BIOTRONICS 13, 29-38, 1984

DIRECT DIGITAL CONTROL OF AIR HUMIDITY

FOR PLANT RESEARCH

Hiromi EGUCHI, Masaharu KITANO and Tsuyoshi MATSUI

Biotron Institute, Kyushu University, Fukuoka 812, Japan

(Received September 6, 1984)

EOUCHI H., KITANO M. and MATSUI T. Direct digital control of air humidity
for plant research. BIOTRONICS 13, 29-38, 1984. Direct digital control
(DDC) system was employed in a growth chamber for accurate control of air
humidity in wider regions. Control elements were composed of cooling brine
coil, dehumidifying brine coil, reheating coil (electric heater), heating coil
(electric heater) and steam boiler: Volumes of circulated brines, electric currents
in heaters and steam volume were manipulated by PID control action (velocity
algorithm). Furthermore, capacities of cooling and dehumidifying coils were
optimalized for high accuracy by the feed-forward control of brine temperatures,
where the optimum temperatures of respective brines were set on the basis of the
desired values of air temperature and humidity. This system made it possible
to control relative humidity from 10 or 20 to 85 % within 3 %RH in an air
temperature control region of 0 to 45 0.3C (the lowest humidity was lO%RH
at air temperatures higher than 10C, but was 20 %RH at OC): This control
region covered saturation deficits of 0.73 to 58.77 g m- 3

INTRODUCTION

Plants grow under various air conditions, and air temperature and humidity
are basal factors in environment control for plant research. Air temperature con-
trol has been mostly generalized, and many kinds of systems have been applied to
various environmental studies of plants (1, 2). On the other hand, it is necessary to
attach more importance to humidity control for developing more reasonable system
of environment control, as humidity is responsible for physiological functions and
plant growth (3-5, 7, 8, 14, 15). Present paper deals with computer control of air
humidity with high accuracy in wider regions.

CONTROL SYSTEM

System diagram is illustrated in Fig. 1. Control elements consisted of cooling

coil, dehumidifying coil, reheating coil, heating coil, and humidifying steam boiler.
A computer system was used for direct digital control (DDC) (6, 13), which was
composed of central processing unit (CPU; TOSDIC-245, TOSHIBA CORPORA-
TION), interfaces and CRT console. Manipulated variable was generated by PID
control action (10, 11): Theoretical PID algorithm of the position form is

29
 w
o

I C~LE ~~~ I
I
I ~NTERFA~ I
::t:
r r-----===-IJ
I -- ----- 1.lf.1
I I I I f I 1= I I , + t tr1
II II r - - - - - - - J I II II II II I, II II II L
I I L __
I I lI l - - -L- - - -==-----
L_-=------=----=-,
I r-.=--J 0
c::
~
I I I Ir r - J- II --=--- I I (j
I -J - I l - - - 1 - -, 1 I ::t:
I ..b I -, I I I I I I ~-

~! I L.J II
I
I I 1 I
s::

I .; hJ
,I
11I
111
I II 1 I
-~
.~..,
0
r---- II M I I I ~
~
CH I

~
0..
, -=-_-...:_.J J I
C'J-J.i
29 I ' - I ""-.,I ~
~
I CR
...,
)-
VJ.
-I
MC 1
c::......
_I
SJ :

Fig. 1. System diagram of air temperature and humidity control: CR, controlled environment room;
MC, mixing chamber; qc, cooling coil; DH/C, dehumidifying coil; E/H-l, reheating coil; E/H-2, heating
o~ coil; SJ, steam jet; F, fan; CH, chiller unit; BR, BRl and BR2, brine tanks; EB, electric steam boiler; M, motor
~ valve; P, pump; TH, air temperature and humidity sensors; Tl, T2, BT, BTl, BT2 and ST, temperature sensors;
o TC, temperature controller; SCR, silicon controlled rectifier.
~
t:1
 AIR HUMIDITY CONTROL 31

Yt=Kp(Zt+ A~Ztdt+TD ~t) (1)

where: Y t , manipulated variable; K p , proportional gain; T I , reset time; T D, rate
time; Zt, controlled deviation (Zt =sv-pv; sv, desired value; pv, controlled variable).
Eq. (1) can be rewritten as following numerical equation,

Yt=Kp( Zt+ ::-itoZt-iAt+TD~LI~t-dt) (2)

where LIt is sampling interval.
In this system, the velocity algorithm of PID control action was employed and
manipulated variable was given by
Y t = Yt-dt+LI Y t
LIt TD }
= Yt-Jt+Kp{(Zt-Zt-Jt)+ -T;"-Zt+ LIt (Zt- 2Zt-dt+ Z t-2Jt) (3)
where Kp>O in heating, humidifying and reheating, and Kp<O in cooling and
dehumidifying.
For rapid compensation of overshooting, manipulation in Eq. (3) was treated
as Yt=O when Zt<Zt-dt<O in heating, humidifying and reheating, and when
o<Zt~dt <Zt in cooling and dehumidifying.
List of K p , T1 and T D were filed in CPU, and these parameters were selected for
optimum settings (11, 12) corresponding to desired values of air temperature and
humidity. The desired values and the other system parameters were set through
CRT console. Air temperature (PVA) and humidity (PVH) were detected in the
mixing chamber by sensors (TH) of Ptl 000 and electric capacitance meter (HMP
15, Vaisala, Oy) (9), respectively. The controlled variables and other signals
detected by TI, T2, BTI, BT2 and ST were transmitted to CPU and used for computa-
tion and monitoring of the system in CRT displays. Figure 2 shows an example of
the CRT display. Yt calculated in Eq. (3) was used for manipulation of the final
control element; respective capacities of cooling, dehumidifying, reheating, heating
and humidifying were manipulated by each Y t at an interval of LIt.

MANIPULATION

The growth chamber was installed in the room where air temperature and
relative humidity were controlled at about 23C and about 60 %. A volume of the
controlled environment room of the growth chamber was 1.5 m (width) X 1.3 m
(depth) X 1.5 m (height), in which controlled air flowed out of left side to right side
with a velocity of about 0.3 m sec- I (circulated air volume, Q was 2000 m3 hr-I);
5.5% of Q was the fresh air introduced from the room (23C and 60%RH) into the
growth chamber. Control regions of air temperature and humidity are shown in
Fig. 3. Air temperature was controlled within a region of -10 to 45C, and
humidity control region distributed from 10 or 20 to 85 % relative humidity (RH)
in air temperature region of 0 to 45C, which covered saturation deficit of 0.73 to
58.77 g m-3 For air control in such wider regions, the control elements were
provided with large capacities. In general, there is the delay in the feedback

VOL. 13 (1984)
32 H. EGUCHI, M. KITANO and T. MATSUI

Fig. 2. CRT display of dew point step control from 35.5 to 1.0 C: PV, con- D

trolled variable CC); SV, desired value CC); MV, manipulated variable of dehumidify-
ing (manipulating percentage of motor valve for brine circulation).

100

80

'<R
>, 60
~
'E
::J
.<::
(J)
> 40
:;::;
'"
lii
0:::

20

0
~1O 0 10 20 30 40 50
Air temperature (Oc)

Fig. 3. Control regions of air temperature and humidity.

control system, and large capacities of the control elements result in decrease in
control accuracy as affected by the overshoot of manipulation, which is caused by
the delay. So, each element in this system was accommodated to the optimalizing
control for high accuracy. On the other hand, in the case of air control at lower

BIOTRONICS
 AIR HUMIDITY CONTROL 33

temperatures or lower dew points, frost forms on the surface of the cooling coil or
the dehumidifying coil and results in decrease in the capacities. For such a case, two
sets of the coils were equipped in two air passes; when one of the coils was frosted,
the other was operated, and the air pass was exchanged into another by driving a
damper with an interval of defrosting.

Preparation of cool brine

A chiller unit (CH) was used for cooling the source brine (aqueous solution of
ethylene glycol) in a brine tank (BR); the lowest temperature of the brine was
- 25C. The source brine was divided into two tanks of BRI for cooling and BR2
for dehumidifying. Respective brine temperatures in BRI and BR2 were controlled
at the desired values for optimalizing capacities of the coil, which were calculated
in feed-forward control system on the basis of desired values of air temperature and
humidity; the brine temperatures were det~cted by BTI and BT2, and volumes of
brine circulated from BR to BRI and to BR2 were manipulated through motor
valves by respective computer signals. Thus, cool brines were prepared at respec-
tive optimum temperatures for cooling and dehumidifying.

Cooling
The brine coil was equipped for cooling, which consisted of 12 (row) X 9
(column) X 600 mm (length) of copper tubes with aluminium fins (fin pitch of 5 mm).
The cooling capacity varies with brine temperature and volume of brine circulated
into the coil. For optimalizing the cooling capacity, the brine temperature was
set 10C lower than the desired value of air temperature in the feed-forward control
system, and volume of the circulated brine was manipulated through a motor valve
by PID control action of Eq. (3). This cooling coil was provided with the maximum
capacity of 10000 kcal hr- I for sensible heat loads brought by introduced fresh air,
heat conduction through the wall, artificial light radiation and humidification stated
below. The maximum capacity was obtained in the case of brine temperature of
-25C and the circulated brine volume of 4080 l hr- I (maximum volume).

Dehumidifying
Dehumidifying load reached a peak at air temperature of 10C and relative
humidity of 10%. At this point of the air condition, specific volume of air (v) is
0.803 m3 kg-I, and humidity ratio (xo) is 0.00076 kg kg-I. Water vapour (L) gained
in controlled environment room was estimated by summation of L r brought by
introduced fresh air and Lp caused by evapo-transpiration in plants in controlled
environment room. Lr was given by

L r = ~r(xr-xo)' .1.3 kg hr- I (4)

v
where Qr (volume of introduced fresh air) is 110 m3 hr- I (5.5 % of Q), v' (specific
volume of fresh air of 23C and 60%RH) is 0.854 m3 kg-I, and Xr (humidity ratio
of fresh air of 23C and 60%RH) is 0.0106 kg kg-I.
Lp was evaluated as 0.3 kg hr-I which corresponded to enough volume of

VO. 13 (1984)
34 H. EGUCHI, M. KITANO and T. MATSUI

evapo-transpiration in 10 young cucumber plants (8). Thus, L was 1.6 kg hr-I.

Therefore, humidity ratio (Xl) of entering air (air entering to the dehumidifying coil)
was given as
L
Xl =XO+ v=0.00140 kg kg-l (5)
Q
So, enthalpy (h) of the entering air was obtained as h =3.2 kcal kg-l from psy-
chrometric chart.
For removing L, it is necessary to reduce the temperature of the leaving air (air
coming out of the dehumidifying coil) to the dew point for Xo by dropping the
temperature of the coil surface to the apparatus dew point. At this point, enthalpy
(i2) of -3.9 kca1 kg- l was obtained from psychrometric chart. Thus, peak heat load
(HL ) of dehumidifying coil was given as

(6)

For dehumidifying, the brine coil was used, which consisted of 14 (row) X 9
(column) x 600 mm (length) of copper tubes with aluminium fins (fin pitch of 5 mm).
Difference (LlTB ) in brine temperature between entrance and exit in this coil was
4.5C at its maximum in the case of brine temperature of -25C (at the entrance)
and circulated brine volume (QB) of 5220/ hr- l (maximum volume). Therefore,
the maximum heat (H) transferred to the brine in this coil was given as
(7)

where p (density of brine) is 1.10 kg /-1, and c (specific heat of brine) is 0.70 kcal kg- l
C-I.
Thus, from Eq. (7), it could be estimated that the capacity of this coil can satisfy
H L For optimalizing humidifying coil, the brine temperature was set about 5C
lower than the dew point at desired air temperature and humidity in the feed-forward
control system; the dew point was calculated from the desired values of air temper-
ature and humidity. Volume of the circulated brine was manipulated through a
motor valve by PID control action of Eq. (3).

Reheating
In order to compensate air temperature drop caused by the dehumidifying coil,
the reheating coil was equipped; the electric heaters were sheathed in stainless steel
tubes with stainless steel fins (fin pitch of 4 mm), consisting of 3 (row) X 6 (column) X
750 mm (length). This coil was provided with the capacity corresponding to the
heat load of Eq. (6), and electric current was manipulated through silicon controlled
rectifier (SCR); the manipulated variable of this reheating coil was calculated from
Eq. (3) where air temperature detected by TI was used as the desired value, and air
temperature detected by T2 was used as controlled variable (feedback signal);
difference in temperature between air entering to the dehumidifying coil and air
coming out of reheating coil was vanished by manipulating the reheating capacity.

BIOTRONICS
 AIR HUMIDITY CONTROL 35

Thus, this reheating coil compensated the disturbance caused by dehumidifying for
air temperature control.

Heating
Heating coil consisted of I (row) X 12 (column) X 750 mm (length) of stainless
steel tubes with stainless steel fins (fin pitch of 4 mm), in which electric heaters were
sheathed. Electric current of this coil was manipulated through SCR by PID
control action of Eq. (3). This heater was provided with the maximum capacity of
8600 kcal hr-I.

Humidifying
Steam was used for humidifying which was generated by an electric steam boiler
(10). Humidifying load reached a peak at desired values of 45C and 85 % RH in
the case of introduced fresh air (23C and 60%RH) of 110 m3 hr-I (Qr=5.5 %
of Q). For this peak humidifying load, water vapour (L') required was given by

L' = Q;
v
(x2-xr)=5.85 kg hr- I (8)

where X2 (humidity ratio of air of 45C and 85 %RH) is 0.056 kg kg-I, and v' and
Xr are defined in Eq. (4).
In practical process, too much water vapour is removed by excessive dehumidi-
fication which is accompanied by cooling in the cooling coil and caused by overshoot
of manipulation and residual cooling in the dehumidifying coil. For these extra
loads, the safety factor of 2.0 was applied to this system, and the boiler was provided
with maximum capacity (water vapour) of 12 kg hr-I. Steam temperature in the
boiler was set within a region of 105 to 120C, and steam was supplied from jet
nozzles in the mixing chamber; volume of steam supply was manipulated through
a motor valve by PID control action of Eq. (3).

PERFORMANCE

Figure 4 shows controlled variables of air temperature and relative humidity

at the lowest and the highest saturation deficits. The lowest saturation deficits (A)
of 0.73 g m-3 was obtained in the control region at air temperature of OC and rela-
tive humidity of 85 %. Controlled air temperature and relative humidity fluctuated
to some extent, but were maintained within 0.3C and 3 %RH~ At the highest
saturation deficit (B) of 58.77 g m-3 (45C and 10 %RH), reliable characteristics
were found in the controlled variables of air temperature and relative humidity.
Figures 5 and 6 show humidity step responses at air temperatures of 10C and
40C, respectively. In the regions at lower and higher humidities, small overshoot
and delay were found, but the controlled variables settled within 10 min. The air
temperature control was not disturbed by the humidity changes, and the accuracy
was kept within 0.3C.
These static and dynamic characteristics indicate that this DDC system makes
it possible to control air temperature and humidity in wider regions with high

VD. 13 (1984)
36 H. EGUCHI, M. KITANO and T. MATSUI

50 100
A B

~Air
90
temperature
80
Relative humidity
70
E 30
60 ~
e
.a >.
t5
~
Q)
a.
20 50 'E
:::l
E ..c
2
... 40 ~
Q)

< ro
Q)
10 0:::
30
Air temperature
/Relative humidity 20
Ot--------~-------~
10
10 min 10 min
I----l /-.j
-10L..- --l
0
Time Time
Fig. 4. Static characteristics of air temperature and humidity controls at the
lowest saturation deficit (A) of 0.73 g m- 3 (OC and 85 %RH), and at the highest
saturation deficit (B) of 58.77 g m- 3 (45C and lO%RH).

50 100

90

40 80

70
---~
30 60
e b
.3 "6
~ 'E
Q)
a. 50 :::l
..c
E Q)
2 Relative humidity >
:p
20 ro
40 Q)
0:::

Air temperature 30

20

I 1 hr I 10
Ol-- ..J
0
Time
Fig. 5. Dynamic characteristics of humidity control at air temperature of 100e.

BIOTRONICS
 AIR HUMIDITY CONTROL 37

50 100

90

40 80

\ Air temperature 70

~
l: 30 60
~ Z'
.3 '5
~ 'E
(I) 50 :J
0- ..c:
E (I)
~ >
... 20 Relative humidity 40 ~
ro
ID
cc
30

20

1 hr 10
I- I
0 0
Time
Fig. 6. Dynamic characteristics of humidity control at air temperature of
40C.

accuracy and can help to analyse various plant responses to air humidity at different
air temperatures.

REFERENCES

1. Downs R. J. and Hellmers H. (1975) Temperature. Pages 7-30 in Environment and the Experi-
mental Control ofPlant Growth. Academic Press, New York.
2. Downs R. J. (1980) Phytotrons. Bot. Rev. 46, 447-489.
3. Drake B. G., Raschke K. and Salisbury F. B. (1970) Temperatures and transpiration resis-
tances of Xanthium leaves as affected by air temperature, humidity, and wind speed. Plant
Physiol. 46, 324-330.
4. Eguchi H. and Matsui T. (1974) Feedback control of leaf temperature. H. Transient
characteristics. Environ. Control in BioI. 12, 5-11.
5. Ford M. A. and Thorne G. N. (1974) Effects of atmospheric humidity on plant growth. Ann.
Bot. 38, 441~452.
6. Funada S., Hashimoto Y., Omasa K. and Toyozumi J. (1973) Fundamental studies on
biological process conltp,l by means of d~gital comp~ter. (H) D~sign of process controller.
Environ. Control in Rip!. 11" 147-lS~ (in Japanese with English~'umlll'lfY).
7. Hoffman G. J. (1979) ''\Humidity. Pages 141-172 in T. W. Tibbitt1 and T. T. Kozlowski (eds)
Controlled Environment Guidelines for Plant Research.. Academic Press, New York.
8. Kitano M., Eguchi H. and Matsui T. (1983) Analysis of heat balance of leaf with reference
to stomatal responses to environmental factors. Biotronics 12, 19-27.
9. Kitano M., Eguchi H. and Matsui T. (1984) Analysis of static and dynamic characteristics of
humidity sensors. Biotronics 13, 11-28.
10. Matsui T., Eguchi H., Hanami Y., Handa S. and Terajima T. (1971) A growth cabinet for
the study on biotronics. I. Design and performance. Environ. Control in BioI. 9, 37-46.

VDL. 13 (1984)
38 H. EGUCHI, M. KITANO and T. MATSUI

11. Matsui T., Eguchi H. and Toh K. (1973) Study on optimum PID controller settings for
temperature control in a growth cabinet. Environ. Control in Bioi. 11, 31-35.
12. Pessen D. W. (1953) Optimum three-mode controller settings for automatic start-up. Trans.
ASME. 75,843-849.
13. Takakura T., Kozai T., Tachibana K. and Jordan K. A. (1974) Direct digital control of plant
growth. I. Design and operation of the system. Trans. ASAE. 17, 1150-1156.
14. Tibbitts T. W. and Bottenberg G. (1976) Growth of lettuce under controlled humidity levels.
J. Am. Soc. Hortic. Sci. 101, 70-73.
15. Woodward R. G. and Begg J. E. (1976) The effect of atmospheric humidity on the yield and
quality of soya bean. Aust. J. Agric. Res. 27, 501-508.

BIOTRONICS