Overpumping emitters.

Henry

I have seen several instances of too much pump. It has happened on several installs of boilers. Some installers put too little pump and others too much. The too much increased gas consumption on an installation that should of decreased by at least 20%. When delta T is less than 10, there is too much velocity and insufficient BTU transfer. By choking the pump to restore at least 20 delta T, there was a better transfer and the following year, over 20% saving compared to the original boilers.

hot_rod

Henry said:

I have seen several instances of too much pump. It has happened on several installs of boilers. Some installers put too little pump and others too much. The too much increased gas consumption on an installation that should of decreased by at least 20%. When delta T is less than 10, there is too much velocity and insufficient BTU transfer. By choking the pump to restore at least 20 delta T, there was a better transfer and the following year, over 20% saving compared to the original boilers.

I'd agree with too much or too little pump sizing issues. The early radiant days with copper tube boilers was a classic example of under pumped systems.

As for over-pumping, I think the type of heat emitter needs to be considered.

Last nights RPA webinar on ceiling radiant panels, the manufacturer suggested an 8°∆T.
With a wide ∆ the ceiling at the front of the room is warm, the tail end cold. Same as the radiant loop example above.

Panel rads are often designed around 30∆ So a wide swing in what is an acceptable design ∆.

Henry

The last case was two Laars 1,500 million each copper fin in P/S. We choked the outlet of each boiler pump. Old ceiling radiant tube designs were of 20F delta T.

Gordy

I really see no need to go less than a 15-20 delta on ceilings unless the tube spacing is huge.......

Gordy

No train hat?

Answer this if I have the burner on the stove on, and swiftly run my hand across it I don't get burned. However if I move it slowly you can bet you will get burned.

Time interface has a role in transfer.

Harvey Ramer

Non of the modeling software packages, that I am aware of, have any built in formulas or calculations to address the alleged change in heat transfer if a turbulent flow goes laminar.

That's what I'm interested in. Ray's situation is the perfect example.

We all know how the UHF works and also average emitter temps and their reaction to space temp. That is neither here nor there and there is no argument about the physics of those processes. This is strictly on what's happening inside the pipe. I can guarantee with 100% certainty, if flow turns laminar, the rate of heat transfer to the inner wall of the pipe will slow down and the delta will also narrow.

hot_rod

Non of the modeling software packages, that I am aware of, have any built in formulas or calculations to address the alleged change in heat transfer if a turbulent flow goes laminar.

The reason for that could be that nobody has defined or proven that it can happen, documanted at what flow rate it can or does happen, or can the phenomena be duplicated across different types of systems and emitters.

We have good, stable, believable proven data on the flow transition shown in Richs post on Reynolds numbers.

What we need and doesn't seem to be available is data like that for flow turning to laminar at high velocities after being turbulent. The engineers I quiz say it is not possible, contractors here claim they have experienced it.

Harvey Ramer

I agree with that completely Bob. I knew this thread would be controversial as hell because it attacks the currently accepted, underlying physics of fluid dynamics.

However, real life experienced phenomenon cannot be ignored and must be explained.

Engineers will typically work out of the toolbox that physicist provide them with.

I would be extremely interested in seeing testing done on the subject. It would have to be extensive though because there are almost an infinite amount of uniquely different scenarios.

SWEI

hot rod said:

I remember back articles of Siggys he FEA modeled the slab temperature difference with tube in different levels in a slab. I linked that article on another thread but for some reason the color FEA didn't show up with the link to the PM archive.

Here's a PDF of the whole article, including pictures.

hot_rod

SWEI said:

hot rod said:

I remember back articles of Siggys he FEA modeled the slab temperature difference with tube in different levels in a slab. I linked that article on another thread but for some reason the color FEA didn't show up with the link to the PM archive.

Here's a PDF of the whole article, including pictures.

Excellent thanks for finding it, vintage 2000 wow!

I think this entire radiant loop could be modeled in that software. The higher flow, smaller ∆ would show as more consistent color, just as I expect it would with an infrared camera.


Let just consider these models to be a 20X15 foot room 300 sq.ft.
4" slab, 300' loop tube 12" oc

Room 1 running a 22∆, .65 gpm transfers 7206 BTU/hr to the room.
Room 2 running 15∆, 1.1 gpm transfers 8378 BTU/hr

Room 1 110 plus 88 divided by 2 = 99° average slab temperature
Room 2 110 plus 95divided by 2 = 102° AST

So the higher flow rate 1.1 gpm provided higher AST and higher BTU/ ft output

With 1/2 pex as the tube I can't realistically run this to 5 gpm, if so the program would show higher AST and more BTU transfer.

How does this not prove that BTU do jump off faster trains?

While it doesn't exactly answer Harveys question about when or how, or if in my mind the BTU /hr transfer starts going the opposite way at higher flow. It does clearly show how increased flow rates, increased AST transfers more BTU/hr.

Gordy

Those are acceptable design deltas. Is Harvey not talking about extreme flows, and extreme narrow deltas.

Is the little gained worth the extra pumping?

I thought there was not a train. Must be an empty subway.....

Harvey Ramer

Another thought.

Depending on the viscosity of the fluid and the smoothness of the pipe, laminar flows can reach a much higher Reynolds number than the generally accepted numbers we design by. Chapter 8 in Fluid mechanics suggests laminar flows can exist with a Reynolds number up to 100,000. That is of course with the fluid and environment in a very stable condition.

One thing the Reynolds numbers don't calculate in their predictions is change in Fluid temperature which certainly causes disturbance inside the tube and would contribute to turbulent flow.

So humor me on this.
Let's say for example that in Ray's case, the conditions were such that it allowed for laminar flows at a much higher Reynolds number or velocity. With the system only displaying a 2-3 degree delta t there would also have been very little change in fluid temperature in the pipe. When he throttled back the flow to a 20 degree delta t, he forced the water in the pipe to undergo temperature change. That may have been enough disturbance to the fluid to cause it to become turbulent and increase heat transfer.
This is of course hypothetical.

hot_rod

Hatterasguy said:

hot rod said:

How does this not prove that BTU do jump off faster trains?

I think it is a question of semantics. The BTU's don't "jump" anywhere. The hotter the tubing or the pipe, the more energy is released. You can use a higher SWT or you can pump faster. Your choice.


What happens if you increase the SWT and keep the train speed constant? More BTU's are "jumping off" the train although the speed of the train has not changed. It's not a good analogy.

But, If it helps you to think in terms of "jumping off", have at it.

I thought the BTU train analogy was something Dan started years ago? I'll blame or credit him for the to explanation Other trainers and industry folks remember that origin also.

I thought it helped visualize the circuit concept. Dan used a semi truck jackknifed in Primary Secondary Made Easy to demonstrate diverter tee concepts, I like a bit of a story line to technical topics.
Sorry if it confuses some.

I'll stick to my point, on just varying flow rates shown in the above examples that faster flows, all things being equal move more BTUs/hr.

Gordy

Hatterasguy said:

hot rod said:

How does this not prove that BTU do jump off faster trains?

I think it is a question of semantics. The BTU's don't "jump" anywhere. The hotter the tubing or the pipe, the more energy is released. You can use a higher SWT or you can pump faster. Your choice.


What happens if you increase the SWT and keep the train speed constant? More BTU's are "jumping off" the train although the speed of the train has not changed. It's not a good analogy.

But, If it helps you to think in terms of "jumping off", have at it.

I think we all understand what happens. It's just an anology that's been used for years. Maybe one that does not suit some though.......

Rich_49

Everybody's time is valuable . Here are graphs , numbers , and opinions that were presented by Iapmo and Belimo . I have posted this before and this conversation has prompted me to post it again . Please take the time to watch it if you can in it's entirety , if you don't have that kind of time maybe the proof is contained from the 30 minute mark throughout the to the end . PRETTY ILLUMINATING

https://www.youtube.com/watch?v=8re7q1Ejg1U&feature=youtu.be

hot_rod

Yep, he agrees with my points show throughout my above posts.

While not exactly identfiying the term, he states "overflowing" a coil will increase performance but may not be worth the additional pumping power to get the additional few % of output.

Agreed, and in the simulation I show we use a high head circ with a balancing valve, maybe Belimo, to show that exact relationship.

Run wide open to show output , choke down the balance to show the lower output a wider delta and how the pump head requirement changes throughout the balance range.

That is why designers use simulation, run all the what ifs. The Belimo webinar this week explained the "patches" they have added to the DOE design simulations software??

This flow vs pumping requirement s shown clearly in the examples I attached pumping 1.1 as opposed to .65 gives addition output but drives up pumping power to accomplish it.

Fairly easy to understand weather you look at the examples I posted or listen to his explanation.

Nobody should "overflow" any emitter as he and most engineers and smart consultants would agree, just find the best design point for coil or emitter balanced against the power required to get you there.

We see this all the time in radiant design with loop lengths. Some installers feel the extra pump power to run 300- 350' loop, or the online hucksters promoting 500 foot lengths! of 1/2" is worth it, others stay at 250 or below to keep the pumping power requirements as low as possible.

His analogy of "linger" time of the molecules in the coil is a bit fuzzy. Keeping molecules in the coil with no energy to transfer does little to increase heat energy transfer, move them out and bring in the higher temperature versions that have heat to give. Again keeping an eye on additional pumping power.

That is whey we look for the sweet spot of flow vs electricity to maintain that.

Keep in mind he is working in commercial applications where they may be pumping through dozens or 100's of coils. There are buildings in Vegas with fan coils in a thousand rooms. We have a rep that has access and security passes to those massive mechanical systems and rooms, some wild stories he has to tell. Start talking 24" diameter piping and pumps get scary big. Probably don't want to run that GPM through a single ZV That could be defined as "overflowing" a coil.

Pump energy, piping size, all that is part of the picture. It becomes a bigger issue in commercial than on a residential system with a 27W circulator running the entire distribution.

He also make it clear several times his products account for and correct for "faulty assumptions at design" With proper concise design, the need to correct with flow regulating "bandaids" as Barba calls them, becomes less.

A top notch consultant spends time on accurate load calc, infiltration data, potential solar gains, owner usage patterns, system simulation, etc when designing a HVAC system. That way boilers, pumps, emitters match and work in harmony from day one instead of adding additional, expensive controls or un-necessary, expensive smart circs to correct for "faulty assumptions at design"

Delta T circulators for example are being promoted to "fix systems that were designed around "faulty assumptions" over pumped, excessive boiler size, not accounted for micro loading, low mass emitter challanges, single fixed high temperature boiler operations required by tankless coils maybe, all sorts of hydronic sins , or old world hydronic thinking of bigger is always better.

Harvey Ramer

This is all good information but it still doesn't quite address the problem we are discussing.

Fan coils are a completely different animal than fin tube. They feature short tubing lengths, lots of u-bends, narrow channels and forced convection.

hot_rod

Harvey Ramer said:

This is all good information but it still doesn't quite address the problem we are discussing.

Fan coils are a completely different animal than fin tube. They feature short tubing lengths, lots of u-bends, narrow channels and forced convection.

Fan coils are merely forced convection. There have been multiple mini tube, headered type of fin tubes tried over the years. Forcing the convection increases the output by increasing airflow. Inside you still have the water to tube wall conduction transfer required.

I guess a good example is the manufacturers of refrigeration coils. They continually look for way to increase turbulence in side the A coils or condenser coils. They twist the tubes, fin the tubes, rifle the tubes, baffle the tubes, insert cork screws inside, roughen the inside surface, anything to increase turbulence and improve conductive transfer. Cost to benefit drives that technology, just like pumped flow and energy cost.

Going back to your original question and discussion with the instructor, it certainly sounds like a heat transfer, fluid dynamics question or puzzle.

Without some, any, data on what he was pumping, through what, what conditions changes, any metering of flow or ∆ change data gathered, etc I don't see how we can come up with theory or answers.

Any way to chat with his son that has the Phd? It sounds like he may already have the explanation your seeking?

jumper

Pumped water goes the easiest way it can. Maybe over pumping causes the water to separate into lower velocity laminar at surface and faster stream through interior?

hot_rod

jumper said:

Pumped water goes the easiest way it can. Maybe over pumping causes the water to separate into lower velocity laminar at surface and faster stream through interior?

Could be, but what is "over-pumping"? In 3/4 CTS for example what velocity causes this observed but undefined decrease in heat transfer. I don't see it in the 2-5 fps range that we consider normal hydronic velocities.

When I speed up my clear tube display to the highest pump speed the water turns cloudy and I cannot see the condition of the stream any more. A 2-4 fps you can clearly see multiple corkscrew type streams or mini tornado looking patterns. The change of speed seems to lengthen or shorten those patterns.

I have a video guy coming by next weekend we are going to try and film the display, he has some ideas for lighting and food coloring to try and film those patterns inside.

Maybe it cavitates and has millions of tiny vapor pockets that reduce conduction transfer, at over pumping conditions? Would it still be considered turbulent by Reynolds number? Maybe turbulent but aerated?

Gordy

Keyword boundary layer......

Rich_49

Could be, but what is "over-pumping"? In 3/4 CTS for example what velocity causes this observed but undefined decrease in heat transfer. I don't see it in the 2-5 fps range that we consider normal hydronic velocities.

When I speed up my clear tube display to the highest pump speed the water turns cloudy and I cannot see the condition of the stream any more. A 2-4 fps you can clearly see multiple corkscrew type streams or mini tornado looking patterns. The change of speed seems to lengthen or shorten those patterns.

I have a video guy coming by next weekend we are going to try and film the display, he has some ideas for lighting and food coloring to try and film those patterns inside.

Maybe it cavitates and has millions of tiny vapor pockets that reduce conduction transfer, at over pumping conditions? Would it still be considered turbulent by Reynolds number? Maybe turbulent but aerated?




Yes , it is still a high Reynolds number . Reynolds numbers can approach or be infinite , at least that was theory presented in several papers , many of which I posted that you probably did not bother to read . Along with MANY others it seems .

Gordy

What I'm seeing here by some is trying to compare accepted design practices in a given range to an over, or under pumping scenerio that falls completely outside design parameters.

Extrapolation of what works well does not always paint the picture beyond its parameters.

I believe Harvey is looking for what happens beyond accepted parameters. Linking that to what problems are actually scene in the field. An over pumped system. Happens all the time with unintended consequences. It's the unintended consequences that creates more questions.

Harvey Ramer

@hot rod
If the water in your clear display turns cloudy at high velocities, that is a clear indication of microbubble formation. I think we all can agree that, that would severely hamper heat transfer. The question becomes, is your pump being exposed to NPSH and would the pump in Ray's project have been exposed to the same condition?

Back to the boiler to look at things from a different angle.

It should be noted that all boilers would not have exactly the same reaction to flow and inlet temp.

Lets start off by examining a large bodied HX such as a cast iron sectional block or perhaps even a firetube. It is not reasonable to expect any flow volume experienced in a practical hydronic application to provide any kind of velocity scrubbing action on the walls of the HX. So lets look at the difference between a boiler with an inlet temp 2° below supply temp and a boiler with an inlet temp 20° below supply temp. Lets consider those temperatures to remain constant and also consider the energy input to the fireside of the HX to remain constant. The boiler with the high inlet temp must have a lower heat transfer coefficient for a couple reasons. Heat flux is greatly affected by temperature difference and also the specific heat (density) of the lower temperature (fluid in this case) that it is being transferred to. Since the fluid velocity in this type of HX can not be expected to scrub the walls of the HX it is reasonable to predict an increase in temperature of the HX walls. This increase in temp of the HX walls will increase the thickness of the adjoining layer of heat-saturated water leading to an increase of microbubble population and increase in bubble size. The thickened layer of microbubbles would also serve as an insulating blanket and further slow down the heat transfer coefficient from the HX walls to the water.

With the entire system running at only a 2 degree temp drop, the reabsorption of the microbubbles will likely be slower than the production. Also at high velocities, any meaningful air separation by currently available mechanical devices, is nonexistent.

Under this proposed scenario, it could be possible for both the boiler and the emitters to drop in output at high flow volumes.

Rich_49

%93Stokes_existence_and_smoothness

hot rod said:

Harvey Ramer said:

This is all good information but it still doesn't quite address the problem we are discussing.

Fan coils are a completely different animal than fin tube. They feature short tubing lengths, lots of u-bends, narrow channels and forced convection.

Bob.
Fan coils are merely forced convection. There have been multiple mini tube, headered type of fin tubes tried over the years. Forcing the convection increases the output by increasing airflow. Inside you still have the water to tube wall conduction transfer required.


Rich .
I agree . The only difference is increased airflow over the active part of the emitter

Bob .
I guess a good example is the manufacturers of refrigeration coils. They continually look for way to increase turbulence in side the A coils or condenser coils. They twist the tubes, fin the tubes, rifle the tubes, baffle the tubes, insert cork screws inside, roughen the inside surface, anything to increase turbulence and improve conductive transfer. Cost to benefit drives that technology, just like pumped flow and energy cost.

Question from Rich .
Could this actually be an attempt to break up the possibility of boundary layers forming and/or making a smooth wall surface no longer smooth ?

Bob .
Going back to your original question and discussion with the instructor, it certainly sounds like a heat transfer, fluid dynamics question or puzzle.

Rich.
I really think we are discussing one of the unsolved problems of condensed matter physics . Turbulence .

Bob.
Without some, any, data on what he was pumping, through what, what conditions changes, any metering of flow or ∆ change data gathered, etc I don't see how we can come up with theory or answers.

Rich.
I think we may have seen a dataset within the Belimo presentation I posted . The output clearly and unarguably began a downward trend at a certain point with a greater flow . Maybe we could look at that and maybe we could actually get more data if available from Dave Kandel . Bob , do you know how to get him ?

Bob.
Any way to chat with his son that has the Phd? It sounds like he may already have the explanation your seeking?

Rich
I don't think he has an answer as much as trying to solve a known problem that nobody has figured out to date . Reason it remains unsolved . He knows a problem exists and work continues .


Theory ;
There is still an issue that we deal with on a daily basis that was , is and will continue to be a mystery . Open minds should take part in discussions and continue to evolve the technology and understanding without agendas getting in the way in the process . Maybe some industry players could share lab findings with schools to further understanding . Something that has remained a mystery for hundreds of years surely could use that attention and cooperation .

hot_rod

As for Rays project, we don't have much, or any really, info on what it was and how it was built? If it was a closed loop, under pressure with a small wet rotor circ for example, a few psi at the pump would be enough to prevent cavitation, or flashing to steam if it is running at high temperatures.

I know Grundfos has numbers for that NPSH in their tech literature, something like 4 psi at 190F in a 15-58 or Alpha sized circulator. Don't quote me on these number, I think it shows 1.3 psi at 60F and required pressure goes up as temperature increases. Maybe 15 psi suggested at 230F fluid temperatures?

If it was a closed loop, was a micro bubbler used to assure it started completely air free? What was the piping and flow rate.

If it was an open, unpressurized system and loop , operating at near boiling temperatures, you are certainly walking a fine line.

If the pump was in a cavitation condition, filling the piping and heat exchanger with vapor pockets that would effect heat transfer and could be mistaken for over pumping and flow going back to laminar, if that is in fact possible, with the pump he was using.. A flow meter would help identify what was going on.

I'm still of the opinion the faster the flow them more heat energy is transferred, up to the point of "flooding" or over pumping. The manufacturer of the product should have that identified in their tech literature.

We know the older high mass high water cast boilers could operate with zero flow and not have over heat problems. I doubt any of the small water tube Sermeta HX can operate under those conditions. As I see it the Loch fire tube states about the lowest minimum flow rate I have seen in the current batch of fire tube HX. You are right in look tab the particular product and what it was designed to do, and limitations of it's design.

Listening to the Belimo video that Rich posted they talk about over pumping conditions and at the point where the juice is no longer worth the squeeze. The small % or fractional % in heat transfer is not worth the pumping power to chase.

The Smart valve they talk about at the end is basically a ∆T limiting device, among other functions. His explanation sounds a bit like marketing was heavily involved, something like "the valve never makes a coil behave better or more efficient then at design condition. It just forces it to behave at design."
A PI or PICCV can accomplish that same function.

It is a very clever valve that brings energy monitoring to the table with flow and temperature montoring and adjusting capability.. At about $9000.00 a pop list cost it may border on that point of the controls not exceeding the cost of the rest of the system concern. A PICCV or E-PICCV that they also manufacture and show in the presentation will get the same balance result for a lot less $$. It may be a solution looking for a problem to correct?

When the energy metering standard is finally in place it, and the valve is certified it can be used to "bill" for energy. That could be a big part of the designers intent. Now you have a bunch of cash registers scattered around the building or campus.

In one of his examples he mentioned a 36 gpm coil and pumping it at 50 gpm flows! That sounds like a fairly substantial over pumping condition. A properly designed and balanced system should generally not be exposed to that large of an over pumped condition. I would suspect a building with those type of flow requirements would have a VFD also to tame the system.

SWEI

hot rod said:

A PI or PICCV can accomplish that same function.

It is a very clever valve that brings energy monitoring to the table with flow and temperature monitoring and adjusting capability.. At about $9000.00 a pop list cost it may border on that point of the controls not exceeding the cost of the rest of the system concern. A PICCV or E-PICCV that they also manufacture and show in the presentation will get the same balance result for a lot less $$.

Those PICCV's are nice valves to be sure, but I'll save them for retrofit problem solving. On a new job, we can accomplish the same thing using CCV's and a ΔP circ with reverse return piping (and save 60% on valve costs in the process.)

jumper

What is over pumping?

Generally heat exchanger manufacturers allow a maximum flow of about three times minimum. Below minimum heat transfer stinks due to laminar flow. Above maximum there's the danger of erosion. So over pumping is over max. Don't do it.

Gordy

https://youtu.be/WG-YCpAGgQQ

Gordy

Pay close attention to the air bubbles at the tube wall, and the center of the tube in the non corrugated tube, and corrugated.

There is no flow rate induced just air released in the vertical column tube.

bob_46

I think we should be very careful when we talk about bubbles. Bubbles can occur at different places in the system for different reasons. When we use the term bubble we should be specific as to what kind of bubble. What is in the bubble air or steam, what caused the bubble, consolidation of air, nucleate boiling etc.

Gordy

In this case it is showing the boundary layer.

Gordy

However any bubbles would reduce heat transfer.

bob_46

Depends. Nucleate boiling on a heat transfer surface can increase heat transfer.( Steam it's Generation and Use Babcock&Wilcox)

Paul48

..........

Harvey Ramer

Bob
Is that with the assumption that the water velocity is sufficient to scrape the bubbles off the HX or break the surface tension popping the bubbles?

I would assume there would be both Nucleate boiling and dissolved gas being released simialtaineously due to heat saturation.

bob_46

When I taught refrigeration I always taught that you can't superheat in the presence of a liquid. A close friend bought a new water cooled BMW bike and was describing how superheated micro bubbles of steam were responsible for it's highly efficient cooling system. I bet him that that wasn't the case because you can't superheat in the presence of a liquid. Well my friend is a prof. at MIT and USC and a senior fellow at Boeing. Guess who won. I did a bunch of reading and this is a really complicated subject but ASHRAE B&W etc agree the faster the flow the better the heat transfer.

Gordy

Interesting.

If we use the universal hydronic formula it paints a different picture. One that shows increasing the delta has far more btu transfer capability than increasing the gpm. So how do we increase a delta?

1 delta X 10 gpmx 500= 5000 btu
2 delta X 10 gpmX500= 10000 btu

1 delta X 11 gpmx500=5500
2 delta X 11 gpmx500= 11000 btu

Gordy

The point is what takes more energy to gain more output? Slowing the pump down, or speeding it up?

A 1 degree step in delta is a very small change. A 20 gpm flow from 10 is a huge change to get the same output.

Gordy

It totally contradicts that more flow is more efficient in transferring btus.