On image analysis and micromechanics

We give an illustration of how quantitative morphological analysis of microstructures may help to develop a micromechanical approach to investigate the mechanical behaviour of materials. * In the first part, we recall the main morphological parameters that account for a quantitative description of microstructures from a statistical point of view : correlation functions of order n, moments Q(B). These parameters may be measured on materials by means of image analysis, or may be calculated for certain classes of random models of structures, in order to sum up available data on the microstructure. * When studying mechanical properties, it may be useful to develop specific measurements. This will be illustrated by three examples in the case of fracture mechanics : - a "self-consistent" approach allowed to relate coke strength to its porous structure ; - spatial relationship between crack paths and underlying mineralogical structure were looked for with appropriate measurements in a study of iron ore sinter degradation during reduction ; - image analysis simulations of crack propagation in porous graphite were developed to explain the role of intergranular macropores on toughness and strength anisotropy.