Main author

Gregor Harvie Architect Website
Last edited 03 Dec 2016

Thermal behaviour of architectural fabric structures


[edit] Introduction

Fabric structures are perhaps the oldest form of human-made shelter. Remains have been found of simple structures constructed from animal skins draped between sticks dating back over 40,000 years, and it is likely these were the first type of dwellings constructed by humans.

Simple tents, such as the black tent, suited a nomadic lifestyle. Lightweight and easy to carry, they could be moved from place to place in harsh environments where it was necessary to keep on the move to stay alive. Where resources were more plentiful, it was possible to settle down and build permanent shelters in the form of huts. In intermediate environments, a whole range of composite structures developed, part tent, part hut, most notably the Yurt, a demountable hut, still in use in places such as Mongolia today.

This combination of a fundamental requirement for shelter, moderated by practicality and resource availability, still drives the design of our buildings today. Just like the earliest tents, modern fabric structures tend to be used in situations where basic shelter needs to be provided with the minimum material. However, with little to separate the inside from the outside, these spaces do not behave like conventional buildings, they provide little insulation and can be expensive to heat or cool. As a result, spaces under fabric roofs are often characterised as ‘enclosed outdoor spaces’.

This quality of ‘enclosed outdoor space’ can be appropriate for stadiums or atria, indoor sports facilities, shopping malls, airports, stations, and temporary, demountable structures like stages, theatres, pavilions, exhibition centres and even military accommodation. In these sorts of spaces, occupants are not expecting a steady 21 °C. They tend to be used by people wearing outdoor clothing, often moving around and not staying for long. If it gets too cold they may be prepared to put on a coat; if it gets too hot, to take it off.

[edit] Thermal design

[edit] What is thermal design?

Designers tune the thermal characteristics of building envelopes so that they moderate external environmental conditions and maintain internal conditions using the minimum resources of materials and fuel.

This is not straightforward however, as external conditions change throughout the seasons and time of day, meaning for example, that the building envelope may sometimes be required to contain heat, sometimes to capture it, and sometimes to reject it. And the building fabric may have to do this at the same time as performing multiple, conflicting roles, such as; allowing access, providing security, creating a suitable acoustic environment, and so on.

In addition, making a building ‘comfortable’ is not as simple as delivering an average internal air temperature of 21 °C throughout the year. Thermal comfort is dependent on a range of environmental factors in addition to air temperature, such as; air velocity, radiant temperature, relative humidity and the uniformity of conditions. It also depends on personal factors such as; clothing, metabolic heat, state of health, acclimatisation, expectations, and even access to food and drink.

[edit] The thermal behaviour of conventional buildings

Conventional spaces, such as offices or houses, tend to tackle the complex issue of thermal comfort by creating fairly uniform, stable conditions throughout the day and year. These stable conditions are created by a combination of thermal mass and thermal insulation, topped up by building services systems. Thermal mass evens out variations in internal and external conditions, absorbing heat as temperatures rise above the average and releasing it as they fall below. Insulation separates the inside from the outside, so that internal conditions can be maintained with the minimum of energy.

[edit] The thermal behaviour of architectural fabrics

Architectural fabrics offer neither thermal mass nor insulation.

Whatever happens on the outside will transmit to the inside almost instantly, and system inputs such as heating and cooling will quickly be lost through the building envelope. Add to this, a very large surface area which gets hot when the sun comes out, gets cold when it goes behind a cloud, and transmits light and solar radiation; and the situation becomes very complex. Thermal conditions in the interior can change quickly, radiant temperatures can be significantly different to air temperatures and both can change from one part of the space to another.

Thermal comfort in a fabric structure.jpg

[edit] The thermal properties of architectural fabrics

[edit] The thermal characteristics of architectural fabrics

Typically, architectural fabrics are around 1mm thick and as a result have a mass of around 1 kg/m². In comparison, a single-leaf brick wall has a mass of approximately 200 kg/m².

The U-value of a fabric (a measure of its thermal conductivity), will be approximately 5 W/m²K. In comparison, the U-value requirement for a roof in the UK is closer to 0.2 W/m²K.

This means that properties such as thickness, mass and thermal conductivity, the bedrock of conventional thermal analysis, have no significance for architectural fabrics. For all practical purposes, the mass of architectural fabrics is zero and so the thermal resistance of their mass is zero.

As a result, architectural fabrics are particularly responsive to changes in conditions, being affected much faster and much more significantly than most other building materials. Even in the relatively benign UK climate, on a sunny day, an architectural fabric can become more than 20 °C hotter than the external air temperature, and surface temperatures of 45 to 50 °C have been recorded for fabrics with high solar absorptance. Conversely, under a clear night sky, they can become 3.5 °C cooler than the external air temperature (Harvie, 1995).

In addition, there is no significant time lag between a temperature change on their outside surface and the resulting change in their inside surface (Harvie, 1995). In comparison, a 25 mm wooden board has a 25-minute time lag and a brick wall a 5-hour time lag (Evans, 1980).

[edit] The optical properties of architectural fabrics

With no thermal mass and no thermal insulation, architectural fabrics affect the spaces they enclose through the amount of solar radiation they transmit directly into a space and the heat they introduce into the space as a result of the temperature of their internal surface.

The amount of solar radiation which transmits directly through architectural fabrics is relatively easy to calculate, based on their solar translucency.

The heat they introduce into the space as a result of their surface temperature is a little more complicated, based on their:

  • Solar absorption.
  • Convective heat exchange with the adjacent air, inside and out.
  • Long-wave infrared radiation exchange with internal and external surroundings.

Solar absorption is an optical property which can be measured.

With a surface which is effectively entirely smooth, convection is a function of surface temperature, air temperature and air velocity.

Long-wave infrared heat transfer is a function of the surface temperature of the fabric, the emissivity (radiant absorptance) of the fabric and the surface temperature and emissivity of the surfaces surrounding the fabric (including the sky).

To understand the behaviour of architectural fabrics therefore, the properties which need to be measured are:

  • Solar transittance.
  • Solar absorptance.
  • Emissivity (or long-wave infrared absorptance).

These are thermal optical properties. They are all dependent on the angle of incidence. As the angle of incident radiation increases, so the reflectance will increase, the absorption (in percentage terms) of the remaining incident radiation will increase and the transmittance will reduce. To accurately model the behaviour of architectural fabrics therefore, it is necessary to know their angular solar absorptance, transmittance and emissivity.

[edit] Manufacturers’ specifications

Detailed information about the thermal optical properties of architectural fabrics can be difficult to obtain. The standard information commonly available on specification sheets may include a U-value, a single figure for solar reflectance, absorptance and transmittance, and perhaps a shading coefficient which describes the amount of solar heat gain through a material compared to the amount of solar heat gain through a standard sheet of glass.

The properties of architectural fabrics which are commonly available are therefore not always those required for thermal analysis. This means that if accurate modelling is being undertaken, either a specific request has to be made to manufacturers, or measurements taken, or assumptions made based on the known properties of similar materials.

[edit] Modelling the thermal behaviour of architectural fabrics

Once the relevant material properties have been ascertained, the thermal behaviour of architectural fabrics can be predicted under different environmental conditions.

The environmental conditions which will affect the behaviour of a fabric are:

  • Position:
  1. The fabric’s location. That is, its orientation and angle of inclination, and if a solar model is being used rather than actual data, its latitude and longitude.
  2. The geometry of external surrounding surfaces (including exposure to the sky).
  3. The geometry of internal surrounding surfaces.
  • External conditions:
  1. The amount of direct and diffuse solar radiation (or the amount of cloud cover).
  2. The temperatures of surrounding surfaces (including the sky).
  3. Air temperature.
  4. Surface air velocity.
  5. Humidity, if the potential for evaporation is considered significant.
  • Internal conditions:
  1. Surface temperatures.
  2. Air temperature.
  3. Surface air velocity.

Pre-existing models such as geometric models, solar models and standard approximations for conditions such as external surface temperatures, sky temperatures and so on can be used to provide much of this information with a reasonable degree of accuracy.

However, there will not be a pre-existing model for the thermal behaviour of the enclosed space. Internal geometry may be known, but internal conditions will not. Perhaps surprisingly therefore, the complexity in modelling behaviour is determining the interior rather than exterior conditions.

The relationship between the fabric and the interior is a dynamic one; for example, on a sunny day, the hotter the fabric is, the hotter the interior will become, and so the hotter the fabric will become, and so on. In assessing the importance of this dynamic relationship, and so how critical it is that it is accurately modelled, it is necessary to assess the relative significance of all the different heat transfer mechanisms that affect the temperature of the fabric.

Heat transfer mechanisms of fabric structures.jpg

It can be seen that the main heat transfer mechanisms affecting the behaviour of fabrics are solar absorption and long-wave infrared radiation exchange with the outside. The most significant impact on the interior is solar transmission, followed by internal long-wave infrared exchange, and then convection.

This means that whilst the relationship between the fabric and the interior is the most complicated to model, the external conditions are much more significant in terms of the behaviour of the fabric. As a result, when modelling behaviour, it may be possible to make some broad assumptions about internal conditions without significantly affecting the final results.

[edit] The thermal behaviour of spaces enclosed by architectural fabrics

These spaces are the product of their contrasting surroundings. Typically, they will have a lightweight fabric roof above them, but they will sit on a thermally-massive base and will have perimeter walls of a relatively conventional construction. This means that the enclosed space is sandwiched between a roof which can change temperature very rapidly, and can be either hotter or colder than the outside air temperature, and a much more massive base which may barely change temperature through the course of the year.

As a result, internal conditions can vary significantly from place to place depending on proximity to either the lightweight roof or the massive base. Even in the UK, on a hot sunny day, temperatures inside a typical fabric structure can be more than 10 °C higher close to the fabric than they are in the area where occupants are likely to be, close to the ground. This variation is a product of both air temperature and radiant temperature. Not only is the fabric itself hotter, and so radiant temperatures are higher closer to it, but as warm internal air tends to rise, so it will accumulate under the fabric roof, producing higher air temperatures as well. This positive stratification can be compounded further by directly-transmitted solar radiation which is more likely to be shaded by obstructions lower down in the space.

Conversely, under a clear sky, with little solar radiation, for example early on a clear winter morning, the fabric temperature may drop below the external air temperature, producing negative stratification; where temperatures are lower near the fabric than they are in the occupied zone. Negative stratification tends to be less significant than positive stratification, partly because the fabric is only likely to become a few degrees colder than the external air temperature, but also because internal air will still tend to stratify positively, despite the cool temperature of the fabric.

Interior behaviour of fabric structures.jpg

It is very important therefore that when internal conditions are modelled, a large number of locations is assessed and that both air and radiant temperatures are considered. Analysis which treats boundaries or internal spaces as single elements, which may be adequate for simple conventional spaces, is not appropriate in these circumstances. The only accurate method for modelling the thermal behaviour of spaces enclosed by architectural fabrics is with the use of Computation Fluid Dynamics (CFD) software.

[edit] The use of computational fluid dynamics

[edit] An introduction to computational fluid dynamics


CFD works by dividing a space into a large number of 'cells' representing the air within the space. The cells are surrounded by a number of boundaries which represent the surfaces that enclose the space and any openings into it. The temperature of the boundaries, air movement at openings, inputs from building services systems, and so on, are then added. The CFD software will then solve equations representing the flow of air from each cell to those surrounding it, and the exchange of heat between the boundary surfaces and the cells adjacent to them.

Simulations might be run for a number of different scenarios, testing the behaviour of a space under different levels of occupancy, different climatic conditions, in different modes of building services operation, with different openings between spaces and so on. This can build up an overall picture of how a building is likely to behave under normal operating conditions as well as during unusual or extreme conditions.

[edit] The limitations of computational fluid dynamics

Carrying out CFD modelling is time consuming and costly, and so it is important to be clear what is being modelled and why. A decision about whether or not to carry out CFD modelling and to what level of detail will need to consider issues such as:

  • What are the design issues being assessed and why?
  • What are the risks of not undertaking analysis?
  • How serious would the consequences be if the completed building performed poorly?
  • Would it be possible to mitigate problems if they occurred?
  • What alternatives are there to carrying out detailed analysis? For example, is it possible to visit at a similar existing structure?
  • Is there the capability, budget and time to adequately model what is required?
  • What potential is there to make adjustments or select different options if problems are found?

Even if a decision is made to carry out detailed thermal analysis, existing techniques have limitations which may influence the accuracy of results:

  • Detailed information about the angular optical properties of the materials being considered may not be available.
  • At present, boundary models suitable for simulating the behaviour of architectural fabrics have not been dynamically linked to CFD models. This means that boundary conditions have to be modelled first using a specialist model, and then the data imported into the CFD model.
  • Computational power and data storage capabilities still determine the amount of analysis that can be done, generally limiting it to snapshots of behaviour under specific steady-state conditions.
  • Little systematic validation has been carried out to verify the accuracy of CFD models in simulating the behaviour of spaces enclosed by architectural fabric structures.

[edit] The future thermal behaviour of architectural fabric structures

[edit] The future of architectural fabrics

This article has focused on simple, single-skin architectural fabrics, however, more complex configurations have existed for some time, and more are being developed:

These developments create exciting new possibilities which significantly increase the range of building types for which architectural fabrics might be appropriate. They have the potential to increase insulation, create an effective thermal mass and tailor optical properties to optimise efficiency without significantly increasing weight.

[edit] The future of buildings incorporating architectural fabrics

As human population grows and we inhabit more and more of the planet, so we will be forced to colonise areas where there are more extreme conditions. This, coupled with the effects of climate change, means we may be subject to increasingly regular and severe weather events. Fabric structures, which have always proved to be effective at the boundaries of conventional environments, are likely to be popular under these circumstances. Despite their thermal limitations, their ability to provide basic shelter at low cost, and the potential to enclose large areas in hostile conditions will doubtless see an increase in their use.

We can begin to see the implications of this sort of change with developments such as the 2022 football world cup in Qatar where extreme conditions have elevated thermal performance from the bottom of the priority list to the top. When international athletes are exerting themselves in a hot climate, providing an enclosed outdoor space and hoping that they will adapt is no longer adequate. If modelling is not accurate enough to give total confidence about the level of performance that will be delivered, designers find they have to resort to other solutions, ranging from the construction of physical mock-ups, to simply hiring in enough building services plant to cope with all eventualities.

Couple these more extreme conditions with increasing regulations and increasing energy prices, and the future for the thermal design of fabric structures may be a challenging one.

[edit] The future of thermally modelling architectural fabric structures

At present the thermal analysis of architectural fabrics is based on a number of approximations and assumptions. For the most part, these simplifications have not proved significant. The types of spaces which architectural fabrics are used for do not demand high levels of thermal performance and other design considerations tend to have a higher priority. Where performance is critical, a margin for error can be built into the sizing of building services systems so that problems which may occur in the finished building can be mitigated.

But with new materials, new building types, a changing global climate, soaring energy costs, and the availability of software to inexperienced practitioners, the totality of issues previously considered insignificant are beginning to matter.

Future developments which could help improve thermal analysis might include:

  • Better availability of manufacturers’ data.
  • The inclusion of more accurate boundary models within CFD models.
  • Increased computational power enabling dynamic modelling of changing behaviour over long periods of time.
  • Accurate modelling of evaporation and condensation.
  • Validation of models with monitored data.
  • Lower-cost modelling to enable more wide-spread use.

[edit] Acknowledgements

The author would like to acknowledge the assistance of Ian Tavener, Matthew Birchall and Bernardo Vazquez from Buro Happold for providing material and advice in the preparation of this chapter. --Gregor Harvie 14:23, 13 June 2014 (BST)

[edit] Find out more

[edit] Related articles on Designing Buildings Wiki

[edit] External references

  • Architen Landrell (2009) Using PTFE glass cloth, Basic information about the properties of PTFE coated glass cloth and how to use it for your project, [pdf] Architen Landrell Associates Limited, Available from:
  • [Accessed 14 January 2014].
  • Birdair Inc. (2011) Performance Characteristics: Tensotherm ™, [pdf] Amherst, NY, Birdair Inc.
  • Devulder, T., Wilson, R. and Chilton, J.C. (2007) The thermal behaviour of buildings incorporating single skin tensile membrane structures, Oxford University Press, International Journal of Low Carbon Technologies, 2 (2), 195-213.
  • Evans, M. (1980) Housing, Climate and Comfort, London, The Architectural Press, 81.
  • Harvie, G. (1995) An investigation into the thermal behaviour of spaces enclosed by fabric membranes, PhD Thesis, Cardiff, Cardiff University of Wales.
  • Meyer, F. (2009) Low-e coatings: Soft coverings for demanding applications, Eggenstein-Leopoldshafen, Germany, FIS Karlsruhe, BINE-projektinfo 05/09.
  • Pause, B. (2008) Improving thermal-regulating properties for membrane structures, Industrial Fabric Structures Association, Fabric Architecture.
  • Poirazis, H., Kragh, M., and Hogg, C. (2009) Energy modelling of ETFE membranes in building applications, Eleventh International IBPSA Conference, Glasgow, Scotland.
  • Semper, G. (1979) The Four Elements of Architecture and other Writings, Cambridge, Cambridge University Press, 254.
  • Solenberger, F.R. (1979) Thermal Gain of Architectural Fabrics, Internal report to P. Biesert, E.I. Du Pont De Nemours & Company, November 14, (Unpublished).
  • Zhai, Z., Chen, Q., Haves, P. and Klems, J.H. (2002) On approaches to couple energy simulation and computational fluid dynamics programs, Building and Environment, 37, 857-864.