output from existing radiators

[ajrobb]

I am trying to tot up the nominal output of my existing radiators. Some are old single or double panel radiators without any convector. I have looked at existing tables for output from single or double panel with one or two convectors and have come up with the following approximations to deal with odd sizes:

single panel, single convector: output = (1.52 x height + 0.120) x length
double panel, single convector: output = (1.9 x height + 0.314) x length
double panel, double convector: output = (2.69 x height + 0.320) x length

output in kW
length and height in m

Does anyone have tables for old single and double panel radiators (no convector)?

Edit: Using Zylo data for finless rads:
single panel: output = (1.27 x height + 0.104) x length
double panel: output = (2 x height + 0.157) x length

 

[D_Hailsham]

Try Guide to Radiator Outputs

Which manufacturer's data did you use to obtain your formulae?

 

[spraggo]

If you could state the actual size of the rads, ie, 450mm X 540mm and whether they are, for instance, Stelrad or Thorn (round top) I could probably answer your query.

Regards

spraggo

 

[ajrobb]

I used Stelrad Quinn - although differences to Zylo aren't huge.

 

[D_Hailsham]

single panel, single convector: output = (1.52 x height + 0.120)*length
double panel, single convector: output = (1.9 x height + 0.314)*length
double panel, double convector: output = (2.69 x height + 0.320)*length

output in kW
length and height in m

There is a major flaw in these formulae: when the height is zero, the radiator will still be giving off heat!!

Better formulae (based on Stelrad Elite) are:

H= height in metres
Output in Watts per metre length.

Single Panel, no convector: 1195.5H - 287.9H²
Single panel, with convector: 1815.6H - 257.3H²
Double panel, single convector: 2796.8H -740.69H²
Double Panel, double convector: 3677.8H - 1166.3H²

 

[D_Hailsham]

I used Stelrad Quinn - although differences to Zylo aren't huge.

Stelrad and Quinn are rival companies!

 

[ajrobb]

There is a major flaw in these formulae: when the height is zero, the radiator will still be giving off heat!!

Nothing wrong with that. A quick look at any radiator table will show that shorter radiators give off more heat per unit area.

Just totted up the radiators (30 kW), de-rating them to 75% (taken from BS conversion chart) for mean temperature difference of 40°C (70°C dropping to 50°C in a room 20°C) instead of 50°C, I get a grand total of 22.5 kW. Boiler size calculator suggested about 44 kW.

 

[D_Hailsham]

There is a major flaw in these formulae: when the height is zero, the radiator will still be giving off heat!!

Nothing wrong with that. A quick look at any radiator table will show that shorter radiators give off more heat per unit area.

Of course it is wrong; a zero height radiator cannot give of any heat.

Just totted up the radiators (30 kW), de-rating them to 75% (taken from BS conversion chart) for mean temperature difference of 40°C (70°C dropping to 50°C in a room 20°C) instead of 50°C, I get a grand total of 22.5 kW. Boiler size calculator suggested about 44 kW.

Do you mean dropping to 60°C?

The BS conversion chart assumes that the temperature differential across the radiator is a constant 75°C/65°C and only the room temperature changes up/down from 20°C

 

[ajrobb]

Of course it is wrong; a zero height radiator cannot give of any heat.

I plotted the data for 5 radiator type into Excel and it shows the regression line for 3 of them. You have to admit, they're pretty close.

Untitled

• ajrobb
• 2 Dec 2010
• 1

Do you mean dropping to 60°C?

The BS conversion chart assumes that the temperature differential across the radiator is a constant 75°C/65°C and only the room temperature changes up/down from 20°C

No, the mean radiator temperature is assumed to be 70°C and room temperature is 20°C. When the BS was brought into line at BS EN422-1997, I thought the standard mean difference dropped to 50°C.

With a 20°C temperature drop across the radiator and a feed temperature of only 70°C, the mean temperature on my radiators is 60°C, so I have to adjust down from 50°C mean difference to 40°C. That means going from a BS-to-EN standard correction of 0.798 down to 0.605.

Now, whether the ratings that are quoted are pre or post 1997, I cannot be certain.

Now, a couple of radiators have reflective foil on the wall behind them. How much does that reduce their output?

 

[ajrobb]

Try Guide to Radiator Outputs

Thanks again for the link. I corrected the tables to a mean temperature of 70°C and plotted them on Excel.

Untitled

• ajrobb
• 2 Dec 2010
• 1

There are three distinct intercept regions for zero height

• cast iron panel line crosses the Y axis close to zero
  double panel double convector radiator line crosses close to 200 W/m
  all other radiator types have lines that cross close to 100 W/m

To accept that the lines don't have to pass through the origin, I suspect that there are two heat transfer mechanism; a higher rate for a small height (less than 300 mm) then a lower rate for any height above this.

 

[D_Hailsham]

I plotted the data for 5 radiator type into Excel and it shows the regression line for 3 of them. You have to admit, they're pretty close.

Untitled

• ajrobb
• 2 Dec 2010
• 1

But the regression lines do not pass through zero on the Y axis when X is zero as there is a constant in each formula. I notice you show the regression formula but not the R² value.

Do you mean dropping to 60°C?

No, the mean radiator temperature is assumed to be 70°C and room temperature is 20°C. When the BS was brought into line at BS EN422-1997, I thought the standard mean difference dropped to 50°C.

I think we are at cross-purposes here. You said:

"Just totted up the radiators (30 kW), de-rating them to 75% (taken from BS conversion chart) for mean temperature difference of 40°C (70°C dropping to 50°C in a room 20°C) instead of 50°C."

If the mean temperature has dropped to "50°C in a room 20°C", the difference is 30°C, not 40°C. I was asking if you meant to write:
"60°C in a room 20°C", which would then give a temperature difference of 40°C.

With a 20°C temperature drop across the radiator and a feed temperature of only 70°C, the mean temperature on my radiators is 60°C, so I have to adjust down from 50°C mean difference to 40°C. That means going from a BS-to-EN standard correction of 0.798 down to 0.605.

Are you talking about the old BS3528 and the current BS EN442?

BS3528 used: Flow=90°C, Return=70°C, Room=20°C. This gives a 60°C mean difference. BS EN442 uses: Flow=75°C, Return=65°C, Room =20°C, a 50°C mean difference. I agree this gives a correction of about 0.798 (I make it 0.791!), but I'm not quite sure why you are using 0.605.

If you know the output of your current rads using BS EN442, the correction figure for Flow=70°C, Return=50°C, Room=20°C (mean difference 40°C) is 0.726.

[quote}Now, whether the ratings that are quoted are pre or post 1997, I cannot be certain.[/quote]
All rads are now quoted using BS EN442 (i.e post 1997). However I have just noticed that the Zylog data is based on a mean water temp of 75°C and a room temp of 20°C; this does not agree with any British Standard!

Assuming that it is based on a Flow temp of 80°C and Return of 70°C, the correction factor to bring them in line with BS EN442 is 0.925, i.e the outputs in the Zylog data has to be multiplied by 0.925 to give the corrected BS EN442 output.

 

[ajrobb]

I use "mean temperature difference" as the mean temperature difference between the water and the air. This is calculated as the average of the radiator flow and return temperature minus the room temperature.

(A better estimate of mean would assume an exponential temperature gradient rather than a linear one.)

 

[D_Hailsham]

I use "mean temperature difference" as the mean temperature difference between the water and the air. This is calculated as the average of the radiator flow and return temperature minus the room temperature.

I am using the same definition

A better estimate of mean would assume an exponential temperature gradient rather than a linear one.

I agree. The formula I use to calculate correction factors is based on this.