Measurement of the Thermal Parameters of Composite Materials During Fire Tests with Quantitative Infrared Imaging

A fire resistance test, as performed on plates of carbon fibre reinforced polymer composites, is conceptually analogous to a step heating, and can be modelled by the one-dimensional heat equation. Thus modelled, the plate operates as a linear system with an impulse response function that relates temperatures at the front face (the one exposed to fire) and the rear face, and depends on two parameters, namely thermal diffusivity (\(\alpha \)) and effective Biot number (Bi). Taking as input the time evolution of temperature at the front face, the output of the model for each (\(\alpha \), Bi) couple is a predicted temperature for the rear side of the plate. At each point of the plate, \(\alpha \) and Bi can be retrieved by fitting the predicted temperature to the experimental one. These parameters are good quantitative indicators of thermal damage, and therefore its measurement has a particular interest to understand degradation processes associated with fire action. To perform their retrieval as described, temperature measurement for both faces of the plate during the experiment are made with a specifically developed infrared imaging system, composed by two synchronized infrared cameras that image both sides of the sample during the fire test and provide surface temperature maps, spatially co-registered and with the flame effects filtered out. Applying the fitting procedure described to these temperature maps makes possible, for the first time, to measure \(\alpha \) and Bi in situ during the fire test. The value of \(\alpha \) obtained by this procedure (varying from \(\approx 0.5 \times 10^{-7} \, \mathrm{m}^{2} \, \mathrm{s}^{-1}\) in the region most affected by fire, to \(\approx 7 \times 10^{-7} \, \mathrm{m}^{2} \, \mathrm{s}^{-1}\) near sample edges) has been compared to those measured after the test, on samples at room temperature, with the classical flash method. A good general agreement has been found, with differences that can be attributed to the temperature dependence of diffusivity. From this comparison a critical temperature of \(T = 450^{\circ }\)C has been identified, that separates two different regimes, probably related to different degradation states, with slopes of \(-\,1.5 \times 10^{-9} \, \mathrm{m}^{2} \, \mathrm{s}^{-1} \, \mathrm{K}^{-1}\) for \(T<450^{\circ }\)C and \(3 \times 10^{-10} \, \mathrm{m}^{2} \, \mathrm{s}^{-1} \, \mathrm{K}^{-1}\) for \(T>450^{\circ }\)C.