Decoding the U‑Value of ETFE Cushions

Abstract

Ethylene tetrafluoroethylene (ETFE) cushions are increasingly used in lightweight roofs and façades because the material is light, transparent and durable. However, ETFE’s thermal performance depends on how many foils are layered together and the air cavities created between them. The U‑value—the rate at which heat passes through a construction—directly influences heating and cooling loads. Standard tables often quote generic U‑values (e.g., a 2‑layer ETFE cushion is about 2.9 W/(m²·K) and a 3‑layer cushion about 1.8 W/(m²·K), but these numbers are typically derived from simplified models of rectangular cavities and may not reflect the curved shapes used in built projects.

This blog post explains why the true thermal transmittance of ETFE cushions can differ from textbook values by drawing on a 2019 research paper that experimentally measured the thermal resistance of double‑ and triple‑layer pneumatic cushions. The study used a bespoke test rig to replicate winter conditions and measured heat flux and temperature across the entire surface of two cushions. The findings show that triple‑layer cushions can nearly double the thermal resistance compared with double‑layer cushions, and that standard formulas tend to underestimate the performance of curved cushions. These insights help designers select the right number of ETFE layers and understand the limitations of simplified calculation methods.

Why U‑Value Matters in ETFE Cushions

The thermal transmittance (U‑value) and solar gain factor (g‑value) are the key parameters that determine how an ETFE envelope performs. A lower U‑value means less heat escapes through the membrane, reducing winter energy demand. In cushions, the U‑value depends on the direction of heat flow and the number of foil layers because each extra layer creates an additional air gap. For example, an ETFE design guide notes approximate steady‑state U‑values of 5.6 W/(m²·K) for a single layer, 2.9 W/(m²·K) for two layers, 1.8 W/(m²·K) for three layers and 1.4 W/(m²·K) for four layers. These improvements arise because each cavity adds a layer of trapped air that resists heat flow.

Yet real cushions are not flat panels. Architectural ETFE pillows are double‑ogive shaped and taper towards the edges, which alters the natural‑convection patterns inside the cavities. Standard calculations, such as those in EN 6946 and EN 673, assume rectangular, parallel air gaps and do not fully account for curved geometries or buoyancy‑driven eddies. Computational fluid dynamics (CFD) studies have shown vortices and stagnation zones inside curved cushions, suggesting that real U‑values could deviate from textbook figures.

Overview of the Experimental Study

The reference study, presented at the TensiNet Symposium 2019, aimed to measure the thermal resistance of two small pneumatic cushions (double‑layer and triple‑layer) under controlled winter conditions. Each cushion measured 1.1 m × 1.1 m and was fabricated from PVC‑coated polyester foils (similar pneumatic behavior to ETFE). Both samples were kept inflated at an extra‑pressure of 300 Pa, mimicking real cushion operation.

Test apparatus and measurement method

·       Dual Air Vented Thermal Box (DAVTB): Two insulated chambers separated by the sample allowed independent control of indoor (Box 1) and outdoor (Box 2) temperatures. A hydronic system provided heating and cooling between 15 °C and 50 °C.

·       Boundary conditions: To replicate Milan winter design conditions, Box 1 was set to 20 °C and Box 2 to 45 °C, yielding a 25 °C temperature difference across the cushion.

·       Instrumentation: 32 T‑type thermocouples and 16 heat‑flux meters were installed on the surfaces. Because the cushions are curved, the surfaces were divided into 49 zones; measurements were taken in four phases so that every zone could be sampled.

·       Data processing: Temperature and heat‑flux data were averaged over time and weighted by zone area to obtain mean surface temperatures and heat fluxes. The overall thermal resistance was calculated from the difference between inner and outer surface temperatures divided by the measured heat flux. The experimental U‑value was determined from operative temperatures and average heat flux. For comparison, standard U‑values were also calculated using conventional surface heat‑transfer coefficients (25 W/(m²·K) outside and 7.7 W/(m²·K) inside).

Results: Double vs. Triple Layers

The two cushions were tested under identical boundary conditions, allowing a direct comparison. Table 1 below summarizes the measured thermal resistances and U‑values.

Table 1 – Experimental results for double‑ and triple‑layer cushions

Cushion	Layers & cavities	Heat‑flux density (W/m²)	Thermal resistance  (m²·K/W)	Experimental U‑value  (W/(m²·K))	Standard U‑value  (W/(m²·K))
Sample 01	Double layer, one cavity	61	0.183 ± 0.009	2.43	2.83
Sample 02	Triple layer, two cavities	42	0.351 ± 0.015	1.67	1.92

The triple‑layer cushion transmitted 31 % less heat than the double‑layer cushion. Because the average cavity thickness decreased when adding a third foil, convective motion inside the cushion was suppressed, nearly doubling the thermal resistance (+91 %). Consequently, the experimental U‑value dropped from 2.43 W/(m²·K) for the double‑layer sample to 1.67 W/(m²·K) for the triple‑layer sample—roughly a 30 % reduction, confirming that more layers offer meaningful energy savings. The standard U‑values calculated with conventional surface coefficients were slightly higher because the controlled environment had lower convective–radiative heat‑transfer coefficients than assumed in codes.

Discussion: Why Standard Models Underestimate Performance

The researchers compared their measurements with calculated thermal resistances based on natural‑convection correlations for vertical rectangular cavities (EN 673 and Ostrach correlations). When the cushion was approximated as a flat cavity with an average thickness of 19 cm (double layer) or 11 cm (each cavity in the triple layer), the calculated thermal resistance was 13–18 % lower than the measured value for the double‑layer sample and 6–9 % lower for the triple‑layer sample. No single equivalent thickness could reproduce the experimental results. The discrepancy likely stems from the curved, tapering cross‑section of pneumatic cushions, which creates stagnation regions and inhibits convective flows. CFD studies on curved ETFE cushions have observed large eddies and secondary flows that are absent in rectangular cavities. These flow structures increase the thermal resistance beyond what is predicted by simplified formulas.

Implications for designers

·       Layer count matters: Each additional foil significantly lowers the U‑value. A 3‑layer ETFE cushion tested under realistic conditions achieved a U‑value around 1.7 W/(m²·K), comparable to the 1.8 W/(m²·K)

·       Geometry cannot be ignored: Curved cushions behave differently from flat panels. Designers should be cautious when applying standard calculations; they may underestimate insulation performance by 10–20 %.

Conclusions

The experimental study demonstrates that triple‑layer pneumatic cushions offer markedly better thermal performance than double‑layer cushions, confirming the value of adding layers in ETFE designs. Under a 25 °C temperature difference, the triple‑layer sample exhibited almost double the thermal resistance of the double‑layer sample. This translated into a U‑value reduction of roughly 30 %. Standard U‑value formulas based on rectangular cavities underestimated the measured performance by up to 18 %, highlighting the need for more sophisticated models that capture curved geometries and internal convection.

For architects and engineers, these findings emphasize the importance of specifying the correct number of ETFE layers and recognizing the limitations of simplified U‑value calculations. A triple‑layer ETFE cushion provides a good balance between weight, transparency and insulation. Future work, including CFD simulations and testing cushions with low‑emissivity coatings, will further refine our understanding of heat transfer in these innovative building skins.

Credit: This blog post synthesizes findings from the paper “Thermal performance of pneumatic cushions: an experimental evaluation” by Andrea Alongi, Adriana Angelotti, Alessandro Rizzo and Alessandra Zanelli. The study was presented at the TensiNet Symposium 2019 and published in the conference proceedings.