Closed Loop Heating Hot Water with 3 MW Boiler

SSN96

Hi People,

I probably should know this but I thought your expertise/input in this matter would be highly helpful. I am currently working on a project where we are installing a 3MW hot water boiler with an operating pressure of 5 bar and a design pressure of 6 bar.

The hot water loop also has a pressurization unit and supplier has selected it to maintain 2.4 bar pressure. (which I think won't work)

2.4 = static head + evaporation pressure + safety margin

I do have a basic understanding of the pressurization unit and it's working but this situation is boggling my mind a bit. So I have selected a setpoint for the pressurization unit to be 4.5 bar (or ideally I think it should be at 5 bar) as I think pressure shouldn't drop below this point in the system. Especially at a point where water is returning to the boiler. which I expect to be at 5 bar currently. However, I read somewhere that water should return to the boiler with a pressure 10% higher than the boiler's operating pressure. Is this right?

Other pressure losses through the system total up to 3 bar. This includes pressure loss through boiler tubes, pipework, valves, etc. Hence differential pressure across the pump is shown as 3 bar. Water leaving through the boiler is at 4.5 bar because of pressure loss through the boiler tube.

I have attached a drawing along with this question on my understanding of how this system should work. I am not sure if am quite right or wrong.

I have pointed out pressure readings at different points in the system wholely on my assumptions and this system is still in the design phase.

Please look at the attached.

Any inputs in this would be highly beneficial.

Image (43) (1).jpg

Jamie Hall

The general layout seems reasonable.

Some of the terminology is a little unfamiliar, however. By "pressurization unit" do I presume you to mean what would be called in North America an expansion tank? Basically a passive device charged with air to the static cold pressure of the system, but larger enough in volume to maintain the hot pressure in the system at a reasonable level? Or is it some active device? I rather hope the former, as it would be quite adequate — and much simpler.

You note that you believe the water should return to the boiler at a pressure 10% over the boiler operating pressure. Not really. What does need to happen is that the pressure at the outlet from the boiler (and the inlet to your circulating pump) must be greater than the vapour pressure of water at that temperature. In fact, it should be significantly greater, to avoid cavitation problems in the pump.

Systems running at a temperature of 82 C, which is regarded as high for many heating applications, run very satisfactorily at a gauge pressure of 1 bar, at which the boiling point of water is 121 C

The other, and much more serious question, is why such a high pressure in the system? The boiling point of water at your static pressure (4.5 bar), assuming you are using gauge, not absolute, pressure is 158 Celsius. What temperature do you plan to operate this system at?

SSN96

Hi Jamie,

Thanks for your comment.

Terminology can be a bit different but you can call it an expansion tank but it is not just an expansion tank. It comes with a control unit including a compressor that controls the diaphragm in the expansion vessel to adjust the system pressure. ( https://cookeindustries.co.nz/products/heating/pressurisation-systems )

For more precise information, It is REFLEX (brand) pressurization unit.

1. They keep the pressure within permissible limits at all points of the system, thus ensuring that the max. excess operating pressure is maintained while safeguarding a minimum pressure to prevent vacuums, cavitation, and evaporation.
2. They compensate for volume fluctuations of the heating water as a result of temperature variations.
3. Provision for system-based water losses by means of a water seal.

I understand and agree with what you're saying. I had overlooked on few things and realized the pressurization unit doesn't need to be set at 4.5 bar. The system needs to run at 105 degrees celsius inflow and 80 degrees celsius return. So around 1.1 degree celsius is saturation pressure or minimum pressure in the system required.

I got in touch with the supplier and he suggested me to consider running the system at 2.4 bar.

Evaporation pressure + static head + safety margin = 2.4 bar

Now I have revised my sketch a bit.

Image (45).jpg

I think this looks more reasonable.

I am still not 100% sure how circulation pumps influence the pressurization unit or maybe I think it has nothing to do with the pressurization unit as it is only circulating the water and compensation for pressure losses in the circuit.

Do you think the pressurization unit will try to adjust the pressure differential created due to the circulation pump?

Specifically talking about pressure at points 3,4 and 5 after the circulation pump.

leonz

UMMM,

Is it too late to ask if a steel compression tank and lower temperatures are out of the question????

Is the ECO unit an economiser? Is the pipe leading to it crossing it or intersecting with a 4 way cross fitting??

hot_rod

how did you come up with all the various pressure numbers? And why🧐

It looks like you are trying to determine the dynamic pressure at various points. If so you would calculate piping lengths and fitting count. The pressure drop of each piping path.

The reflex is basically doing what an expansion tank and fill valve does, just packaged together.

Unless it has sensors at all those points to monitor pressure differentials.

I do see some value in the version that does deareation, on large systems.

Jamie Hall

Interesting device, that pressurization unit. Frankly, I'd avoid it and simply specifiy a large enough expansion tank — unless the system is very large indeed. But I am adverse to active devices when passive ones will do — personal philosophy, I guess you could say.

Now as to how it will interact with the rest of the system. Just like a passive tank, the best place to connect to the rest of the system will be on the inlet to the circulating pump. Not right at the inlet, but with minimal piping between it and the inlet. The reason for placing it there is two fold: first, that location is the lowest pressure point on the system, and the pressure there is held constant, the pressure at all other points will be greater — and thus no opportunity to get air into the system. Second, that location ensures that the inlet to the circulating pump always has the pressure it needs to avoid cavitation.

As I understand the principle of operation of the pressurization unit (either flavour), it is sensitive to the system pressure at its location — so the pressure at other locations such as 3, 4, and 5 will not affect it. Pressures there will be governed by the flow conditions in the system and the head added by the pump.

SSN96

Hi @leonz

Yes it is out of the question at this stage. We have to install Pressurisation Unit. I am not that familiar with this arrangement at all.

ECO is an economiser and pipe leading to it is crossing. It's not four way. Apologies, i should have shown it in a right way.

Hi @hot_rod

I am just trying to visualise pressure reading at different points in the system. I have calcuated pressure losses already and this indeed is a very LARGE system with longer pipe runs, three heat exchangers. Pressure loss throughout is around 3 bar. Hence pumps duty is shown as 28 l/s@3 bar in the sketch above.

Hi @Jamie Hall

Thank you!

Yes alot of people do question the need of pressurization unit (rightly so) but it has been considered and decided upon after a long discussion. Yes it is a very large system serving a very crucial facility.

I do agree with your advise on unit's connection in the system and it will be connected to what they call as "most vulnerable point" as you rightly described.

Also, from what i read pressurisation unit maintains set pressure at all points in the system and not just the point it is connected to. it maintains a lower pressure limit (as pressure cannot go below 1.1 bar) and pressure increased in the system due to fluid expansion caused by temperature rise.

Though i am not sure if it only governs pressure increased by the expansion but it makes more sense if it does so. As you said yes i thought that circulating pump is only forcing the liquid, there is some increased pressure there to compensate for dynamic losses but it is not related to fluid expansion. This pressure gradually decreases while returning to the boiler (point 6 in my sketch) and pressurisation unit will read that only.

This is just a thought though. I can be wrong. This is the main conern i have and that lead me to this sketch.

hot_rod

Let talk static pressure first. The goal is to have 5 psi positive pressure at the top of the system, with no circulators running. This assures the system is in fact filled and allows float type air vents located at the top to seal tightly.

I suppose a sensor at the top would assure you always have this positive pressure. it could increase as temperature rises and it could drop if the system cools below the fill water temperature. Or if a chiller is used on the piping.

The next pressure we look at is differential pressure ∆P head energy added by the spinning circ. I think that is what your numbers show?

This varies based on the pressure drop in the various circuits or loops. A simplified drawing attached. This also shows the importance of pumping away from the expansion vessel connection. The pumps ∆P must show up as positive pressure throughout the system.

I don't see a benefit to knowing or adjusting pressure at different points in the system. If a sensor allowed the system to always be at 5 psi at the uppermost point, why would that not be adequate?

Now on a system that switches to chilled water a fill device would want to maintain pressure as fluid temperature drops, to prevent cavitation if nothing else. So installers over-charge the upsized expansion tank to have a small amount of fluid volume ready as the water contracts from cooling. Same with solar thermal.

hot_rod

Building a graph like this maybe helps visualize the differential pressure throughout the system.

Notice how a circuit with high pressure drop devices can allow ∆P to drop with an improperly plac

ed expansion tank. The dotted red lines show the profile of the differential pressured added. That should always be at or above the static fill pressure, the blue line. The PONPC point of no pressure change is, or should be :) the only point where the spinning circ cannot change the pressure

Jamie Hall

Someone perhaps misunderstands. (" from what i read pressurisation unit maintains set pressure at all points in the system and not just the point it is connected to"). This is not true. The system — while working properly, will maintain the set pressure at the location at which it is connected. However, the pressure elsewhere in the system will vary depending on flow conditions in the system. This cannot be avoided.

Now you mention that an active pressurisation unit was chosen because this is a very large system and a very crucial facility. This leads me to a critical question: is this system intended to be fail operational or fail safe? If it intended to be fail operational, which it may be, the design you have shown isn't. To be even close to fail operational, you must have at least two active pressurisation units connected at the same location, either of which on its own can handle all possible variations. Further, they must be alarmed and controlled in such a way that a failure of either one will not affect the ability of the other to maintain control — and the alarm must be attended to immediately. If there are any special controls or equipment involved, replacements for all such equipment must be on site.

I might mention — though it is hardly modern — that there is a passive pressure control system which is incredibly reliable (mtbf on the order of decades): a standpipe. A simple pipe extending vertically from the connection point high enough to maintain the desired pressure at the connection point. You mention a tagret pressure of 2.5 bar. If I assume that that is a gauge pressure, you would need a pipe extending vertically 25 metres (if your quoted pressure is absolute, it would be about 15 metres).

I realise that this may not be feasible — I have no idea what your layout and plant is like — but I mention it as it is, as I say, it is extremely easy to make it fail operational, but has an almost incredibly low failure rate.

It is not part of this discussion, but I feel obliged to add as a design professional, that if this is to be a fail operational system for a crucial operation, you have other single point failure potentials in there which must be considered, such as the pump and, for that matter, the boiler itself.

Lastly, if this is to be a fail safe, rather than fail operational, system, consideration should be given as to how to arrange an orderly shutdown in the event of a failure (and yes, I could tell you some horror stories of fail safe failures where consideration of this contingency had NOT been made — I was involved with one (didn't design it — I was picking up the pieces and redesigning it) where the failure of a single ten amp instrumentation fuse caused a loss of a week's production — and the destruction of almost a million dollars of machinery.)

SSN96

@hot_rod

Thank you. Yes, I get your point.

As you mentioned, Reflex (pressurization unit) is just a combination of an expansion tank and fill valve. Are you 100% certain about this?

Do expansion tanks come with a compressor, a control unit and a diaphragm for pressure maintenance?

Sorry for asking this. I am new to this field. I have only been involved with the mechanical maintenance side of things in my life and not the design. This is my first major project.

@Jamie Hall

That's a really good point. Something to ponder. However, this heating hot water system is not alone. There's a duty and standby system. That's how everything works at this facility. If something fails, there is something else running to cover up. There are a lot of safety devices/measures that are installed/taken. Everything is automated and BMS monitored.

This system has a lot of sensors. So does this pressurization unit. It comes with a low-level sensor and a pressure switch that communicates to the boiler PLC. There's even a bladder rupture sensor. If something goes wrong, everything will turn off. Stand-by system will start operating until all faults are fixed.

Also, i took this screenshot from the pressurization unit tech data. This is how it reads.

As it clearly says "to keep pressure within permissible limits at all points of the system"

Reflex_Dynamic Pressurisation Systems_RE1879en-9127864 (002).pdf

hot_rod

no problem in using a system like that if that is what the job specs. But I don’t think it will do exactly what you are thinking, nor does it need to

Being a dynamic device it seems as it is adding more fail points not eliminating them. It does have some intelligence that a bladder tanks lacks to alert of a failure. Maybe that makes it worth the $$ and added complexity

Jamie Hall

I think the main point I was trying to make, @SSN96 , if nothing else, is to take nothing for granted — and most essentially in the automation part, which is by far the most failure prone.

Failure tree/consequence analysis is a field rather unto itself and was once described to me as requiring an odd combination of paranoia and imagination. The basic idea is to analyse ever single component and figure out what might happen if that component failed or glitched, and what other components must NOT fail or glitch at the same time. Then figure out how to cope.

Again, I'm not sure what you are up to — in fact, I have no idea — but I'm beginning to get a feeling that applying a pessimistic, aerospace approach to your failure analysis might be needed.

And never rely on a single computer or power source! Just don't.

leonz

I agree with Jamie, a standpipe like a steel compression tank or open to air expansion tank whether it is being vertical or a horizontal mounted saddle tank is simpler and requires only enough water to maintain system pressure and would be much more effective and of course trouble free with nothing more than a tube water type level gauge to monitor the water level or a petrometer and it is easy to isolate when needed by using a repackable gate valve using a second one in line as a backup or a vacuum breaker piping arrangement.