Energy Efficiency in Chilled Water Systems.

Constant flow systems have limited applications for multiple chillers piped in parallel and
serving multiple cooling loads. When the system operates near full load, performance is
satisfactory as all chillers and pumps are operating. However, constant flow systems have
problems during part-load or off-peak conditions. Consider a constant-flow primary-only chilled
water system (Figure 1) with three chillers fed by two pumps and a part load condition small
enough so that two chillers can handle the load.

Fig No-1. Constant-flow primary-only chilled water system.

By turning off one chiller and allowing the flow to continue through the down
machine, the supply temperature from the plant increases due to the mixing of
water from the chiller that has been off, with the water of the chillers that remain
“on-line”. This creates a step reset of the supply water temperature and at the
same time, pumping energy is wasted unnecessarily in the evaporator of the chiller
that was stopped.

The only way to counter the supply temperature degradation is to drop the temperature of the
water leaving the active chillers, but this complicates the operation and/or controls, making it an
unsatisfactory option. This scenario shows the severe limitations of the constant-flow primary-
only scheme to face real life thermal load fluctuations in an efficiently manner.

On the other hand, a primary/secondary constant flow scheme with each chiller having a
dedicated primary pump (Figure 2), allows shutting down a chiller and its pump without
affecting the secondary flow. It is possible because both circuits are hydraulically independent.

1 |Page
This way the system face the variable thermal load without having to keep on-line all chillers and
pumping energy saving are realized during periods of low loads.

Now consider a similar system with three identical chillers operating in parallel
designed to cool 55 F (12.8 C) chilled water return to 45 F (7.2 C) chilled water
supply. A 50 % drop in load could be faced with two chillers, provided that some
supply temperature degradation is allowed.

Fig. Nº 2. Constant flow primary/secondary chilled water system

It means facing a 50 % cooling load with around 67 % of the total production flow of
the plant. It is of course a more efficient strategy than keeping in service all
production units and theirs pumps at all loads.

However, chiller sequencing in a constant flow system does not change really the
primary loop into a true variable distribution because the flow remains a higher
percent than load. Additionally, although some primary flow variation is allowed, it
can be done only in a stepped way with as many steps as chillers in the plant.
Shutting down a chiller in a primary/secondary constant flow system with multiple
chiller piped in parallel means to break the design condition balance between
production flow (primary loop) and distribution flow (secondary loop). As secondary
circuit flow remains constant, the excess flow will run through the common pipe in
the direction towards the secondary pumps creating a mixing point and further
degradation of the supply temperature.

2 |Page
This reduces the cooling capacity of the coils, especially latent cooling capacity
which could mean a loss of humidity control in the zones. Additionally, the negative
effect of capacity reduction due to higher supply temperature, demands sometimes
the selection of the next higher available size for the coils.

Fig. Nº 3. Constant-flow primary/variable-flow secondary chilled water system.

The current standard design for central chilled water plants with multiple chillers
and multiple cooling loads is the constant-flow primary/variable-flow secondary
scheme (Figure 3). The primary pumps are typically constant volume, low head
pumps intended to provide a constant flow through the evaporator of the chiller.
The most common arrangement in the primary loop is chiller-pump in tandem, so
primary pumps are sequenced with chillers. Chiller capacity is staged in response to
rising leaving chilled-water temperature (T1 in figure 3). A chiller is staged off when
flow in the bypass exceeds the design flow of one chiller.

This can be determined by monitoring the bypass temperature (T2), the return temperature (T3),
and mixed return temperature (T4). Since the chilled-water-flow rate in the primary loop is
known to be a good approximation, the flow rate through the bypass can be calculated using
these temperatures. Alternatively, a flow meter in the bypass line can be used to measure the
excess primary flow directly.

The secondary pumps deliver the chilled water from the common pipe to coils then back to the
common pipe. These pumps are variable-speed pumps controlled from differential pressure
sensors located remotely in the system or from cooling coil valve position, thus secondary flow
tracks the continuously variable thermal load. Pumping energy saved this way provides
significant energy savings relative to constant-flow schemes (primary-only and
primary/secondary). However, primary flow variation remains stepped which avoids achieving
the ideal correspondence between load and flow.

3 |Page
Low delta-T syndrome in chilled water plants
When the secondary circuit return water temperature is lower than design temperature, chillers
can not be loaded at their maximum capacity. If the chillers in a chilled water plant designed to
cool 55 F (12.8 C) chilled water return to 45 F (7.2 C) are receiving their design flow rate at 52 F
(11.1 C) rather than the design temperature of 55 F, the chillers will be loaded at the ratio of:

Where:
CHL (%): Percent chiller loading
CWRTR: Real chilled water return emperature
CWSTD: Design chilled water supply temperature
CWRTD: Design chilled water return temperature

In this case:

The delta-T (the difference between return and supply chilled water temperature) in
the plant has been lowered from 10 F design condition to 7 F, then chillers will be
unloaded a 30 %.

In variable flow systems, it is assumed that delta-T will remain relatively constant at
all loads. Then the flow must vary proportionally with the load. Most variable-flow
systems are designed based on this assumption and fail to perform well if the delta-
T does not stay relatively constant. The fact is in almost every real-world chilled
water plant, delta-T falls well below design levels, particularly at low loads. The low
delta-T “syndrome”, as it is known, causes the plant operators to run extra pumps
and chillers to meet the load, which in addition to reducing the plant’s cooling
output capacity, wastes energy. The system may be keeping the building cool but it
is inefficient and a lot of chiller capacity is being wasted. Table 1 shows some
causes and mitigation measures for low delta-T syndrome.

4 |Page
Table 1 Some causes and measures against low delta-T syndrome:

Cause
Improper setpoint or controls calibration: A
modest drop in supply air temperature of an
AHU can cause coil flow rate to be doubled
and delta-T to drop in half.
Using three-way valves: Three-way valves by
their nature bypass supply chilled water into
the return line, causing chilled water return
temperatures lower than design. This
exacerbates low delta -T problem.
Poor two-way valves selection: An improperly
sized two-way control valve may consume
more water flow when open than the design
calls for. With full flow through the coil, at
partial loads the delta-T will invariably be
lower than design.
Reduced coil effectiveness: Coil transfer
effectiveness is reduced by water side fouling, treatment. Reduce air side fouling by cleaning
air side fouling, air side deterioration, non
uniform air distribution across the cooling coil,
and coil bypass air. Any reduction in coil
effectiveness increases the flow rate of water
required to deliver the desired leaving water
temperature, thus reducing delta-T.
Measure
Controls must be calibrated and setpoints must
be checked regularly. Use of pressure
independent delta-P control valves or automatic
flow control valves.
Do not use three-way valves in variable flow
system. Two-position bypass valves across
supply and return lines, strategically situated and
controlled properly is preferred to ensure
minimum flow.
Select the control valves considering the
pressure drop of the load served and the
available differential across the supply and
return mains. Manufacturers usually recommend
that wide-open control valve pressure drop be
equal to or grater than the pressure drop of the
coil plus the pipe and fittings connecting them to
supply and return mains.
Control water side fouling by proper chemical
coil faces and filters periodically.

Variable primary-flow scheme, potential benefits

Among the variable flow schemes, variable-primary flow (VPF) chilled water systems are of
much current interest. A VPF system consists of single or multiple chillers with a unique set of
pumps that moves the water through the chillers and distribution system to the cooling load. The

5 |Page
cooling output at each coil is controlled with two-way valves. A bypass line with a two-way
control valve diverts chilled water from the supply into the return line to maintain either a
constant or minimum flow through the chiller(s).

The VPF approach has three main potential advantages over the primary/secondary system:
energy and operating costs savings, first cost savings, and better ability to deal with low delta-T
syndrome. Energy savings are possible because no excess flow recirculates from supply to return
through decoupling lines or three-way valves. In theory, every bit of supply water, without any
mixing, must pass through a load before returning to the plant. Energy savings are also possible
when conditions permit flow to one or more chillers to exceed design flow. If outside wet bulb
temperature is below the design value, as it is over 95 percent of the year, the condensing
temperature will be also lower, giving each chiller additional capacity. If more water can be put
through the chiller, this extra capacity can be tapped.

The first cost of a VPF system is likely to be lower than that of a primary/secondary system
simply because two set of pumps are replaced with one.

As regards dealing with low delta-T syndrome, VPF controls permit flow through evaporators to
be increased above design value, making it possible to adjust to less than ideal chilled water
return temperature.

However, VPF approach is not a panacea. Chiller sequencing requires more care in order to
achieve stable operation during simultaneous flow and load changes. Additionally, the low
chiller flow bypass control adds further complexity. In short, the VPF system is both less
expensive and more efficient than a primary/secondary system, provided that control issues
associated with variable primary flow are handled properly.

6 |Page
7 |Page