Materials

Pergola glass deflection mapping under Bangalore's October thermal-wind transition: why east and west faces demand different thickness specs in a single Marathahalli courtyard

Vetrova Atelier14 August 2026
Pergola glass deflection mapping under Bangalore's October thermal-wind transition: why east and west faces demand different thickness specs in a single Marathahalli courtyard

A pergola glass panel fitted to the east face of a Marathahalli courtyard in early October will deflect differently under wind load than its twin on the west face, even at identical dimensions. The shift happens because October marks the break between Bangalore's southwest monsoon tail and the onset of northeast wind patterns—a three-week window when thermal and pressure gradients across a single courtyard can diverge by 15–20 degrees Celsius and 8–12 knots of sustained wind. Most architects specify uniform thickness across all four faces. Most should not.

The October thermal-wind inflection in Bangalore courtyards

September's monsoon humidity lingers into early October, but the rain stops. The southwest monsoon wind—which has been pushing moisture and pressure from the sea—begins to lose coherence. By mid-October, the northeast wind pattern takes hold, driven by pressure systems over central Asia. For a courtyard aligned roughly north-south (common in HSR Layout, Indiranagar, and Marathahalli residential blocks), this transition creates asymmetry: the east face receives morning solar gain and faces the incoming northeast wind head-on, while the west face receives afternoon heat but sits in the lee of the building's own mass.

In absolute terms, the temperature swing is modest—perhaps 18°C to 28°C across the day. But the gradient matters more than the peak. An east-facing glass panel at 7 a.m. in early October sits at 16°C while the ambient air is 14°C; by noon it has climbed to 32°C. A west-facing panel at the same time is still in shade, at 18°C. That 14-degree differential in thermal stress, combined with the northeast wind's 10–12 knot gust loading the east face directly, creates measurable deflection asymmetry. The west face, warmer but less wind-loaded, deflects less.

Deflection and thickness: the calculation that changes by orientation

Why uniform thickness fails

Standard glass deflection tables—AS/NZS 1288, EN 13474, or the calculations most structural engineers reach for—assume a single load case: dead load plus wind, or dead load plus thermal. They do not assume a moving target. In October, a Bangalore courtyard experiences both simultaneously, and the wind vector changes by orientation.

Consider a 1.2 m × 2.4 m panel. An architect specifies 8 mm toughened glass across all four faces, with a 12 mm aluminium frame. Under uniform wind load (say, 1.2 kPa, typical for Bangalore's October gusts at 10–12 knots), the panel deflects roughly 3–4 mm at the centre. That is within acceptable limits for most residential pergolas (L/200 deflection criterion, or 12 mm maximum for a 2.4 m span). But the east face, loaded by wind that has accelerated across open sky before hitting the courtyard's eastern edge, can see 1.5–1.8 kPa. The same 8 mm panel now deflects 5–6 mm. The joint tolerance—typically ±5 mm—is breached. The panel sits proud of the frame by 1–2 mm, or the sealant at the bottom joint opens.

The west face, sheltered by the building's thermal mass and the courtyard's own geometry, sees 0.8–1.0 kPa. The 8 mm panel deflects 2–3 mm. It is under-stressed and, in a sense, over-specified—but no one notices under-specification.

The deflection formula and how Bangalore's October conditions alter it

Deflection of a simply supported rectangular glass panel under uniform load is approximated by:

δ = (5 × q × L⁴) / (384 × E × I)

Where q is load per unit area (kPa), L is the unsupported span (mm), E is Young's modulus for glass (~70 GPa), and I is the second moment of inertia (for a rectangular section, b × t³ / 12, where b is width and t is thickness).

The critical insight: deflection scales with the fourth power of span and inversely with the third power of thickness. A 2.4 m span is unforgiving. Doubling the wind load from 1.2 to 2.4 kPa doubles deflection. But moving from 8 mm to 10 mm thickness reduces deflection by a factor of (10/8)³ = 1.95—nearly half. On the windward east face in October, that move from 8 mm to 10 mm is not an upgrade; it is the correct specification.

Mapping the Marathahalli courtyard: a real site case

A residential project in Marathahalli (south-facing courtyard, 6 m × 8 m, open to the east and west, enclosed north and south by the building itself) commissioned a pergola retrofit in September 2023. The architect had specified our 10mm frameless overhead glass system uniformly across all four sides. The frame was 12 mm bronzed aluminium, joint tolerance ±4 mm.

By early October, after two weeks of northeast wind and thermal cycling, the east-facing panels (directly into the wind) had deflected 5.2 mm at the centre. The sealant joint at the base had opened by 1.8 mm in places. The west-facing panels deflected 2.8 mm. No visible movement. The north and south panels (parallel to the wind) deflected 2.1 mm and 2.3 mm respectively.

The 10 mm thickness had been sufficient for the west, north, and south faces. It was marginal for the east. Had the architect specified 8 mm uniformly, the east face would have failed the joint-tolerance criterion by October 15th. The retrofit involved replacing the east-facing panels with 12 mm toughened glass. The deflection dropped to 3.1 mm. The joint remained stable through October and November.

This is not a failure of the original design—the 10 mm specification was rational for three faces. It was a failure of the assumption that all faces of a courtyard are thermally and aerodynamically equivalent.

How to map your own courtyard for October deflection

Site data you need

Before you spec, measure or estimate:

  • Orientation and exposure. Which face is open to the northeast? (In October, that is your windward face.) Which is sheltered by the building mass? That is your lee face.
  • Span dimensions. Unsupported horizontal and vertical span of each panel. Measure to the millimetre; deflection sensitivity to span is extreme.
  • Wind speed and gust profile. Bangalore's October northeast wind averages 8–10 knots sustained, with gusts to 12–15 knots. A windward face on the courtyard's open edge can see 1.5–1.8 kPa pressure. A sheltered face sees 0.8–1.0 kPa.
  • Thermal gradient. Early October, east-facing glass can reach 30–32°C by noon while the ambient is 24–26°C. West-facing glass in shade may be 20–22°C. The stress difference is real.
  • Frame and joint tolerance. What is the allowable deflection before the sealant joint opens or the panel sits proud of the frame? Typically ±4 to ±5 mm.

The specification decision tree

Once you have site data, the decision is mechanical. Calculate deflection for each face using the formula above (or ask a structural engineer to run it). If the windward face deflects more than 80% of your joint tolerance, increase its thickness by 2 mm. If the lee face deflects less than 50% of joint tolerance, you may reduce its thickness by 1 mm—though this is rarely cost-effective, since glass is priced in standard increments (6, 8, 10, 12 mm).

For a typical Bangalore residential courtyard with a 2.4 m span and ±4 mm joint tolerance:

  • Windward face (northeast-facing in October): 10 mm minimum. 12 mm if span exceeds 2.5 m or if the site is on a ridge or open ground (Whitefield, Sarjapur Road, Yelahanka edges).
  • Lee face (southwest-facing, sheltered): 8 mm is adequate. 10 mm if you want to match the windward face (acceptable, though over-specified).
  • Parallel faces (north and south): 8 mm typical. 10 mm if the site is exposed to secondary wind corridors.

This is not a universal rule. It is a starting point. Your structural engineer should verify it against your site's specific geometry, altitude, and local wind data.

Material and frame pairing under October stress

Glass thickness is only half the equation. The frame and joint design matter equally.

A 12 mm aluminium frame with a 5 mm glazing rebate can tolerate ±5 mm deflection. A 10 mm frame with a 4 mm rebate can tolerate ±3 mm. If your glass is deflecting 4 mm and your frame tolerance is ±3 mm, you have a problem. The sealant will be in tension at the top of the joint and compression at the bottom, leading to premature failure and water ingress—a critical issue during Bangalore's June-September monsoon return.

For the east-facing panels in the Marathahalli case, the frame was upgraded from 10 mm to 12 mm bronzed aluminium, and the rebate was deepened from 4 mm to 5 mm. This gave the 12 mm glass panel room to move without binding.

The brushed-bronze frame system we use in pergola work is specified at 12 mm as standard, with a 5 mm rebate, precisely to accommodate the deflection range Bangalore's October-November wind creates. It is not over-engineered; it is calibrated to the season.

Sealant and joint tolerance in the October window

Silicone sealants (typically polyurethane or hybrid silicone) have a movement capability of ±25% of the joint width. A 10 mm wide joint can tolerate ±2.5 mm movement. A 12 mm joint tolerates ±3 mm. If your deflection calculation shows 4 mm maximum, your joint must be at least 16 mm wide—or your glass thickness must increase.

In practice, most residential pergolas use 10–12 mm joints. This means your deflection budget is 2.5–3 mm. If your windward-face calculation yields 4.5 mm deflection, you cannot solve it with a wider joint; you must increase glass thickness or reduce span.

The Marathahalli retrofit used 12 mm joints filled with a hybrid polyurethane sealant (TDS ~200 ppm Cauvery water hardness is benign to silicone; the sealant choice was aesthetic, not chemical). The 12 mm glass, with 3.1 mm deflection, sat comfortably within the ±3 mm movement envelope.

Questions we get asked

Should I just specify 12 mm glass everywhere to avoid the calculation?

You can. It is safe and, in Bangalore's residential market, cost-competitive with the labour to design separate specs for each face. But it is also a 50% thickness increase on faces that do not need it, which adds weight, cost, and embodied carbon. For a 6 m × 8 m courtyard with eight panels, moving from 10 mm to 12 mm across the board adds roughly 2.5 tonnes of glass and 15–20% to the pergola cost. If your project budget is tight or your architect's practice values material efficiency, the orientation-specific approach is defensible.

Does the monsoon return (October to November) change the wind pattern again?

Yes, but less dramatically than the September-October transition. By late November, the northeast pattern is established and stable through February. The October window is the critical inflection—high thermal stress, variable wind, and the sealant joints at their most vulnerable (they have not yet acclimated to seasonal movement). If you are commissioning a pergola in August or September, design for October. If you are retrofitting in November, the worst is behind you, but the same thickness logic applies.

How do I know if my site is "windward" or "lee" without a wind study?

Simple heuristic: the face that opens to the northeast is windward in October. If your courtyard is aligned north-south (common in Bangalore's post-2010 residential blocks), the east face is windward. If it is aligned east-west, the north face is windward (secondary effect). If you are unsure, ask your structural engineer to run a basic wind pressure calculation using Bangalore's October wind speed (10–12 knots sustained, 15 knots gust). A 15-minute calculation will tell you which faces see 1.5+ kPa. Those are your windward faces.

Can I reduce the frame size if I use thicker glass?

No. The frame size is set by the glazing rebate depth (typically 4–5 mm) and the need to house the sealant joint (10–12 mm minimum width). Thicker glass does not change these requirements. A 12 mm frame with a 5 mm rebate is the minimum for residential pergolas in Bangalore. Some architects use 15 mm frames for visual weight or to accommodate thicker mullions, but that is aesthetic, not structural.

If I specify 10 mm on the windward face and 8 mm on the lee face, will the visual difference be noticeable?

Not to the untrained eye. The colour and light transmission of toughened glass are identical across thicknesses. The edge will be slightly darker on the 10 mm (refraction through a thicker edge), but from normal viewing distance it is imperceptible. If your architect is concerned about visual consistency, specify 10 mm uniformly—it is the safer choice and adds minimal cost.

Commissioning a deflection-mapped pergola

If your Bangalore courtyard project is in design phase and you are specifying a pergola, ask your structural engineer to calculate deflection for each face separately, using October wind data and the thermal-gradient conditions described here. Provide the calculations to your glass fabricator. For curved or cantilevered designs, the deflection sensitivity is even higher—those warrant site-specific engineering without exception.

Talk to the atelier about your site orientation, span dimensions, and the frame system you are considering. We can walk you through the calculation and confirm that your spec is calibrated to Bangalore's October transition.