water pumps and cooling tower fans;
( 2) fully automated network controls;
( 3) new relational control algorithms to
optimize all system components for
energy efficiency; and ( 4) optimization
software. Striving for ultra-high performance using this combination of
technology represents a new paradigm
in district cooling plant optimization.
Peter (the chiller energy) to pay Paul
(condenser pump energy).
Although there is nothing new about
the concept of employing variable-frequency drives on tower fans, unfortunately there is a new design and control
trend not to include them, especially in
tropical climates. This design method
assumes the objective is to ‘make the
cooling tower water as cold as possible,’
which would require the tower fans to
run at full speeds anyway. Variable-frequency drives, again, mitigate tower
energy at low chiller capacities – without
compromising the amount of tower cells
kept on line – and allow new control
strategies for optimal chiller, pump and
tower cell sequencing (add/shed).
chilled-water valve positions (0 percent
to 100 percent). Networking all this
information back to the plant control
system can hugely affect the algorithms’
ability to automatically control and significantly improve system performance
and delta T.
Figure 1 shows delta T results before
and after improved cooling demand visibility at the University of North Carolina
(UNC) at Chapel Hill. Jim McAdam, PE,
energy engineer with UNC Energy
Management, further explains, “We have
begun operating the bridge pumps in
series with the mains pretty much
everywhere now and are very pleased
with the energy savings, reliability
improvements and delta T improvements we are seeing. We have also piloted resetting each building’s end-of-line
differential-pressure setpoint using
building chilled-water valve positions.
This has significantly lowered building
pumping heads and allowed us to run
the buildings off the campus differential pressure much more than we
thought possible. The available pressure
increases nicely as building load
increases.”
Variable-Frequency Drives
Variable-frequency drive retrofits
for large-tonnage chillers are becoming
more and more commonplace. Retrofits
have been done on both 4,160-V and 12-kV
chillers by major chiller manufacturers
and local contractors in almost all major
world markets. Variable-speed medium-voltage new chillers have been available
for a few years now and are offered by
most manufacturers. Variable-speed dual-compressor chillers are also available
from at least two major manufacturers.
There is nothing new in primary-only variable-speed chilled-water pumping. Variable primary-only plants are
being implemented around the world
with plant tonnages well exceeding
50,000 tons. The key to successful variable-speed chilled-water pumping is getting
away from solely relying on the ‘blind’
differential-pressure control and blaming the problem on the ‘building guys’
and their low delta T. New relational
control methodologies with extensive
networked controls need to employed.
An energy management and control system network down to the building connections – bridges, booster pumps,
plate-and-frame heat exchangers and
even air-handler chilled-water valve
positions – is mandatory. Open conversations with building owners and managers about improving delta T are also
strongly encouraged (or put into the
chilled-water contract language).
(add/shed). The key to success in employing variable-frequency drives on condenser water pumps is not robbing 15 10 5
Variable-frequency drives on condenser water pumps significantly contribute to lowering a plant’s annual kilowatts per ton. Variable-frequency drives
mitigate condenser pump energy at low
chiller capacities and also allow new
control strategies for optimal chiller,
tower and condenser pump sequencing
Delta T
Fully Automated Networked Controls
Many district chilled-water plants
depend on operators to make efficiency
control decisions. Optimization is
deployed manually based on operator
experience, or trial and error, and can
be complaint-driven (i.e., ‘tweak something until a customer complains’). This
type of optimization is unsustainable
and can never achieve the full potential
for energy savings.
For truly effective optimization, the
district plant should have as much cooling demand visibility as possible.
Cooling demand visibility is defined as
energy management system points that
extend to individual building booster
chilled-water pump variable-frequency
drive speeds (0 percent to 100 percent),
building chilled-water flow (gpm), building chilled-water supply and return temperatures, commanded heat exchanger
valve positions (0 percent to 100 percent)
and commanded building air-handler
Relational Control Algorithms
Relational control algorithms are
defined as control sequences that use
mathematical relationships between
subsystems [as opposed to proportional-integral-derivative (PID) independent
temperature or pressure setpoint con-trol]. Examples of relational control
would be using orifice position from air-handling unit valves to control chilled-water distribution pump speed or using
chiller load to control cooling tower fan
speed.
Figure 1. Chilled-Water Delta T Improvement at the University of North Carolina at Chapel Hill, 2004
vs. 2007.
2004
2007
15
Delta T
10
5
0
Jan
Source: Jim McAdam, University of North Carolina at Chapel Hill.
Dec
0
Jan
Dec
