IUNIrI MITY CU AII DISTIIUIIUIIN
IN A IUMI U L IDRY I\ILN
Information Iuvictk l dtid Ultlirmiitd
UNITED STATES DEPARTMENT OF AGRICULTURE
FOREST PRODUCTS LABORATORY
In Cooperation with the University of Wisconsin
Digitized by the Internet Archive
U1IFORMJITY OF AIR DISTRIBUTION IN A LUMBER DRY KILN1
By 0. W. TORGESON, Engineer
Forest Products Laboratory,- Forest Service
U. S. Department of Agriculture
The advantages of optimum rates of air circulation in a lumber dry kiln
are often lost by permitting nonuniformity of air flow through the load.
Any pronounced irregularity of air flow will cause a lag in drying in
some areas and will govern either the drying time or the uniformity of
Recently the Forest Products Laboratory started a study of aerodynamic
problems within a kiln which included a few experiments to determine the
degree to which certain factors influence the uniformity of air flow
through a pile of 1-inch boards. Three factors were studied: (1) width
of entering air space, (2) projecting edges on the entering air side,
and (3) thickness of stickers. Factors (1) and (2) were determined
simultaneously so that the results presented here portray the combined
effect rather than the independent effect of these two factors.
The kiln used is 18 feet long and 8 feet wide, and is equipped with
three 36-inch overhead disk fans. These were driven by individual 1.5
hp. motors at a speed of 550 revolutions per minute. No heat wras
supplied and all air velocity readings were taken at ordinary room
temperatures by means of a velocity meter equipped with an 18-inch
(1) Width of Entering Air Space
The lumber pile in this test consisted of 1-inch lumber 16 feet long
piled on 1-inch stickers to a width of 3 feet and a height of 7 feet.
The entering air edges projected or recessed from zero to three-fourths
of an inch. A movable partition was placed in the entering air space
and shifted so as to provide entering air spaces 6, 12, 18, 24, and 30
inches wide, respectively. Another position was provided by resting the
top of the partition against the kiln wall and the bottom against the
Publi.hd in the Southern Lumbern-.an, April 15, 1940.
-ainrained at liaison, Wis., in cooperation with the University of
Agriculture -';ad ison
Rept. No. R1267
pile, thus forming a deflecting baffle with an entering air space 30
inches wide at the top and 1 inch wide at the bottom. The leaving air
s:-ace was 24 inches wide throughout the series of tests.
Velocity readings were taken in every sticker space along three veritcal
lines on the leaving air side of the load. For each position of the
partition, an average of all readings was obtained and in addition read-
ings above and below this grand average were averaged separately, the
difference between these latter representing, approximately, the range
between which 50 percent of the readings fell. This range was divided
by tro to obtain the plus and minus average variation from the grand
T'.e results are given in table 1. These indicate that the width of the
entering air space and the resulting air velocity within that space
greatly influenced the uniformity of air velocity through the load of
lumber. Apparently the sloping position of the partition was best.
Under this condition the air entered the pile at approximately the same
velocity from top to bottom and was partly deflected into the openings
between layers. As will be illustrated later, much of the variation
wias caused by the projecting edges, but when the circulation was reversed
and thie entering air edges were even, yet considerable variation
Table l.--Effect of width of entering air space on uniformity of air flow
Width of :Average down-: Average : Average variation through lumber
entering : ward air : air : pile above and below average
air : velocity in : velocity : velocity (includes approximately
space :upper area of: through : 50 percent of readings)
:entering air: lumber --------------------------------
space : pile
(1) (2) (3) (4) (5)
-------- ----------- ----------------------- --------------
Inches : Ft.per min. :Ft.per min.: Ft.per min. : Percent
6 1,840 : 269 : 225 : 84
12 : 860 : 251 170 68
1 620 272 : 123 45
24 : 490 : 286 90 32
30 390 : 286 69 24
30 top ): 400 294 57 19
Column 2 of table 1 gives the velocity of the air as delivered from the
f:-.< into the space between the kiln wall and lumber pile. The velocity
wiithin this -:ace has a direct and most important effect upon the uni-
formity of air flow throughout the lumrer pile. With a constant volume
Rept. 'P1. PI267
of air delivery, the entering air velocity is inversely proportional to the
width of this space. Even with the comparatively ample width of 18 inches,
the uniformity of air flow left much to be desired, being less than half
of that secured with the 30-inch space and the sloping baffle.
Column 3 gives the average air velocity, which is a measure of the total air
delivery of the fans. Eliminating the average for the 6-inch width, where
slight errors of measurements may have been caused by turbulent.fl',j, and
using the average for the 12-inch width as a base, the fan delivery was in-
creased 8 percent using an 18-inch space, 14 percent using 24 or 30-inch
spaces, and 17 percent when the side wall sloped from a width of 30 inches
at the top to 1 inch at the bottom. These added amounts of air, plus
greater uniformity, indicate the desirability of using an entering air
space at least 24 inches wide.
One other effect, not shown in table 1, is that with the 6-inch srace the
highest velocities were through the lower layers, but as the width increased
the distribution vertically cl.anged until at the 30-inch width the highest
velocities were through the top layers.
(2) Projecting Edges on Entering Air Side
Much of the nonuniformity sh,-.'T. in table 1 was due to projecting edges,
which deflected air into the spaces above and prevented air from entering
the spaces below. In some cases the direction of flow in the spaces be-
low was reversed, that is, air returned from the leaving air side to the
entering air side. In many other cases the air in spaces below project-
ing edges was practically stagnant. This effect was very pronounced when
the air velocity dowmward in the entering air space was high, but became
very much less as this velocity decreased.
Figure 1 shows layers 16, 17, and 18 (numbering from the top) and the
relative position of their enterin- edges. The air velocities obtained
under entering air space widths of 6 and 24 inches, respectively, are
given in the separate sketches "A" and "B3."
In case "A"l it is ap:-arent that, in drying, the lower surfaces of layer 17
will lag behind the upper surface resulting in uneven moisture content and
increased warp. In case "B," with an entering air space 24 inches wide the
uniformity of air flow is much better, but even here the air velocity in the
space below the projecting edge is only 76 percent of that in the space above.
In other words, uniformity of air circulation is increased considerably when
the air flows through a lumber pile having the entering edges even; and this
effect becomes more pronounced as the air velocity in the enterir.n air srace
(3) Thickness of Stickers
The pile arrangement and the procedure for this study were practically
the same as for the studies on entering air width and projecting edges,
except that in this case the entering air edges were vertically .linr.ed
Rept. 1o. R1267 -3-
and the sticker thicknesses used were 1/2, 3/4, 1, and 1-1/4 inches. The
width of both entering and leaving air spaces was 30 inches. The air
velocities were read on the leaving air side in each sticker space at
five locations along the length of the load.
The fo.lc'ling results favor the use of thin stickers: first, the kiln
capacit"- is increased. Second, the air velocity is increased due to the
decrease in total air space between layers. This increased air velocity
may preda.-e a faster drying rate. Third, the uniformity of air flow is
increased, as shown by the data presented in table 2.
Table 2.--Effect of sticker size on uniformity of air flow
Thickness: Average air : Average variation through lumber pile above
of :velocity through: and below average velocity (includes
stickers: lu-er pile : approximately 50 percent of readings)
Inches : Ft. per min. : Ft. per mmin. : Percent
1/2 : 454 : 41 : 9
3/4 : 39 : 49 : 13
1 367 : 60 : 16
1-1/4 : 330 75 23
One objection to the use of thin stickers is the additional back pressure
caused by the entrance and frictional losses and the resulting effect of
air delivery and power consumption. It might be thought that with more
lumber in the kiln more air would be needed to provide heat for the
additional amount of evaporation and to replace the amount lost due to
increased resistance. With a possible increase in drying rate and a more
uniform distribution of air, however, the time needed to obtain a given
uniform moisture content might not be increased even if no additional air
is sup-plied. T7. economics of the use of thin stickers has not been
determined, but in this particular study an estimate of the power and
air delivery losses was obtained by the use of a watt-hour meter and an
inclined manometer and pitot tube.
Tet.'lirs of pressure difference were taken on each side of one fan at a
point 6 r:!: auay from the blades and 5 inches in from the trim. These
readin)s of pressure differences are coni:-ar:able within themselves, but
are only approximately correct because of the turbulence of flow especially
on the discharge side where the air is turned abruptly by a baffle.
Readin -s :r:.e taken when the kiln was empty and when completely baffled
as representing the two limits of air delivery and power consumption.
Additional readings were taken with the lumber pile in place and piled on
1/2-inch and 1-1/4-inch stickers, re!jectively, as representing the two
e.-tremes of the stickers used. The results are given in table 3 and
Tept. No. R1267
indicate that by using 1/2-inch in place of 1-l/4-inch stickers (thus in-
creasing th-e kiln capacity 52 percent) the air deliv:.-ryr was reduced 16
percent'and the power consumption increased 1 percent. Under these con-
ditions, the capacity of the fans would have to be increased 68 percent
to provide the same amount of air per board foot as available to the
lumber when piled on 1-1/4-inch stickers.
Table 3.--Effect of sticker size on total air delivery and power
Th:ickness : Static pressure Total air :Total power
of stickers :difference across fan: delivered :consumption
Inches Inch of water : Cubic feet : Kilowatts
1/2 ................... 0.26 13,200 3.45
1-1/4 .............. .24 15,700 3.41
::iln empty.......... .23 18,000 : 3.37
Kiln fully baffled..: .52 1,000 : 4.10
It is obviously not possible, from the data at hand, to determine the
proper sticker thickness for any given set of conditions. The very evi-
dent advantages of greater kiln capacity and greater uniformity of air
flow resulting from the use of thin stickers must be balanced against
decreased air moving efficiency. A reasonable balance for one set of dry-
ing conditions :,.i be far different from a reasonable balance for an en-
tirely different set of drying conditions.
In regard to the fan delivery and power consumption, one other item of
interest w:as disclosed during these tests. The kiln used has a baffle
which deflects the air from the fan 180 uri.ird causing it to nm'k.C another
180 turn before reaching the lumber below. with the kiln empty, the
dropping of this baffle caused the static pressure difference to drop
from 0.23 to 0.09 inch of water, the power consumption to drop from 3.37
to 2.E3 kilowa-ts, and the air delivery to increase from 18,000 to
slightly over 20,000 cubic feet per minute. This shows the importance
of keeping any sudden changes in air direction to a minimum.
The following conclusions are applicable only -..en conditions are similar
to those of these e:.:reriments.
1. The ridth of the entering air space should i..t be less than 24 ':.cs.
A tentative rule mi-ht be set up that tLA-s width of the enteri:-.r air space
Rept. '.J, R1267
should not be less than one-half of the sum of the spaces between the
layers of boards of a full kiln load of 1-inch lumber. This minimum
width of space needs to exist only at the top of the load and from there
down a sloping baffle is beneficial by deflecting the air into the spaces
between I he boards.
2. Projecting edges deflect large volumesof air into the spaces immedi-
ately above them and cause local variations in air velocity even to the
point here the direction of air travel is reversed through the spaces
Liediately below them. The higher the air velocity down the entering
air space the g".ater was this action, but in all cases good returns in
uniforri-ity c~n be secured simply by aligning the entering air edges.
3. Uniformity in air distribution was improved by reducing the thickness
of stickers. This procedure has the added advantage of increasing the
`iln capacity with only a negligible increase in power consumption. It is
believed that these advantages overbalance the fact that some reduction
in air delivery is caused by the additional frictional losses.
4L. Baffles causing sudden changes in the direction of air movement
result in considerable increases in power consumption and decreases in
P.rc.t. No. P1267
480 FT. PER Il/N.
17 FT. PER MIN.
0307 FT. PER MIN.
10233 FT. PER MIN.
U . ^ .. -i1 8m i j_ j
EFFECT OF PROJECTING EDGES ON ENTERING /IR SIDE OF LAYERS
/6,1/7 AND /8 OF I-INC/ LUMBER ,4T TWO VELOCITIES.
LOCATION OF THESE LAYERS IS 0.4 THE TOTAL HEIGHT FROM THE TOP OF THE PILE.
1 'O5701 F
UNIVERSITY OF FLORIDA
3 1262 08866 6234III III
3 1262 08866 6234