No abstract

No abstract

No abstract

This work discusses issues concerning the implementation of scanning of unknown engineering objects containing just simple (i.e. no freeform) surfaces with touch probes on three-axis computer numerical control (CNC) measuring machines in order to reconstruct their shape in a computer aided design (CAD) system. Several ideas are put forward, e.g. scanning along vertical slicing planes, adaptive point sampling distances, in-process`proactive' segmentation of points into curve sections and probe radius compensation in two directions, as well as limited remedy of edge scanning ambiguities. Most of the suggested algorithms are implemented as parametric numerical control (NC) programs on an OKUMA machining centre.

Drying of ceramic parts with complex geometry has been studied with emphasis on the effects of shrinkage on their final dimensions. The first step was to understand the processes governing moisture loss from a porous medium through experimental measurements yielding the coefficient of moisture expansion. As a result, non-homogeneous dimensional changes (shrinkage) occur in three-dimensional artefacts with varying cross-section. The moisture diffusion problem governed by Fick's laws was solved numerically by analogy to the heat conduction problem. To this end the correspondence was established between physical parameters, variables and boundary conditions of heat flow and diffusion, characteristically the coefficient of moisture expansion being analogous to the coefficient of thermal expansion. The numerical model established and solved using finite element analysis predicted moisture distribution inside the part as well as the resulting change in its shape. Validation of numerical predictions was first ensured in two dimensions by modelling a simple slab and comparing with experimental measurements. Validation for a fully three-dimensional shape required use of the wellestablished iterative closest point algorithm for surface point matching and subsequently the creation of an error map. The reverse of this numerical model was used to predict the appropriate die geometry starting with the shape of the desired part by taking into account the variable drying shrinkage allowance, and the relevant steps are outlined.