Determination Of Thermal Properties Of Mineral Wool Insulation Materials For Use In Full-Scale Fire Modelling

Abstract

Temperature-dependent materials properties are required for use in many contexts in fire safety engineering. While property values for many materials do exist, we often are limited in our understanding of how representative a given set of materials properties is for the application of interest. Thus, more work is needed to critically evaluate the measurement methods used, data obtained, and interpretation of the values in terms of their use in subsequent engineering applications. This research evaluates methods for determining thermal conductivity, density, mass loss, emissivity, porosity and specific heat capacity as functions of temperature of mineral wool insulation materials. These thermophysical properties will be applied to detailed heat and mass transfer modeling of the response of wall assemblies to realistic fire exposures. Use of the properties in more detailed models will, in turn, provide additional insight into the potential behavior of structural components during a fire for the purpose of occupant egress planning and firefighter safety.

The ability to model thermophysical degradation of materials is extremely important when assessing response of assemblies to the wide range of temperatures characteristic of a real fire exposure. Development of consistent methods for analysis and interpretation of the thermophysical properties of each element of the assembly will also guide testing of new, or previously untested, construction materials. Further, in order to develop a model of the response of the full assembly, it is necessary to determine appropriate parameters and properties required as input to submodels developed for the behavior of each material and those must accurately reflect the thermal and mass transfer processes taking place in that material. The objective of this study is to establish a set of methods for accurately characterizing the thermophysical response of construction materials as a function of temperature across a range of temperatures that would be encountered during exposure to real fire events.

Currently, one of the most common practices is to model heat transfer in a material using “effective” thermophysical properties for that material. This can involve estimating one or more properties at a set value of temperature (often room temperature or an average value between room and fire temperature) or combining two or more properties into a single “effective” value as needed for input to the model. Such treatments are approximations intended to simplify the modeling process. It is well known that material properties change as a function of temperature, however, determination of properties such as density, specific heat, and thermal conductivity as functions of temperature is time consuming and is oftentimes inconsistent, depending on the material of interest and application. Some problems include the difficulty in preparation of specimens that are representative of the actual application of the material, as well as the variation of property data depending on the methods and/or heating regimes used in their determination. The wide range of materials commonly used in construction, each with distinct temperature response characteristics, also presents a challenge, making it difficult to develop universal methods for characterization that can be easily applied to every material.

In this research, physical and chemical properties of mineral wool insulation are first obtained using common practices listed in the literature, such as thermo-gravimetric analysis (TGA), and differential scanning calorimetry (DSC), as well as experiments designed by the writer. These reference values are then compared to values obtained using alternative or modified methods, as well as different test parameters, such as heating regimes or specimen preparation, designed to explicitly measure phenomena that are not specified through current property values. For example, from TGA experiments mass loss as a function of temperature is measured, which provides an estimate for density of a material as a function of temperature. Through the mass loss rate curve, temperatures at which thermally induced reactions are taking place can be identified, which advises on how to model that particular material in the temperature range of interest. From DSC test data, the specific heat capacity is calculated. The specific heat and density will then be used in combination with other results to estimate the thermal conductivity. Finally, properties measured using the various methods will be used as input into one-dimensional or more complex models of small-scale tests and full-scale wall fire experiments to predict the response of the assembly. New methods will be further interpreted in terms of their differences to current methods and those methods that provide the best representation of observations seen in the tests will be recommended as the methods to be used in future characterization of construction materials in this context.