PSI - Issue 10

A. Hein et al. / Procedia Structural Integrity 10 (2018) 219–226 A. Hein and V. Kilikoglou / Structural Integrity Procedia 00 (2018) 000 – 000

222

4

the Roman period onwards (Freestone (1989)). Only then it was possible to manufacture thin-walled crucibles, which could be externally heated, while before other types of crucibles were used, primarily heated internally or from above and thus requiring a heat insulating ceramic matrix. Heat resistance was achieved with the construction of comparably thick walls, so that extreme heat was restricted to the internal layers, which were in contact with the heat source. The ceramic matrix was furthermore stabilized by non-plastic inclusions, which were not affected by high tem peratures. The reduction of heat transfer in crucibles similarly as in furnaces was achieved by the fabrication of a highly porous ceramic matrix. Common practice was the addition of organic materials to the clay paste, with an ob served preference for fibrous materials, such as plant fibers or animal hair, and a typical orientation of the resulting elongated pores perpendicular to the assumed heat flux. It has to be considered, though, that any increase of porosity has decreased mechanical properties (Wachtman et al. (2009)). In order to assess the performance of different types of ceramics the verification of simulation results by material testing is essential. The mechanical performance under load, in terms of strength and toughness, can be measured with flexure tests, either in one axis, commonly by 3-point bending, or bi-axially, for example with a point load on a ring supported disk (Morrell (1998)). Apart from the maximum load at fracture the recorded load-displacement curves provide also information about the elasticity of the materials. For the mechanical tests, specimens with typical dimen sions of a few centimeters can be cut out of sufficiently large fragments of archaeological ceramics in order to assess the material performance of specific ceramic wares. Furthermore, on the basis of experimental specimens, fabricated and fired in the laboratory, parameters such as the clay type, the micromorphology, which depends basically on the firing conditions, or the addition of different amounts of particular temper materials, can be investigated (M ü ller et al. (2015)). Thermal properties of different materials can be assessed with laboratory tests as well. Thermal conductivity of disk shaped samples for example can be determined with a steady state hot plate setup (Hein et al. (2008b)). The sample disk is heated on top of a hot plate with a stabilized arbitrarily chosen temperature and the temperature development in a brass disk on top of the sample disk is recorded. In this way experimental materials as well as archaeological fragments can be examined for their heat transfer properties (Hein et al. (2013)). The simulation of the performance of modeled microstructures under defined loads offers the opportunity to investigate specific material parameters in a systematic and reproducible way. Scope is to estimate effective material properties of microstructures, exhibiting specific pore structures or inclusions. For a complete assessment multi-scale modeling will be necessary. Starting point of every simulation is the generation of a digital model, representing the essential characteristics of the ceramics’ microstruct ure. Commonly, models are based on microscopic observations which provide, however, two-dimensional images of the microstructure. These images can be used to generate straightforwardly two-dimensional models, extracted by image processing, which yield basic information for example about the effect of pore shape and orientation (Hein and Kilikoglou (2007)). For more realistic simulations three-dimensional digital models are necessary. One way would be the generation of models based on examinations of the ce ramics’ microstructure by micro-scale X- ray computer tomography (μCT) ( Machado et al. (2017)). Alter natively, three-dimensional models can be generated by placing randomly distributed pores or inclusions of specific shapes in a ceramic body (Roberts and Garboczi (2000)). Shape, orientation and total amount of pores and inclusions can be adapted to the observation under the microscope. For the assessment of the material performance the models are investigated using the finite element method (FEM). For this, typical material properties, based on former tests, are defined for different materials comprised in the model. The model is then meshed, dividing the three-dimensional geometry in small elements, which are connected among each other by nodes. At last, suitable constraints are defined, such as pressure on specific surfaces or spatial fixation of others. In the case of simulations of heat transfer particular temperatures or modes of heat exchange are defined on surfaces. The model then is evaluated by solving the entire system of equations, which describes the interaction among the elements. Based on the results concerning simulated deformation, strain or stress 1.2. Material testing 1.3. Simulation of material performance