In my previous post where I was discussing the retrocommissioning opportunities in a heating hot water system I am currently working with, I mentioned the term “pump heat”. Subsequently, one of the students in a class I was teaching asked
what that was. So, I thought I would do a short post about it.
If you do a Google Search for the term pump heat, you get a whole bunch of hits about heat pumps; not the same thing. A heat pump is a machine that uses an energy input of some kind to move heat from a low temperature source and reject it to a high temperature sink. Chillers and refrigerators are common examples of heat pumps were we take advantage of the heat removal aspect of the cycle.
But it is also possible to take advantage of the heat rejection side of things and use the heat
that is rejected in some sort of HVAC process while using the atmosphere or ground water as the low temperature source. In industry jargon, the term “heat pump” typically refers this type of application. It can also refer to an application where somebody has been really clever and used the low temperature source to cool a process where cooling is required and the high temperature sink to provide heat to a different process; i.e. a heat recovery process.
O.K., so having established that you may not find out what you wanted to know about pump heat from a Google Search (although we will have to see what happens there after I post this), lets talk about pump heat.
When a pump moves water from a lower pressure to a higher pressure or from a lower elevation to a higher elevation, it is doing work on the water. This is mathematically evident from the pump power equation:
BHPPump = (Flow x Head)/(3,960 x EfficiencyPump)
BHPPump = The pump brake horsepower; i.e. the energy input to the shaft of the pump for the flow and head under consideration.
Flow = Pump flow rate in gallons per minute
Head = Pump head in feet of water column
3,960 = A units conversion constant, which varies a bit with the state of the water
EfficiencyPump = Pump efficiency, which can be as low as 40% or less for a small (less than 100 gpm) centrifugal pump and as high as 85% or more for a large (1,500 gpm +) centrifugal pump
Incidentally, there is a similar equation for fan power;
BHPFan = (Flow x Static)/(6,356 x EfficiencyFanStatic)
Because the pump (or fan) is doing work on the water (or air) passing through it, the work will show up as a temperature rise across the system. For pumps, where the fluid is incompressible, the temperature rise shows up across the distribution network. For a fan, where the fluid is compressible, the temperature rise shows up across the fan.
Fan and pump heat can both represent significant loads on a refrigeration process if the fan or pump is located in the fluid stream that is to be cooled. And, this stuff can add up, as I attempt to illustrate in this slide.
Specifically, for the air handling system out at the end of the loop served by the water cooled chiller plant.
- The fan heat associated with moving air through a distribution system to serve a cooling load becomes an additional load on the cooling coil. This means the chilled water pump has to move just a bit more water than it otherwise would if there were no fan heat.
- The pump heat associated with moving water through the chilled water coil and piping network gets added to the actual cooling load and fan heat. This means that the chiller has to work just a bit harder than it otherwise would if there were no fan or pump heat.
- The compressor heat; i.e. the power put into the chiller compressor to remove the cooling load, fan heat, and chilled water pump heat from the chilled water and place
it in the condenser water shows up as a an additional load on the condenser water system. This means that the condenser water pump has to move just a bit more water than it would have to move if there were no fan heat, pump heat, or compressor heat.
- The condenser pump heat associated with moving condenser water through the condenser, condenser water system, and cooling tower is added to the cooling load, chilled water pump heat, and compressor heat that must be rejected by the cooling tower.
- The compounding effect of all of this, in addition to the energy alluded to in the bullets above, is that the components in the system need to be marginally larger to handle the additional load represented by the fan heat, pump heat and compressor heat relative to the size they would be if there were no such thing.
Kind of gives a whole new meaning to the phrase “no free lunch”. And, it can be very inspirational in terms of making you want to eliminate pressure drops from systems,
especially pressure drops from air handling systems where the penalty is compounded several times over; details matter!
If you dead head a pump, an additional pump heat consideration comes into
play which is typically not a concern with a fan. The term “dead heading” a pump is actually not a term used to describe working with pumps while listening to music by Jerry Garcia and Co. Rather it is a reference to operating the pump in a manner that forces it up its pump curve to a condition where there is no flow produced, only pressure.
Typically, this can occur for a number of reasons.
- An operator or HVAC technician intentionally throttles the discharge side of the pump to 0 gpm to determine the impeller size. (For more information on why and how this works, see the Energy Design Resources Design Brief titled Pumping System Troubleshooting. )
- In a variable flow pumping system (typically characterized by loads equipped with two way control valves) at low load conditions, the valves can drive closed with the combined effect being very similar to what happens when the discharge valve on the pump is throttled. Frequently, these systems are equipped with a bypass valve piped across the mains and controlled to modulate open as the differential pressure rises. This maintains some minimum level of flow, with the level set by the requirements of the pumps and prime movers (usually the boilers or chillers).
- In a system with multiple pumps in parallel, where the pumps have significantly different pump curves, pumps with higher head capabilities can interact with pumps with lower head capabilities, forcing them up their curves just as if a valve had been closed. I illustrate this in a previous post titled Parallel Pumps; An Indicator of a Retrocommissioning Opportunity.
Dead heading a centrifugal pump (but not a positive displacement pump) is safe for a brief period of time, which is why its usually O.K. to do a pump shut-off test to verify impeller size1. But, if you dead head a pump, their is no flow through the pump, and the work the pump does while dead headed shows up as heat in the water trapped in the volute. After a while, the water can actually reach the boiling point, which triggers cavitation, causes seals to fail, warps flanges, discolors paint and generally does significant damage.
If you are curious about this, fan heat is easier to witness in the field, mostly because the magnitude tends to belarger and because it occurs directly across the fan. For most HVAC systems it will be in the .5 to 1.5 °F range and it can be several degrees for a system that has a lot of static or is moving a lot of air or both. For instance, in the large make-up air systems I worked with during my days as a facilities engineer for Komatsu Silicon, the fan heat was 3-5°F or more. But these systems were operating at static pressures in the 5 to 7 in.w.c. range vs. the 2-4 in.w.c. range that is more common in commercial HVAC.
Its important to realize that the temperature rise I am discussing here has nothing to do with the heat generated by motor efficiency losses when the motor is in the air stream. It
will exist across a fan with the motor out of the air stream. If the motor is in the air stream, the temperature rise will reflect both the brake horse power into the fan as well as the motor efficiency loss.
So for those who are really curious (this is how I first learned about this), some time you are around a fan with a motor that is not in the air stream, take a thermometer that can
resolve fractions of a degree and measure the temperature entering and leaving the fan. (Note that its important ot use the same thermometer so that any calibration error cancels out when you subtract your two readings.)
If you do this accurately enough, you will discover that you can measure a temperature rise that is exactly equal to what you get if you take the sensible heat equation …
Q = 1.08 x cfm x Δt
Q = Load in Btu/hr.
1.08 = Another one of those pesky units conversion constants that are dependent on the conditions of the fluid associated with the equation (in this case, air, and the constant is good for typical HVAC systems at or near sea level)
cfm = System flow rate in cubic feet per minute
Δt = Temperature rise or drop in °F
… and insert
the fan brake horsepower at the current flow rate in Btu’s for Q,
the actual flow rate for cfm,
…and then solve it for Δt.
Jerry Williams wrote a really great two part article on fan and pump heat that appeared in the August and September 2005 issues of Heating, Piping, and Air Conditioning. You used to be able to find it in their article archive, but they changed the web site and I can’t
figure out how to do it now. So, I contacted one of the editors I know there to find out how to access it and will let you know what I find out in a subsequent post.
Meanwhile, he discusses it in more general terms as part of his article Air System Basics, starting on page 72. Even if you are not curious about fan heat, this article is a great one to add to your library of resources since it covers all of the important basic physics of air systems, including fluid statics, conservation of mass and energy, conservation of momentum, measurements, and system losses and system curves.
Senior Engineer – Facility Dynamics Engineering
1. This assumes that the pump can not generate a pressure that is higher than the rated pressure of the pipe, valves, fittings, flex couplings, etc. located between the discharge valve and pump inlet. In other words, it is not safe to do so a shut of test if you do not know for a fact that the peak on the pump curve for the pump with its biggest impeller installed is below the pressure ratings mentioned above.