1. Introduction
1xxx series wrought aluminum alloys are used in a wide range of applications and product forms, such as foil and strips for packaging and heat-exchanger tubing, cable sheathing and fin stock, where excellent formability, corrosion resistance and electrical and thermal conductivity are required [
1,
2,
3]. Typically, the levels of iron and silicon in a specific alloy are controlled to provide the required performance characteristics, such as strength, formability or corrosion resistance, which resulting in many commercial variants within a given AA specification. The effects of iron and silicon levels on hot workability in 1xxx alloys were reported in our previous work [
4]. Increasing both iron and silicon contents generally increases the high-temperature flow stress, which can negatively impact the hot workability [
4,
5,
6,
7]. For example, the extrusion speed and corresponding productivity of thin-wall tubing can be reduced. Thus, there is a trade-off between room temperature mechanical properties and hot workability. An alternate approach to increase room temperature strength is to add copper for solution strengthening, as is the case in the widely utilized AA1100 alloys. In the current work, the impact of copper content on hot workability is assessed with the long-term view of optimizing 1xxx alloy design for both strength and hot processability.
The flow behavior of aluminum alloys during hot deformation is complex. The work hardening and dynamic softening mechanisms are both significantly affected by a number of factors, such as chemical composition, forming temperature, strain rate and strain [
8,
9]. Constitutive models are extensively employed to describe the correlation between flow stress, strain, strain rate and temperature during hot deformation. In addition to the deformation parameters, the alloy chemical composition also has a significant impact on the high temperature flow behavior of aluminum alloys. Therefore, it is of great interest to develop a model that represents the relationship between chemical compositions, deformation variables and flow stress. This model can be employed to optimize the chemical composition of dilute Al-Fe-Si-Cu alloy as well as process parameters to obtain deserved strength and processability. Constitutive models are either analytical [
10,
11,
12] or phenomenological [
13,
14,
15]. Analytical constitutive models are based on the physical aspects of a material’s behavior and require comprehensive understanding of the underlying mechanisms that control the materials’ deformation. In addition, there are many independent parameters in analytical constitutive equations that require experimental determination. These features make this type of model difficult to apply. An alternative approach would be adapting a phenomenological model based upon empirical observations. However, these models are typically restricted to certain processing domains where a specific deformation mechanism operates and the accuracy of the flow stress predicted by these regression methods is low.
In recent years, artificial neural networks (ANNs) have provided a fundamentally different approach for material modeling and material processing control techniques [
16]. The most important advantage of ANNs is that they do not require postulation of a mathematical model at the outset or the identification of its parameters. ANNs learn from examples and recognize patterns in a series of input and output data without the need for any prior assumptions about their nature and interrelations [
17,
18]. Recently ANNs have been successfully applied to model the high-temperature flow behavior of stainless steels [
18], aluminum alloys [
19,
20], magnesium alloys [
21], titanium alloys [
22,
23] and Al-base metal matrix composites [
24]. However, to date, no ANN model has been developed to simultaneously include the effect of both chemical composition and deformation variables in aluminum alloys.
In the present study, the capability of the ANN approach to predict the high-temperature flow behavior of Al-0.12Fe-0.1Si-Cu alloys was examined as a function of chemical composition and process parameters. An ANN model has been proposed to predict the flow behavior of Al-0.12Fe-0.1Si alloys with various levels of Cu addition (0.002–0.31 wt %) under different deformation conditions. Sensitivity analysis was carried out to quantify the relative importance of Cu addition and individual deformation variables on the flow stress.
2. Experimental Procedures
Al-0.12Fe-0.1Si alloys with different Cu contents ranging from 0.002% to 0.31% were investigated (all alloy compositions in this study are given in wt % unless otherwise indicated). Materials were prepared from commercially pure aluminum (99.7%), Al-50%Si and Al-50%Cu master alloys.
Table 1 provides the chemical compositions of the experimental alloys used. For each composition, approximately 3 kg of material was melted in an electrical resistance furnace and then cast into a rectangular permanent steel mold. Prior to casting, the melts were grain-refined by the addition of 0.015% Ti in the form of an Al-5Ti-1B master alloy. The cast ingots of these alloys were homogenized at 550 °C for 6 h, and then water quenched to room temperature.
Cylindrical samples (10 mm diameter and 15 mm height) were machined from the homogenized ingots. Uniaxial hot compression tests were conducted using a Gleeble 3800 thermomechanical testing unit (Dynamic Systems Inc., Poestenkill, NY, USA) at strain rates of 0.01, 0.1, 1 and 10 s−1 and temperatures of 400, 450, 500 and 550 °C. To minimize the friction between the sample and anvil during test, thin graphite foils were placed on both ends of the samples. Specimens were deformed to a total true strain of 0.8 and then immediately water-quenched to room temperature to retain the deformed microstructure.