You are here

Maintenance on site: Impact of fire lighting on thatch temperature

Wednesday, 29 February, 2012
Steve Burrow
Ian Daniel
Number of Participants / Visitors / Audience: 
This survey was designed to test the potential of a thermal imaging camera as a tool for investigating and managing the environments of reconstructed buildings.

The Moel y Gerddi roundhouse at St Fagans was constructed by Peter Reynolds in 1992. It was built in a woodland clearing and, thanks in part to the shielding effect of the surrounding trees, its thatch has suffered from recurrent problems. In 2000 the decaying outer layer of thatch was scrapped away and a new covering layer was put on; by 2010 this new thatch was also in a state of decay. At this time it was decided not to refurbish the roof as plans were already being developed for a new village to be built on a different site. This decision provided an opportunity to study the environment of a roundhouse at the end of its life cycle, and at a time when its thatch was essentially permanently saturated with rain water. This research objective is in line with that initially established for the site by Peter Reynolds who saw the monitoring of decline at Moel y Gerddi as a means of understanding the original excavation plan. But in the context of the current project, the main aim of the monitoring was to identify ways of improving site management, and to provide a base line for studying the environment of the replacement roundhouses.

A full suite of environmental monitoring equipment was installed in and around Moel y Gerddi, and this report focuses on the efficacy of using a thermal imaging camera in this context.

The survey consisted of repeatedly photographing the exterior of the roundhouse both before and after a fire was lit in the central open hearth. Photographs were taken at two minute intervals, with a selection of these photographs being shown below to indicate the general trend in the results.

Weather conditions
On the day of the survey the sky was overcast until 11:37AM when the sun broke through the clouds. Ambient temperature was around 9˚C and relative humidity was 100% at the start of the survey, rising to 12˚C and declining to 93% at the end. There was little or no wind throughout, and there had been no rain since the 27th February. All photographs were taken with the sun behind the camera.

Equipment and analysis
The thermal imaging camera used was a Testo 875-1, with emmissivity set to 0.95 and TRefl reset after every photograph using a tin foil marker. The Testo 875-1 has a rated accuracy of +/-2˚C, +/-2% of m.v, meaning that it is suitable for recording trends rather than precise values.

After the survey, data was analysed in Testo's IRSoft with all images being allocated the same max and min values (10 - 25˚C) and colour range. The temperature range of the thatched roof, excluding other parts of the image, were also analysed as bar charts showing the % of pixels exhibiting each temperature. Prior to the survey the principle recorder, Steve Burrow, had attended a training course at the Property Care Association. This course aimed to provide practical instruction in the basics of thermal imaging survey, but not detailed expertise in the interpretation of thermal imaging data. For this reason the interpretation of results presented here is likely to be the subject of review as more experience is gained in the use of the thermal imaging camera.

Inside the house were three TinyTag temperature and relative humidity data loggers, with additional TinyTags outside the village.

Recording began at 9:11AM. The previous day's fire had been extinguished 16 hours before with the hearth recording a core temperature of 24˚C at the start of the survey. Ambient temperature in the roof space of the house was 9˚C and relative humidity 100%.

At 9:47 the exterior surface of the thatch began to warm a few degrees above that of the external temperature, presumably a consequence of the diffused light falling on this face of the roundhouse.

The fire was lit at 10:05AM, at which time St Fagans opened to the public, and the effect on the external surface temperature of the thatch was rapid and pronounced. Temperatures at the top of the thatch cone rose to 25.5˚C within a few minutes. Internally, air temperature in the lower roof space stayed around 9˚C.

Over the course of the next 30 minutes, maximum surface temperature stayed at around 25˚C, but the area subject to this warming increased, spreading down the thatch. This is presumably a consequence of the interior of the roof space filling with warm air but, unfortunately, no temperature recording devices had been placed high enough in the roof cone to record this spread.

Surprisingly these raised temperatures were not maintained after this time, with both maximum and average temperatures declining thereafter, despite continued fuelling of the fire. Indeed by 11:19AM, just one hour after the fire was lit, the external surface temperature of the thatch had returned to the same temperature that was exhibited prior to the fire lighting. Internally the house was still heating at this time, with the highest TinyTag (c. 2m off the floor) only just beginning to record a temperature rise. It was a further hour before temperature in the lower levels of the house rose to around 20˚C – the same temperature as was recorded on the upper parts of the outer thatch surface. Relative humidity inside the house remained at 100% throughout the survey.

At first glance, these results suggest that the roundhouse thatch became more effective as an insulator during the course of this trial, trapping heat inside. But there is no reason to believe this is the case.

More likely, the saturated thatch exhibited evaporative cooling - effectively sweating. In the initial heating of the thatch the warm air escaped, heating the trapped water as it passed. Once the surface water had heated sufficiently it evaporated, leaving behind the cooler water and thatch surface.

This early temperature spike is an unexpected result and one which, if it has been interpreted correctly, should not be apparent in a roof with a drier thatch. As a working hypothesis, it would be expected that a drier thatch would reach a standard temperature and maintain this unless the weather conditions changed or the fire was damped down.

While the interpretation presented above is only tentative, the survey was sufficient to demonstrate the value of a thermal imaging camera in monitoring building environments. The possibility that the camera could be used to identify saturated thatch in other cases is an intriguing one.

However, this conclusion comes with caveats. The thermal imaging camera is a technical piece of equipment and requires a skilled and experienced operator. In addition, it is most useful when there is a significant temperature difference between the inside and outside of a house and when weather conditions are stable throughout the survey. The decision to hire or buy one, should therefore be weighed against the ability to use it, and the cost of the purchase.

All bar charts show temperature range on the roof surface alone.

Fig 1. Thermal image taken 9:11AM.
Fig 2. Thermal image taken 9:47AM.
Fig 3. Thermal image taken 10:13AM.
Fig 4. Thermal image taken 10:21AM.
Fig 5. Thermal image taken 10:30AM.
Fig 6. Thermal image taken 10:49AM.
Fig 7. Thermal image taken 10:59AM.
Fig 8. Thermal image taken 11:09AM.
Fig 9. Thermal image taken 11:19AM.
Fig 10. Thermal image taken 11:29AM.
Fig 11. Thermal image taken 11:39AM.


Thermal image taken 9:11AM.
Thermal image taken 9:47AM.
Thermal image taken 10:13AM.
Thermal image taken 10:21AM.
Thermal image taken 10:30AM.
Thermal image taken 10:49AM.
Thermal image taken 10:59AM.
Thermal image taken 11:09AM.
Thermal image taken 11:19AM.
Thermal image taken 11:29AM.
Thermal image taken 11:39AM.