Hot water temperature reset and VSP

zeke

Brad,
Excellent points, but do you really think laminar flow in the radiators is a big problem? i would think that if you look at the heat transfer across the radiator crossection, the dominant series thermal resistance is the film between the radiator surface and the room, with say a conductance of about 1.5. while the water to inside surface coefficient would be an order of magnitude higher.
I think the idea of constant differential temp is very good and with proper control of the ECM , you may not have to worry about laminar flow, just control delta T with motor speed.
I have no field experience on these systems, so please correct any of my misobservations.

Been Kuo

Hot water temperature reset and VSP

Knowing that resetting hot water temperature will save energy and slowing down pump at low heating load will save energy too, how will a system designer take advantage of both strategies? Shall the temperature reset be on the boiler loop supply or return or on the building loop or both? How can the temperature reset and the pump speed be controlled in a closed loop, meaning with a temperature reset the pump operates at the minimum speed too?

Brad White

I am currently writing a white paper

(come to think of it, all my papers are "white papers" on a similar tangent of the subject and inspired by John Siegenthaler. Actually it was a simplified control method to maintain the wire-to-water efficiency and BTU delivered per Watt as constantly as possible across the operating range.

What I am writing is for more commercial and institutional work (germane to my normal job) but applies as does any principle and this is a synopsis.

The method I would use (excepting and accepting the limitations thereof) would be "Constant Delta-T", a "stock" control strategy available across a number of circulator manufacturers. This is as opposed to the more commonly applied "constant differential pressure" method of variable speed control. After all, pressure is great but it really is a Delta-T we seek in the end. Why not capture that as the primary objective?

All it takes is a variable speed circulator (or on more commercial pumps, a variable frequency drive) with supply and return water temperature sensors.

For example, a 100 MBH/10 GPM load at peak will have a 20 degree delta-T. Stipulated is that the load is based on 70F indoors and 0F outdoors for discussion.

Say it is 180-160 with an average water temperature (AWT) of 170F. at peak load (0F outdoors)

At 35 F. outdoors, the load is 50 MBH (halfway from 0F to 70F and 0 to 100 MBH).

With the same 10 GPM as a constant value, the Delta-T drops to 10 degrees from 20, (50,000/10/8.33/60)=10 degrees.

You are moving twice the water you need to.

With outdoor reset, your average water temperature would be 120F, presumably though, your HWS would enter at 125 and leave at 115.

To illustrate this further, as the ODT rises, Room T. and Water T. begin to intersect, you are moving the same water volume for a rapidly diminishing effect. Your Delivered BTU's per Watt would soar as your Delta-T narrows. Why not lower flow accordingly?

Using a constant Delta-T setup (I just bought a Taco 007VS for this purpose) would allow me to reduce flow in proportion to load and maximize the extraction per unit of fluid circulated.

The downsides: At very low flows, laminar flow may kick in, especially with larger CI radiators, less so with small-tube fin-tube. With CI radiators, it may require an accepted lower Delta-T at very low loads periods. The good news is, the cube-root effect of Watts required to flow and head. A small flow change begets an exponential (^1.75) head reduction and a ^3 power reduction.

Such a deal.

This may not respond directly to your question, but it was a collection of thoughts which I thought to share with my associates at work. How might it apply to yours?

Mike T., Swampeast MO

Constant Delta-T

Before I installed my Vitodens, I did my best to calculate a heating curve that would do just that--result in a nearly constant 30F delta-t.

If you neglect the plus/minus 10F bounce around the setpoint, my old W-M (mechanical reset, TRVs, constant circulation) system came pretty close to this in all but the mildest weather. It was all about the reset curve and bypass via the differential pressure valve. Just the right amount of bypass really seemed to help keep delta-t quite stable.

Vitodens did not however act ANYTHING like that... (Of course I still have very little idea how it modulates its circulator since it has no sensors either for delta-t or delta-p.)

In periods of very low load, delta-t drops to an extremely low value--3F-4F is common but again you have to neglect the "bounce" around setpoint as the Vitodens is "pulsing" (e.g. producing heat in little batches).

Once the boiler is barely and truly modulating (supply/return temps won't change more than 0.1F within 5 minutes for at least two hours), delta-t is typically 10F. As system load and burner output increase, so does delta-t. Highest I've seen sustained (and with TRVs unchanged for days), is about 40F when it was about 0F outside for a few hours.

I [believe] this increasing delta-t with system load allows a condensing boiler to have the maximum condensation and heat recovery from the condensate.

Mike D_7

vs circulators

Brad, I tried the same thing last year. I've got a Teckmar controled primary/secondary HWBB system I put in my house.
I first ran it with a fixed speed pump on the secondary side and found Delta T's as low as 8-10 degrees at lower temperatures.
I then bought a Taco VS circulator and set it for a 20 deg Delta T. It didn't work! Because I have night time setback, the system water can get very cold. I'm talking near freezing in the coldest weather.
The cold water came back and spazed out(a technical term) the VS pump control. The pump would slow to it's lowest speed and stay there because of the huge temp difference. At the pump's lowest speed it couldn't move water through the system fast enough to significantly raise the temp!
I spoke to engineering at Taco and they sent me a revised pump control with a delay in speed response when first running, but it still didn't work. I tried every possible dip switch setting including reverse response, but the thing never worked out. I gave up and re-installed the original fixed speed pump.
I'm not knocking Taco, they gave it their best shot and were very helpful. It just didn't work for reheating after a night setback in cold weather. F.Y.I., in mild weather the revised control worked fine.
Also, I asked Taco about energy savings. I was told that with the PSC motor they use, the energy saving at lower speeds/loads was negligable.

Mike T., Swampeast MO

Great point Mike D!

I'm so happy with my barely adequate heating curve that I've nearly forgotten about daily setback.

My first season with the Vitodens used a quite generous curve. I also conducted numerous setback recovery tests.

If I increased the setting of TRV(s) considerably, delta-t would drop almost immediately and considerable, but since the boiler was still operating on the same heating curve, only the return temperature increased--the supply temperature would only show a short drop until the burner ramped up--but once back to normal, the supply temp stayed perfectly steady.

As the room(s) recovered, delta-t gradually increased. Once fully recovered, delta-t was again stable but lower than before as overall system load had now increased.

If you modulate strictly via delta-t you have a fixed amount of energy available to the system unless you increase the supply temperature. The VS circulator really needs to know what is happening both in the boiler and in the space. TRVs will provide an indirect feedback but it still won't help unless the circulator also knows what the boiler is doing...

My system has now operated with a barely adequate heating curve for slightly over two heating seasons. Setback recovery is torturously slow. Even increasing the sun dial setting significantly does very little to speed recovery. It's really as if the boiler knows it is operating the entire system with maximum efficiency and virtually refuses to change its ideas. The ONLY way I can get reasonably fast setback recovery now is to press the "party" button.

Mike D_7

Mike T, my system recovers OK because the Tekmar control has an Optimum Start/Stop mode that boosts the water temperature when recovering from a night time setback.
I suspect that this feature agrivated the problem when the Taco VS control "saw" 35 deg water returning and 160 water going in, for a Delta T of around 125 deg!
I would have thought the control would have ramped up to full speed, but it SLOWED DOWN to minimum speed.
I watched the process, viewing the temperature guages I have in the system plus an infared heat sensing gun.
At first, a big slug of cold water came arround in the secondary loop, with hot water being injected by the mixing pump. The supply sensor first saw 160 degree water, but then saw the supply side plunge when the cold slug mixed in and that must have made the pump speed drop since the control was reading a sudden DROP in Delta T!
I signed up for John Seigenthaler's on line seminar about variable speed pump. I'm hoping there are some improvement in the controls (such as ignoring Delta T for the first 5-10 minutes). I also hope some of the newer pumps save more electricity at part load.

ALH_4

VS Pump Control

Brad, with a VS pump providing constant dP and TRV's (or even zone valves to a lesser extent) controlling radiators, wouldn't the pump speed decrease to provide the same dP as the valves close? I'm pretty sure it would not be as efficient as dT control, but I think it might behave in somewhat the same way?

Brad White

That is interesting, Mike D!

Had not thought of that and the behavior makes some sense as you describe it, yet there is a comprehension gap I mention below. (I am not in favor of such deep setbacks and that is another issue entirely but I can see how that "drags" the operation down.)

This is all open discussion so I am very glad to read your experience. None of this is "settled" by any means, just something I am putting out there.

Here is my comprehension gap: I am curious why the circulator did not speed up as the delta-T was so far out of spec (as if it should think it was running far too slowly already) and did not speed up to narrow the gap. That is what I would expect.

The pump savings is disappointing if as you say it is negligible. But if it really is that negligible then the pump laws are wrong. There has to be at least some appreciable savings. Granted, my ideal situation is an ECM motor.

I have run VS circulators based on constant Delta-T and it worked fine.

What I am coming up with in my white paper research is from a graph I made of constant circulation with variable delta-T. The average water temperature line crosses to a point of intersection at "0-0" (say 68 degrees common temperature of water, space and outdoors, less offset of internal gains). What is becoming apparent is that keeping a narrow Delta-T is not a bad thing, if forces a rise in AWT the warmer it gets outside. This in turn lessens that "shift" where one feels chilled (or at best just below neutral) even though the system is at temperature in mild weather.

Brad White

Laminar flow

Hi Zeke-

No misobservations that I can tell- heck, if you saw it with your own eyes, who am I to say otherwise?

As far as laminar flow goes, I do not see it as a show-stopper. Yes, in CI you have the retention time but with such low velocities, what is to keep the core of warmer water from passing through as much as mingling with colder water near the outside of the radiator? There will be transfer of course, this is not insulation. But there also has to be less of a motive force than when the inside of the radiator is turbulent and breaking up that party on the outside wall.

The difference is probably best measured as "BTU per Watt" or "BTU per GPM" as a delivery model. Not that you are "not delivering heat" so much as doing so less efficiently per unit of flow generated.

I picture eddy currents, little micro-thermal currents like solar flares in reverse, as warmer water moves to displace spent water along the radiator walls. A thermography video in a glass radiator would make me very happy. (Spoken like some kind of engineer type...) Maybe a CFD model (Computational Fluid Dynamics) would make a nice project, but I cannot afford $20K software at the moment...

In my thinking, I am settling on a bottom limit on constant delta-T to avoid or at least stay above any pretense of laminar flow.

Mike T., Swampeast MO

Here's a great demo of Laminar Flow. Not really germane to the topic, but pretty cool. Make sure you watch the entire video.

I can't say for certain, but know there are times when flow through my iron rads is extremely low with quite high delta-t (across the TRVd rad) and they've always heated fully and evenly. Only time I [might] have observed a problem is with the old boiler and B&G 100. If I opened a TRV fully, would almost swear that it took longer for the rad to heat fully than with the relatively tiny circulator built into the Vitodens. My only explanation is that higher velocities a good chunk of the supply water was passing directly across the bottom-bottom connections to the return without having enough time inside the rad to distribute itself evenly.

jp_2

no fault of Taco

the pump was doing what it was designed to. you were just trying to make it do what it incapable of doing. to deal with excessive delta T it would have to know "why", need more info than just delta T at that particular point, ODT, indoor temp etc..

another words, it would need more intelligence.

Brad White

That was part of what I am exploring, Andrew

For years, dP was "the way" and the delta-T became an incidental that followed, theoretically, but was not specifically "controlled to". (Excuse Preposition).

In the end, there may not be as much difference as I make it out to be, but given that we are all in the temperature business, the notion of controlling flow to an incidental (and necessary) pressure differential) seems a tad oblique to me.

By controlling to a given delta-T, it simply maintains the ratio of delivered BTU's to flow generated.

With TRV's closing, I can see the delta-T widening (less flow) hence the circulator speed would also back down because it's delta-T needs are met. So, I see your point! But the respondent variable is temperature.

I am thinking a "greater need of" control, such as DP or DT, just to make sure that my furthest emitter has adequate flow. Maybe this will dictate a certain "floor" limit for circulator RPM.

As a practical matter, if my furthest emitter is calling for full flow and the one's closest to the circulator are also, the DP to my furthest emitter will be very low (less flow in all of those mains even over that distance) that it should work out fine.

Mike D_7

night setback and Delta T's

Brad, my night setback is 6 degrees F, from 70 degerees down to 64 degrees. What do you recomend?
The pump would initally speed up, the cold slug of water passes the point where water is injected from the primary side, and now the SUPPLY sensor is sensing a mix if warm injected water and a larger volume of very cold water and slows down. Then, since the return water is losing heat faster than the suppy side is replacing it, it stays slow. It drove me crazy, and I couldn't visualize what was happening. You had to see it to believe it. Everything was double and tripple checked, I tried every combination of setting I could think of, and finally gave up. The spirit of experimentation always yeilds to a wife who's tired of waking up to a cold house.

The house in question is a 70's post and beam (Deck House) raised ranch if about 2,500 sq ft. The upper level is cantilevered over the lower level like an upside down wedding cake. It projects out about 24" with a solid wood (1/2"plywood over 3x6's) floor with no insulation.
Also the supply and return piping run under the lower level's concrete floor. System has a high head, I've never seen so many el in a system, 1" copper splits to feed the main floor with 2 3/4" loops that are too long, etc.

The house is weather tight and has a well insulated roof, so the temp drops off at night and then the system would normally be cycling on and off during the night in cold weather. With the VS pump control this didn't happen because of the previously described problem, so we woke up to a cold house.

As to the electrical efficency of the motor at lower loads/speeds, I'm just repeating what I was told by Taco. They said it was becaue the controler they used dropped in efficency as the speed went down. I would also have expected a noticable drop in power consumption at lower speeds, but my knowledge of electric motors is limited.

The whole concept of minimizing electric consumption while improving comfort and fuel efficency is really interesting. I expect much more to be done on these fronts.
I'm laying out a heating system update for my son in law that will be using a buffer tank to (among other things)seperate primary and secondary sustems. Also using big pipes, big check valves, etc to minimize head so we can use the smallest possible pumps.

jp_2

radiators

boy, I would think a cast iron radiator would be full of eddy current making formations? but, I've never been inside of one:)

jp_2

neat!

i wonder what the temperature of the water and syrup were? I'd guess quite cold.

ALH_4

dT and Proportional Flow

It seems to me that as TRV's close, the pump could ramp up to try to keep the dT tighter while the TRV's are making it wider. Would they fight one another? Maybe this isn't applicable to TRV's? I have to admit, without spending some time drawing this out I can think myself in circles.

Brad White

Ed. E. Current

Actually, J-P, you picked up on exactly what I was getting at. The only eddy currents would be the hot water making inroads by lateral convection (not to mention vertical convection) as warm moves toward cold.

With say 0.15 GPM in a radiator with passages the equivalent of 2-inch iron pipe (and many such passages in parallel, not series) you can see how the dominant motive force can hardly be the circulator. It has to be something. A thermal bumblebee!

Mike T., Swampeast MO

Just Throwing Out Ideas, Brad

I see another potential pitfall with simple delta-t modulation of a circulator:

What happens when your desired delta-t is impossible due to low load and/or greatly oversized emitters? If the difference between the your current reset target and the minimum possible return temp (ambient or very near ambient) is less than your target delta-t, won't the circulator just keep trying to get slower and slower? Go too slow and flow balance through the emitters may be destroyed--especially without TRVs. Even with TRVs, if the circulator goes slow enough that all of the head is eaten up by the piping itself, the TRVs won't have sufficient head to operate properly.

There's a potentially serious problem if the VS circulator is in the primary loop or if this is a direct-drive, single circulator system. Let the flow get too low in an attempt to maintain some level of delta-t and the boiler temp may rise high enough that the burner shuts down--or in extreme conditions kicks out on high-limit. The high-limit thing might sound ridiculous, but what about mod-cons that begin firing at full output and then gradually ramp down?

Seems to me that in any of these cases, some minimum circulation speed will have to be established and that such will be HIGHLY related to every other element of any given system.

Brad White

You know

there is a circular logic here, Andrew. As usual, you get me thinking.

I am on vacation this week (ha!) so have some time to think of such things... let me work on this. I can see it both ways, not to say either of the ways I am thinking of this are right....

Mike T., Swampeast MO

Did they say it was water? It looked awfully viscious. Perhaps glycerine?

Brad White

You raise good points as always, Mike

I would not substitute VS for a primary (boiler side) circulator unless it's control was integrated with burner output (much as we suspect or Vitodens are). As you say, you run the risk of shut-down on high limit.

Rather, I am thinking of VS piping in the realm of the secondary (radiation side) of things.

Naturally if P/S piping, this raises the issue that the boiler side flow rate had better be less if a Mod Con so as not to short-circuit back to the boiler and lose efficiency. But on a purely "radiation side" equation, I do not see much down-side, provided that the furthest emitter's needs are met. (Reverse Return piping, as I have in my own house, may be the perfect test-bed for this.

Direct return can work especially with TRV's but only if flow is constant and the emitters are proportional to heat loss. Each TRV will be the gatekeeper as the system ramps up and down. Those which are satisfied will defer excess flow to those less fortunate, shall we say.

As you your point that as flow ramps down and all of the head eaten up by piping- if flow goes that low, the head goes down practically by the square, the head becomes negligible. If a given radiator needs flow, it's TRV will reduce the resistance and open up- just as it should.

Thanks for the brain tweaker! Not the last word by any means.

Back to my attic- sealing for a later blower door test here in rainy Boston.

Brad

jp_2

good point

i assumed water at first then was really puzzled when it nearly came back to the original state, so I guess they were incompatible fluids.

zeke

My guess as to the failure of the system to respond to nighttime setback is that the reset curve which is based only on outside temp is flawed since it should depend on the load and that load is a function of the DIFFERENCE BETWEEN THE OUTSIDE TEMP AND THE ROOM TEMP. If that were the case, on morning startup, the boiler would immediately be on high fire and the circ pump be on high until the room temp approaches the setpoint giving maximum response, and then you initiate control.

zeke

These systems are ( or should be ) designed to respond to the vast variety of load changes, and before they are implemented , software analytical solutions using proportional and integral control make sure the "ramping up" response of the pump is sufficient to reach equilibrium without substantial temperature overshoots. Modulation of the TRV's in normal operation should not pose any significant problem to the control. It has problems only when loads are abruptly changed, like after overnight setback when in the interest of speed of response you may get some initial overhoot. So the system bandwidth should be wide enough to accommodete most load variations but not too wide to cause excessive overshooting for most load variations.

Weezbo

i am certain that timing is everything....

provided that you have enough emitter and heat source.

the heat source would have to be taken into consideration as a condensing system source will be able to see far lower temps and still "survive". Outdoor weather information that could be channeled into your setback formula would allow the best operation of the source, to me , off at X a.m. on at Y A.M. doesn't work quite right... modulation of the boilers circ would have to be something that was lashed up to the burners modulation or step fire or minimum setpoint requirement.(as in partial differential set back). Your question also asks how to control these two aspects within the recovery of room setback... well, exercising specific zones,... for now i will call that a micro zone pulse injection within the field, will increase recovery* i call that (*)," Quicker Ramp Time", within specific areas.

This probably sounds dumb buh the closer you live to the equator the less important any of this is,...the variables change as it were to sand storms ,wind speed ,topographical surroundings , man made shading, and bodies of water.

Today it is very chilly here about 10 below zero F, which makes me consider the fluid that you are using to control the distribution, here is something that has always been a bit of a stickler for me....everything is wonderful when it "Goes" should you have burner difficulties or need replacement parts.... a designer needs to consider Protection of the system and that includes the field. so, depending on where the system is located on Terra firma, certain contingencies need to at least be considered, some low temp over ride and signal (communication) might be a Good Thing.

Waiting till the return temps are 14F some 28degree corner of the building with sensors on the supply header or return header mangles the mind(bit of a stickler). what i mean by that is you really wouldn't want the system to shut down the flow in the fields or boiler or any bridging or bypasses, shunt pumps that back off flow from the field might not be my first choice in controls.

i have some faith that boiler manufacturers will see the benefit of condensing equipment that was designed with variable flow the century before last, someday, and incorporate it into their designs of new boilers.

that way, every last brass farthing will be squeezed out of the combustion side efficiency. As a Thought....

Mike T., Swampeast MO

Complicated subject with lots of dynamics involved, but I still see a "running out of head" problem even if head losses in the pipe are dropping with reduced flow.

In the straight open (non-TRVd) system, you want to keep circulator RPM high enough to get enough flow to ensure the farthest rads have flow. Right? You also see it possible that circulator RPM could go so low in an attempt to maintain some set delta-t that you don't get flow to the farthest rads. Doesn't this mean that the nearest rads have "eaten up" the available head with none left for circulation for those towards the end? While it won't happen immediately, the same thing could occur with a TRVd system as more and more TRVs open wider and wider with the nearest "eating" the available head.

As I understand "proper" European design where TRVs are the norm, it is all about delta-p. I understand that my own observations are highly limited and confined to a system where delta-p only occurs in the boiler's HX and across the TRVs, (e.g. zero delta-p in the piping and rads), but I'm fairly confident in saying that a reset curve with reasonably proper slope results in a quite stable delta-p with increased shift of the curve merely resulting in a fairly constant upwards shift in delta-p.

I have a really hard time anymore thinking of residential and commercial systems without a condensing and modulating heat source. The dead men's systems did this and even if they did have poor combustion efficiency and often required careful tending they did at least maintain a relatively high system efficiency via their inherently proportional nature.

Mike D_7

Re: no fault of Taco

J-p I agree 100%. The pump just didn't have the programing to deal with what was going on.
No way do I fault Taco.
I put this info out because others may run into the same problem.
Both Taco and I thought my system aplication was just what this VS Delta T control was made for. I had checked with their tech dept before trying it out.
The VS pump is a fine product, but it can be tricked under the right system conditions.
I still believe this is the way to go. I just need to figure out how to apply the concept.

Brad White

Absorbing Head Losses

One thing to keep in mind is the laws of balancing. When a system is balanced or at least has a static operating point (RPM, flow, pressure, HP) whether air or water, you have a starting point regardless if you know it or not

Say you have such a system at a certain operating point (which we assume allows adequate flow to all emitters, air registers, what have you).

When you change the RPM or throttle a damper or valve in the main (common to all branches and terminals), all terminals will change proportionately with insignificant differences.

It is in this way that air-side VAV systems operate or for that matter, any variable flow system.

When these systems have TRV's (analogy: VAV box in an air system), these take care of local needs. Any back-pressure and excess flow should one device close is distributed throughout the system. Remember, we already established that the system was capable of reaching the furthest termination.

What we have to do is to ensure that this furthest termination also has sufficient flow at the lowest operating point. I have yet to see this not occur, but then again, I do dictate the low flow point

Mike T., Swampeast MO

I do dictate the low flow point

Yet you know nothing of what the heat provider is doing and have no idea of the load on the system and you want to assume order based on a single element when the target is constantly moving.

Andrew Hagen_2

Delta-T

But wouldn't outdoor reset give the DT controller some idea of the load on the system? As the fluid gets warmer, the pump will speed up to attempt to maintain the DT.

With outdoor reset controlling the fluid temperature, TRV's/ZV's controlling the flow based on the indoor temperature, and the pump varying the flow based on the dT, it all gets fairly complicated.

If the emitters are "oversized" and the DT pump control cannot maintain the desired DT, the reset curve is set too hot. What is the ideal DT?

Weezbo

Brad, you need to take a peek at Segami's Sumingashi .*~/:)

you would find her work inspirational:)

Look for eddys everywhere

i like this work of Amy's *~/:)

Brad White

I think that is what started

much of this, at least in my mind, your first sentence especially.

The Delta-T would remain constant (by definition) regardless of what the actual temperatures were. The circulator, ideally, would ramp up and down to maintain that.

The TRV's would tailor radiator output to local conditions (compensating for insolation, internal gains, over-sized radiators relative to heat loss).

On this last item, over-sized emitters, I suppose if they all are equally over-sized (proportional to heat loss hence allowing operation at a constant, shared hot water temperature) the Delta-T "could" remain constant, at least in our starting hypothesis.

If one or more radiators were grossly over-sized out of proportion to the others, then I agree, the Delta-T would be greater for those radiators. If the radiator selections were all over the place, the system would not be self-balancing if not for the TRV's compensating as I see it. But by then, the notion of a constant temperature drop as a control strategy becomes moot.

As for the reset curve being set too hot, I graphed some hypothetical data by hand, (soon to be in a form to share, I hope).

What came of that is that the average water temperature curve in a constant flow system is naturally shallower than the supply water curve (of course- because it is cooler), but it intersected the origin point as it would.

With constant Delta-T, you have two parallel slopes but that also means that your average water temperature is lower at lower temperatures, hence less effective. Sort of a non-parallel shift at low system temperatures, by default.

This in turn got me thinking that narrowing the Delta-T (i.e. constant circulation) may be a benefit best left alone in this case. My own white paper premise started with an internal discussion of the cost of delivering heat (transport costs per BTU) by the way. What a wonderful thing this discussion group is.

Brad White

Not sure I follow, Mike

All I was saying is that when setting up an actual variable flow system for conventional "constant Delta-P" control, I have the opportunity during Testing, Adjusting and Balancing (TAB) to set a minimum speed (Hertz) at the variable frequency drive.

This is to document that the furthest emitter has adequate flow proportional to all of the others as they were at 100% speed.

I would have an idea what the heat provider is doing and the load on the system (the heating curve being established by this time).

The remaining variables do tend to move a bit (they are variables, that is what they do! but less so and with less effect on the system as a whole.

Really enjoying this discussion.

Brad White

Now that

is something, Weezbo! Reminds me of something the Hubble telescope would have taken.

Come to think of it, for the billions I helped pay for the Hubble, that these things could have been created on earth...hey, wait a minute!

mtfallsmikey

Allow me to jump back to basics for a moment.

I personally believe in the laminar flow theory in CI rads. I have seen this effect in the past, and was stumped, but these kinds of issues were never pondered on by the Dead Men. Brad, I was wondering...about this white paper you are writing; Are you going to look at the concept of water cooling as well as heating,ie:systems like that are used in my buildings using open-loop condenser water for cooling in the "economizer" mode? These are commonly used systems in commercial buildings now, and I have been toying around with this concept. Our self-contained units have separate economizer coils, the EMS has changeover,and mechanical setpoints that are adjustable, but I believe that the pumps,that are single speed, are still throwing too much water thru these economizer coils to effectively do a great deal of cooling, hence the need for VFD's on the condenser pumps. The only potential rotten fish in the barrel would be a power glitch that trips the VFD's, and stops flow, which shuts down all of my units on all floors. We use VFD's on one of my two tower fan motors now. Interesting concepts!

Brad White

Hi Mike

I am going to contain it to heating for now just to keep it simple. This particular discussion though is helping out a great deal, to exercise my thoughts and stretch into areas I had not considered. All good stuff!

HW systems also have the advantages of higher delta-T's between the space and water temperatures, versus chilled water which has narrower delta-T's between CHW and leaving air temperature. Sometimes they can cross as you know, with counterflow in deep coils.

The constant delta-T on cooling is as critical if not more so. Having a chiller system operate with a narrow delta-T kills chiller efficiency and leads to trip-outs/shut down on satisfaction of CHWR setpoint if referenced.

With CHW systems though, it is all about humidity. I would rather watch the temperature than delta-T (as a first variable). Say I have a system with a design delta-T of 12 (44F in the coil, 56F out. If that drifts up at all and I want extraction (absorption) of heat by widening my delta-T at the coil, I might run the risk of losing humidity control.

Constant delta-T in CHW systems makes sense more on the system side (where I control the chiller) but if I am more in control of my CHW coils, dang it, I will run more water through to get my average CHWT down!

The District CHW System Manager's Dilemma.

mtfallsmikey

Straying off-topic just this time

I actually do not have any chillers...these are Trane 100 ton self-contained units, one on each floor, five scroll compressors in each one. Economizer operating setpoint is detrmined / set by using outdoor WB, which corrolates to your statement about the humidity control. Bad part is that the heat load generated by the tenant's "equipment" is so great that I can only economize to about 42 degrees ODWB in one building, but the other building has less heat load, can economize to about 55 degrees. Would be a nice research project by itself...many thanks!...mfm

zeke

As I see this, it is the supply water temperature, Tin that determines the delta T for the TRV, since the load is satisfied by the difference of some average of the water temp and the room and the thermal resistance across the radiator, no controller can override this.

Therefore, one must change the supply temperature and the flow rate to accommodate a constant dT.

Simplistically:

Q=W'*c*dT

and

Q=[(Tin-dT/2)-Troom)]*radiator area*/Rt(W')

where

Q= load

W'= flow rate

c= specific heat

Rt(W')= thermal resistance of radiator(s) as a function of W'

For a fixed Tin, the two eaquations will yield a value of dT and a W' and there can be no control over the dT. or if you fix dT, your case, then you get one value of Tin.

I assume in the real world, another control loop involving the Tin is used.

Mike D_7

Variable speed motor controls and their efficiency

Brad, I thought you might find this link interesting
http://www.controlres.com/techeval.htm

It's a report on variable speed drives on fans with constant Delta P. The inverse square rule applies to fans as well as hydronic pumps.

Although the pump or fan may requre a given level of power to operate it it at a given level of resistance and RPM, both an AC motor's efficency and the controller's efficency determine power consumption.

I think this is why we don't always see the energy savings we expect with VS pumps.
I havn't seen any data on the Wilo Ecco pumps, but their pre release lit. seems to sugest a better corespondence between load and power requirements.