Abstract
Although extensive experimental studies have underpinned random variable characterisation of composite material properties, random field models are so far based primarily on engineering judgement. In this paper, modelling strategies are presented for characterising and specifying the spatial randomness in terms of random fields, utilizing the results of extensive experimental studies at macro scale. To assess the spatial variability in composite material systems, experiments were conducted on two different types of Glass Fibre Reinforced Polymer (GFRP) composite panels. The first set of experiments was carried out on aged and weathered hand lay-up panels, made with Chopped Strand Mat (CSM) glass reinforcement in a polymer matrix while the second set comprised structural grade pultruded GFRP plates supplied in pristine condition. Typical auto and cross correlation patterns for a range of mechanical properties are discussed. It is concluded that manufacturing and in-service conditions play a major role in determining random field parameters, with correlation lengths in appropriately selected functions ranging from 20 to 100 mm © 2011 Taylor & Francis Group, London.