Abstract:
Modeling of high strength and wear resistance aluminum alloy based casting of
composite material developed via conventional foundry method which is one of the
most economical versatile and active research area and so for has not been thoroughly
investigated.
Due to complex nature of the composite materials and their related problems such
as the nonlinear relationship between composition, processing parameters, heat
treatment with the strength and abrasive wear, resistance can more efficiently be
modeled by artificial neural networks. The artificial neural networks modeling requires
sufficient data concerned with chemical composition , processing parameters and the
resulting mechanical properties which were not available for such type of modeling.
Therefore, a wide range of experimental work was conducted for the development
of aluminum composites using conventional foundry method. Alloy containing Cu-Mg-
Zn as matrix and reinforced with 1- 15 % Al 2 O 3 particles were prepared using stir casting
method. The molten alloys composites were cast in metal mold. More than eighty
standard samples were prepared for tensile tests and sixty samples were given solution
treatment at 580 0 C for 1⁄2 hour and tempered at 120 0 C for 24 hours.
Various characterization techniques apparatus such as X-ray Spectrometer,
Scanning Electron Microscope, Optical Metallurgical Microscope, Universal Tensile
Testing Machine, Vickers Hardness and Abrasive Wear Testing Machine were used to
investigate the chemical composition, microstructural features, density, tensile strength,
ductility (elongation), hardness and abrasive wear resistance.
xixThese investigations including the material development and characterization
were used for data generations as needed for modeling of high strength and abrasive wear
résistance aluminum cast composites.
For modeling purpose a multilayer perceptron (MLP) feedforward was developed
and back propagation learning algorithm was used for training, testing and validation of
the model.
The modeling results shows that an architecture of 14 inputs with 9 hidden
neurons and 4 outputs which include the tensile strength, elongation, hardness and
abrasive wear resistance gives reasonably accurate results with an error within the range
of 2-7 % in training, testing and validation. The modeling results shows that an alloy
contents 2-3 % Cu, 2-3 % Mg, 3-5 % Zn reinforced with 10 % Al 2 O 3 can successfully be
developed for highest strength (297 MPa) and highest abrasive wear résistance (0.4 gm
weight loss /15 minutes using stir casting method. The modeling results also suggest that
it is possible to develop the highest strength 466 MPa tensile strength and highest
abrasive wear resistance aluminum alloy based casting composite materials having the
matrix composition of 6 % Si, 2 % Mg with 3 % Zn reinforced with 2-5 % Al 2 O 3
particles.