MODELLING OF AIRFLOW

1
MODELLING OF AIRFLOW IN WOOD KILNS
UBC Mechanical Engineering CFD Modelling
Process Simulations Ltd.
by E. Bibeau
Kiln Drying Course UBC June 1, 2000
2
CONTENTS

• Airflow in kilns
• Factors affecting airflow
• Airflow modelling
• Airflow results
• Plenum design, sticker thickness, and roof design
• Wood drying model
• Conclusions

3
Research Group
4
PROCESS MODELS
5
NUMERICAL MODEL

• Developing wood kiln model
• Predict airflow, mass transfer, and heat transfer

License agreement
UBC
PSL
Service agreements
Custom agreements
Consulting agreements
License agreements
Government labs
Industry
Other Institutions
6
DRYING KILN
Reversible Fan
Fan Deck
Automatic Vents
Steam Spray
Heating Coils
Top Load
Baffle
Lumber
Stack
Bottom
Booster Coil
Load Baffle
7
DRYING CYCLE
Stage I Convection
Stage II Convection- Diffusion
Stage III Diffusion
Bound water
Drying
Free water
Time
8
KILN OPERATION
KILN OPERATOR CONTROL STRATEGY
o
Dry Bulb T
o
Wet Bulb T
WATER
HEAT
Heat Transfer
Mass Transfer
WOOD
STRESSES
9
IMPORTANCE OF AIRFLOW
FLUID DYNAMIC CONTROL STRATEGY
AIRFLOW
Mass Transfer
Heat Transfer
o
Wet Bulb T
o
Dry Bulb T
WATER
HEAT
WOOD
Valid in
Stage I II
STRESSES
10
IMPORTANCE OF AIRFLOW
AIRFLOW
Relationship
Relationship
MASS TRANSFER (DRYING)
HEAT TRANSFER
Valid in
Stage I II
11
KILN AIRFLOW CONTROL

• Fan speed (not always an option)
• Fan reversal
• Fan positions and ducting
• Packaging (sticker, aligning, boxing)
• Airflow devices (baffles, door strips)
• Kiln geometry
• Minimize leakage
• Lumber size control

12
SOME PARAMETERS AFFECTING AIRFLOW

• DEVELOPING FLOW
• GAPS BETWEEN BOARDS
• LUMBER IRREGULARITIES
• TURBULENCE LEVELS

LITERATURE
13
DEVELOPING FLOW

• Airflow between 2 plates creates a profile
• Air sticks to the wall thus slowing down the
  airflow at the wall

14
DEVELOPING FLOW

• The profile changes as the air travels through
  the wood stack
• Shear varies along wood stack
• Flow is turbulent

15
DEVELOPING FLOW

• Air detaches from leading edge
• Further increases shear and non-uniformity near
  leading edge

16
DEVELOPING FLOW

• Combined effect
• Flow sticks to the wall
• Airflow detached from wood at the start
• Increase in drying rate gt 100
• Region of influence Sticker/L lt 50

17
DEVELOPING FLOW

• Strategy to avoid non-uniformity caused by
  developing flow
• Fan reversal
• Especially important in first stage of drying

18
SMALL GAPS BETWEEN BOARDS

• Cause airflow exchange between the air in the
  channel and the air trapped between the gaps
• Cause increase in shear

Airflow
Increase
Shear
Wood
Wood
Gap
19
SMALL GAPS BETWEEN BOARDS

• Unsteady flow (period of 2 to 7 sec)
• Literature reports overall mass transfer increase
  of 17 to 32 for 1 to 5-mm gaps
• Influence felt 20 to 40 mm
• Large increases at leading edge

20
SMALL GAPS BETWEEN BOARDS

• Gaps are beneficial
• Helps reduce drying time
• Offer more surface area to remove water
• Strategy to avoid non-uniformity caused by gaps
  between boards
• Proper stacking of wood
• Fan reversal (Stage I and II)
• Gaps should be approximately equal and
  distributed evenly throughout charge

21
BOARD IRREGULARITIES

• Unevenness in lumber height
• Caused by improper size control
• Leads to additional shear upstream and downstream
  of the variation

22
BOARD IRREGULARITIES

• Thick to thin
• Up to 100 increase initially in mass transfer
  rate
• Lower than normal afterwards (15-30 mm)
• Thin to Thick
• Larger influence
• Lower than normal afterwards (15-30 mm)
• Board height irregularities gt gaps
• Superposition of effects

23
BOARD IRREGULARITIES

• Irregularities help reduce drying time in Stage I
  and II
• Strategy to avoid non-uniformity caused by board
  irregularities
• Fan reversal (Stage I and II)
• Minimize irregularities
• Irregularities should be evenly distributed
  throughout charge as much as possible

24
Gaps and Board Irregularities
25
TURBULENCE LEVELS

• Turbulence Levels small velocity fluctuations
  in the mean flow
• The free stream turbulence of the airflow can
  affects the mass transfer significantly
• Turbulence Level Turbulent Flow

Mean flow
Fluctuating component
26
TURBULENCE LEVEL

• Increasing the turbulence level increases the
  mass transfer rate
• 55 increase for 8 increase in turbulence for
  flat plate
• Influences the velocity profile
• Turbulence in wood kilns are relatively high
• Turbulence level may decrease inside the wood
  stacks

27
KILN GEOMETRY

• Plenum width / roof height
• Study show gt 1
• Plenum width / (sticker x lumber pieces)
• Experience claim approximately 1
• Sticker thickness
• Between 1/2 to 1 1/4

28
AIRFLOW MODELING

• Plenum Design
• Sticker Thickness
• Roof Design

Numerical Simulation (CFD)
29
Some Examples of CFD Applications
Weather
Computer
Jet engines
Harrier jet
Automotive
30
Mathematical Modelling
IN
OUT
OUT
31
KILN SIMULATED
32
KILN SIMULATED SUMMARY

• Inlet Velocity 3 m/s (381 ft/s)
• Sticker 3/4
• 2 wood stacks (30 rows/stack)
• 4 gap between stacks
• Opening roof / stickers 2.0
• Opening stickers / plenum 1.2
• Rough walls and fully turbulent
• No leakage, perfect packaging
• Model half of kiln

33
KILN SIMULATED (GRID)
Base Case
34
BASE CASE-FLOW VELOCITIES
35
BASE CASE
36
BASE CASE

• Uneven flow distribution
• Lower velocities at top
• Higher velocities at bottom
• Velocity in gap between stack increases because
  of lower resistance
• Flow circulation at entrance of plenum
• Vertical flow reduces the flow entering the top
  flow channels

37
BASE CASE

• Velocity distribution influenced by plenum
  entrance geometry
• Baffle and fan deck design
• Elbow effect
• Bottom design of baffle causes non-uniformity
• Flow recirculates in lower plenum cavity
• Flow is reduced in first channel
• Larger flow in second channel

38
THREE PLENUM DESIGNS
39
PLENUM DESIGN WIDE PLENUM
40
PLENUM DESIGN TAPERED
41
PLENUM AVERAGE VELOCITY
42
PLENUM DESIGN (VELOCITY)
43
PLENUM DESIGN (PRESSURE)
44
PLENUM DESIGN

• Influence of plenum is related to the flow
  resistance through plenum and wood stack
• Kplenum smaller Ksticker
• Kplenum approximately equal to Ksticker

K
s
t
i
c
k
e
r
K
p
l
e
n
u
m
45
PLENUM DESIGN RESULTS

• Slanted plenum does not offer the best flow
  distribution
• Pressure buildup Bernoulli
• Wider plenum causes a better distribution
• Better entrance effect with wider plenum
• Improvement is based on 900 roof angle
• Better even downward flow velocity
• All 3 designs have elbow effect

46
DOUBLE PLENUM DESIGN

• May want to add vertical plates to obtain uniform
  flow

Add Vertical Plates
47
DOUBLE PLENUM DESIGN
48
DOUBLE PLENUM DESIGN
49
STICKER THICKNESS (MESH)
Base Case
50
STICKER THICKNESS (1)
51
STICKER THICKNESS (1 1/4)
52
STICKER THICKNESS
53
STICKER THICKNESS-PRESSURE
54
STICKER THICKNESS

• Main flow characteristics do not change
  significantly with the sticker thickness
• Choice of sticker thickness is dependent on all
  the other parameters affecting airflow
• Need better geometrical control for small sticker
• Small gaps
• Height irregularities
• Missing boards

55
STICKER THICKNESS

• Decrease in sticker thickness
• Increase in flow resistance
• Increase or decrease in flow velocity in the
  channels
• Reducing sticker thickness increases kiln
  capacity but longer drying times
• Smaller sticker is risky
• Kiln more prone to flow variations
• Some mills found reduced drying using 1/2 rigid
  stickers
• Report an increase in moisture variation

56
STICKER THICKNESS

• How is the moisture variation in a channel
  affected by change in sticker
• Answer Depends
• Did you preserve same
• mass of air per channel
• air velocity
• Related to shear stress at the wall
• If shear and air mass are similar
• No real effect on moisture variation expected
• Provided excellent geometry control

Airflow
Wood
57
ROOF DESIGN (MESH)
58
ROOF DESIGN (45o Baffle)
59
ROOF DESIGN (30o Baffle)
60
ROOF DESIGN (VELOCITY)
61
ROOF DESIGN (VELOCITY)
62
ROOF DESIGN (PRESSURE)
63
ROOF DESIGN

• Roof design affects how the flow enters the kiln
• The baffle affects how the flow distributes in
  the top wood stack
• The slanted roof causes the flow to accelerate
  before entering the plenum
• Velocity distribution in the top part of the
  plenum is velocity dependent

64
DOUBLE TRACK KILN
65
Wood Drying Model

• The lumber is assumed to be a porous, homogeneous
  solid
• There are three kinds of water inside the lumber
  free water, bound water and water vapor
• Moisture content at the surface of the lumber is
  in equilibrium with the air
• Shrinkage of the lumber during drying is
  neglected

66
Wood Drying Model
Mass balance
n mass flux density m phase change term M
Moisture Content
Liquid phase Vapor Air
Energy balance
Three parameters are retained M Moisture
Content T Temperature P Total Pressure in
gaseous Phase.
67
Wood Drying
Wood
M, T
Shear stress (Result of Airflow)
Airflow
M, T
Wood
M, T
Two-way coupling
Shear stress (Airflow)
Heat and mass transfer (wood surface)
Temperature, Moisture (wood surface)
Temperature, Humidity (Airflow)
68
Moisture
69
Pressure
70
Temperature
71
Develop Tools
Operator experience
Process knowledge
Operational Simulators
Simulator Core
Training Simulators
Measurements
Virtual Cameras
Physical Model
72
Process Simulators
Numerical
Model
Training Data
Neural network
S
u
p
e
r
-
G
e
n
e
r
a
t
o
r
h
e
a
t
e
r
B
a
n
k
B
a
n
k
S
p
r
a
y
73
VIEWER EXAMPLELook at different states
interactively
74
CONCLUSIONS

• Importance of airflow
• Factors affecting airflow
• Numerical simulations of airflow
• Plenum designs, roof shapes, and sticker
  thickness
• Airflow model can constitute a powerful tool
• Optimize functional and design kiln parameters
• Help operators better operate kilns without
  adding major costs

75
COPY OF PRESENTATION

• Go to www.psl.bc.ca
• Press on Public Download button
• Go to directory Woodkiln
• Download file kiln_course.ppt