The research and analysis of mixing process parameters with application of fuzzy logic method for the improvement of properties of PM composites
Marek Wojtaszek, Jarosław Durak, Filip Pernal
Quarterly No. 4, 2009 pages 327-331
DOI:
keywords: composites reinforced with particles, aluminium alloy, silicon carbide, mixing process, powder metallurgy, fuzzy logic
abstract The fuzzy logic method was used as a tool for quick choice of optimal and economic parameters of the mixing process in aluminium alloy particles-ceramic particles system. Experiments were the source of the input data for computer analysis. Al17Si5Fe3Cu1.1Mg0.6Zr alloy powder, with particles of diameter below 40 μm, was used as a matrix. Particles of silicon carbide (10%) were used as a reinforcing phase. Mixing of components was carried out using several rotation speeds and mixing times. The samples of the composite were prepared from obtained mixtures by hot-compaction in a closed die at 500°C in isothermal conditions. Microstructural examinations of obtained composites were carried out and selected properties of samples were evaluated. The correct distribution of particles of the reinforcing phase in the matrix was reached at rotational speed 0.9 s−1 and mixing time 45 min. For the composites prepared by hot compaction with these parameters, favourable properties and their smallest dispersion were found and the particles of reinforcing phase were uniformly distributed. Prolongation of mixing time, at the rotation speed mentioned above, didn't lead to amelioration of the product's properties. For the rotation speed equal to 0.2 s−1, an attempt of optimal time determination wasn't successful. For this purpose, a method based on knowledge engineering was proposed. The obtained relations between the parameters of mixing process and properties of the final products were used as input data for the analysis. Information about mixing process characteristics, in knowledge engineering called expert’s knowledge, were the complementary data. The source of this information was the experience based on previous experiments and the data found in the literature. A set of data was prepared, which, from the point of view of knowledge engineering, was incomplete and uncertain, because a part of those data wasn't verified by experiments. Those data could be used for quick estimation of the mixing process parameters for future experiments with compound characteristics similar to those used in this experiment. A fuzzy reasoning was carried out using Fuzzy Toolbox program, which is a part of Matlab software. The mixing time for the rotation speed 0.2 s−1 was estimated to be about 80 minutes.