Feasibility of Integrated Insulation in Rammed Earth

Building Codes in Europe stipulate strict thermal performance criteria which any traditional rammed earth recipe cannot meet. This does not infer that the material itself is inferior; it has many other face saving attributes such as low embodied energy, high workability, sound insulation, fire resistance, aesthetics, high diffusivity and thermal accumulation properties. Integrated insulation is experimented with, to try achieve a 0.22 [W/(m2.K)] overall coefficient of heat transfer for walls required by 2015 Slovak standards, without using external insulation or using technologically complex interstitial insulation. This has the added aesthetic benefit of leaving the earth wall exposed to the external environment. Results evaluate the feasibility of this traditional approach.


Introduction
With the advent of thermal criteria for building codes, thermal comfort as a criterion is superseded by efficiency measured in kWh/m² p.a. which is a product of a buildings shape, size, airtightness and thermal resistance of the envelope and its parts. Contemporary stabilized rammed earth (SRE) is an evolutionary product of traditional rammed earth (RE) methods and materials, often incorporating steel reinforcing and rigid insulation, enhancing the structural and energy performance of the walls while satisfying building code parameters. Interstitial thermal insulation is known to be effective for many types of building envelopes made of sandwich or composite structures. The implementation of interstitial insulation in stabilized rammed earth (SRE) has received significant attention in Research and development as stringent thermal requirements for building codes continue to increase. Thermal conductivity tests, conducted by M.A. Hall, on composite SRE walls with extruded polystyrene (XPS) insulation demonstrated that the combination of a high mass wall with the low conductivity of foam 3 To whom any correspondence should be addressed. insulation resulted in a wall that had a lower thermal conductivity than a solid earth wall or an earthen wall with insulation located at only the internal or external face, while improving the mass performance of the wall as a whole [1,2]. However, due to the complexity and technological limitations of implementing (SRE) interstitial sandwiches in situ, the authors were inspired to analyse if it is holistically feasible to integrate bulk insulation in rammed earth.

Sample production
All tested soil components were obtained locally and oven-dried to a constant mass at a temperature of 60 °C as opposed to 110 °C to prevent potential changes in the material structure. The Portland cement (type III) content was 10 % by weight. The clay and silt (30 % content) was pulverized into a coarse powder, passing through an 8 mm sieve, and then mixed with sand and gravel using a variable speed mixer to ensure uniformity between batches. Eleven kilograms of soil were used to make one test specimen that measured 32.5 x 32.5 x 5.5 cm. A controlled amount of water was added to raise the soil to optimum moisture content (OMC) which is the moisture content at which a material reaches its maximum dry density (MDD) for a given compactive effort. This is said to influence the strength and durability of the material. Table 1 shows the MDD and OMC of the reference sample. Columns one through five represent the five test points on the line curve and represent increased moisture content for the same recipe [3]. The material for all test samples was compacted using a pneumatic tamper with a 55 mm head. Component soils were blended in ratios and then graded to obtain their particle-size distribution parameters plotted logarithmically with respect to percentage (by dry mass) of the total specimen on a linear scale, see

Selection of thermal insulation
Insulation had to fulfill several criteria. These included that the material be ecological; natural based; inherently hydrophobic, but not a vapour barrier; be relatively incompressible and withstand the impact of ramming. Foam glass, clay pellets and cork were amongst those considered, but hydrophobic expanded perlite seemed the most promising. Batches were defined by the percentage of bulk volume of thermal insulation that was put in the form. Reference batches had no insulation, while each subsequent batch had an increase of 20 % bulk insulation while the maximum feasible case had a 100 % volume of bulk insulation relative to the volume of the form. Figure 2 illustrates a freshly  Figure 3 overleaf shows the preparation of the test specimens, equipment used and formwork for the test specimen.

Determination of the coefficient of thermal conductivity
Thermal conductivity was measured using a calibrated Stirolab heat flow meter with compliant Lm 305 software. Accuracy was set to 99 % with only a 1 % permissible deviation. Each test lasted 3600 seconds and every specimen was tested twice. The first test stabilized the high mass material while the second test produced the final measurement results.

Test Procedure
The Stirolab lambda meter works by placing a test sample between a warm and cold plate. The standard setting is 0 ºC for the bottom plate and 20ºC for the top plate producing a 20 K temperature difference. The temperatures of the plates are kept constant while the amount of energy required to maintain constant temperatures depends on the conductivity of the material. The conductivity of the material is therefore directly proportional to the amount of energy required to maintain a constant temperature between the plates. Thus thermal conductivity is calculated as: Where j -is the heat flow (W/m 2 ) d -is the thickness of the test specimen (m) ∆T -is the temperature difference (K) Figure 4 shows the simultaneous thermal stabilization curve of the reference specimen for the warm and cold plates.   Figure 5 below depicts the heat flow needed by the warm and cold plates to maintain a constant temperature. Table 2 was drawn to define the dry mass, density, quantity of insulation expressed as a percentage of total volume and statistical thermal conductivity of each test specimen. These are statistical values averaged over multiple samples. The volume of each sample was approximately 0.0058 m 3 . Conditions of the feasibility study were that the walls should have a coefficient of thermal conductivity that is low enough to meet thermal transmittance targets without sacrificing load bearing ability. Therefore, multiple Schmidt hammer tests were performed as an indicator of strength for each corresponding test sample. The Schmidt hammer was used because the shape of the specimens did not conform to compression tests via crushing. The tests were performed in the direction of the ramming and showed that each 20 % increase in perlite dramatically reduced the strength. The reference sample had a characteristic strength of almost 4.94 MPa, the 20 % had strength of 2.45 MPa and all subsequent samples had strengths lower than the minimum required for identification. Ignoring this fact and focusing only thermal conductivity and Thermal Admittance related values [7], the results were still disappointing. Figure 6 shows the heat flows due to unit swing in the internal environmental temperature that would occur over a 24 h period. Figure 6. Heat flow of the 100% bulk volume specimen for a 1 K temperature variation over a 24 hour period.