You can of course make an appointment, that we also have negative heat transfer coefficients for linear values and mathematically it will not be a mistake. It is allowed by the PN-EN ISO standard 14683:2007. This standard allows to consider thermal bridges in relation to the external surface, but it does not impose it, unlike the Regulation.
Such imposition causes a certain awkwardness, as some line bridges will be negative (corners, ceilings, etc.), some positive line bridges (e.g. inclusion of steel sheet passing through the wall, locks in panels of lightweight enclosures, joints of window and door joinery with the wall or bolts and posts in mullion and transom facades). Additionally, we have point bridges, which are also rather positive, I guess, that they appear in the area of the inter-story ceiling or corner.
In my opinion, there is too much inconsistency and deviation from the traditional understanding of the thermal bridge. Thermal bridges are always associated with something negative, that is, increasing U (causing greater heat loss), here they turn out to be something positive (reducing losses). The share of U itself in heat loss seems to be preliminary (tj. before considering bridges) distorted by overstatement, counting it not only along the partition itself, which corresponds to the value of U, but also an element that is not actually this partition, but takes part in the heat exchange indirectly – as heat sink on hot or cold side. I mean an inter-story ceiling or an internal wall on the hot side or an additional heat transfer area in the corner of the building on the cold side. Another quite significant reason, To stay with the traditional understanding of linear and point heat transfer coefficients, the results are comparable. Two results with the same sign can be easily compared and related to a given U-value. Two results with different signs, need an analysis of the geometry of the area of occurrence, to compare them.
The simplest reason for calculating the resistance / thermal conductivity characteristics of building partitions in relation to the internal surface is to follow the side of the heat source. The heat source is inside – from ceiling to ceiling, wall to wall. It seems to be the most natural and probably most used. Then the bridges are always positive and increase U in relation to the basic value. U0 jest najmniejszą możliwą wartości U dla danej przegrody. I imagine the linear bridge with a negative transmission coefficient as a band of increased insulation.
In the event of the need to obtain the same parameter, i.e.. U for the outer surface (for example, to calculate gains by insolation) simple conversion factors can be used:
U ’= U · Aw / Az
U = U ’· Az / Aw
where Aw - inner surface [m2];
Az - outer surface [m2];
U - heat transfer coefficient calculated for the internal surface [W/(m2K);
U’ - heat transfer coefficient converted to the external surface based on U,
calculated for the inner surface [W/(m2K)].
The third issue is the notation with an indication of the sources of information about linear permeation coefficients in the form:
„a) documententację techniczon budywings,
b) tablice mosties ciePLNand ch,
c) formczenia szczeheadline mostcenters ciePLNych.”
Indication of the technical documentation of buildings designed so far will in most cases be ineffective, as the design documentation usually includes only an indication of the U value of partitions without a method of calculating this value. Unfortunately, most often it is U0 lub U0 z dodatkami normowymi na mostki, based on old standards. The result can be very inaccurate. Documentation created in the future based on ORDINANCE MINISTER INFRASTRUCTURES, changing ordinance w case detailed range i forms project construction, as long as it is created reliably, may be the basis for the author of the energy certificate. What though, when the designer makes a mistake in his study and an incorrect certificate is issued on this basis? Who will bear the civil consequences - designer, or the author of the certificate? Seems, that it will not happen without verification.
Thermal bridge tables, are the simplest and most common source of data on the effects of thermal bridges. The problem is that, that these are standard details for standard structures. Distant analogies can lead to significant errors. PN-EN ISO 14683:2007 determines the accuracy of this method on 20%, when used appropriately.
Finally, it was defined as the source of data on linear heat transfer coefficients "detailed calculations of thermal bridges". However, it was not specified, what are "detailed calculations". You can guess, that they are suitable 2D numerical models for linear bridges and 3D for point and spatial bridges. From my point of view, it is the most precise way of determining the insulating properties of partitions, therefore the best from the point of view of e.g.. property owner, to which the certificate under development relates. The previously mentioned standard specifies the accuracy of this method on 5% and it coincides with my experience – 3 do 5% error for working calculations in relation to the previously defined strictly results of the same models.
A few years ago, there were no tools available on the Polish market for numerical analyzes of heat flow in building partitions. This problem no longer exists today. Anyone can get KOBRA programs (international product) or Thermal Analysis System (domestic product). Point 3.2.3., as one of the most important in the Regulation, should be clarified.
First of all, the provision of the basic formula and the source of data on bridges should be clarified and their consequences should be considered – in particular, imposing the outer surface as obligatory for partition analysis and the source of data on bridges.