Application of fuzzy logic to determine suttable parameters of mixing process for composites obtained from powder reinforced by ceramic particles by use of pm method
Marek Wojtaszek, Jarosław Durak Akademia Górniczo-Hutnicza, Wydział Inżynierii Metali i Informatyki Przemysłowej, al. Mickiewicza 30, 30-059 Kraków
Quarterly No. 4, 2008 pages 390-396
DOI:
keywords: powder metallurgy, fuzzy logic, numerical analysis, mixing, composites reinforced by particles
abstract An attempt to assessment the possibilities of application of a fuzzy logic aided quick determination of mixing process parameters in a metal powder particles-ceramic particles system was made. A set of composites was obtained by use of the powder metallurgy method, then a set of chosen parameters was determined. Two different powders were used as a matrix material, the first one was the aluminium particles with expanded surface, the second one was subieve-size particles fraction of the Al17Si5Fe3Cu1.1Mg 0.6Zr alloy with a spherical shape. Particles of silicon carbide were used as a reinforcement phase. The different volume fraction depending on the matrix material was used. Such approach made possibility to analyse two systems with different mechanisms determining uniformity of the reinforcement phase distribution. Samples were obtained using mixing, densification at the room temperature or hot densification, concurrent extrusion with fixed process parameters for each pair of the components. The influence of the volume fraction of the reinforcement phase on the chosen parameters and structure of the product was determined. The influence of the powder composition on the relative porosity of composites obtained of the extruded at the temperature of 500C and with extrusion ratio = 4.1 compacted semi-finished products of aluminium based powders and mixtures of this powder with silicon carbide SiC1000 particles with several volume fractions was estimated. The influence of the composition of powders on the abrasive wear was evaluated. The results of the experiment with addition of the information about the process (expert's knowledge) were the base for a numerical analysis of the mixing process. In a result of such data acquisition a set of information was obtained. From the engineering point of view this is a incomplete knowledge because it still contains information non verified by experiments. This knowledge could be useful in combination with the artificial intelligence methods like the fuzzy logic. For the numerical analysis the Fuzzy Toolbox from Matlab package was used.