Dear Delphin Team,

I would like to evaluate the effective U-value of a wall from my Delphin simulation.

1. Is there a direct output of the U-value in Delphin?

2. If the U-value cannot be directly exported from Delphin, then I assume that it should be calculated from the temperature difference and heat flux. If this is the case, is there a certain recommendation for the threshold of the temperature difference for the U calculation? (Naturally, simulation results in the days with small indoor/outdoor temperature difference cannot be used for such calculation).

3. Is there a recommendation on how to average the transient hourly calculated U-values into a single U-value? This would depend on how the thermal mass is considered in Delphin.

Thank you very much for your help.

HA

## U-value from Delphin

**Moderatoren:** ghorwin, fechner, ruisinger

### Re: U-value from Delphin

Hello,

1. No, we haven't a direct output of the U-value

2. You should use mean values over a longer period for all values. This you can easy calculate by using the setting time integration in ouput formats. A mean value is in this case the difference of two integral values divides by the period between them

3.I would never calculate transient U-values. because the thermal storage suc values makes no sense. Keep in mind a U-value is designed for representation of a steady state heat loss.

1. No, we haven't a direct output of the U-value

2. You should use mean values over a longer period for all values. This you can easy calculate by using the setting time integration in ouput formats. A mean value is in this case the difference of two integral values divides by the period between them

3.I would never calculate transient U-values. because the thermal storage suc values makes no sense. Keep in mind a U-value is designed for representation of a steady state heat loss.

### Re: U-value from Delphin

Hi,

"Heatflux across unit surface vs. given constant temperature difference".

An "effective U-value" in a transient simulation would probably be defined by

"Average heatflux across unit surface vs. given average temperature difference"

Of course, solar or longwave radiation are not considered and shouldn't be taken into account.

So, in DELPHIN the straight-forward way to get this quantity is to define a heat conduction flux output with time handling type "Average values in time [Mean]". Simularly, the boundary temperatures should be averaged. This can be done separately (Excel etc.) or you can log the mean surface temperatures.

Mind, that the output interval must be large enough to capture the transients of the construction itself. Hourly outputs means also "average in hourly steps" and that will likely give wrong values. For typical constructions I'd recommend output intervals of 2 days or more. Also, it would be an error to simply take normal "moment" heat flux values and temperatures and use those for calculation.

With the mean temperature values computed, you can get the temperature difference and compute the average/effective U-value. Mind: if you are logging mean surface temperatures from DELPHIN, you need to add the surface heat conduction resistances as well, i.e.:

First get the mean heat transfer resistance:

1/R_wall = Flux/DeltaTSurface

Then add the surface resistances:

R_ges = R_i + R_wall + R_e

And get the U-value:

U = 1/R_ges

Final note: the reason why we do not have this output in DELPHIN natively is that it is generally a bad idea to mix physical quantities defined for

U-value is defined as steady-state quantity used to express the

"Heatflux across unit surface vs. given constant temperature difference".

An "effective U-value" in a transient simulation would probably be defined by

"Average heatflux across unit surface vs. given average temperature difference"

Of course, solar or longwave radiation are not considered and shouldn't be taken into account.

So, in DELPHIN the straight-forward way to get this quantity is to define a heat conduction flux output with time handling type "Average values in time [Mean]". Simularly, the boundary temperatures should be averaged. This can be done separately (Excel etc.) or you can log the mean surface temperatures.

Mind, that the output interval must be large enough to capture the transients of the construction itself. Hourly outputs means also "average in hourly steps" and that will likely give wrong values. For typical constructions I'd recommend output intervals of 2 days or more. Also, it would be an error to simply take normal "moment" heat flux values and temperatures and use those for calculation.

With the mean temperature values computed, you can get the temperature difference and compute the average/effective U-value. Mind: if you are logging mean surface temperatures from DELPHIN, you need to add the surface heat conduction resistances as well, i.e.:

First get the mean heat transfer resistance:

1/R_wall = Flux/DeltaTSurface

Then add the surface resistances:

R_ges = R_i + R_wall + R_e

And get the U-value:

U = 1/R_ges

Final note: the reason why we do not have this output in DELPHIN natively is that it is generally a bad idea to mix physical quantities defined for

**steady-state**conditions (like U-value, PSI-values etc.) with**transient simulations**. Often it is better to define meaningful equivalent quantities, like*mean heat loss across the room surface over a certain cold period under defined indoor and ambient conditions*. This heat loss maps directly to energy costs and CO2 emissions and is often much more expressive than U-value comparisons (that neglect any thermal storage characteristics of the selected construction).### Re: U-value from Delphin

Thank you very much for the extensive replies!