Spray3D coalescence
s/model(Simultaneous Particle Tracking)

• The parcels of larger diameter collect their
  smaller counterparts if a random number Y,
  generated in the range 0,1, does not exceed the
  minimum surface energy required to reform the
  individual colliding drops, i.e.
• The droplet parcels reduce by one if a small size
  parcel is collected by its larger counterpart

26
Particles classification for the (novel)
agglomeration sub-model
27
15 categories of possible interactions for the
collision agglomeration models
Yes Implemented collision agglomeration
models
28
Evaporation -vs- drying sub-models(Diff. eqns.
droplets external heat transfer)
29
Evaporation -vs- drying sub-models (Diff. eqns.
droplets internal mass transfer)
Spray3D approach for evaporation of droplets of
pure liquids
D - diffusivity Dd droplet diameter Pt total
pressure Pv,? - partial pressure of vapour far
from the droplet Pv,s - partial pressure of
vapour at droplet surface Rf vapour gas
constant Tm man film temperature
30
Spray3D evaporation sub-model(Discretized
equations pure liquid droplets)
31
Drying sub-model from WP 4(Discretized
equations droplets w/ solids contents)
32
(No Transcript)
33
Example drying technology 2
Strength durability
Out
In
Ceramic Kiln
Brick
Wet clay slab
Brick industry