T50 on new install?

[SJB060685]

What I'm finding hard to reconcile is that the same does not apply to the heat loss from a radiator to the room. Because all being equal, the same relationship should apply.

All I can imagine is that all is not equal because the house has plenty of natural air movement from the external wind whereas a radiator creates a convection current, so the availability of cold surrounding air for a radiator to heat is partly driven by the temperature difference, whereas in the case of a house, this factor is hardly worth taking into consideration?

I'm not sure of what you're asking mate. Sorry, I'm absolutely shattered. Only got a couple hours sleep last night

 

[Ric2013]

I'm not sure of what you're asking mate. Sorry, I'm absolutely shattered. Only got a couple hours sleep last night

Not asking anything, really, just sympathising with John.g. Really I think we should continue this as a private chat, if you both want, as I feel we're massively OT and it's not helping the OP.

 

[siricosm]

One you've got the pipes warmed up, it's very hard to have a 10°C loss between the boiler and the cylinder (boiler flow 82 and reaching cylinder at 72), unless your pipe is 100m long?

Sorry, maybe I wasn't clear. I meant the two pipe thermometers gave readings of 72 for flow, and 60 for the return, but my guess is that it was more like 77 for the flow and 65 for the return, so a 12C delta across the boiler.

 

[John.g]

Not asking anything, really, just sympathising with John.g. Really I think we should continue this as a private chat, if you both want, as I feel we're massively OT and it's not helping the OP.

Still keeping Siricosm happy I think but won't dwell too long more now on the above.

A 150 litre insulated cylinder has 90W loss stamped on it which is a loss of 0.6C/hr, AFAIK, this is obtained by maintaining the cylinder at 60C with a surrounding air temp of 20C, deltaT of 40C, a 150 litre cylinder will roughly have a surface area of 1.33 M2 so the "U factor" is 90/1.33/40, 1.692w/m2/degC......check 1.692*1.33*40 = 90watts. If this loss is linear then the cylinder, in falling from 60C to 55c will lose 1.692*1.33*5,11.25 watts and will lose the same in falling from 25C to 20C, both obviously equate to a loss of 90 watts/hr. or a loss of 0.6C/hr.

Now, its patently obvious that the cylinder will not lose heat at a rate of 0.6C/hr, 11.25watts, in falling from 25C to 20C but it will be close to losing it at a rate of 0.6C/hr, 11.25 watts in falling from 60C to 65C.
If these losses are/were constant why do they keep the cylinder at a set temperature and why not just monitor the heat loss in falling from 60C to 20C and convert this to the loss, a good bit less than equating to 0.6C/hour I would guess.

 

[siricosm]

If this loss is linear then the cylinder, in falling from 60C to 55c will lose 1.692*1.33*5,11.25 watts and will lose the same in falling from 25C to 20C, ...

OK, I'll chime in. I think there is confusion about the terms "linear" vs "constant". It is linear with respect to the gradient of temperature. So will lose heat at a rate of 8x more at 60C vs 25C. The hotter it is, the faster it loses heat, hence the desire to design for as cool as practical.

The same applies to the output from the plate exchanger. It appears that most commercial heat banks have undersized DHW plate exchangers to run at lower temperatures. With condensing boilers and heat pumps so popular, I am a bit surprised this is still the case.

 

[John.g]

OK, I'll chime in. I think there is confusion about the terms "linear" vs "constant". It is linear with respect to the gradient of temperature. So will lose heat at a rate of 8x more at 60C vs 25C. The hotter it is, the faster it loses heat, hence the desire to design for as cool as practical.

The same applies to the output from the plate exchanger. It appears that most commercial heat banks have undersized DHW plate exchangers to run at lower temperatures. With condensing boilers and heat pumps so popular, I am a bit surprised this is still the case.

The Heat store has potential to give very high boiler efficiencies by increasing the condensing effect with reduced return temperatures as you have good control over that, just assume you set the cylinder stat to 40C, it will then have a cut in somewhere around 35C, the stat should be located a bit above the return to the boiler, as you start using hot water; cold water will start replacing the hotter water above it and you could have water as low as 35C or even lower to the boiler return which with a HE (condensing) boiler will give very high efficiencies, you can still get a deltaT of say 25C/28C by reducing the pump speed, this will still give a minimum store temp of 60C/65C.

 

[John.g]

The Heat store has potential to give very high boiler efficiencies by increasing the condensing effect with reduced return temperatures as you have good control over that, just assume you set the cylinder stat to 40C, it will then have a cut in somewhere around 35C, the stat should be located a bit above the return to the boiler, as you start using hot water; cold water will start replacing the hotter water above it and you could have water as low as 35C or even lower to the boiler return which with a HE (condensing) boiler will give very high efficiencies, you can still get a deltaT of say 25C/28C by reducing the pump speed, this will still give a minimum store temp of 60C/65C.

Apologies, that's incorrect as you are not drawing any water off the store as you have a HX for that!!

 

[John.g]

The Heat store has potential to give very high boiler efficiencies by increasing the condensing effect with reduced return temperatures as you have good control over that, just assume you set the cylinder stat to 40C, it will then have a cut in somewhere around 35C, the stat should be located a bit above the return to the boiler, as you start using hot water; cold water will start replacing the hotter water above it and you could have water as low as 35C or even lower to the boiler return which with a HE (condensing) boiler will give very high efficiencies, you can still get a deltaT of say 25C/28C by reducing the pump speed, this will still give a minimum store temp of 60C/65C.

You wondered what flow/temp you might get at a lower cold water temp, I found my old calcs from years ago and should give a fair indication of what you might expect, the heat input shows practically the same (at your primary flow rate of 33 LPM estimatd) and shows ~ HW flow rates of 7.4LPM at 72C & 7.23LPM at 57.4C.
You may be interested in the calcs I used.

 

[siricosm]

You wondered what flow/temp you might get at a lower cold water temp, I found my old calcs from years ago and should give a fair indication of what you might expect, the heat input shows practically the same (at your primary flow rate of 33 LPM estimatd) and shows ~ HW flow rates of 7.4LPM at 72C & 7.23LPM at 57.4C.
You may be interested in the calcs I used.

Thanks. I would target an output temp of 48C on the hot side of the DHW.

The main question is that if the plate exchanger has a rating of 165Kw at a delta of 40C, then what is the minimum tank flow temperature required to get a flow rate of 10l/m at 48C if the cold water feed is 5C?

The delta of 40C is for rating the exchanger, I would expect the delta to be considerably lower in operation. The Grundfos Alpha is rated at a max of 50l/m (I think), so 33l/m is probably a good estimate for the max primary flow because there will be some resistance in the plate exchanger.

 

[John.g]

IF only using HW, then, theoretically, a store temp of 48.6C will give ~ 10 LPM at 48C with a return of 25C (with a reduced pump flow of 18.6LPM) see below, but its not as simple as that because the rad return temperatures will influence the return since it will be mixed with that 25C and also the total energy required will increase due to the rad requirements, so realistically you might require a store temp of 60C which should give a output of 50kw, 30kw for HW + say 20kw for rads.
If installing new rads, I would aim for T30 rads, a factor of X1.9 of a T50, and I would suggest calculate each room energy requirement separately in watts, multiply this by a factor of say 2, so if you require say a 2500 watt rad for a room, you look up your brochure for 5000 watt rad (T50) and you can then pick the appropriate rad, single, double, or triple and see how it might look in each room.

HOT WATER ONLY

 

[siricosm]

IF only using HW, then, theoretically, a store temp of 48.6C will give ~ 10 LPM at 48C with a return of 25C (with a reduced pump flow of 18.6LPM) see below, but its not as simple as that because the rad return temperatures will influence the return since it will be mixed with that 25C and also the total energy required will increase due to the rad requirements, so realistically you might require a store temp of 60C which should give a output of 50kw, 30kw for HW + say 20kw for rads.
If installing new rads, I would aim for T30 rads, a factor of X1.9 of a T50, and I would suggest calculate each room energy requirement separately in watts, multiply this by a factor of say 2, so if you require say a 2500 watt rad for a room, you look up your brochure for 5000 watt rad (T50) and you can then pick the appropriate rad, single, double, or triple and see how it might look in each room.

I don't see where the size of the heat exchanger factors into the calculation above. Forgetting about the radiators for a moment, at a lower tank temperature the size of the heat exchanger needs to increase to maintain the desired flow and temperature of DHW.

The question is how low can you go with a 165Kw exchanger?

 

[John.g]

AS above (if my calcs correct) 48.5C with cold water at 5C (winter). The hx heating surface is taken into account, it surface area is 0.92M2, first yellow up from bottom. The HX output required to heat 10.18LPM from 5C to 48C (30.54kw) which is equal to LMTD*heat flux*Hx surface area, 5.53*6000*0.92/1000, (30.54kw).

Your query was I think how low can the store temperature go to heat 10 LPM from 5C to 48C?? and of course the area of the heat exchanger matters. if you halved your heating surface you would only get 5 LPM at 48C.

 

[siricosm]

AS above (if my calcs correct) 48.5C with cold water at 5C (winter). The hx heating surface is taken into account, it surface area is 0.92M2, first yellow up from bottom. The HX output required to heat 10.18LPM from 5C to 48C (30.54kw) which is equal to LMTD*heat flux*Hx surface area, 5.53*6000*0.92/1000, (30.54kw).

Your query was I think how low can the store temperature go to heat 10 LPM from 5C to 48C??

In that case, assuming all is correct, T30 is a possibility with this exchanger (ie a 50C tank), although the rads would need to be 2.4 times oversized.

Where did you get the heat transfer coefficient (6Kw/m2.degC)? I found references that said .4-.57 for copper coils in water, and .9-1.6 for corrugated stainless coils. Working backwards from the Nordic Tec's specs I get 165kw/.92m2/40C = 4.48Kw/m2degC, but maybe I am missing something.

 

[John.g]

Ca n you forward those specs to me, I worked backwards as well from the info I thought you gave me, what's the 40C referring to? if the flux indeed is 4480watts/M2/degC, can redo the calcs no problem.

there you go, just up the store flow to 33 LPM.

T30 rad require X 1.94, =(50/30)^1.3

 

[siricosm]

Thanks again.

I got the .92m2 from the description of the NordicTec Ba-23-40 (165Kw). I seem to remember reading a year ago when I was looking that the rating was assuming a 40C average difference from the primary to secondary sides, as that is what I have written in my notes. However now all that I can find is a description reading "The nominal power is 165kW for high parameters.", whatever that means. The Nordic Tec site says one can find the data sheet in the Download area, but there is no download area. You would think that people selling/making heat exchangers would give the specifications. There are a few other brands, Hrale, Wiltec, but it is the same story. I emailed Nordic Tec to ask. I'll post back if I get an answer.

A corrugated stainless coil has a heat transfer coefficient 2-3x that of a smooth copper coil, presumably because of the turbulence on the inside. With a PHE you have turbulence on both sides, I believe the channels are designed to maximize the turbulence. Also on one side of a coil the water might not be moving much, whereas on PHE it is moving on both sides. So maybe it makes sense that the heat transfer is 3-5x that of a corrugated coil.

That looks correct for the rads, thanks. I think I made the mistake of comparing to T60, not T50.

 

[John.g]

Agree, why they don't give the spec, ie primary and secondary flow/return temps and flow rates, surface area and output, I found one 100kw that gave all of the above, except the surface area., so for that I assumed a surface area of 0.5M2 so this gave me the 6000 watts/M2/degC. (see below)

 

[siricosm]

Update from Nordic Tec on the Ba-23-40 (165Kw) heat exchanger.

My question was:
"I am using a thermal store (buffer tank) to heat domestic hot water. On the primary I have a pump capable of around 33litres/minute to feed water from the buffer tank. On the secondary, the incoming cold feed temperature is 5C and the output temperature needs to be 48C at 10litres/minute. I am trying to calculate the minimum tank temperature (primary feed temperature) required to do this"

I also asked for the heat transfer coefficient.

Their reply was:
"Ba-23-40 seems to be a very correctly sized unit for this. You didn't give the water tank temperature - with Ba-23-40 - 53oC would be needed to heat up the 2nd side from 5 to 48oC, basing on the flow rated you gave. Heat exchanger coefficient in case of this serie is 6200 W / m2 / oC."

I notice that 6.2Kw/m2.degC is very close to what John calculated.

 

[John.g]

Thats excellent but I have just found the details that we were looking for in a swimming pool Hx spec file and I attach it and also calcs based on their numbers, their Hx heat flux (calculated) comes out at 4690 watts/M2/degC, I suppose different manufacturers use different designs to give the highest number, I have seen coefficients as high as 8500 watts/M2/degC.
Anyway, you can now do the numbers based on hard numbers from the manufacturers.

 

[John.g]

Update from Nordic Tec on the Ba-23-40 (165Kw) heat exchanger.

My question was:
"I am using a thermal store (buffer tank) to heat domestic hot water. On the primary I have a pump capable of around 33litres/minute to feed water from the buffer tank. On the secondary, the incoming cold feed temperature is 5C and the output temperature needs to be 48C at 10litres/minute. I am trying to calculate the minimum tank temperature (primary feed temperature) required to do this"

I also asked for the heat transfer coefficient.

Their reply was:
"Ba-23-40 seems to be a very correctly sized unit for this. You didn't give the water tank temperature - with Ba-23-40 - 53oC would be needed to heat up the 2nd side from 5 to 48oC, basing on the flow rated you gave. Heat exchanger coefficient in case of this serie is 6200 W / m2 / oC."

I notice that 6.2Kw/m2.degC is very close to what John calculated.

I get double the HW flow at 48c using 53C as the store temp, be interesting to see what the lowest required temperature actually is.

 

[siricosm]

be interesting to see what the lowest required temperature actually is.

In a few months I should be in a position to do that experiment, although I am not sure how low the water temperature will go in Norwich this winter.

Currently I have it set at just over 50C. I suppose the tank is warmer at the top, but that might be cancelled out by the thermostat differential. I can't find any specs on the cylinder stat. I am not sure what brand it is, RM Cylinders? they seem to be used by a few manufacturers.

So it looks like I can plan for a T40 system when purchasing radiators. I figure I can run at T30 (55C) most of the year and just turn up the tank temperature for the coldest periods. Too bad they don't seem to make weather compensating dual cylinder stats.

 

[John.g]

If you go for T40 Richard then you will only have to oversize by around 35%

Controlling your required store temp and still having a "full" cylinder is a bit of a challenge I suppose., you may be able to use the boiler SP temp as the required store temp. I think you stated in your other thread that you have a 17.6KW Boiler, assuming a boiler circ pump flow of 20LPM then the boiler deltaT is 12.6C, say 13C at full output, so at stat cut out (50C) the store temp is 63C but if the stat hysteresis is say 8C then the hot water to the store is only 55C (42+13) but ideally you are looking for a constant store temp, just thinking out loud, if you set the boiler SP to say 60/65C and set the store stat to 80C how would that work out?, the boiler will then fire as required and depending on the store bottom temp will fire either flat out or modulate (if you have a modulating boiler).

 

[siricosm]

If you go for T40 Richard then you will only have to oversize by around 35%

Controlling your required store temp and still having a "full" cylinder is a bit of a challenge I suppose., you may be able to use the boiler SP temp as the required store temp. I think you stated in your other thread that you have a 17.6KW Boiler, assuming a boiler circ pump flow of 20LPM then the boiler deltaT is 12.6C, say 13C at full output, so at stat cut out (50C) the store temp is 63C but if the stat hysteresis is say 8C then the hot water to the store is only 55C (42+13) but ideally you are looking for a constant store temp, just thinking out loud, if you set the boiler SP to say 60/65C and set the store stat to 80C how would that work out?, the boiler will then fire as required and depending on the store bottom temp will fire either flat out or modulate (if you have a modulating boiler).

That is encouraging about T40, because a 35% oversize for the radiators is manageable (affordable), even in cast iron, which is what I am planning to do.

Yes, it is an old Ideal FF360, so 17.6Kw output. I think the idea of using the boiler stat to control the temperature would indeed keep the store at a more uniform temperature, however there are a few problems with the equipment I have. The Ideal FF360 is pretty simple and it doesn't modulate. It doesn't even have a pump, it is external, and from the circuit diagram in the manual, there looks to be no way to control it with the boiler stat. This means that when the cylinder stat calls for heat, the pump is always running. So if I were to rely on the boiler stat, the pump would be running all the time. I am guessing it would also cycle, when it started to get close to the SP. Another problem is that the cylinder stat I have only goes to 65C.

Currently when I get up in the morning to have a bath/shower before heading off to work, the store is in a random state. It may have just heated, or it could be cooled off a bit with the boiler just about to fire. Either way, the water temperature at the hot tap is still too hot hold your hand under. After a short time with the hot running, the boiler will fire, and by the time I am finished the boiler is usually off. Currently it must be close to being able to replace the heat as fast as it is lost. The cold feed is around 19C or so now, and the rads are off, so maybe 17.6Kw is enough for infinite hot water in current conditions. This will surely change in the winter when the incoming cold water temperature is lower and the rads are on. Then I might need to rely more on the heat storage in the tank itself, and I will probably need to raise the temperature. I will experiment and post results here.

Another option of course is a bigger boiler. That is another advantage of a thermal store, they directly heat the tank water, and do not have to rely on a coil. With our previous vented cylinder, the boiler used to cycle for some time after one was out of the tub. With a normal vented/unvented cylinder, a bigger boiler wouldn't necessarily give you more hot water, as you are limited by the cylinder's input coil size.

 

[John.g]

Re temperature control, there must be plenty of systems about like yours with oil fired boilers which obviously also don't modulate so wonder where the installed stat is located.
I think you also said you have "only" 9kw of installed rads so that store should provide plenty of buffer even with rads i/s (assuming boiler cuts in early) except you are running baths/showers all day and that boiler should have no problem in topping up

 

[siricosm]

In a few months I should be in a position to do that experiment, although I am not sure how low the water temperature will go in Norwich this winter.

A quick update. It looks like with the cylinder stat set to 55C the 165kw Nordic Tec pex is plenty to keep up with our hot water demand, although as I mentioned before our flow rate is not so great. With some pipe thermometers under insulation on the pipes, my measurement of the temperature at the top of the tank goes between 55-65C when the thermostat cuts in and out. We no longer run out of hot water, but to be fair, our previous vented cylinder was only 120l whereas this store (heat bank?) is 300l.

The house feels warmer with the buffered central heating setup and TRVs controlling all the rads. It seems that in some rooms the rads are oversized, and others undersized, and the rooms that were cold before are much better now, even though the flow temperature is a bit lower. I think it is because the rads are always on when they need heat, whereas before they would cycle with the wall thermostat. All the rads now have TRVs with the lockshields wide open and the wall stat is gone. It is now very easy to control the temperature in each room individually.

The boiler cycling has stopped. It comes on for around 15min at a time throughout the day depending on demand. When there is heavy hot water usage, it just stays on continuously until it catches up.

 

[John.g]

Hi Richard, Thanks for reporting back and glad that your system is performing well.
Will read back through the posts tomorrow but as a quick update, what is the boiler stat set to? and where is the cylinder stat located?.

 

[siricosm]

The cylinder stat is in a dry pocket at the bottom of the cylinder and the boiler stat is set to max, which according to the manual has a maximum of 82C. However, the water coming from the boiler doesn't seem to get much over 70C by the time it gets to the tank, but it could be that I just don't have an accurate way to measure the temperature. If I turn the boiler stat down, it cycles towards the end when reheating the store. The boiler is quite some distance away from the cylinder with a lot of 90 degree bends along the way. It is piped in 22mm pipe, with maybe 20 or so bends and 15m of pipe (one way), and is currently pumped with a Grundfos Alpha 2 set on fixed speed 3 (top setting). A stronger pump might make a difference, but my plan was to eventually get the boiler out of the kitchen and move it next to the cylinder when the boiler dies.

The pump on the pex is a Grundfos Alpha 2L at fixed speed 2 (middle setting), so I thought it might need to be increased to 3 when it got colder, but apparently not. The pump for the central heating is a Grunfos Alpha 2 on constant pressure speed 2 (middle setting) which I though also might need increasing since it is a microbore system, but again the middle setting seems to work OK.

 

[John.g]

Very interesting as you are essentially controlling the store temperature with a cyl stat located in a cool part of the cylinder where the temperature will vary with PHE only in service (summer) or the rads only (winter) or rads+PHE (winter), the store (top) temperature will then depend on the cyl stat setting and hysteresis and the dT through the boiler, in turn which depends on boiler output and pump circulation rate. You are quite correct IMO to keep the boiler SP at 82C as it can't cycle then except that the flow rate was drastically as low as ~ 9 or 10LPM with that 17kw non modulating boiler?.

I think I asked before but what is the cylinder stat hysteresis?.

 

[siricosm]

The cylinder stat is made by RM Cylinders, but there look to be a lot around with the same design, probably made in China. There a picture of the stat a few posts back. I estimate the differential to be around 10C. In the warmer weather it was taking around 13-15min to re-heat 300l with no heat load (ie no hot water draw off and no central heat) with the 17kw boiler. It is an old Ideal FF360 so it does not modulate at all, it is either on or off. The thermostat differential is larger than I expected, the stat came as a package with the tank, along with an electric immersion heater. The electric heater is not connected, I suppose one could use it as a backup heat source in a pinch.

You are correct about the hot water. Because the return from the pex enters the tank at the bottom where the cylinder stat is (stat is a few inches above), it only takes a few minutes for the boiler to kick in when we start using hot water. The boiler will then stay on continuously during the shower or bath, and then stays on afterwards until the store reheats. This is one area where the thermal store appears to differ from an unvented/vented cylinder. When our previous cylinder ran out, the water was cold, and the boiler would cycle trying to re-heat it because the indirect coil was so small. It would take quite some time after a bath before you could get any hot water. With the store, the boiler directly heats the store water, and starts shortly after you start using hot water. So the water never gets cold, but cools down gradually as you run out. Since it reheats from the top, and doesn't cycle, you start to get hat water again very quickly. With our 17kw boiler and our rather poor flow rate, this means we have limitless hot water in the summer. With a 9kw heating load on a very cold day, it should eventually run out, but we have not experienced that yet.

I am pleased with the hot water performance, but most of the time the vented cylinder we had before was OK. The buffered heating however is much nicer, and the reduction in cycling has to be a good thing for the boiler.

 

[John.g]

The stat positioning is interesting, you say that "With a 9kw heating load on a very cold day, it should eventually run out", is that because the rads return are higher up in the cylinder than the stat and the return is relatively hot so the stat will not kick in until the whole cylinder cools down and the rads return will then also be cooler?.

 

[siricosm]

I was thinking that with a significant heating load, and colder incoming water in the winter, our 17kw boiler will not keep up heating the water forever, like it can in the summer.

The central heating and plate exchanger return to the tank in the same connection. This Telford tank only has two connections towards the top, and two at the bottom, so one pair are dedicated to the boiler, and the other pair are shared for DHW and CH. The cylinder stat is in the upper pocket of the two. The stat pockets look farther apart because the tank in the picture is only 90l whereas mine is 300l and taller, both cylinder pockets are actually quite close to the bottom. They could have put a few more connections on this tank, as it is only a buffer with no coils, but at £430 with a 3kw immersion heater, cylinder stat and TPRV+tundish included, this stainless tank seemed like a deal. That was the pre-pandemic price.

 

[John.g]

So you are thinking that a 17 kw boiler can,t keep up with the combined demand of both CH & DHW even if the Boiler is running continuously?

 

[siricosm]

I suppose in the winter the incoming water temp could be as low as 5C, so at 10 lpm with a 40C rise it would require around 28kW with the hot water running full blast. Add another 7-8kW for CH on a cold day and it could get up to 35kW peak demand.

In the summer, if the incoming water was 18C, 17kW could provide 9lpm at 45C, so it could likely keep up forever.

The winter numbers above are a bit pessimistic, but with a more reasonable 10kW deficit, the tank should last around 40 minutes, or 400 litres if it starts at 65C and drops to 45C.

 

[John.g]

The combination of CH demand + mains water at 5/6C certainly requires a lot of energy especially if running HW demand at 10LPM for 30/40 minutes, the 120 litre cylinder wasn't much use with that sort of use.

What ~ is the DHW temp at 10LPM with cylinder temp at 50C. (current mains temp ~ 7.5C) as I'd like to do another LMTD calc on it.

 

[siricosm]

I will see if I can measure that. I'll have to find a day/time when I can drop the store temperature without the family noticing.

 

[John.g]

No need to drop it, just give as accurately as you can any current readings, the important ones are the store temp, the DHW temp and estimated DHW flow, you can always take a another set of readings sometime when the store temp is at its lowest due to a big demand for both DHW & CH.