Dynamics of Hydronic Systems - JCI

7/24/2019 Dynamics of Hydronic Systems - JCI

1/20

Valve and Actuator Manual 977

Hydronic System Basics Section

Engineering Bulletin H112

Issue Date 0989

1989 Johnson Controls, Inc. 1Code No. LIT-351H112

A hydronic system can be configured in many different ways. This report

will discuss the following types:

1. Reverse Return

2. Direct Return

3. Primary/Secondary

The variant configurations of the hydronic systems noted above will be

analyzed as well as the relative advantages and disadvantages of each.

Part-load characteristics will also be considered.

Any of the three types of hydronic systems could be utilized in either a

heating or cooling system. To simplify the remainder of this report only

the case of a cooling system will be considered. Most of the concepts

covered however, would be applicable for either heating or cooling

systems.

It is strongly recommended that the concepts discussed in Engineering

Reports H111 and H110 be understood before preceding with this report.

Engineering Report H111 includes a detailed discussion of the

performance of control valves. Engineering Report H110 discusses the

relationship between fluid flow and pressure drop as it pertains to typicalhydronic systems.

Accurate, stable control is dependent upon two factors. The first is the

ability of the controller and control valve to compensate for process

nonlinearites. A coil is a good example of a nonlinear process. If the

capacity of a coil is plotted for each flow rate through the coil, the result is

a logarithmic relationship. Secondly, the physical limitations of the

control valve must be considered. As discussed in Engineering Report

H111, all control valves have some amount of uncontrollable flow when

the valve plug is initially lifted form its seat. The magnitude of this

uncontrollable flow is related to the valve rangeability and the pressure

differential across the valve. Rangeability values are limited by machining

process tolerances. Therefore, it becomes important to minimize the

variation in pressure differential across the valve to minimize the

uncontrollable flow through it.

Dynamics of Hydronic Systems


2/20

2 H112 Engineering Bulletin

Even small amounts of uncontrollable flow can compromise the ability of

a hydronic system to be properly controlled. This is the result of the

extremely high coil gain at low flow rates. It is important to properly size

the control valves. It is also necessary for the consultant to design a

hydronic system which is able to maintain a relatively constant branch

differential pressure, Pb, regardless of building load. This reportdiscusses how well the standard types of distribution systems are able to

maintain a constant Pb.

In a reverse return system the pressure drops in the distribution

piping between the circulating pump and each air handling unit are

equal. This is true because the system is designed so that the

length of the water circuit is the same for each air handling unit

regardless of its location with respect to the circulating pump. As

a result the Pbfor every branch is equal regardless of the

magnitude of the building cooling load. If the air handling unit

cooling coils are all selected for the same pressure drop this type

of system is self-balancing. Unfortunately the installed cost of a

reverse return piping system is higher than that of the other types

of hydronic systems. This is due primarily to the cost of the

additional chilled water return piping (See Figure 1).

Reverse ReturnDistributionSystems


3/20

H112 Engineering Bulletin 3

When 3-way valves are installed in a reverse return system, each valve can

be optimally selected based upon the same criteria. Changes in the

magnitude of the cooling load will have little effect on the chilled water

flow rate. As a result, the value of Pbfor every branch will remain

constant regardless of changes in the magnitude of the building cooling

load.In the case of the reverse return system shown in Figure 1 the Pbfor

every branch will alwaysbe 50ft H2O. Since Pbis always

50ft H2O, every control valve can be optimally selected, based upon the

above criterion. If a valve authority of 50% is desired, all of the control

valves can be selected for a 25ft H2O pressure drop at design flow. Since

the branch pressure drop does not deviate from 50ft H2O, there should be

no increase in uncontrollable flow nor will there be any loss in effective

valve travel as the building cooling load decreases.

As with any system which utilizes 3-way valves, the flow rate through the

chiller(s) can always be maintained at a constant level without therequirement for a large central plant bypass with its associated controls.

Because the 3-way control valves shown in Figure 1 are mixing valves

piped in a bypass configuration, problems associated with control valve

actuator spring range shift or the inability to shut off flow through the

cooling coil will be minimized.

The disadvantages of utilizing 3-way valves are increased installation and

operating costs; specifically, a more complex (costly) coil piping

arrangement, a high continuous building flow requirement and a lower

overall chiller coefficient of performance (C.O.P). When 3-way control

valves are installed, the building demand for chilled water does not vary

with diversity in the building thermal load. Remember, the flow rate

through each branch does not change with the cooling load requirements

of the space served by the air handling unit. Only the chilled water flow

rate through the cooling coil changes with the load. Consequently, chilled

water pumps must often be operated without their associated chiller to

insure each branch will have chilled water available to meet the space

cooling load. Otherwise the branches nearest the chilled water pumps will

consume all of the available chilled water.

In the case of a chilled water system the operation of extra chilled water

pumps not only increases pumping horsepower requirements but also

reduces the overall chiller plant C.O.P. The lower overall chiller plantC.O.P. is the result of mixing relatively warm water passing through a

nonoperating chiller with exceptionally cold water leaving a operating

chiller. This is required to meet both the building flow and chilled water

supply temperature requirements. The C.O.P. of the operating chiller

decreases since it is forced to provide a lower chilled water supply

temperature for a given building thermal load.

This problem is illustrated in Example 1.

Reverse ReturnSystem with3-way Valves


4/20


Refer to Figure 1. This distribution system was designed for a total

cooling capacity of 400 tons. There are four air handling units (AHU),

each with a design capacity of 100 tons. Each AHU requires 150 gpm of

chilled water at a 16F rise (45-61F) when fully loaded.

Assume that the AHU is Branch #4 serves a computer facility and has a

constant cooling load of 100 tons regardless of outdoor conditions. Alsoassume the AHUs in branches #1 through #3 serve exterior zones in the

building. Finally, assume it is a temperate spring day, such that the

cooling load on the AHUs located in branches #1 through #3 is 20 tons per

AHU.

The total building cooling load is 160 tons. It would seem, since each

chiller has a capacity of 200 tons only one chiller will need to operate.

Unfortunately things arent that simple. Because this distribution system

has 3-way control valves, both chilled water pumps must operate to

maintain flow in all branches.

The cooling coil in branch #4 is fully loaded. Therefore 150 gpm ispassing through this coil and is being heated from 45F to 61F.

Meanwhile in branches #1 though #3 only 30 gpm of the chilled water

actually passes through the cooling coil. The other 120 gpm of 45F

chilled water bypasses the coil. These two streams then mix resulting in a

branch chilled water return temperature of 48.2F. The temperature of the

chilled water returning to the central plant will be 51.4F after the water

returning from all four branches is mixed.

Here is the problem. A single operating chiller must produce 38.6F

chilled water if 45F water is to be supplied to the building. This

extremely low chilled water temperature requirement is dictated by thefact an equal volume of 51.4F water is simply passing through the

nonoperating chiller and the AHU in branch #4 requires 45F to meet its

cooling load.

When a chiller designed to produce 45F chilled water is forced to produce

lower temperature chilled water its C.O.P. drops in a nonlinear manner.

For each additional degree drop in chilled water supply temperature

increasing amounts of compressor power is required. In this case it is

unlikely that the one operating chiller could even provide 38.6F chilled

water. In this extreme case both chillers would be required to operate even

though the actual cooling load is less than the capacity of one chiller. Thisis very inefficient.

Example 1


5/20


When 2-way valves are installed in a reverse return system the Pbfor

every branch will the same regardless of load, but the magnitude of the

Pbwill vary with changes in the building cooling load. This occurs

because the pipe-friction factor changes with the variable chilled water

flow rate (See Figure 2).

A reduction in cooling load will cause one or more of the 2-way control

valves to reduce the system chilled water flow rate. The hydronic system

depicted in Figure 2 does not have a central plant bypass. Without a

bypass the pressure developed by the chilled water pump(s) will increase

with the reduction in the cooling load. The pumps operating point will

ride up its pump curve as the chilled water flow rate decreases. Thisincrease in pressure developed by the pump, as well as the reduced pipe-

friction factor, will increase the differential pressure across the branches.

This is undesirable. The numbers shown in Figure 2 without parenthesis

represent the system pressures at maximum design cooling load. The

numbers shown in parenthesis represent system pressures when each

branch is at (1/2) of its design load. Notice how each Pbincreased from

50ft to 120ft H2O.

Reverse ReturnSystems With2-way Valves


6/20


If the distribution system had a central plant bypass, the magnitude of the

Pbcould have been held constant (See

Figure 3).

The control valve located in the central plant bypass will be controlled by

a differential pressure controller. The high and low reference sensing lines

for the controller will be connected across one of the branches in the

distribution system. If the sensed differential pressure deviates from the

controller set-point, the control valve in the bypass piping will modulate as

required to maintain the desired Pb.

Since each branch has the same Pbin a reverse return system, common

sense would dictate that the sensing location for the differential pressure

controller would be across the branch closest to the bypass. This would

reduce the length of the controller sensing and/or output lines, thereforereducing installation costs. This system has all of the advantages of a

reverse return system with 3-way valves, but does not have its

disadvantages.


7/20


As an alternative to using a central plant bypass to maintain a constant

Pb, a variable frequency drive could be installed on the chilled water

pump motor. The differential pressure controller would then vary the

speed of the chilled water pump to maintain the desired Pb. The main

problem with using a variable speed drive in lieu of a central plant bypass

is that the flow through the chiller(s) will now vary with changes in thecooling load. This potential problem must be carefully considered, and it

would be wise to consult the chiller manufacturer to obtain acceptable

guidelines.

The major advantage in using a variable speed drive in lieu of a central

plant bypass is the potential for energy savings. The variable speed drive

will save energy in two areas. First, any energy losses which normally

occur within the bypass valve will be avoided. In addition, the reduction

in the chilled water flow rate will proportionally decrease the chilled water

pump brake-horsepower requirement. Remember, in a system with a

central plant bypass the flow rate to the building is directly proportional to

the cooling load, but the flow rate through the chilled water pumps

actually increases with reductions in the cooling load. In a system

utilizing variable speed drives the flow rate through the pumps will vary

directly with the cooling load.

Keep in mind however, that in a properly sequenced multiple chiller

central plant the flow rate through the bypass line will be less than the

water flow rate associated with only one chiller. In other words, the

excess amount of chilled water produced in the central plant above and

beyond the demand required by the building would be less than the

amount of water provided by one chilled water pump. This assumes there

is one chilled water pump for each chiller. Therefore, the savingsprovided by using a variable speed drive compared to a central plant

bypass is not a function of the total plant chilled water pumping capacity.

Instead the potential savings is a function of the capacity of only one, not

all, of the chilled water pumps. If the central plant has a large number

(greater than four) of chillers and chilled water pumps, variable speed

drives will save very little energy over a properly operated constant speed

system utilizing a bypass valve.

In a direct return distribution system the length of the distribution piping

between the chilled water pump and each air handling unit will vary (See

Figure 4). As result the Pbfor every branch will be different. The closer

an air handling unit is located to the chilled water pump, the larger the

Pb. The magnitude of the Pbwill also be affected by the chilled water

flow rate. Remember the pipe friction-factor changes with the square of

the flow rate through the pipe.

DirectReturnDistributionSystems


8/20


The following four items can be utilized to minimize the variation in Pb:

1. Properly sizing the distribution piping can help to maintain a

constant Pb. This piping should be sized for small (less than 4 ft

head loss per 100 ft of pipe) friction-factors. The change in

pressure drop of the distribution piping, as flow is varied from

maximum to minimum, then should not exceed the design pressuredrop across a properly sized control valve. This is true for all but

the largest systems. Some central campus systems, and the like,

would require booster pumps in remote locations.

2. Utilize a central plant bypass or variable speed pump. The Pb

across one of the branches can then be maintained at a constant

level. This will prevent unwanted changes in the pressure

developed by the chilled water pump from being shifted out to the

branches.

3. Utilize control valves with large design pressure drops.

4. In some cases, utilize balancing valves.


9/20


When 3-way valves are installed in a direct return hydronic system the

magnitude of the Pbfor each branch will be constant. This is true

because the chilled water flow rate is always constant and is not affected

by the load. However the magnitude of the Pbwill be different for each

branch. This occurs because the length of water circuit between the

branch and chilled water pump is different for each branch (See Figure 4).The control valves for each branch can, therefore, be optimally sized for

any cooling load condition. Unfortunately each branch will have a

different sizing criterion since the Pbis different for each branch. In

comparison, the 3-way valves in the reverse return systemcan all be

sized for the same pressure drop.

As discussed earlier, the 3-way valves will insure a constant flow rate

through the chiller(s). Problems with actuator spring range shift, and

cooling coil shutoff will also be minimal. Unfortunately, the same

disadvantages are still apparent. The installed cost of a system with 3-way

valves is higher due to more complex coil piping arrangements. Also asdiscussed previously the chiller plant C.O.P. will be lower when compared

to a system with 2-way valves. This results from the high continuous

building demand for chilled water dictating operation of chilled water

pumps without their associated chiller.

When 2-way valves are installed in a direct return system, the Pbwill be

dependent upon the location of the branch relative to the chilled water

pump and the magnitude of the cooling load (See Figure 5). The numbers

in parenthesis indicate the system pressures when each branch is at (1/2) of

its design cooling load. The numbers without parenthesis indicate the

system pressures when the system is at maximum design cooling load.

If the system does not have a central plant bypass the value of Pbwill be

extremely dynamic. These extreme changes in differential pressure occur

for three reasons: the chilled water pump will ride up its curve, the pipe-

friction factor in the distribution piping will change with load, and the

length of the water circuit is different for each branch. This system is as

far from optimal as any can get.

Direct ReturnSystem with3-way Valves

Direct ReturnSystems with2-way Valves


10/20


If the system does have a central plant bypass, the pressure fluctuations

across each branch are reduced (See Figure 6). The numbers in

parenthesis indicate the system pressures when each branch is at (1/2) of

design cooling load.

Adding the central plant bypass keeps changes in the pressure developed

by the chilled water pump from being shifted out to the branches.

However, since the pipe-friction factor in the distribution piping will

change with load, the value of the Pbwill be affected.

Whenever the load in the building decreases and the bypass valve

opens, the flow rate in the central plant will increase. This occurs

because the pressure drop across the bypass will decrease as it is

opened. In turn the pressure developed by the pump will decrease.

As the pump operating point moves further out on its pump curve,

its flow rate will increase. This increased flow rate will, in turn,

increase the pipe-friction factor of the central plant piping.Eventually a balance point between the increased central plant

piping friction losses and the decreased pressure across the bypass

will be found. This is graphically shown in Figures 7, 8 and 9.

Notice how the slope (pipe-friction factor) of the line representing

1/2 design load changes within the central plant portion of the

graph.


11/20


Assuming the branches are uniformly spaced in a distribution

system, the best differential pressure sensing location for the

bypass controller is in the middle of the system. Figure 7 relates

the change in Pbwith location as a function of cooling load. The

graph illustrates the possible variation in Pbas the building load

changes from no load to maximum design load. The slope of each

of the different load lines is equal to the negative of the respective

piping friction factors. The area underneath the lowest load line

represents the pressure available for the branch piping, coil and

control valve. Depending on the location of the branch in the

system the lowest load line could be either the design or no loadline. In this case the control valves in the first half of the branches

which are closest to the pump should be selected for a higher

design pressure drop than those in the second half of the system.


12/20


By sensing differential pressure in the middle of the distribution system

the magnitude of the maximum variation in branch differential pressure is

minimized when compared to other sensing locations. Compare the

maximum variation of Pbin the building portion of Figure 7 to that

shown in Figures 8 or 9.

The bypass control scheme shown in Figure 8 will provide good pressure

control at the end of the system. Unfortunately the variation in Pbat thebeginning of the system may be excessive.

The bypass control scheme shown in Figure 9 can provide good pressure

control at the beginning of the system. Unfortunately the variation in Pb

at the end of the system may be excessive.

In spite of some inherent problems, direct return hydronic systems with 2-

way control valves are very popular. This is likely due to the fact it has a

lower initial cost relative to other types of hydronic systems and it is a

very energy efficient pumping system. Lastly, a direct return system

incorporating either a central plant bypass or variable speed pumping can

provide acceptable pressure control when distribution piping is properlysized. Proper distribution pipe sizing is critical to insure trouble free

control. If the distribution piping is undersized the Pbcan vary

significantly with the load.


13/20


As a rule, the potential changein Pbmust be less than the designpressure drop of the control valve, or problems can occur. These problems

can manifest themselves in the following ways: unacceptable

uncontrollable valve flow, problems with valve shutoff and spring range

shift, valve cavitation, and overflowing marginally sized cooling coils.

A primary secondary distribution system utilizes two sets of pumps. The

first set is used to pump water through the chillers. These pumps are

called the primary pumps. The second set of pumps are used to pump

water through the building. These pumps are called the secondary pumps.

The primary and secondary pumps are hydraulically isolated from oneanother. This isolation is accomplished by installing a bypass line

between the primary and secondary pumping systems as shown in

Figure 10.

Primary/SecondaryDistributionSystems


14/20


The bypass line is also often called a decoupler line since it isolates or

decouples the primary and secondary systems. The direction of flow in

the decoupler piping can be in either direction. The direction of flow will

depend upon the amount of chilled water produced in the central plant

(primary flow) and the amount consumed by the building (secondary

flow). If the primary loop produces a greater volume of chilled water than

the secondary system consumes, flow in the decoupler will be from supply

to return. If then secondary system consumes a greater volume of water

than the primary system produces the water in the decoupler will flow

from return to supply. Normally flow within the decoupler should be from

supply to return. Otherwise the chilled water supply temperature in the

secondary loop will rise to unacceptable levels.

The hydraulic isolation allows both the primary and secondary pumping

systems to function as if the other was not present. The primary pumpsare, therefore, able to maintain a constant flow rate through the chiller

regardless of the building load.

The secondary system can be configured as either a direct or

reverse return distribution system. The same advantages and

disadvantages discussed earlier will still apply. However, in

either case only two-way control valves should be installed.

The chilled water flow rate in the secondary system will then

vary in relation to the building load. This is desirable for two

reasons.


15/20


First, a variable flow secondary system will require less pumping energy

than other types of systems. The secondary flow rate, in a primary

secondary system, can be reduced to extremely low flow rates. The

minimum secondary flow rate is not fixed by flow limitations imposed on

the system by the chillers. If a variable speed secondary chilled water

pumps is installed, the amount of pumping power required will beminimized. In a variable speed pumping system, the pumping power

requirement will vary with the cube of the flow rate as dictated by the

pump affinity laws. The sensing location for the differential pressure

transmitter should be near the middle of the secondary system. Once

again, this location is the best compromise for minimizing the variation in

the magnitude of the Pb.

Secondly, if both the magnitude and direction of the flow rate in the

decoupler are known, this information can be utilized in various building

automation schemes. Particularly for chiller sequencing. Chiller

sequencing is discussed in Engineering Report H324.

It is important for the decoupler line which separates the primary and

secondary systems to be properly sized. It must be sized so that its

pressure drop at full flow is kept very small (i.e., less than 1 psig).

Normally full flow is about 115% of the flow rate associated with the

largest chiller. Flow rates in excess of this value would indicate an

operating chiller should be stopped. The decoupler should be a straight

length of pipe with no restrictions. It is very important for the pressure

drop in the decoupler to be kept very small to provide the hydraulic

isolation between the primary and secondary systems. If the size of the

decoupler is too small, pressure changes in the secondary system will betransmitted into the primary system. This interaction can cause stability

problems and undesired flow variations through the chillers.

Occasionally, a primary secondary distribution system with a check valve

installed in the decoupler line is encountered (See Figure 11).

Decoupler LineSizingConsiderations


16/20


17/20


Note: Do not use square law type flow meters to measure the flow rate in

a decoupler line. Square law type flow meters include annubars,

orifice plates, flow nozzles, and venturi tubes. The output signal

cannot be accurately measured with these flow meters at

turndowns greater than 3:1. A bidirectional turbine flow meter

should be utilized because it can indicate both direction and

amount of flow accurately at much lower flow rates. Turbine

meters with turndown ratios of 10:1 are typical, but are available

with ratios up to 30:1. Remember water must be able to flow bothways through the decoupler line.


18/20


Notes


19/20


Notes


20/20


Notes

Controls Group

507 E. Michigan Street

P.O. Box 423

Milwaukee, WI 53201 Printed in U.S.A.

Dynamics of Hydronic Systems - JCI

Documents