Performance Testing: What the Numbers Actually Prove

Published 8 October 2026 — Hanart Facade Engineering Team

Why testing matters more than the brochure

The curtain wall is the building's outer skin. It determines how much heat escapes in winter, how much solar gain enters in summer, how much water penetrates during a typhoon and whether the glazing survives a design-level wind event. Yet many procurement decisions are still made on the strength of product datasheets that were written for a generic 1.5 m by 1.5 m test specimen in a temperate laboratory.

The reality is that every project condition is different. The facade grid on your tower is not the grid in the type-test report. The wind pressure at your site is not the wind pressure assumed in the catalog rating. The thermal exposure of your climate zone is not the thermal exposure of the test chamber. A standardized product does not automatically deliver standardized performance — and that is why performance testing exists.

Testing proves three things that no brochure can: (1) the system performs under the specific loads your building will experience; (2) the materials and seals retain their properties after thermal cycling, UV exposure and mechanical fatigue; and (3) the quality of assembly — in the factory or on site — meets the design intent. Without test data, you are buying a promise. With test data, you are buying proof.

The three core tests every facade must pass

Every credible curtain wall system is validated against three fundamental performance criteria: air infiltration, water penetration and structural capacity. These are codified in ASTM E283, E331 and E330 in North America, and in GB/T 21086 in China. Hanart tests to both standards, depending on project jurisdiction.

Air infiltration — ASTM E283 / GB/T 21086

Air infiltration testing measures how much air passes through the curtain wall under a specified pressure differential. The test specimen is sealed into a chamber, and air is drawn across the exterior while the interior is depressurised (or pressurised) to a standard pressure difference — typically 300 Pa (1.2 in H2O) for field quality assurance, or 75 Pa for baseline comparison.

The result is expressed in different units depending on the standard. GB/T 21086 classifies air infiltration into four grades: Grade 1 (less than 0.5 m3/m/h at 10 Pa), Grade 2 (less than 1.5), Grade 3 (less than 2.5) and Grade 4 (less than 4.0). ASTM E283 reports in CFM per square foot at 1.57 psf (75 Pa), with typical high-performance targets below 0.06 CFM/sq ft.

Why does this matter? Because air infiltration directly drives heating and cooling load. Industry studies consistently show that uncontrolled air leakage through the building envelope is responsible for 30–40% of total HVAC energy consumption in commercial buildings. A facade that moves from Grade 4 to Grade 1 air tightness can reduce the heating load by 25% or more — not through better glass, but through better seals. At Hanart, every unitized module is pressure-tested in the factory before shipment, and every stick-built installation is site-tested with a calibrated smoke generator to locate leaks before handover.

Water penetration — ASTM E331 / GB/T 21086

Water penetration testing simulates wind-driven rain by spraying water on the exterior of the specimen while applying a calibrated static air pressure on the interior. The test pressure is increased in steps — typically 250 Pa, 500 Pa, 700 Pa and 1,000 Pa — until either water appears on the interior surface or the maximum test pressure is reached.

GB/T 21086 classifies water penetration into five grades, from Grade 1 (no water penetration at 700 Pa) to Grade 5 (water penetration below 250 Pa). ASTM E331 does not assign grades; it records the pressure at which leakage first occurs, and the test is typically deemed successful if no uncontrolled water appears on the interior at the specified test pressure.

A critical distinction exists between static and dynamic water penetration testing. ASTM E331 is a static test — the pressure is held constant. ASTM E547, the dynamic test, applies cyclic pressure pulses to simulate gusts. Real-world rain is dynamic, and a system that passes static testing at 1,000 Pa may leak at 700 Pa under dynamic cycling. Hanart's test protocol includes both static and dynamic phases, because the difference between the two often reveals design weaknesses that static testing conceals.

The most common failure modes in water penetration testing are not exotic: insufficient gasket compression (the seal is too soft or too thin for the design deflection); drainage path blockage (silicone overspill or debris in the weep channels); and cavity interconnection (drainage channels that link adjacent chambers, allowing water to travel horizontally rather than draining out). Each of these is detectable in testing, invisible in a catalog photograph and catastrophic in a building.

Structural performance — ASTM E330 / GB/T 21086

Structural testing subjects the curtain wall to positive and negative wind pressure, applied incrementally to a full-scale mock-up. The test measures deflection, permanent deformation and ultimate failure load. ASTM E330 specifies a test pressure of 1.5 times the design wind load (the "test load factor"), with a maximum deflection criterion — for mullions, typically L/175 of the span, though many specifications tighten this to L/250 or L/360 for high-rise towers where occupant perception of movement matters.

GB/T 21086 uses a similar approach, with performance grades from 1 to 9 based on the pressure differential the system can withstand without failure. The standard also includes an inter-storey drift angle test, in which the specimen is racked to simulate seismic or differential settlement movement. A typical seismic design target is an inter-storey drift angle of 1/100 — the specimen must retain its structural integrity and weather-tightness after being displaced by 1% of its height.

What the structural test proves is not merely that the framing is strong enough. It proves that the glass-to-frame interface retains sealant adhesion under full design deflection; that the pressure plate screws do not strip at peak load; that the thermal break does not fracture under combined thermal and mechanical stress; and that the weathertightness system (gaskets, drainage channels, baffles) continues to function after the structure has moved. These are integration tests, not component tests — and integration failures are what cause facades to leak in year five.

Beyond the big three: the tests that separate good facades from great ones

The three core tests are the minimum. Projects that demand 25-year performance — which is most projects Hanart works on — require a broader validation regime.

Thermal performance. U-value measures heat transfer through the entire assembly, not just the glass. A low-U-value glass in a high-U-value frame delivers mediocre overall performance. Hanart tests complete assemblies in a guarded hot box per ISO 8990, and measures the temperature factor (f_Rsi) to ensure that interior frame surfaces remain above the dew point under design conditions. Condensation on the frame interior is a thermal bridge; it is also a mold and corrosion risk.

Acoustic performance. The weighted sound reduction index (Rw) and the American STC rating quantify how much exterior noise is attenuated by the facade. A difference of 5 dB in Rw is perceptible to occupants; a difference of 10 dB is the difference between a quiet office and a space where conversation is difficult. Acoustic testing requires not just the right glass (laminated interlayers add mass and damping) but the right seals — leaky gaskets destroy acoustic performance faster than they destroy thermal performance.

Seismic performance. Beyond the GB/T 21086 inter-storey drift test, some projects require full-scale seismic simulation with cyclic racking to 2% or 3% drift. The facade must not shed glass, must not buckle its framing and must remain weathertight after the seismic event. These tests are expensive — but so is post-earthquake remediation of a failed facade.

Fire resistance. Curtain walls are required to maintain compartmentation. Fire tests per EN 1364 or GB/T 9978 measure integrity (E) and insulation (I) for 30, 60, 90 or 120 minutes. The failure mode is typically not the glass itself but the sealant and gasket system, which degrades at temperatures above 200 degrees Celsius and allows flame and smoke to bypass the fire barrier.

Impact resistance. Soft-body impact (a sandbag) tests resilience against accidental collision or wind-borne debris. Hard-body impact (a steel ball) tests resistance against deliberate damage or vandalism. These tests matter for ground-floor glazing, atria and public-facing facades.

Durability and accelerated ageing. Salt spray testing (ASTM B117) validates corrosion resistance of anodised and coated aluminium in marine environments. UV exposure chambers validate gasket and sealant longevity. Thermal cycling between -30 and +70 degrees Celsius, repeated 1,000 times, reveals fatigue failures in seals and drainage paths that static testing cannot predict.

What a test report actually tells you

A test report is a contract between the laboratory and the specifier. To read it correctly, look for four things:

1. Test conditions versus project conditions. What was the specimen size? What was the glass thickness? What was the frame section? If the test used a 1.2 m by 1.8 m specimen with 6 mm monolithic glass and your project uses 2.4 m by 4.2 m units with laminated insulating glass, the test results are guidance, not proof.
2. Sample representativeness. Was the test specimen built by the same factory, with the same materials, by the same workers, that will build the project? A hand-built laboratory sample with perfect craftsmanship tells you nothing about production-line quality control.
3. Boundary conditions. Was the specimen supported in a way that matches the real structure? Rigid laboratory supports overstate stiffness; overly flexible supports understate it. The boundary condition matters for deflection and for the stresses transferred to the anchors.
4. Accreditation. Was the test laboratory accredited to ISO/IEC 17025? Does the report carry the laboratory's accredited scope? An unaccredited report is an opinion, not evidence.

Why per-project testing beats catalog claims

Type-test reports — the certificates that manufacturers produce to show that their system "meets the standard" — have a fundamental limitation. They test the standard configuration of the system under standard conditions. Your project is not standard.

The grid dimensions on your tower may be 50% larger than the type-test specimen. The glass may be thicker and heavier, changing the deflection characteristics of the frame. The mullion section may be deeper to span the larger grid, but the type-test report used the shallow section. The wind pressure at your site, calculated from local meteorological data, may be 20% higher than the design pressure assumed in the type test. Each of these differences erodes the relevance of the catalog claim.

Per-project testing builds a mock-up using the actual project materials, dimensions and details, and tests it under the actual project design loads. The result is a project-specific validation that eliminates the extrapolation risk. It also provides a quality benchmark: the production units must match the mock-up, and the mock-up is the reference against which site samples are compared.

At Hanart, we do not ask clients to trust our type-test reports. We offer project-specific testing as a standard part of our quality assurance program. The cost is modest compared to the value of knowing — with documentary proof — that the facade will perform as specified.

Next step

If you are specifying a curtain wall for a project where performance matters — and performance always matters — ask for the test data, not the marketing story. Hanart's engineering team can review your project conditions, recommend the appropriate test regime and manage the full testing process from mock-up construction through laboratory supervision to report interpretation.

Request Testing Assessment   Weather-Tightness Systems →