Steady-state and transient thermal measurements of green roof substrates

Highlights

• First time thermal conductivity of Australian green roof substrates reported.

• New steady-state and standard transient methods are compared.

• Steady-state results are more congruous and accurate than transient results.

• Normalized substrate thermal conductivity is correlated with moisture content.

• Scoria substrate has the lowest thermal conductivity among the substrates tested.

Abstract

There has been growing interest in using extensive green roofs for commercial and residential buildings in urban areas. Green roofs provide many benefits, including adding an additional insulation layer. The potential of this benefit depends on many factors, including the thermal properties of the green roof substrate. Thermal conductivity values of three substrates comprised primarily of scoria, crushed roof tile and bottom-ash were measured with steady-state and transient techniques under three moisture conditions. Specific heat capacities of the green roof substrates were also measured with a transient technique. Steady-state measurements were performed with a “k-Matic” apparatus while transient measurements with KD2 Pro needles. In general, the steady-state measurements showed more consistency than transient measurements. Thermal conductivity differed among the three substrates: crushed roof tile had the highest conductivity values across all moisture contents. Substrate moisture content consistently increased thermal conductivity across all substrates, but this was significantly greater for the crushed roof tile substrate. Steady-state thermal conductivity curves were fitted using the thermal conductivity model for green roof substrates adopted by Sailor (2011). The coefficients obtained are presented and can be used in green roof models to quantify the thermal performance of green roofs and building energy savings.

Introduction

Green roofs are becoming more common in cities and towns due to the multiple ecosystem services they provide [1], [2]. Green roofs can limit some of the environmental and societal problems urban areas are facing. They can reduce storm water runoff [3], increase urban biodiversity [4], [5], mitigate the urban heat island effect [6], improve air quality [7], reduce building energy consumption [8] and may even improve worker attention and productivity [9], [10] helping to sustain their health and wellbeing [11]. However, in new and emerging green roof markets, such as Australia, there is often a lack of quantifiable evidence for these benefits in a local context [12]. While evidence of the benefits of green roofs to urban stormwater retention in Australia is mounting [13], [14], this is not the only potential driver of their uptake. Reduction of building heating and cooling loads, and the resulting building energy savings, can help promote green roof installations [15], as buildings consume about 20% of the total Australian annual energy consumption [16]. However, many studies have shown that building energy savings from green roofs are strongly dependant on climate and location [17], as well as building characteristics and existing insulation [18], [19]. Given that Australia is conventionally divided into eight different climate zones [20], the thermal performance and consequent building energy savings would be different across the country, and specific recommendations on the best green roof build-up should be tested and provided locally [21].

A number of energy balance models have been developed and refined to estimate heat fluxes through green roofs and then estimate building energy savings [22], [23], [24], [25], [26]. Energy savings from green roofs can be predicted from models if we know specific features of the two main layers which form a green roof: the substrate (or growing medium) and the vegetation.

Although many studies have investigated the thermal performance of vegetated green roofs, few have focused specifically on the thermal performance of the substrates. We do this here to isolate the relative effects of substrates and the vegetation (paper in preparation) and gain a better understanding of the heat transfer across the substrate. Physical properties such as composition, water holding capacity, moisture content and porosity are important parameters which dictate heat transfer through green roof substrates. Generally, substrates with high porosity are good insulators [27]. This is because, in general, heat flows more easily through solid particles (by conduction) than through pores, and higher porosity implies a lower effective thermal conductivity (or higher thermal resistivity). The lower the porosity, the tighter the particles are compressed together and the more contact points exist among particles. The contact points then facilitate the conduction heat transfer [28], [29], [30].

Thermal conductivity data is available for a number of green roof substrates, including those whose main aggregate materials are pumice (volcanic porous rock), porous silica, expanded clay, expanded shale (sedimentary rock) and expanded slate (metamorphic rock) [31], [32], [33], [34]. Presumably, these substrate components were chosen due to their high porosity, which is natural or, in the case of the expanded components, derived after heating at very high temperatures, their lightness, availability and cost. Their use varies. For example, expanded shale is a very common and cheap substrate component in western U.S.A., while expanded clay is more commonly used on green roofs in mid-western and eastern U.S.A. [32]. These components have different thermal properties. Shale has a higher thermal conductivity than many other substrates, such as pumice [32] because of its mineral composition, particle size and porosity (Table 1 [28]). In contrast, expanded clay, due to its engineered porosity obtained by heating clay to about 1200 °C, has a lower conductivity than other green roof substrates (Table 1 [28,33]).

In south-eastern Australia scoria, crushed terracotta roof tile and bottom-ash (non-combustible residue of burning black coal also known as Horticultural Ash) have been used as the primary components of green roof substrates [13], [14], [35], [36]. These materials are inexpensive and readily available but their thermal properties have not been previously assessed. However, this information is required to predict the thermal performance of Australian green roofs, estimate potential energy savings and facilitate the inclusion of green roofs in building energy rating systems such as the Green Star Energy scheme [37].

The moisture content of a substrate is one of the key variables that determines the thermal resistance of a green roof. Kotsiris et al. [26] stated that the heat transfer in a substrate is linearly dependent on the substrate’s moisture content. However, studies of soil thermal conductivity [38], [39], [40], [41] show that a linear dependency between soil moisture content and soil thermal conductivity is not realistic. These findings also apply to green roof substrates [31]. Consequently, quantification of substrate thermal conductivity values at different substrate moisture contents is needed.

Studies of green roof thermal properties conducted so far have predominantly used transient techniques to measure the thermal conductivity of green roof substrates [32], [34], [42], [43]. They are often made using either single or dual needles that measure thermal conductivity along the needle itself or between the two needles. Transient measurements, which imply the temperature of the samples varies with time [41], may be inconsistent because they are affected by external factors such as ambient air temperature and relative humidity. Green roof substrates are also normally quite porous with many voids that allow water to drain freely. Transient measurements for wet substrates are therefore strongly influenced by the point where the measurement is taken because it can contain uncertain moisture content. These factors can cause transient measurements with thermal needles to be inaccurate and may also explain widely varying thermal conductivity values reported for wet green roof substrates.

In contrast to transient techniques, measurements in steady-state require that the thermal properties of the substrates are recorded without any external influence and after the sample has reached the steady-state conditions. Using thermal conductivity values measured under steady-state conditions should therefore make green roof thermal models more realistic. However, existing steady-state methods such as ASTM C518 [44] may produce results with high uncertainties due to the nature of green roof substrates which are highly conducting but have varying contact resistance due to their porous nature. This study employed a new steady-state technique specifically developed to better measure the thermal conductivity of porous substrates [45] and therefore improve the validity and accuracy of steady-state measurements.

The main objectives of this research were to quantify the thermal conductivity values of three green roof substrates that are commonly used on green roofs in south-eastern Australia but whose major components are also readily available elsewhere. Thermal conductivities obtained using the steady-state and transient methods were compared and correlations were developed between the thermal conductivities and substrate moisture content. In addition, we discuss the relative merits of the methods used and the specific heat capacities of the green roof substrates tested are reported.