Thermal: U-value & thermal bridges
The wall has a number that decides how much heat walks through it - and a single steel bracket can quietly cancel out a whole layer of insulation.
You can specify 100 mm of perfect insulation and still lose the heat through a bolt the size of your thumb.
Heat does not care about your elevation. It moves through the path of least resistance, every hour of every day, and a facade is a leaky dam holding it back. The headline number for that dam is the **U-value** - watts of heat crossing one square metre of wall for every degree of temperature difference. A good opaque wall is around 0.3 W/m2K; a bare single glass pane is nearly 6. But the U-value of the flat wall is the easy part. The real losses hide at the **thermal bridges** - the brackets, slab edges and frames where the insulation is punctured and heat takes a shortcut. In an Indian cooling-dominated tower this runs in reverse, but the physics is identical: control conduction and the junctions, or pay to pump the heat back out for forty years.
Conduction, resistance, and the U-value of a build-up
U-value is conductance: low is good, and it is built up from layer resistances
Conduction is heat marching through solid material, fastest through dense conductive stuff (aluminium, steel, glass, concrete) and slowest through trapped still air (insulation, cavities). We measure a material's resistance to that flow as its thermal resistance R, in m2K/W - thickness divided by conductivity (R = d / lambda). Thicker and less conductive both mean a higher R, which is good.
A wall is just layers in series, so you add the resistances: the outside air film, each material layer, any cavity, and the inside air film. The total is Rtotal. The U-value is simply its reciprocal: U = 1 / Rtotal, in W/m2K. Low U means a well-insulated wall.
The intuition that saves you: resistances add, so the weakest (thinnest, most conductive) layer barely helps and the insulation layer does almost all the work. Doubling 100 mm of insulation roughly halves the U-value; doubling the plaster does almost nothing.
R-value is a layer's stubbornness against heat. Add up the stubbornness, flip it over, and you have the U-value.
A thermal bridge is a conductive shortcut - and it gets its own number, the psi-value
A thermal bridge is any place where the insulation layer is bypassed by a more conductive path: a steel bracket spearing through the insulation to reach the slab, the un-insulated concrete slab edge of a curtain wall, the aluminium frame of a window, a balcony slab projecting straight out of the floor plate. Heat funnels through the shortcut, and the clean U-value you calculated for the flat wall no longer describes reality.
Junction bridges are quantified by the linear thermal transmittance, psi (W/mK) - the extra heat flow per metre of junction, over and above the plain wall on either side. A poorly detailed slab edge can carry psi = 0.5 to 1.0 W/mK; a thermally broken one, under 0.1. Point bridges (a single bracket) get a chi-value (W/K).
The honest, uncomfortable arithmetic: brackets at 600 mm centres, each leaking a few tenths of a watt per kelvin, can degrade a wall's effective U-value by 20 to 40%. The fix is a thermal break - a low-conductivity spacer (reinforced polyamide, structural thermal pads) that interrupts the metal path without losing the structural connection.
Cooling-dominated, but the maths is identical - and the code sets the ceiling
In most of India the heat flows inward, from a 40 C afternoon into a 24 C office, but U-value is direction-agnostic - the same low number that keeps a London flat warm keeps a Chennai office cool. The Eco-Niwas Samhita sets a residential opaque-wall U-value ceiling (generally U <= 1.0 to 1.2 W/m2K depending on assembly), and ECBC sets envelope U-values for commercial buildings. These are the targets your build-up must beat.
The trap is that Indian construction loves heavy, conductive materials - dense concrete, solid brick, stone - which have poor R-values for their thickness. A 230 mm solid brick wall sits near U = 2.0 W/m2K, roughly double the code ceiling, so you almost always need a deliberate insulation layer (often AAC block, EPS or rock wool) to comply. And every bracket that pins the cladding back through that insulation is a bridge waiting to be detailed out.
The U-value is set by your section, not your elevation. When you choose a material and its thickness you are choosing an R-value, and when you draw a thin, elegant slab edge with cladding hard against it you are drawing a thermal bridge. Two habits pay off: leave real depth for insulation (you cannot conjure R out of 25 mm of render), and treat every junction - slab edge, parapet, reveal - as a place the thermal line must continue unbroken, the same discipline as the control layers in Module 0.
You own the U-value calculation to ISO 6946 and the psi-values at every junction. Build the wall up layer by layer (R = d/lambda), add the surface films (Rsi, Rse), invert for U, then apply a correction for fixings and bridging - a flat-wall U-value with no bridging allowance is a fiction. Specify the thermal break explicitly: a structural thermal pad or polyamide isolator at brackets, and a back-pan or insulated spandrel at the slab edge. Model the worst junctions in 2D (THERM or equivalent) and report effective, bridged U-values, not clean ones.
Insulation only works if it is continuous and dry. The two field killers are gaps (insulation cut short around a bracket, boards not butted tight) and compression (squashed mineral wool loses R fast). Every bracket that comes through the insulation should land on its thermal pad - if those pads are 'saved for later', that is a thermal bridge being installed permanently. And keep insulation dry: wet insulation can lose most of its R-value, turning your carefully calculated wall back into a conductor.
ISO 6946
U-value of opaque elements
The global method for calculating thermal resistance and U-value of walls and roofs from layer R-values and surface films. It governs the clear-wall number but does NOT itself fully resolve point/linear bridges - those need ISO 10211/14683.
ISO 10211 / ISO 14683
Thermal bridges (psi-values)
ISO 10211 is detailed 2D/3D numerical modelling of junctions; ISO 14683 gives tabulated default psi-values. As of 2026 most Indian projects use catalogue/default psi-values rather than full modelling - conservative, but it can miss a badly detailed bridge.
Eco-Niwas Samhita 2018 / ECBC 2017 (India)
U-value ceilings
ENS caps residential opaque-wall U-value (about 1.0-1.2 W/m2K by assembly) and ECBC sets commercial envelope U-values. They set the target but leave the bridging allowance largely to the designer's judgement - a real-world gap.
“If I specify thick, high-performance insulation, the wall's U-value is sorted - the junctions are a detail.”
The flat-wall U-value is only the performance in the middle of the panel. Thermal bridges at brackets, slab edges and frames carry heat around the insulation and can degrade the wall's effective U-value by 20 to 40%, sometimes erasing the benefit of the upgrade you paid for. Modern envelope assessment counts the psi-values of every junction (ISO 14683) alongside the clear-wall U-value - the bridges are not a detail, they are often the bigger half of the loss.
Worked example - U-value of an insulated cavity wall
Let us calculate the clear-wall U-value of a realistic Indian insulated wall build-up to ISO 6946, then see what one row of steel brackets does to it. All numbers are conventional; the method is the skill.
A calculator (or spreadsheet), the layer conductivities (lambda) from manufacturer data or ISO 10456, and the ISO 6946 surface-film values (Rse 0.04, Rsi 0.13 for a wall).
GIVEN - wall layers (outside -> inside), each R = d / lambda: LAYER d (mm) lambda (W/mK) R = d/lambda (m2K/W) outside surface film - - Rse = 0.04 20 mm cement render 20 0.72 0.028 100 mm concrete block 100 1.10 0.091 50 mm rock wool insul. 50 0.038 1.316 115 mm clay brick 115 0.81 0.142 12 mm gypsum plaster 12 0.40 0.030 inside surface film - - Rsi = 0.13 R_total = sum of all R ; U = 1 / R_total Bridged U (brackets) = U_clearwall x (1 + bridging factor)
- 1Compute each layer R = d(in metres) / lambda. Render: 0.020/0.72 = 0.028. Block: 0.100/1.10 = 0.091. Rock wool: 0.050/0.038 = 1.316. Brick: 0.115/0.81 = 0.142. Plaster: 0.012/0.40 = 0.030. (Already given above - confirm them.)
- 2Add the surface films: Rse = 0.04 (outside), Rsi = 0.13 (inside), standard ISO 6946 values for a vertical wall with horizontal heat flow.
- 3Sum the series: R_total = 0.04 + 0.028 + 0.091 + 1.316 + 0.142 + 0.030 + 0.13 = 1.777 m2K/W.
- 4Invert for the clear-wall U-value: U = 1 / 1.777 = 0.563 W/m2K. Comfortably under the ENS 1.0-1.2 ceiling - on paper.
- 5Now the thermal bridge. Add a row of steel cladding brackets through the rock wool, contributing an estimated chi/psi that works out to a bridging uplift of about +25% on the clear-wall U (a typical penalty for unbroken steel fixings at ~600 mm centres). Bridged U = 0.563 x 1.25 = 0.70 W/m2K.
- 6Read the lesson: the brackets alone pushed U from 0.563 up to 0.70 - a 24% rise, eating most of the margin under the code ceiling. Specify a thermal-pad break and the uplift might drop to +8%, giving U = 0.61 - the difference between a wall that complies comfortably and one that scrapes by or fails on as-built verification.
You’ll walk away with
A clear-wall U-value of 0.56 W/m2K rising to a bridged 0.70 W/m2K - the exact ISO 6946 method, plus the visceral proof that an un-broken bracket can swallow a quarter of your thermal performance.
Two quick ones to make the maths stick.
- 01Recalculate the worked example with the 50 mm rock wool swapped for 100 mm. The insulation R doubles to 2.63; recompute R_total and U and see how much the clear-wall U improves - then ask whether the bracket bridge now matters _more_ as a share of the total loss.
- 02Find a single 6 mm glass pane's U-value (about 5.7 W/m2K) and compare it to the 0.56 wall above. One square metre of bad glass loses roughly ten times the heat of one square metre of this wall - which is why glazing ratio, not wall build-up, usually dominates a glass tower.
A wall's U-value is just the reciprocal of its summed layer resistances - easy maths that the insulation layer dominates. But the clear-wall number is a half-truth: thermal bridges at brackets, slab edges and frames carry heat around the insulation and can degrade effective U by 20-40%. Engineer the build-up AND break every bridge, or you pay for insulation you never get.
U-value (W/m2K) = 1 / sum of layer R-values plus surface films (ISO 6946); low is good. Insulation dominates the total; dense Indian materials need a deliberate insulation layer to meet the ENS/ECBC U ceiling. Thermal bridges - brackets, slab edges, frames - get psi-values (W/mK) and can erase 20-40% of the wall's performance unless broken with a thermal pad.
What is a good U-value for a facade wall in India?
For an opaque wall, the Eco-Niwas Samhita generally caps residential walls at roughly 1.0-1.2 W/m2K, and ECBC sets commercial envelope U-values; well-insulated walls reach 0.3-0.6 W/m2K. Lower is better. A bare 230 mm brick wall is around 2.0 W/m2K - about double the code ceiling - so a deliberate insulation layer is almost always needed to comply.
How do you calculate the U-value of a wall?
Add the thermal resistance R (= thickness / conductivity, in m2K/W) of every layer, plus the inside and outside surface air films (ISO 6946 gives Rsi = 0.13, Rse = 0.04 for a wall). That sum is R_total; the U-value is its reciprocal, U = 1 / R_total, in W/m2K. Then apply a bridging correction for fixings and junctions - the clear-wall number alone overstates real performance.
What is a thermal bridge and why does it matter?
A thermal bridge is a place where the insulation is bypassed by a more conductive path - a steel bracket, an un-insulated slab edge, a metal window frame. Heat funnels through the shortcut, so the wall's effective U-value can be 20-40% worse than the clear-wall calculation suggests. They are quantified by a linear psi-value (W/mK) and fixed with thermal breaks: low-conductivity pads or isolators that interrupt the metal path.
Peer-reviewed journals & authoritative standards
- 01Rawal, R. et al. Development of RETV (Residential Envelope Transmittance Value) Formula for Cooling-Dominated Climates of India for the Eco-Niwas Samhita 2018. — peer-reviewed (BEEP), 2020.
- 02Li, X. & Wu, Y. A review of complex window-glazing systems for building energy saving and daylight comfort. — Journal of Building Physics (SAGE), 2025.
- 03Eco-Niwas Samhita 2018 (ECBC for Residential Buildings), Part I: Building Envelope. — Bureau of Energy Efficiency, Govt. of India, 2018.
_Conduction through the opaque wall is only one way the sun gets in - the bigger, faster route is radiation straight through the glass, which brings us to SHGC, shading and glare._