A guide to TBD - a Ruby gem and OpenStudio Measure that auto-detects major linear thermal bridges (like balconies, parapets and corners), then derates outdoor-facing opaque surface constructions (of walls, roofs and exposed floors). The guide is mainly written for architects, technologists and envelope consultants who are new to energy simulation, and to OpenStudio in particular.
In a nutshell …
Thermal bridges are structural elements that interrupt the continuity of insulation in building envelopes. A curtain wall spandrel back pan may hold in place an R17 (RSi 3.0) batt of insulation, yet the spandrel’s overall R-value would likely trickle down to R5 (RSi 0.9) - less than a third of its nominal value. This drop in performance is due to spandrel area (vs its height-to-width ratio), how the back pan is held in place continuously along spandrel edges, and its use of highly conductive materials (e.g. galvanized steel, aluminium).
Minor thermal bridges are regularly-spaced supports or framing elements (such as studs, Z-bars and fasteners). The initial derated R-value stemming from minor thermal bridging, generally known as a construction’s clear-field effective R-value, is largely independent of a surface’s actual geometry or its proximity to neighbouring surfaces (e.g. around windows, along corners). This means design changes to surface geometry (e.g. floor-to-ceiling height, number of windows) can be made without having to update surface clear-field effective R-values - very practical! For a few simple 2D framing configurations, the ASHRAE Fundamentals and ISO standards support established hand-calculations like the parallel-path method. Yet in most cases, designers are better off consulting published collections of common configurations (with unique clear-field effective R-values), such as BETBG or thermalenvelope.ca.
Major thermal bridging instead relates to a surface’s geometry and shared edges with neighbouring surfaces (e.g. along roof parapets, slab perimeters, cantilevered balconies). While U-factors (in W/K per square meter) are suitable metrics for surface area heat loss (under standard winter rating conditions), linear conductances from major thermal bridges are commonly annotated using the greek letter psi (units in W/K per meter) - khi for point conductances (e.g. a cantilevered beam or a column, in W/K per point). Both BETBG and thermalenvelope.ca links above provide useful psi and khi data for common cases.
Contrary to minor thermal bridging, changing a room’s height or adding windows should trigger a revised calculation of major thermal bridging effects. This can be quite daunting, time-consuming and error-prone to do by hand (per design iteration), given the hundreds (if not thousands) of major thermal bridges in a building model. The simple US DOE Commercial Reference Warehouse Model has over 300 of such major thermal bridges - mostly around fenestration.
Commercial Reference Warehouse
Relying on the OpenStudio SDK and the Topolys gem, TBD automatically - and pretty instantaneously - identifies and manages major thermal bridges behind the scenes for OpenStudio users. For ASHRAE climate zone 7, enabling TBD should increase annual heating requirements between 5% to 15% (depending on building type) for “poor” to “regular” thermal bridging details in an otherwise well-insulated envelope.
Building energy modelling
While materials, constructions and envelope surfaces are well-defined variables in OpenStudio (and energy simulation engines like EnergyPlus), shared edges and points are simply not! TBD automatically factors psi and khi losses from major thermal bridges it manages, by further derating a construction’s clear-field effective R-value - more specifically by further decreasing its insulating layer thickness. If the insulation layer becomes too thin, TBD then increases insulation conductivity. This approach, in line with published research and standards such as ASHRAE’s RP-1365, BETBG & thermalenvelope.ca, as well as ISO 10211 and 14683 Standards, is best summarized as follows:
Ut = Uo + ( ∑psi • L )/A + ( ∑khi • n )/A
… where:
Ut : derated construction transmittance
Uo : initial clear field transmittance
psi : linear edge transmittance
L : length of the edge
khi : e.g. column point transmittance
n : number of similar columns
A : opaque surface area
TBD users are required to initially select generic psi and khi factors that best characterize the major thermal bridges in their project (see Settings). TBD will apply these values against individual edge occurrences Topolys identifies in the OpenStudio model. Geometric variables like L and A are also automatically extracted from the model (… n requires special treatment, discussed in the Customization section).
Each OpenStudio construction (that TBD may alter) is comprised of multiple material layers (typically 2 or 3 at a minimum), each of which has a thermal resistance. Users are expected to have already factored in minor thermal bridging effects, for instance by decreasing the nominal thickness of the insulating layer of each construction - a common technique in building energy modelling. Uo is simply the inverse of the sum of resulting layer resistances and surface air films. Behind the scenes, TBD automatically generates new derated materials and surface-specific constructions - each surface now having its own unique Ut.
As discussed in greater detail in Settings, users are required to have:
- a fully-enclosed OpenStudio model
- materials and layered constructions
- a shortlist of psi and khi factors
… from there, TBD & Topolys will do the heavy lifting!
Next?
Support
Merci aux gouvernements du Québec et du Canada (CNRC, CanmetÉNERGIE).
As with many publicly available OpenStudio Measures, TBD is open source, MIT-licensed and so provided “as is” (without warranty).