In chilled water systems,
pumping normally draws from around 6 to 12 % of the total
annual plant energy consumption. Older chilled water plants
designs circulate constant volume of chilled water through
chillers (primary loop) and building (secondary loop), no
matter if the cooling load is large or small. If loads are
small, the constant volume of chilled water is diverted
around the cooling coils by three-way valves. As thermal
loads are essentially variable, variable flow systems, where
flow tracks the thermal load, have emerged as the most efficient
pumping strategy that allows considerable reduction of the
total energy consumption of the plant.
Selecting Pumps Always select a pump at, or close
to, its "best efficiency point" (BEP). The "preferred
operating region", headed by the BEP, is established
on pump-specific speed basis. Specific speed (Ns) is a dimensionless
index used to describe the geometry of pump impellers and
to classify them as to their type. Such index allows explaining,
among other things, the wide variation in the value of efficiency
at the BEP for different pumps.
The BEP represents a combination of head and flow produced
by the pump where the efficiency is at its highest. All
other things being equal, high efficiency also represents
the minimum horsepower for that head-and-flow combination.
Because it is rare that an application duty point will just
happen to land on the BEP, there are guidelines to help
you make the best choice.
Because head calculations are more uncertain than flow calculations,
and because designers are likely to add safety factors,
it makes sense to pick a constant-speed pump with the required
duty point to the left of the BEP. By doing that, the pump
will move toward higher efficiency as it reacts to lower-system
head loss.
By selecting pumps that operate in the center or high efficiency
part of the curve most of the time, we can lower operating
costs and lengthen the life of pump seals and bearings.
This is because the pump impeller generates increasingly
high radial loads as its point of operation moves away from
BEP toward lower efficiency points.
The pump net positive suction head (NPSH) is a term that
has extremely limited usage for closed systems as chilled
water circulation loops. Most closed air-conditioning system
pump application does not require NPSH evaluation because
it is possible to obtain adequate pump suction pressurization.
Larger pumps can be equipped with a range of impeller diameters.
It is not good practice to choose a pump with either the
largest or the smallest impeller. In doing this you will
limit your ability to fine-tune the pump's performance after
it is installed. Once installed you can measure the pressure
difference across the pump and convert it to head being
produced by the pump (and demanded by the system). If an
impeller is larger than required for the system, it can
be "trimmed" by putting it on a lathe, reducing
the diameter, and rechecking the balance.
Trimming an impeller is easy and inexpensive. It is an excellent
way to counteract the oversizing that might have resulted
because of errors in early calculations. If a pump was selected
with the smallest diameter impeller and it is still too
large, it can not be trimmed. A similar situation exists
if the largest diameter impeller has been selected. If the
pump is undersized, there is no way to install a larger
impeller and increase capacity. This situation is seldom
encountered, but it could happen. And the solution is probably
pretty expensive.
The horsepower savings between the different operating points
can be determined by looking at the pump curve, and if you
know kilowatt-hour costs you can figure your savings. Once
you get the cost savings per year for trimming the impeller,
a simple payback period can be determined to see if trimming
is economically feasible. Payback periods of less than a
year are not uncommon.
Distribution
Pumps Arrangements
The distribution loop (secondary
loop) of a chilled water plant uses any of several
possible primary-secondary pumping situations or smaller
pumps in series or parallel arrangement. These applications
offer a great deal of flexibility, redundancy, and
possibly cost savings if the flow requirement varies
with time.
The distribution loop in Fig.
60 has three circuits. The circuit 1
has two identical pumps, one operating and the
other in standby. Such arrangement ensures a
100 % back-up pumping capability in case of
active pump failure. On the other hand, circuit
2 and circuit 3 have a standby pump that is
common to both circuits.
Fig. Nº60.
Two secondary pumps
arrangement.Circuit 1
has one pump operating and one standby.
Circuit 2 and circuit 3 sharing a common
standby pump.
Click on
image to enlarge
It means both
circuits have a back-up pumping capability of 50 %
respect the total flow handled by both. It is obvious
that sharing pumps allows lower initial pumping investment,
which could be demanded by a limited budget, but the
maintenance staff or contractor must be able to face
the damage as soon as possible or chilled water feeding
to both circuits will collapse if the standby pump
brought on line trips. Chilled water plant designers
must then face the trade-off between initial pumping
costs and back-up pumping requirements in each particular
project.
Sometimes a secondary pumping scheme is not the most
efficient pumping distribution strategy. For example:
In a big system as a campus, the pressure needed to
pump the entire hydronic net is large and it is advantageous
to place a third set of pumps out in the system to
avoid imposing high pressure on the equipment close
to the pumps' discharge.
In certain secondary pumping scheme the same head
may be imparted to all chilled water passing through
the distribution pumps, whether it is making the short
trip through the closest building or the longest trip
through the hydraulically most distant building. The
extra head imparted to chilled water passing through
the closer buildings must be wasted across balancing
valves and/or throttling valves at those buildings.
Only the small
fraction of the total chilled water flow going to the most
distant building is produced without wasted energy.
A better strategy to pump distant loads is via distributed
pumping. Distributed building pumps assume the function
of secondary pumps, so each pump is sized to deliver its
building's flow at just the head needed to pump the building
hydronic loads and draw the chilled water through the mains
from the center plant.
The practices used in selecting pumps for series or parallel
operation are roughly similar to those discussed above,
but there are some important differences. The motors provided
for parallel pumps must be chosen for the single-pump operating
mode because the single pump will run out on its curve and
draw more horsepower than when both pumps are operating
together. Motors for pumps in series must be selected to
handle the horsepower requirements when both pumps are running.
Parallel constant speed (C/S) pumps have been used
for many years in a simple method for achieving
pump power reduction for variable volume (two-way
valve controlled) HVAC systems. Paralleled C/S pumps
compete favorably with variable speed (V/S) pumps,
particularly in smaller systems where variable system
piping head loss is a small proportion of total
system head needs.
Parallel C/S pumps are usually selected so that
each pump provides half design flow, which means
that each pump provides half of design HP needs.
The parallel C/S pump power correlation with a single
C/S pump, assuming similar pump efficiencies, is
shown in Figure 61.
A similar advantage for parallel C/S pumps in comparison
with V/S pumps occurs when the variable head loss
in a V/S pumped system is a low percentage of total
pump head. This is shown in Figure
62.
Fig. Nº61.
Parallel
C/S pump power saving advantage over single C/S
at low flow.
Click on image
to enlarge
Fig. Nº62.
Best application area
for parallel equally sized C/S pumps in comparison
with V/S occurs when system variable head loss
is low.
Click on image
to enlarge
An
interesting variation of parallel pumping is one V/S
pump in parallel with a similarly sized C/S pump.
The principal advantage of such arrangement is that
only one variable speed drive (VSD) is needed to obtain
a variable flow pumping arrangement. Two equally sized
pumps and motors with a single VSD that can control
either pump. At start-up, the VSD controls one pump
and increases its speed in accordance with system
need until full speed is reached. At this time, the
pump is kept at full speed and VSD control is changed
to the other pump, which now changes its speed in
accordance with system needs. A similar reversed cycle
operates as system load decreases. When properly selected
and controlled, HP savings will correspond with the
savings potential for a single V/S pump handling the
total flow and head required but also using a lower
size VSD, which means lower initial control cost.
A similar, but less costly control scheme establishes
one pump as C/S only, the second as V/S only. In this
case, at start-up, the V/S increases pump speed in
accordance with system needs until a set control point
is reached, at which time the C/S pump is started.
The V/S continues
in operation, meeting increasing system flow needs as required.
As before, a similar but reversed cycle occurs as load decreases.
Variable Speed Pumps Variable speed pumping by the use
of VSDs, has grown common in recent years as people have
become more familiar with the technology and as costs for
VSDs have gone down. With a VSD, pump head as well as flow
are both reduced to handle part load situations. In the
right kind of application, this can result in large savings
in operating costs.
The V/S pump must operate at the intersection of its "speed"
curves with the system curve. Thus, at 50 % speed, the V/S
pump will provide 50 % design flow and 25 % design head
and reduced power draw of 12.5 % of design point HP, due
to a square relationship between flow-head, and a cubic
one between pump speed-HP draw. It should be noted, however,
that variable speed pumping is not a cost saving panacea.
A meaningful energy savings payback with V/S systems is
only possible if the designer has followed time tested design
fundamentals in the application of a V/S pumping system.
Some designers attach overriding importance to pumping costs,
to the extent that the design of the plant appears to be
primarily determined by this consideration, without tacking
into account the effects these choices have on comfort.
It is true that pumping costs may be fairly accurately estimated,
which is a powerful incentive to take them into account.
In a balanced constant flow distribution system, the real
pumping costs, in percentage of the production unit’s
seasonal consumption, are around 6 to 12 % in cooling applications.
In order to achieve greatest pumping energy saving, a constant
differential pressure control should be kept at the most
remote unit terminal. On the other hand, if in a direct
return distribution circuit, all terminal units work al
50 % load, except terminal closest to the pump that needs
full flow, the differential pressure available will be insufficient
at this point and design flow will not be attained at this
terminal unit. The closest unit will then be underfed. However,
coil performance is inherently non-linear so a reduction
in flow is not matched by an equivalent in coil capacity.
From figure 63
it can be seen that variations from the design flow rate
down to approximately 50 % of the design requirement have
a relatively negligible effect on the overall heat transfer
or system performance, assuming supply water temperature
is at design condition. These capacity shortages are most
likely within the error (or safety factor) of the engineer’s
design load calculations and are unlikely to affect the
space comfort levels.
In case of such flow variation would affect the
space comfort level, relocation of the controlling
pressure differential taps from the end to the
center of the circuit, will provide higher differential
pressure available and thus flow closer to design
in the terminal nearest to the pump. In other
words, it means to reduce energy saving in order
to ensure the required comfort level.
Several approaches about where to establish the
differential pressure control in the circuit could
emerge based on the type of the building, comfort
standards required, and other subjects. It seems,
we are again facing an engineering trade-off between
comfort and energy saving that must be carefully
evaluated at each particular project.
