Debating All Aspects of Passive Houses - And May the Best Argument Win!
You are not logged in.
Trying to determine the psi-value at the ground/perimeter wall junction would appear to be the proverbial "Can of worms." How then can two dimensional methods (PHPP and EN 13370) be satisfactorily harmonised when seeking to determine meaningful psi-values? Here are some thoughts and a proposed solution. I've also attached a PDF that diagrammatically walks though the issues and the proposed solution.
1) EN 13370 states that the thermal bridges at edge of floor are “based on an isolated floor that is considered independently of any interaction between floor and wall. They also assume uniform thermal properties of the soil (except for effects solely due to edge insulation). In practice, wall/floor junctions for slab-on-ground floors do not correspond with this ideal, giving rise to thermal bridge effects. These shall be allowed for in calculations of the total heat loss from a building, by using a linear thermal transmittance.”
On this basis thermal bridges are excluded from the basic calculation however, the standard does provide “typical psi-values for slab-on-ground floors” (Table 3)…. It goes on “This table may be extended on a national basis to include specific wall/floor details, and for a particular dimension system, *provided* that these values have been obtained in accordance with Annex A.
2) ISO 13370 Annex A: sets out a procedure for establishing the psi-value at the wall/ground junction. Importantly all dimensions are measured internally! There is no method for external dimensions. Due to the external measurement protocol of PHPP this raises questions about how to estimate the psi-value for a PassivHaus. The concern arises from the fact that unlike other corner elements there is, due to a clash in protocols (13370 vs PHPP), a dimensional gap occurs that is the thickness of the wall and height of the floor.
3) EN 13370 also states that the linear thermal transmittance associated with basements is small and may be neglected” [Is this true for a PassivHaus?].
4) In PHPP all heat losses are based upon external dimensions.
5) The PHPP user guide states that EN ISO 13370 is used to determine the U-value of the floor. It also notes that the algorithms in ground sheet are an "improvement" upon EN ISO 13370. It then references AkkP 27 for further information. The user guide does not discuss the protocol for determining the psi-value at the wall/ground junction.
6) Internally consistent: To ensure accuracy all calculations must be internally consistent and this relates particularly to the way in which thermal bridging is calculated. To this end the modelling U-value (say THERM U-factor) should presumably be used in all calculations and should not be mixed with EN U-values.
7) ) In PHPP, on the Ground worksheet states an “Average Ground Temperature” and a “Ground temperature amplitude.” What impact will this have upon calculating the psi-values? In THERM should the boundary conditions for ground be given this temperature? (rather than ambient as per the assumptions in EN 13370)
…..If the boundary condition is to be set what should be the assumed film coefficient?..
Summary and Conclusion:
In light of the above, when determining the External psi-value for the wall/ground junction, my thoughts are that one should use the internal modelling U-value of the floor (THERM U-factor.) The modelling U-value of the floor may be determined by calculating L2D,2 in accordance with the EN ISO 13370 (fig A.2). This modelling U-value may then be used in the psi-value calculation to determine the notional external heat losses. (The modelling U-value of the external wall, the externally measured losses of the wall and the subsequent external psi-value would be determined in the usual manner.)
To my mind this approach would seem to be internally consistent as it is solely reliant upon the U-factors developed by the modelling tool (thus avoiding external distortions), it allows the conceptual framework for floor U-value EN ISO 13370 to be included, it is capable of including the PHPP external calculation methodology and finally if over comes the internal EN ISO 13370 compliant psi-value methodology.
This approach would seem to be appropriate as no errors are raised in the "check tool" …..so it seems to give the appropriate accuracy. (The "check tool" compares the total heat losses from dimensional differences of the building elements with the total difference between the internal and external psi-value. If they agree at zero then accuracy is achieved.)
The Questions:
1) How then can two dimensional methods (PHPP and EN 13370) be satisfactorily harmonised when seeking to determine meaningful psi-values?
2) For internal consistency should the modeling U-value be used in all calculations? (i.e. Should the modelling U-value of the floor as determined by calculating the L2D,2 in accordance with the EN ISO 13370 (fig A.2) be used in the calculation or should it be the EN 13370 U-value?)
3) Should ground temperatures be applied to more closely mirror the assumptions in PHPP?
4) In a PassivHaus can the linear thermal transmittance associated with basements be ignored (as EN 13370 suggests)?
5) Over all is the appraoch that I propise the right one or should the external psi-value be determined in another manner?
I look forward to hearing your thoughts.
Mark
Last edited by Mark S (Thu, 05/02/2009 22:29:24)
Offline
thank you for all this,
i want to reread it carefully.
but where should i look for that PDF ?
someone please help
Offline
Romeo,
Currently the forum does not allow documents to be uploaded. I found after having written the post I could not attach the document. ;-( Patience. This may get fix sometime soon....
Mark
Offline
Sorry, spamers are already a pain. If I'd allow uploading files, I would spend my days eradicating inappropriate files, and - given current legislation - would probably even be liable in some cases. Only way out would be to make this an invitation only forum. I don't want to do that at this stage, but consider it for the future.
Offline
at this moment i am stuck because of the same problem : use of the internal dimensions in ISO 133370 when computing psi for the wall-ground floor junction.
but if this is not too late, i have an answer to no 7) in Mark's post: it does not matter what temp you use for the ground surface boundary.
i noticed the same L2_D all the time, whatever that temp.
re the film coefficient i am using the same value as for other exterior surfaces of the building, that is h= (1 / (Rse=0.04) ) = 25.
is there any specification to use a different value for the ground surface ?
by the way, when computing ground floor slab thermal bridges (for its "central" zone not the perimeter), what is the boundary condition you use for its down side (the one toward the ground) ?
romeo
Offline
Mark, if you are still around,
could you please upload the promised pdf (on rapidshare for example) so that it can be downloaded and seen ?
i am not very sure that i understand you right.
at this moment, my opinion (not final though) is the following:
L_2D is the same when computing a detail having several adiabatic boundaries and two oher boundaries, one internal (Ti, Rsi) and one external (Te, Rse) , irrespective of the dimensions you use (internal or external).
if we keep this "rule" and apply it to Fig. A.2 in EN ISO 13370, then we know L2_D of that detail (in Fig. A.2) even if it was calculated using the internal dimensions.
you denoted L2_D of that detail by L2D,2 in your post (as in the standard).
if we again keep the "rule" when calculating the detail of the wall-ground floor junction, we get its L2D,1 which again is characteristic to the detail, it does not matter that we got it by using internal dimensions.
all we have to do now is calculating psi for the junction, that is subtracting the 1D contribution of the wall and the floor.
i guess the 1D contribution of the wall should be:
U-wall*l_ext_wall where l_ext_wall is the external dimension of the wall
re the 1D contribution of the ground floor slab, that is specifically defined by the standard, namely L2D,2.
so if i am not wrong, we get what we look for in the case of external dimensions as
L2D,1 - L2D,2 - U_wall*l_wall_ext
romeo
Offline
Hi Romeo,
Still around don't worry. If I could get on rapidshare or similar I would. Access is currently denied by the offices flamewall! (A major frustration.) I'll try to talk our IT guy into playing ball.
Mark
Offline
hi Mark,
you have a question about the modelling U-value of floor in the sent pdf: "derived from model based on Fig A2 ?"
how do you think to derive U-value for floor from Fig. A2 ?
i am not sure what to understand by the "modelling U value of the wall". i understand this is not its 1D U-value, am i right ?
the key point seems to me accepting L2D,2 or not as the right measure.
L2D,2 derived from Fig. A2 stands (if i understand well, please correct me if i am wrong) for the contribution of the floor alone (no wall).
this is the convention made in the standard (a convention which i think is linked to the assumption made in Fig. A1, namely those adiabatic boundaries in the ground. because of those special boundaries, the heat-transfer-to-the-ground case is not like a "common" case, where the exterior face of the opaque envelope element faces the exterior air. when that opaque element is a floor, such an example is an overhang. we would use then l_ext_floor*U_floor, where U_floor is the 1D U-value of the floor. but that's not the case for a ground floor).
what other way of defining the contribution of the floor alone could we imagine ?
i guess a modification might be made to Fig. A2 in which the whole external surface of the floor is present. but what should we do about the internal surface of the floor in that case ? should we "glue" to the floor a portion of the wall having the thickness of the floor ? in order to have the internal surface equal to the external one (as it is in Fig.A2 where both surfaces of the floor are equal to its internal surface). what should be the boundary condition imposed on the upper surface of the "glued wall" ? maybe we could similarly assume an adiabatic boundary there. that would be possible.
one possible objection to such an approach could be the fact that we mixed a generally different structure (that of the wall) with the structure of the floor in order to define the contribution of the floor alone.
it seems to me, the most "natural" way of defining the contribution of the floor alone is the one illustrated by Fig. A2:
its contribuiton is the heat flux "collected' by its internal surface. an adiabatic boundary is considered at its junction with the wall meaning the heat exchange between the two is "stopped', no interference of the wall in the heat transfer is allowed.
so if WE MAKE THE CONVENTION
L2D,2 is taken as the contribution of the floor alone (irrespective of the way we compute the heat losses for the building envelope, either by using internal dimensions or external ones),
than
we have to subtract L2D,2 from L2D,1.
and we also have to subtract the contribution of the wall. things seem clear for the wall. that contribution would be
l_wall_int * U_wall in case we use internal dimensions
and
l_wall_ext * U_wall when we use external dimensions.
the idea of this approach is the contribution of the floor alone is THE SAME
but the contribution of the wall IS NOT when we use internal respectively external dimensions.
i also have a different but connected problem:
did you compute ground floor slab thermal bridges for the "central" region of the slab ?
what should be the boundary condition for its side adjacent to the ground ?
PHPP reads " the surrounding ground is not considered in this calculation" (p.61).
romeo
Last edited by incepator (Mon, 27/04/2009 20:20:11)
Offline