An Investigation into the thermal behaviour of spaces enclosed by fabric membranes

 

 

A thesis submitted to the University of Wales for the degree of Philosophiae Doctor by GREGOR HARVIE. Welsh School of Architecture, University of Wales College of Cardiff, March 1996

 

 

The findings in brief:

  • Boundary models and CFD models need to be dynamically linked to properly represent the impact of thin boundaries whose temperatures might change rapidly.

  • Analysis of comfort must include radiant temperatures as well as air temperatures, particularly in spaces where there is a significant difference in temperature between the air in the space and the surfaces enclosing it, or where solar radiation penetrates the space.

  • Where there is likely to be a significant difference in temperature between a boundary and the air adjacent to it, a very small cell size (40mm or less) is required to properly simulate the increased air velocity at that surface (and so the increased heat transfer). An example of this increased air velocity is the downdraft generated by a cold window. Failure to properly simulate this increased air velocity will result in an underestimation of the contribution the boundary makes to internal conditions.

  • In non-cartesian spaces (ie where surfaces are not all vertical or horizontal, but may be inclined or curved) it is necessary to use a body-fitted cell grid (i.e. one in which the grid is distorted to follows the contours of the surface) in order to allow a small enough cell size adjacent to boundaries to properly simulate the flow of air across those boundaries. If a cartesian grid is used, a refined grid is required throughout the space just to simulate the flow of air across the non-cartesian boundaries and this is computationally impractical.

 

To download the thesis click on the links below

 

This thesis describes a programme of research the aim of which was to investigate the thermal behaviour of spaces enclosed by fabric membrane envelopes.

 

Initial analysis of the overall situation suggested that a fabric membrane can affect conditions within a space enclosed by it as a result of its internal surface temperature and the amount of thermal radiation it directs into that space. In order to investigate these two parameters, a test cell was constructed which allowed the thermal behaviour of a range of fabric membranes to be monitored.

 

The monitored data revealed that the thermal behaviour of fabric membranes is only significantly affected by their angular thermal optical properties. These properties were then measured and a dynamic spread sheet model was developed which was able to simulate the monitored behaviour fairly accurately.

 

In order to investigate the thermal behaviour of spaces enclosed by such membranes, conditions within four existing fabric roofed buildings were monitored. The monitored data revealed that comfort temperatures could vary significantly from place to place within such spaces. These variations were produced by both the stratification of internal air temperatures and differences in internal radiant temperatures.

 

An attempt was made to simulate the behaviour of the buildings monitored, using a general applications CFD code in conjunction with information generated by the spread sheet model. Whilst simple behaviour patterns could be simulated accurately using this approach, it was apparent that over simplistic boundary specification options left the CFD code unable to accurately predict strong internal stratification.

 

It was proposed that improving the reliability of this process would require the development of a more holistic CFD model which should be able to accurately predict the thermal behaviour of fabric membranes itself.

 

download

Contents

Chapter 1

Chapter 2

Chapter 3

Chapter 4

Chapter 5

Chapter 6

Chapter 7

Chapter 8

Chapter 9

Chapter 10

Chapter 11

Appendices

Surface model

 

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

PDF

XLW

 

(163 KB)

(54 KB)

(1.6 MB)

(120 KB)

(2.5 MB)

(419 KB)

(264 KB)

(640 KB)

(6.5 MB)

(1.4 MB)

(188 KB)

(34 KB)

(269 KB)

(221 KB)

mail@gregorharvie.com