Dynamics of Primary and Secondary Chilled Water Systems
Dynamics of Primary and Secondary Chilled Water Systems
Dynamics of Primary and Secondary Chilled Water Systems
Abstract
he purpose of this paper is to bring some additional understanding to the dynamics of variable flow chilled
water systems. The literature has had several articles in recent months on this subject with some
disagreement regarding a check valve in the bypass line. Reference (1) gave an example of a very positive
result after the installation of a check valve. This paper will show a similar result and try to illustrate the reasons
for the improvement. The Reference (1) article also presents information on load degradation as a chilled water
system ages and, therefore, requires colder supply water and greater flow. This paper will present a study of the
dynamics of a new and aged primary/secondary system based on a computer simulation which includes the
characteristics of a specific chiller and the laws of thermodynamics.
We will analyze the effects of restricting flow in the bypass line and also look at the dynamics of adding thermal
storage to the system.
BALANCED SYSTEM
We define the balanced system to be essentially zero flow in the bypass line at peak cooling load. We will assume
the load is designed for a delta temperature of 16F and a peak design load of 15,150 tons. The primary pumps
provide constant flow through the chiller with variable flow to the load. Our selected chiller is a 2,165-ton unit at
peak conditions. Therefore, seven chillers will meet the peak load. Figure 1 illustrates the temperatures and flows
at this near-balanced condition. Note that there is a small negative flow in the bypass line of 305 gpm due to the
difficulty of exactly matching a chiller system to the load. Table 1, Column 1 shows the same data for this
balanced condition. (See References 2 and 3 for additional discussion.)
The chiller is modeled as providing about 2,165 tons for a chiller entering temperature of 60F and
1,700 tons for an inlet temperature of about 48.5F.
Nomenclature
Load in Tons
Delta T
Fce
Tls
Tce
Fls
Tcl
Tlr
Tpl
Fb
Tb
BYPASS VALVE
A valve to stop the negative flow in the bypass line (Fb = 0) will improve the efficiency of the chillers as illustrated
by Table 1, Column 5. Forcing all flow through the chillers increases the temperature entering the chillers from
50.08F to 52.62F which results in increased chiller capacity; and therefore, chiller loading is decreased from 96%
(Column 4) to 92% (Column 5). For this system at these conditions, stopping negative flow in the bypass improved
plant performance, i.e. the chillers are unloaded about 4%. References 1, 4, 5, 6 and 7 are recommended reading
on this subject.
EXCESS FLOW
A condition to consider is that plant operators may have a tendency to pump more flow to the load than required.
If this condition exists, the result is a decrease in load delta temperature and decreased negative flow or positive
flow in the bypass line. Column 6 of Table 1 assumes about 35% over pumping which results in 3,445 gpm of
positive flow in the bypass line and a T of 11.0F. The nine chillers are 99% loaded and must supply 36.5F water
to mix with the 48.78F bypass water to provide the required 37.78F supply to the load as shown by Table 1,
Column 6. Note that for this condition, if the plant operators increased flow a little more, the load could not be met
and reducing flow would reduce the load on the chillers and improve plant performance. A shut-off valve in the
bypass line will force all excess flow to the chillers and improve performance just as it did in Column 5 of Table 1
for negative flow in the bypass line. We must, of course, be sure that the flow to the chillers does not exceed the
manufacturers requirement, nor reduce the flow below the chiller manufacturers requirement.
OPTIMUM OPERATION
We suggest that the procedure for defining optimum operation of an existing system consists of first defining the
dynamic characteristics of the plant equipment, and then by observing the dynamic characteristics of the real
system, incorporate these characteristics into a system computer model. This procedure will define the
characteristics of the load which will typically be the source of making major improvements to the efficient
operation of the system. As changes are made on the load side, the computer simulation would be updated; and
therefore always provide the plant operators direction on how to optimally control the plant.
Simply stated, we are suggesting that without a computer simulation of the existing system that is updated as the
system changes, optimum operation of the plant is not achievable for reasons including the tendency of plant
operations to pump excess flow and to turn on more chillers than required.
For the purpose of this paper, lets assume that the plant operation as defined by Column 8, could be returned to
the conditions of Column 4, wherein the flow and supply temperature to the load are optimized to the load we
have assumed without a shut-off valve in the bypass line. (See References 8, 9, 10, 11, and 12.)
FIGURE 2
Figure 3 illustrates the 24-hour operation of the system. The seven chillers can meet the cooling load for all but
four hours; therefore, ice melting occurs during four hours for a total melt of 7,144 tons. The 306-ton ice slurry
generator makes ice all 24 hours, with storage peaking the 14th hour at about 6,200 ton-hours. The storage tank is
essentially depleted the 18th hour and begins to build up on the 19th hour. (See References 18, 19, 20, and 21.)
LOAD DYNAMICS
This paper illustrates the positive effects offered by thermal storage, especially in a system that has experienced
decreased delta temperature. However, it must be remembered that fixing the load is also an alternative. Reference
3 presents some basic concepts and Reference 22 is a case history of a successful effort to fix the load along with
decreasing the supply temperature to the load. Also, References 23, 24 and 25 present an alternative approach to
thermal storage.
FIGURE 3
Figure 4
FIGURE 5
FIGURE 6
FIGURE 7
The following table illustrates the 13 conditions we have considered. The author suggests making copies of Figures
1 and 2 and writing in the values of Table 1 for each condition for ease of studying the dynamics of this assumed
system.
Table 1
Load
Balanced
Balanced
+ One
15,150
15,150
Load
Load
Degradation Degradation
-8 Chillers -9 Chillers
14,000
15,150
Shut-Off
Valve
15,150
Excess
Excess
Flow with
Flow
Shut-Off w/Shut-Off
Valve
Valve
Excess
Flow
15,150
15,150
15,150
Thermal
Storage
Thermal
Thermal
Thermal
Thermal
Aged
Storage
Storage
Storage
Storage
System
New
-6 Chillers -5 Chillers
Aged
w/Shut-Off
System
Aged
Aged
System
Valve
6 Chillers
System
System
15,150
15,150
15,150
15,150
15,150
Delta T
16.0
16.0
14.8
14.8
14.8
11.0
11.0
11.0
14.8
14.8
16.0
14.8
14.8
Tls
44.03
44.03
37.36
37.80
37.82
37.78
37.07
37.82
37.80
37.80
44.00
37.80
37.80
Tlr
60.03
60.03
52.16
52.60
52.62
48.78
48.07
48.82
52.60
52.60
60.00
52.60
52.60
Fls
22,725
22,725
22,703
24,568
24,568
33,055
33,055
33,055
24,568
24,568
22,725
24,568
24,568
Fb
-305
-3595
-3617
-5042
3445
1538
2985
4828
8118
Tb
44.03
44.03
37.36
37.80
48.78
52.60
60.00
52.60
52.60
Fce
23,030
26,320
26,320
29,610
24,568
29,610
33,055
33,055
23,030
24,568
19,740
19,740
16,450
Tce
59.81
57.84
50.12
50.08
52.62
48.78
48.07
48.82
52.6
52.6
60
52.6
52.6
Tcl
44.03
44.03
37.37
37.8
37.82
36.5
37.07
37.82
39.06
39.77
44.15
39.06
39.06
Tpl
44.03
44.03
37.37
37.8
37.82
37.78
37.07
37.82
39.91
39.77
46.24
41.72
43.53
Chillers On
% Load
100
89
100
96
92
99
99
97
100
100
100
100
100
Peak Ice
Tons
2159
2016
2117
4015
5871
24-hr
Ice Melt
7144
6794
6975
15,543
32,897
Tank
Storage
Ton-Hours
6200
5900
6200
12,400
19,100
Ice Slurry
306
288
306
670
1400
Column
10
11
12
13
SUMMARY
The paper makes three points:
Point One:
A shut-off valve in the bypass line can improve system performance, but the amount of
improvement is a function of the characteristics of the system and how it is being operated.
Point Two:
This paper suggests that a dynamic math model simulation of a system, verified by existing plant
operational data, is the first and necessary step in improving operational performance or
modifying an existing system for increased capacity. (See References 26, 27, 28, 29 and 30.)
Point Three:
We have shown that thermal storage can offer some significant system characteristics for new or
modified systems. For the system of this study, we have shown the following:
10
Table 2
Tons Chilling
Required
Aged System
Chilling Tons
Not Purchased
with Thermal Storage
Required
Ice Generating
Capacity Tons
Required Tank
Thermal Storage
Ton-Hours
System without
Thermal Storage
9 Chillers
19,476
System with
Thermal Storage
and
7 Chillers
6 Chillers
5 Chillers
15,148
12,294
10,820
4328
6492
8656
306
670
1400
.07
.10
.16
6200
12,400
19,100
For the systems studied here, unless the cost of space for a storage tank is very high, the five-chiller system with
thermal storage should have an attractive return on investment and the six-chiller and seven-chiller systems will
have an even more attractive return on investment.
CONCLUSION
Simulating the dynamics of physical systems by applying the laws of physics and thermodynamics is a well
established procedure in the design of military products which the author participated in until the mid-1970s. The
oil embargo in the fourth quarter of 1973 presented the challenge of simulating energy systems, including buildings
and its contents, for the purpose of defining energy waste and systems to eliminate that waste. A fundamental rule
of system simulation is that the simulation must give the same results, within reasonable accuracy, as the real
system.
This paper presents a simulation based on the first law of thermodynamics. Other features provided by a computer
simulation of a central chilled water system include:
The system energy use and KW demand
Chiller dynamics of various type chillers
Pumping dynamics
Load dynamics
The purpose of a computer simulation of a physical system is to understand the dynamics of the system so that
correct decisions about the system can be made. A computer simulation is not required to understand many HVAC
systems; however, the author suggests that a primary/secondary chilled water system is sufficiently complex to
require a simulation for purposes of operating, changing and expanding the system. Reference 31 is an example of
the approach.
References
(1)
(2)
(3)
11
(4)
(5)
(6)
(7)
(8)
(9)
12
13