Quantifying uncertainties in the thermo-mechanical properties of particulate reinforced composites

Recently, there has been a growing interest in the use of particulate reinforced composites in a wide variety of applications. Particle reinforced composites are candidate materials for structures where shock or impact loading is expected. The ICAN (Integrated Composites Analyzer) computer code developed at NASA-Lewis Research Center has been modified to include micromechanics equations for particle reinforced composites. In this work, the simplified micromechanics of particulate reinforced composites and a fast probability integration (FPI) technique have been used to obtain the probabilistic composite behavior of a A1/SiCp (silicon carbide particles in aluminum matrix) composite. The probability density functions for composite moduli, thermal expansion coefficient and thermal conductivities along with their sensitivity factors are computed. The effect of different assumed distributions in constituent properties and the effect of reducing scatter in input random variable on a select composite property - namely the thermal expansion coefficient are also evaluated. The variations in the constituent properties that directly effect these composite properties are accounted for by assumed probabilistic distributions. The results show that the present technique provides valuable information about the scatter in composite properties and sensitivity factors, which are useful to test or design engineers.