Storage As Hydraulic Separator: mini-tutorial

[Nofossil]

I have to explain this idea to a customer, and I couldn't find anything at the right level of clarity. Here's my first cut at explaining it. Any comments or thoughts?

Storage tanks can serve as a hydraulic separator. There are are two independent heat loops with the tank serving as part of both loops. The first loop takes cool water from the bottom of storage through a heat source (wood boiler, for instance) and returns it as hot water to the top of storage. This creates a layer of very hot water at the top of the tank which is pushed downwards.

Figure 1: Boiler to Storage Loop​

The second loop takes hot water from the top of storage, passes it through the heat load(s), and returns it as cool water to the bottom of the tank. This creates a layer of cool water in the bottom of storage that’s pushed upwards.​

Figure 2: Storage to Heat Load(s) Loop
​

Of course, both of these loops can and often do operate at the same time. The beauty of this approach is that either loop can operate whenever it needs to, and each loop can operate at an optimal and independent flow rate for current conditions.

Figure 3: Combined Loops​

If the heat source has a higher flow rate then the heat loads need, then the additional hot water goes into storage, pushing the hot/cold boundary downwards. If the heat loads need more flow than the heat sources are providing, then the hot/cold boundary moves upwards. All of this behavior happens automatically, irrespective of the number and types of heat sources and heat loads.

 

[ewdudley]

An often perpetuated misconception is that separate pairs of ports for supply and load are needed or beneficial.

 

[Nofossil]

An often perpetuated misconception is that separate pairs of ports for supply and load are needed or beneficial.

Yeah - I didn't draw separate ports in my 'Simplest Pressurized Storage' sticky (above) and it seems like that confuses people. Perhaps I should add a fourth image that's the same as the third, but with a single tee.

Does anyone have any first-hand experience with ghost flow issues in this type of installation when separate ports are not used?

Here's a possible addition to the tutorial:

The inlets and outlets on the storage tank can be connected to a single port as shown here:

Figure 4: Combined Storage Port​

While this is identical from a hydrostatic point of view, there are two potential problems. First, the path of lowest pressure drop across the tees is in a straight line. This creates the possibility of 'ghost flow' through the heat loads when the heat source is operating, and vice versa. Second, water is injected vertically into the tank. If the velocity is high enough, this can contribute to thermal mixing in the tank, decreasing stratification.

A better approach is to use side ports and bring incoming water through a tee oriented to not favor flow past the storage tank, as shown here:

Figure 5: Side Ports and Ghost Flow reduction​

 

[velvetfoot]

I think the vertical injection of supply water leading to non-stratification is what is happening in my buffer tank. It seems to be more stratified when just drawing from it.

 

[Nofossil]

I think the vertical injection of supply water leading to non-stratification is what is happening in my buffer tank. It seems to be more stratified when just drawing from it.

I'm working with a college here in Vermont, and this is an area where I would absolutely *love* to have students do some experiments and get some hard data. I'm sending the professor a link to this thread.....

 

[jebatty]

I would strongly favor separate ports not in close proximity to each other, the reason being that it is possible for the system to return "hot" water to the wood boiler (supply = 185, return = 170, for example) which, depending on boiler sizing and plumbing can result in the boiler btu output exceeding the btu demand and causing the boiler to idle, even though cool water is available at the bottom of storage. With separate ports, the boiler always is assured of the coolest (bottom of storage) return water and there is maximum ability of boiler + storage to accept boiler output. I will assume that it is possible with careful system design to avoid this result, but a separate port design leaves more room for error in system design and still achieve a satisfactory result.

What I have described is what happened on a Froling install. A separate port plumbing change made after two heating seasons of extreme frustration with the Froling completely solved the problem. With separate ports it is easy to design the wood boiler/loading unit/circulator side of the system and it is equally easy to design the demand side of the system because each operates independently.

Also, a single port design with a wood boiler with its own circulator/loading unit and a system with separate circulator supplied zones gets complicated on the hydronics because the boiler circulator/loading unit and the zone circulator(s) operate in series or series/parallel, and as of yet I have not grasped very well exactly what happens with flows in a series or series/parallel operation. I think there is too much opportunity for unexpected (disappointing) results in a single port design. It can work, but in the Froling example with an experienced installer the results were disappointing. Separate ports solved the problem.

 

[Beagle Dad]

Separate ports are a benefit as jebatty mentioned because the boiler pump loop is decoupled from the loads. I am currently piping my new system which consists of 4 tanks for a total of almost 1,000 gallons using this approach. The 4 tanks are connected in series so that the effective height is 28 feet (4 tanks at 7ft each) the load draws from the top of the fourth tank 180F and returns to the bottom of the first tank 90F. The boiler draws from the bottom of the first tank and pushes into the top of the fourth tank using separate connections.

 

[ewdudley]

I would strongly favor separate ports not in close proximity to each other, the reason being that it is possible for the system to return "hot" water to the wood boiler (supply = 185, return = 170, for example) which, depending on boiler sizing and plumbing can result in the boiler btu output exceeding the btu demand and causing the boiler to idle, even though cool water is available at the bottom of storage. With separate ports, the boiler always is assured of the coolest (bottom of storage) return water and there is maximum ability of boiler + storage to accept boiler output. I will assume that it is possible with careful system design to avoid this result, but a separate port design leaves more room for error in system design and still achieve a satisfactory result.

What I have described is what happened on a Froling install. A separate port plumbing change made after two heating seasons of extreme frustration with the Froling completely solved the problem. With separate ports it is easy to design the wood boiler/loading unit/circulator side of the system and it is equally easy to design the demand side of the system because each operates independently.

Also, a single port design with a wood boiler with its own circulator/loading unit and a system with separate circulator supplied zones gets complicated on the hydronics because the boiler circulator/loading unit and the zone circulator(s) operate in series or series/parallel, and as of yet I have not grasped very well exactly what happens with flows in a series or series/parallel operation. I think there is too much opportunity for unexpected (disappointing) results in a single port design. It can work, but in the Froling example with an experienced installer the results were disappointing. Separate ports solved the problem.

As I was saying, an often perpetuated misconception is that separate pairs of ports for supply and load are needed and/or beneficial.

 

[Nofossil]

Does my 'Figure 5' above provide a satisfactory approach that covers all the bases mentioned here? It has widely separated ports, one near the top and one near the bottom, oriented to minimize internal mixing.

 

[ewdudley]

Does my 'Figure 5' above provide a satisfactory approach that covers all the bases mentioned here? It has widely separated ports, one near the top and one near the bottom, oriented to minimize internal mixing.

Injecting sideways is certainly ideal, but for the most part all the stratification and porting strategy discussions fall into the overthinking it category. If we were talking chilled water storage then it would be important to sweat every detail, but hot water sure does rise so for hot water storage it's hard to go wrong.

Vertical injection will generally work just dandy, a diffuser can be added if someone is worked up about it.

The "high flow" situation is when storage is being filled and some mixing at the top is no big deal. By the time the burn cycle is completing the mixing will be hot water mixing with hot water at the top of the tank, so mox nix, or mix nox as the case may be.

The "low flow" situation is when drawing from storage, so the velocity of the return water entering the tank is much slower, so much less mixing.

As far as the supposed problem of water 'shooting past the tee' in figure 4, there would be no such problem.

 

[Nofossil]

Thanks. According to my good friend the (broken link removed to http://www.engineeringtoolbox.com/resistance-equivalent-length-d_192.htm), straight flow through a 2" tee is equivalent to 7.7 feet of pipe, where branch flow is equivalent to 12 feet. With two tees oriented with storage as a branch, that would give the storage path an additional 9 feet of head loss. That *should* be trivial (less than .2 psi), but sometimes the real world doesn't match theory as well as I'd like. I also don't know what the cracking pressure is on the integral check valves on smaller circulators, so I don't have a good context for how much forward pressure an inactive zone can tolerate before ghost flow sets in.

I'd love to have hard numbers on how much mixing to what depth for a vertically introduced stream of hot water into a cold tank at a given velocity. If 4 ft/sec creates mixing that only extends 6", then it's truly no problem. If it's 2', then probably OK for a vertical tank, but a real problem for a smaller horizontal tank. I really don't know the hard numbers. If my college connection comes through, I'll post real data when they get it.

 

[jebatty]

To assure that there is no confusion, by "separate" ports I mean that the boiler supply to storage and return from storage are separately ported in the storage vessel and the load supply and return also have their own set of ports in the storage vessel. Figures 1, 2 and 3. By single ports I mean T ports as shown in Figures 4 and 5.

Does my 'Figure 5' above provide a satisfactory approach that covers all the bases mentioned here? It has widely separated ports, one near the top and one near the bottom, oriented to minimize internal mixing.

My answer is "no" and this is why with T ports. I'm going to use a northern Minnesota example of extended periods, often days, when the outside temperature never gets above 0F and night temps into the -30'sF. The heating system is demanding nearly continuous hot water supply, but the btu demand is less than the btu output capacity of the boiler. The boiler has its own loading unit/circulator with a flow rate of 12 gpm (typical maximum flow for some loading units). Each system demand zone also has it own circulator.

1) Spec loading unit flow of 12 gpm from boiler to the storage T. Boiler is loaded with wood and boiler output is near rated capacity of 185,000 btuh.
2) Flow from the storage T to meet demand could be more or less than 12 gpm depending on the number of zones calling for heat.
3) Initial boiler supply = 180F, system return = 160F.
4) Initial storage top = 180F, bottom = 120F.
5) Demand circulators now are in series or series/parallel with the loading unit (this is where the hydronics are complicated). It is unknown what the actual flow rate was through the boiler and through system demand.
6) Actual experience resulted in system 160F return going directly to the boiler return and bypassing storage (what flow leaves the boiler has to return to the boiler).
7) At boiler output of 180F with 160F return, it would take a flow rate of 18.5 gpm for the flow to move the entire boiler 185,000 btuh output. Actual experience resulted in boiler temperature rising to idle and boiler cycling; therefore, actual flow rate had to be less than 18.5 gpm and would have been equal to minimum 12 gpm (loading unit spec) or somewhat greater due to series or series/parallel operation of demand circulators.
8) Little or no cool water from bottom of storage flowed to the boiler. Storage tank heated little or none even though boiler btuh output exceeded demand.

... now

9) If not T ports and boiler and system each had separate ports (Figures 1, 2 and 3), boiler output at 185,000 btuh and 180F, boiler flow at 12 gpm maximum, boiler return from storage at 120F: at 12 gpm and delta-T=65, boiler to storage capacity = 390,000 btuh.
10) Actual flow would have been divided between the loading unit mixing return to the boiler and cool storage water return to boiler. EW is good at computing how much flow is going to storage, how much is being returned directly to the boiler to maintain 160F return water protection. In all events, with 120F bottom of storage mixing the result is no boiler idling/cycling.
10) System demand now also draws 180F from top of storage and returns 160F to bottom of storage. Storage will mix and increase tank temperature.
11) The boiler/storage system is dynamic during operation based on flows, boiler output and system demand; all variables are subject to change.
12) Regardless, in the Froling example I mentioned, the actual result of eliminating the T ports and plumbing separate ports (Figures 1, 2 and 3) was that boiler idling/cycling was eliminated. No other change was made in the system.

And I agree with EW with respect to a discussion on stratification, that

for the most part all the stratification and porting strategy discussions fall into the overthinking it category.

 

[jebatty]

I'd love to have hard numbers on how much mixing to what depth for a vertically introduced stream of hot water into a cold tank at a given velocity.

Can't answer this one for you, but on a horizontal 19' 1000 gal tank, 3' diameter, stratification is extreme with hot water injection horizontally into the end of the tank down 6" from the top of the tank, with return to the boiler at the same end of the tank and up 6" from the tank bottom. Estimated/calculated flow rate 14 gpm.

 

[Nofossil]

Thanks, Jim - that's a very complete writeup. I understand your point about the single tee port now. My question is, though, if zone return is 160 then how did the bottom of storage get down to 120? I'm working on cleaning up the design on a really non-functional system with multiple heat sources (biomass, fossil, and solar) and multiple heat loads (9, of 8 different types), each of which has different supply / return temperatures. It's an interesting optimization problem, especially since there's a desire to minimize plumbing changes as much as possible.

I really need to get the idea of storage stratification across to others involved in this project, and I have limited time in which to accomplish this. I need them to understand 'this is good because...' and 'this is bad because...'

I'm afraid that there are nuances that will not be conveyed, but at least the main idea needs to get across.

 

[jebatty]

if zone return is 160 then how did the bottom of storage get down to 120?

Variable weather, variable firing of the boiler, variable demand. A night of low outside temps and not keeping the boiler fired will draw down storage.

The point I am making is that with T ports and low temp water in storage, that low temp water may not be available to the boiler to eliminate idling/cycling, and also that warm system return water may not be returned to heat storage but flow directly to the boiler, also resulting in idling.

 

[jebatty]

I really need to get the idea of storage stratification across to others involved in this project, and I have limited time in which to accomplish this. I need them to understand 'this is good because...' and 'this is bad because...'

Stratification is not necessarily the best objective in all applications. If a low temp radiant system (in floor pex), supply water may need to be only 100-120F, and mixing poses little if any impediment to meeting demand. On the other hand, mixing can adversely impact a wood boiler and increase the possibility of idling by not making the coolest water available to the boiler. And on one other hand, in a low temp radiant system and draw down of storage to 100-120F before firing the boiler, mixing during boiler operation is unlikely to cause idling as the whole tank is cool.

 

[Nofossil]

Stratification is not necessarily the best objective in all applications. If a low temp radiant system (in floor pex), supply water may need to be only 100-120F, and mixing poses little if any impediment to meeting demand. On the other hand, mixing can adversely impact a wood boiler and increase the possibility of idling by not making the coolest water available to the boiler. And on one other hand, in a low temp radiant system and draw down of storage to 100-120F before firing the boiler, mixing during boiler operation is unlikely to cause idling as the whole tank is cool.

In this instance we have solar hot water panels that want to be supplied with cool water, and we have baseboards that need really hot water - and everything in between. Stratification is critical here.

 

[Coal Reaper]

i am Ted top and bottom (fig. 4) and i do see "shooting past the T". not only do i have some ghost flow (at the end of a full load burn the PEX coming into my house 200' away from boiler is warm), but this does in fact allow water to return to my boiler hotter than the bottom of storage tanks. this can certainly cause boiler to overheat, especially if one were to be increasing house temp such that the zones were cranking for an extended period of time. for this reason i do not set back the thermostats. what it does nice is allow hot hot water to get to my baseboards soon after lighting a fire, even if the tanks are cold. if i were to do things over i think i would try T into top of tanks just as i did and sepereate ports on bottom that way the house return mixes with bottom of storage before going to boiler.
heres my storage tank connection. rear connection from boiler, foreground to house. its similar at bottom.
also diffusers for top. the bottom dips down on an angle and pulls water from about 4" above the bottom of the dome. stratification is fantastic. i have 5 temp sensors and with no load i can see each one jump to about 155* (loading unit doing its job) then 175* then 195* with each pass through the boiler.

 

[Coal Reaper]

this is charging. it is not exactly as my previous post described because i was still messing with the flow.

 

[Bob Rohr]

i am Ted top and bottom (fig. 4) and i do see "shooting past the T". not only do i have some ghost flow (at the end of a full load burn the PEX coming into my house 200' away from boiler is warm), but this does in fact allow water to return to my boiler hotter than the bottom of storage tanks. this can certainly cause boiler to overheat, especially if one were to be increasing house temp such that the zones were cranking for an extended period of time. for this reason i do not set back the thermostats. what it does nice is allow hot hot water to get to my baseboards soon after lighting a fire, even if the tanks are cold. if i were to do things over i think i would try T into top of tanks just as i did and sepereate ports on bottom that way the house return mixes with bottom of storage before going to boiler.
heres my storage tank connection. rear connection from boiler, foreground to house. its similar at bottom.
also diffusers for top. the bottom dips down on an angle and pulls water from about 4" above the bottom of the dome. stratification is fantastic. i have 5 temp sensors and with no load i can see each one jump to about 155* (loading unit doing its job) then 175* then 195* with each pass through the boiler.

There will always be the potential to have blending as the two flow rates change, there is a mixed flow formula to predict that. Also it is not goof piping practice to "bullhead" a tee. That is flow from to directions in the run combining at the branch, various flow conditions will be possible.

I'll see if we can get Siggy to comment, he has done a lot of design and research on hydraulics in separators.

Also idronics 15, at www.caleffi.us is another update to air, dirt and hydraulic separation, along with formulas.

 

[Coal Reaper]

image 1216 is the temps the following morning. i have setpoint circ at 140* return.
and the 7 Jan shot shows how i can get warmer water at the bottom T than the middle of the tank. note it was 12*F in the barn.

 

[goosegunner]

These are the diffusers I put in my tank for vertical ports. Don't know if it helped or not but my 1000 gallon horizontal tank stratifies nicely.