Artificial neural network (ANN) prediction of kinetic parameters of (CRFC) composites

Carbon materials exhibit outstanding thermo-mechanical properties and low density (<2 g/cm3), which make them useful for structures submitted to high temperatures (T > 1000 °C). Solid rocket motor throats, thermal barrier protection systems, turbine rotors and high temperature process components are typical applications for these materials. The manufacture of carbon materials, such as carbon fibers and carbon reinforced fiber composites (CRFC), undergo a controlled pyrolysis process, where organic precursors such as resins, pitches, and polymeric fibres are converted over 800 °C into a solid carbon material. In the present work the activation energy from a carbon reinforced carbon composites were obtained by Thermogravimetry Analysis. The experiment involved heating samples of the mentioned materials at fixed constant heating rate schedule up to 1000 °C. The kinetic parameters were obtained by analyzing the weight loss curves. The non-linear fitting method based on Levenberg–Marquardt approach was used to fit the weights loss curves at five heating rates.