06-10-2012, 04:10 PM
Modeling Mechanical Properties of Aluminum Composite Produced Using Stir Casting Method
Modeling Mechanical.pdf (Size: 872.55 KB / Downloads: 70)
ABSTRACT
ANN (Artificial Neural Networks) modeling methodology was adopted for predicting
mechanical properties of aluminum cast composite materials. For this purpose
aluminum alloy were developed using conventional foundry method.
The composite materials have complex nature which posses the nonlinear relationship
among heat treatment, processing parameters, and composition and affects their
mechanical properties. These nonlinear relation ships with properties can more
efficiently be modeled by ANNs. Neural networks modeling needs sufficient data base
consisting of mechanical properties, chemical composition and processing parameters.
Such data base is not available for modeling.
Therefore, a large range of experimental work was carried out for the development of
aluminum composite materials. Alloys containing Cu, Mg and Zn as matrix were
reinforced with 1- 15% Al2O3 particles using stir casting method. Alloys composites
were cast in a metal mold. More than eighty standard samples were prepared for
tensile tests. Sixty samples were given solution treatments at 580oC for half an hour
and tempered at 120oC for 24 hours.
The samples were characterized to investigate mechanical properties using Scanning
Electron Microscope, X-Ray Spectrometer, Optical Metallurgical Microscope, Vickers
Hardness, Universal Testing Machine and Abrasive Wear Testing Machine.
A MLP (Multilayer Perceptron) feedforward was developed and used for modeling
purpose. Training, testing and validation of the model were carried out using back
propagation learning algorithm.
INTRODUCTION
The applications of neural networks modeling was adopted
by Badeshah [1], Shah, I., [2], and Genel, et. al. [3] who
had well documented neural networks applications in
materials science concerned with the microstructural
evaluation of steels, processing and properties of steels
as well as conducted study on ductile cast iron. The
materials science based work using neural networks which
had brought attention of researchers recently conducted
by Sha and Edwards [4].
Microstructural features such as amount of austenite
retained in ductile austempered cast iron was estimated
by Yascas, et. al. [5] using artificial neural networks. The
features classification of alloy steel microstructures
consist ferrite and pearlite was investigated using back
propagation algorithm which can effectively be used for
the features classification by Martin and William [6]. The
microstructure image analysis of complex systems have
been determined by using neural networks technique as
reported by Maly, Harck and Novotny [7].
Strategy for Alloy Development
For this purpose pure aluminum, magnesium, zinc, copper
metals as ingots and Al2O3 as powder were purchased. Six
master alloys were prepared through conventional foundry
method. The detailed compositions in weight percentage
in grams are given in Table 1.
For manufacturing of alloys pure aluminum was melted in
a pit furnace using graphite crucible at Mehran University
of Engineering & Technology, Workshop as shown in
Fig. 1.
Chemical Analysis
Samples size 15mm dia and 15mm length were prepared
for chemical analysis of specimen using Spectrometer
available at Dawood College of Engineering &
Technology, Karachi and Scanning Electron
Microscope which is available at Mining, Engineering
Department Mehran University of Engineering and
Technology Jamshoro (Fig. 9). Aluminum oxide particles
size was also analyzed using Horiba particle size analyzer
(Fig. 10).
Measurement of Density and Porosity
Densities of all specimens were measured through weight
in air divided by volume method. The porosity of samples
was calculated through the theoretical density of
specimens which were calculated from the average
composition from the elemental density of metals as given
in Tables 3-4.
Metallography
Optical Metallurgical Microscope at Mehran University
of Engineering & Technology, Jamshoro, and Pakistan
Steel Mills Laboratories were used for metallographic
studies. Scanning Electron Microscope (Fig. 9) was also
used for few samples. The samples were metallographically
prepared after grinding at 120-1200 mesh emery papers
and samples were polished using alfa aluminum oxide
powder. Samples were etched in Keller's solution.
CONCLUSIONS
(i) Mechanical properties of aluminum cast
composite materials can successfully be modeled
using artificial neural network which is an initial
attempt to correlate the non linear behaviour
between chemical composition and properties of
cast aluminum composite developed via
conventional foundry method carried out at
Mehran University of Engineering and
Technology, Jamshoro.
(ii) A more comprehensive model was successfully
developed with 80 dataset considering nonlinear
relationship between the composition, processing
parameters, and solution treatment with tensile
strength, elongation, hardness and abrasive wear
resistance using multilayer perceptron network.
(iii) A well trained model with 9 hidden neurons giving
smaller training error 2-7% and has better
performance as compared to lesser or higher
number of neurons The suggested network
model contain 14 inputs, 9 hidden neurons with
4 outputs. The model can be trained within 30
second having good generalization ability.
Present proposed model predicts accurately the
output of the unseen test data.