Published in
ASHRAE Transactions, Volume 117, Part 1. For personal use only. Additional reproduction, distribution, or transmission in
either print or digital form is not permitted without ASHRAE'S prior written permission.
LV-11-027
Song Deng, PE
Member ASHRAE
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
ABSTRACT
INTRODUCTION
Zhiqin Zhang is a PhD student in the Department of Mechanical Engineering and a graduate research assistant in the Energy Systems Laboratory, Hui Li is a post-doctorate and Song Deng is an associate director in the Energy Systems Laboratory, and William D. Turner is a professor in the Department of Mechanical Engineering, Texas A&M University, College Station, TX.
934
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
CW loop of centralized heating, ventilating, and air-conditioning (HVAC) systems. A modified generic algorithm for this
particular problem was proposed to obtain the optimal
setpoints of the process. Simulations and experimental results
on a centralized HVAC pilot plant showed that the operating
cost of the CW loop could be substantially reduced compared
with conventional operation strategies. Yet they did not
explain what setpoint should be used for the CWLT to control
the fan speed for optimizing the system. Yu and Chan (2008)
presented the use of load-based speed control to enhance the
energy performance of water-cooled chiller systems. The optimal cooling tower CWLT and CWP speed were expressed as
a function of ambient WB temperature and chiller part load
ratio (PLR). The system performance under the optimal
control could increase by 1.4%16.1% relative to the equivalent system with fixed temperature and flow rate controls for
the cooling water leaving from cooling towers.
This paper presents the optimization of the cooling tower
CWLT by coupling an effectiveness-NTU cooling tower
model and a Gordon-Ng model for vapor-compression chillers
with variable-condenser flow. An example chiller system is
used and the total power of the chiller compressor, condenserwater pump, and fan motors are minimized by selecting an
optimal cooling tower approach setpoint. The factors that
could affect the form of the optimal reset schedule are
discussed, such as CW flow rate, chiller PLR, chiller chilledwater (ChW) leaving temperature, chiller and tower characteristics, and climate zones.
CHILLER-COOLING TOWER SYSTEM
System Configuration
Figure 1 shows an example condenser-water loop considered to characterize and compare annual electricity energy and
cost savings when applying various tower CWLT control
methods. It consists of one constant-speed water-cooled
centrifugal chiller, one constant-speed condenser-water
pump, and two VSD-equipped induced draft-type cooling
towers. The two towers are staged on at the same time, and the
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
2011 ASHRAE
935
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
(1)
(3)
(4)
(5)
V CT, CW = V CHLR, CW 2
(6)
Upper and lower limits are defined for the cooling tower
CWLT according to the system chiller requirements:
T CHLR, CW, E, min T CHLR, CW, E T CHLR, CW, E, max (7)
If the tower airflow rate found by the solver corresponding to the optimal CWLT is higher than the maximum airflow
rate of the fan, the following equation is used to find the cooling tower CWLT at the maximum airflow rate:
T CT, CW, L =
F 2 ( V CT, CW, T DB, T WB, T CT, CW, E, V fan, max )
( c0 + c1 x1 + c2 x2 + 1 )
5
(T
-------------------------------------------------------------------------------------------------P
= 3.517Q
CHLR
ChW
3.517c Q
3 ChW
5
-(
T
32
)
+
+
273
CHLR, ChW, L
9
3.517Q
ChW
------------------------------------------------
236.34V
c
CW w pw
(10)
The trended historical data for the chiller are used to identify the coefficients of the model with the ordinary leastsquares linear regression method. Figure 2 is a comparison
between measured and predicted motor power using the
Gordon-Ng model. Statistical analysis shows that the rootmean-square error (RSME) of the predictions is 102.5 kW and
the coefficient of variation (CV) is 2.96%. Table 1 shows the
rated chiller parameters as well as the chiller model coefficients.
Cooling Tower Modeling
The mass and heat transfer process in a cooling tower is
fairly complicated. The effectiveness-NTU model is the most
popular model in CT simulations, but iterations are required to
obtain a converged solution (Braun 1989). The details of the
model can be found in the original paper. Particularly, the
overall number of transfer units (NTU) can be correlated with
the following form:
m w 1 + n
-
NTU = c ----- m a
(8)
(11)
m w, i ( T w, i T ref )c pw m w, o ( T w, o T ref )c pw
= 60 ------------------------------------------------------------------------------------------------------------------------------ (12)
a
a ( h a, o h a, i )
936
(9)
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
Table 1.
Variable Name
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
w, o
SI
Symbol
Value
Unit
Value
Unit
Capacity
QCHLR,rate
5500
ton
15,474
kW
TCHLR,ChW,L
36.0
2.2
CW flow rate
VCHLR,CW
10,000
gpm
2271
m3/h
Chiller coefficient 0
c0
0.28100
0.28100
Chiller coefficient 1
c1
10.20000
10.20000
Chiller coefficient 2
c2
1740.00000
1740.00000
Chiller coefficient 3
c3
0.00271
0.00271
Vfan,rate
650,000
cfm
306.8
m3/s
Pfan,rate
150
hp
111.9
kW
CT coefficient 1
3.6152
3.6152
CT coefficient 2
0.6667
0.6667
Reference temperature
Tref
32.0
0.0
PCWP,rate
400
hp
298.3
kW
motor
0.98
0.98
shaft
0.97
0.97
Pump head
HCWP
Head curve
ft
Head curve
Pump efficiency
cwp
Efficiency curve
Efficiency curve
= T
m w, i ( T w, i T ref )c pw a V a ( h a, o h a, i )
- (13)
+ ---------------------------------------------------------------------------------------------------------------ref
m w, o c pw
(14)
Figure 2 Comparison of chiller measured and predicted
motor powers.
Va
PLR = ---------------------V fan, rate
(15)
2011 ASHRAE
937
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
Pump Modeling
Weather Conditions
Weather conditions play an important effect on the operation of a water-cool chiller plant. Six cities representing six
typical climate zones in the United States are selected for
simulation: Houston, TX (hot and humid), Phoenix, AZ (hot
and dry), Chicago, IL (cool and humid), Denver, CO (cool and
dry), Los Angles, CA (warm and dry), and Miami, FL (hot and
(a)
(b)
0.746V H
P pump = ---------------------------3960 all
(16)
Figure 3 Comparison of CT condenser water measured and predicted leaving temperatures; (a) I-P, (b) SI.
Table 2.
Dry Bulb,
F (C)
17
(8)
24
(4)
21 (6)
30
(1)
36
(2)
43
(6)
49
(9)
55
(13)
62
(17)
68
(20)
74
(23)
81
(27)
29 (2)
26
66
37 (3)
118
227
25
147
395
78
18
158
503
272
19
202
432
681
18
42
152
380
915
239
48
111
467
1476
73
13
154
644
228
95 (35)
29
258
108
103 (39)
14
46 (8)
54 (12)
62 (17)
70 (21)
78 (26)
87 (31)
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
938
Copyright American Society of Heating, Refrigerating and Air-Conditioning Engine
Provided by IHS under license with ASHRAE
No reproduction or networking permitted without license from IHS
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
OPTIMIZATION RESULTS
T App = T wb + 55 if T wb 47F
T App = 0.1325T wb + 13.56 if T wb > 47F
(a)
(17)
(b)
Figure 4 Optimal CT approach temperature versus ambient WB temperature; (a) I-P, (b) SI.
(a)
(b)
Figure 5 Cooling tower approach temperatures under various fan control strategies; (a) I-P, (b) SI.
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
2011 ASHRAE
939
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
Baseline
Energy
savings
percentage
CHLR Power,
kWh
CT Fan Power,
kWh
CW Pump Power,
kWh
Total Power,
kWh
24,611,417
1,325,921
2,245,375
28,182,713
TApp,sp = Optimal
5.8%
19.7%
0.0%
4.1%
TApp,sp = Near-optimal
5.7%
18.6%
0.0%
4.1%
3.7%
12.8%
0.0%
3.8%
TApp,sp = 4F (2.2C)
5.8%
20.8%
0.0%
4.1%
TApp,sp = 6F (3.3C)
4.1%
1.3%
0.0%
3.6%
TApp,sp = 8F (4.4C)
2.0%
18.3%
0.0%
2.6%
(a)
(b)
Figure 6 Optimal CT approach temperature under different CW flow rates; (a) I-P, (b) SI.
940
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
Table 3.
(a)
(b)
Figure 7 Optimal CT approach temperature under different chiller loads; (a) I-P, (b) SI.
(a)
(b)
Figure 8 Optimal CT approach temperature under different chiller ChW leaving temperatures; (a) I-P, (b) SI.
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
2011 ASHRAE
941
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
(a)
(b)
Figure 9 Optimal CT approach temperature under different CT coefficients; (a) I-P, (b) SI.
(a)
(b)
Figure 10 Optimal CT approach temperature in different climate zones; (a) I-P, (b) SI.
942
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
The scatter plots of the optimal CT approach temperatures at six typical climate zones are shown in Figure 10. The
chiller is loaded at 80% PLR, the chilled-water leaving
temperature is 36F (2.2C), and the condenser-water flow
rate is 10,000 gpm (2271 m3/h). Except for the plots for
Denver, these plots are overlapped with each other and a
significant correlation can be observed between the optimal
cooling tower CWLT setpoint and the ambient WB temperature. The atmospheric pressure in Denver is 0.824 atm and it
is cool and dry. It is easier for water to evaporate. For the same
airflow rate or fan power, a lower tower CWLT can be
achieved.
SUMMARY AND CONCLUSION
Resetting the cooling tower CWLT is one of the most
popular measures to improve the performance of a chiller
plant. It plays opposite effects on the efficiencies of the chiller
and cooling tower. An optimal value exists for specific operating conditions to minimize the power consumption of the
chiller, cooling tower, and CW pump.
This paper introduces the optimization of the cooling
tower CWLT using a component-based model. The model is
applied in an example chiller CW system and the cooling
tower approach temperature setpoint is optimized to minimize
the total power of the chiller, pump, and fans at given weather
conditions, chiller load, chilled-water leaving temperature,
and condenser-water flow rate. Simulation results show that
the optimal cooling tower approach setpoint reset schedule
can be approximated with two straight lines. Significant
energy savings could be achieved if compared with the
scenario with a constant cooling tower CWLT. Further simulations show that chiller PLR, chiller ChW leaving temperature, and climate zones the plant locates in play minor effects
on the coefficients of the optimal CWLT reset schedule. A
higher condenser-water flow rate per cooling ton leads to a
higher optimal cooling tower approach, but this effect can be
neglected for a system with a constant CW flow rate. The form
of this reset schedule is determined by the performance characteristics of the chiller and cooling tower.
gpm
GRG
h
H
HVAC
m
n
NTU
P
PLR
PPMP
Q
RSME
T
V
VSD
WB
x
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
Greek Symbols
= efficiency
= density, lbm/ft3 (kg/m3)
Subscripts
a
App
E
i
L
max
min
o
ref
sp
tot
w
wb
=
=
=
=
=
=
=
=
=
=
=
=
=
air
approach
entering
inlet
leaving
maximum
minimum
outlet
reference
setpoint
total
water
wet bulb
NOMENCLATURE
REFERENCES
c
cp
ChW
CT
CV
CW
CWP
DB
=
=
=
=
=
=
=
=
2011 ASHRAE
943
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST
--`,,,`,,``````,``,````,``,,,`,`-`-`,,`,,`,`,,`---
Graves, R.D. 2003. Thermodynamic modelling and optimization of a screw compressor chiller and cooling tower
system. M.S. thesis, Department of Mechanical Engineering, Texas A&M University.
Jiang, W., and T.A. Reddy. 2003. Re-evaluation of the Gordon-Ng performance models for water-cooled chillers.
ASHRAE Transactions 109(2):27287.
Kreider, J.F., P.S. Curtiss, and A. Rabl. 2002. Heating and
Cooling of Buildings: Design for Efficiency, 2d ed. New
York: McGraw-Hill.
Lu, L., W. Cai, Y.C. Soh, L. Xie, and S. Li. 2004. HVAC system optimizationCondenser water loop. Energy Conversion and Management 45(4):61330.
NREL. 2008. National Solar Radiation Data Base 19912005 Update: Typical Meteorological Year 3. From
http://rredc.nrel.gov/solar/old_data/nsrdb/1991-2005/
tmy3/.
Schwedler, M. 1998. Take it to the limit ... or just halfway?
ASHRAE Journal 40(7):3239.
Stout, M.R. 2003. Cooling tower fan control for energy efficiency. M.S. thesis, Department of Mechanical & Aerospace Engineering, North Carolina State University.
Yu, F.W., and K.T. Chan. 2008. Optimization of watercooled chiller system with load-based speed control.
Applied Energy 85(10):93150.
944
ASHRAE Transactions
Licensee=Istanbul Teknik Universtesi/5956919001
Not for Resale, 12/26/2014 01:00:35 MST