Thermal bridging

[aaron-boranian]

Problem: EnergyPlus provides no method to describe thermal bridges in envelope components.

Solution 1: A very basic approximation would be to allow independent parallel path calculations for the same construction (e.g., wood stud and insulation cavity).

Solution 2: Develop adjustments to the conduction calculations for linear and point thermal bridges, similar to Psi factors.

Migrated from UserVoice feedback from @nealkruis

[aaron-boranian]

Multi-Dimensional Heat Transfer through the Building Envelope

Problem: EnergyPlus models heat transfer through the building envelope as an one-dimensional problem. There are almost no exceptions to this (low temperature radiant systems can model 2-D heat transfer). While thermal bridging modeling would be a good step forward, ultimately more accurate modeling of edges, corners, etc. would need either a full 2-D or 3-D transient conduction model.

Rationale: The assumption of EnergyPlus that all conduction through the building envelope is one-dimensional is limiting and is not as accurate as possible when walls are not infinitely long without any breaks and the accuracy diminishes as the wall shortens in either dimension, windows are added, etc.

Solution: Modify EnergyPlus to allow upon request at least 2-D modeling of the building envelope. This would require the developer to not only modify the current conduction transfer function method of determining CTF coefficients but also come up with both a method for inputing these constructions and interfacing with the zone air heat balance which is somewhat hard-wired to interface with a single temperature per surface.

Context: See legacy enhancement list item Envelope_2007_22.doc

Migrated from UserVoice feedback from @RKStrand

[rd2]

(it's been 4 years since the previous comments were added, yet this New Feature Request remains open - a good sign).

The Thermal Bridging & Derating (or TBD) OpenStudio Measure (initially released late 2020, under semi-continuous maintenance) addresses a large number of points raised here. TBD (relying on the Topolys gem) automatically tracks (shared) edges of above-grade envelope surfaces (linear thermal bridges), tags them based on topological relationships (e.g., concave vs convex corners, subsurface perimeters, floor-linked shading surfaces as inferred balconies), estimates related quantities (e.g., edge lengths, number of point conductances), and finally thermally derates linked opaque constructions, as per ISO 10211/14683 and/or BETBG methodologies (they largely overlap). TBD also offers a number of OpenStudio related features, like PSI-set management/inheritance (similar to default construction sets), auto-generating Kiva inputs, etc. - but these do not pertain to the core issue discussed here.

We're not suggesting that TBD should be considered as a definitive answer. Au contraire, we would prefer seeing future, built-in EnergyPlus features that could do much of the heavy lifting, namely automated Topolys processes: maintaining edge objects in memory, tracking linked surfaces, etc. This opens the door to other enhancements down the line (AFN-related, edge-linked costing outputs). While TBD is an OpenStudio Measure, Topolys remains by design standalone and OpenStudio-agnostic. Both are distributed freely (MIT licensed), and so can be freely ported into EnergyPlus (if so desired).

We do suggest that potential enhancements remain compatible (as much as possible) with ISO/BETBG methodologies, given the increase in related requirements in codes/standards. But there's wiggle room, in part due to inherit limitations of the methodologies. In a nutshell, ISO/BETBG methodologies are (more) suitable in relatively cold climates (i.e. noticeable heating loads) on envelope assemblies with one or more (~massless) insulation layers. In increasingly colder climates (e.g. going from Climate Zone 6 to Climate Zone 7), one typically sees more clearly defined insulation layers in envelope constructions (and of course higher R-values), yet with thermal inertia (in envelope assemblies) playing an increasingly marginal influence on building energy use. ISO/BETBG methodologies start to lose relevance when moving further South, depending on the constructions. For instance, it is doubtful that derating wall layers (i.e. changes in layer thicknesses) is a good idea when faced with uninsulated solid masonry assemblies or similar (e.g., adobe construction, rammed earth walls, warehouses with painted CMU walls [with little to no insulation], uninsulated CLT construction). In such cases, the solution may be to extend ISO/BETBG methodologies to factor-in multiple layers at a time, and more importantly factor-in proportionate adjustments in layer density and specific heat - à la RP-1145. Or to track dynamic edge RF/loss/gain distinctly (yet in relation to) linked surfaces (instead of a strictly 1D approach, as suggested in previous comments), opening the door to localized surface temperature assessments (e.g., comfort, condensation). Options to consider.

PS - It would be nice to have access to the "Envelope_2007_22.doc" (or updated version).

[the-donn]

In New Zealand we have required thermal bridge calculations using the Isothermal Planes approximation which area weights the heat transfer through framing and insulation portions of a wall for around 30 years. However, as we examine mould formation and condensation risk, within a still 1 Dimensional calculation, making the modelling of these two parallel heat flow paths - accounting for the thermal inertia in each seems increasingly important in our relatively mild, high humidity climates. Accounting for this is a high priority for users of E+ in our situation.