Exploring the intricacies of machine learning-based optimization of electric discharge machining on squeeze cast TiB2/AA6061 composites: Insights from morphological, and microstructural aspects in the surface structure analysis of recast layer formation and worn-out analysis

Abstract

Aluminium (Al) Alloy-6061/TiB2 was developed with Squeeze casting while varying composite quantities with titanium diboride (TiB2). The metallographic structure of the composite was investigated using EDS and scanning electron microscopy (SEM). The machining of developed composite is challenging with the conventional methods due to higher tool wear and surface roughness (Ra). This study analyses the output responses, and machinability of TiB2 of 15% volume reinforced AA6061 composites using Electrical Discharge Machining (EDM). The EDM operation variables such as current, pulse on time, and voltage gap were utilized to examine material removal rate (MRR), tool wear rate (TWR) and Ra employing Box–Behnken design of experiments. The MRR, TWR and Ra data acquired from various experiments were optimized considering single and multiple objectives by a hybrid approach. Machine learning (ML) was applied to predict responses using linear regression (LR), decision tree (DT) and random forest (RF). The samples' recast layer was examined using SEM, and the thickness of the recast layer was also documented. The Al6061 alloy's SEM micrographs reveal deep equiaxed dimples on its fracture surface, indicating a high level of plastic deformation before failure. TiB2 is evenly distributed as darker particles in the Al-matrix alloy with no directional-orientation. The Al6061-15 wt% TiB2 composites' fracture surfaces display deep dimples, ductile-fracture for Al6061 alloy due to higher plastic deformation, and cleavage facets in some areas. The aluminum alloy's wear pattern depicts long, extensive ploughing grooves with short cracks parallel to the sliding direction at the grooves' bottom. SEM micrographs in recast-layer formation have reported that molten metal ligaments break into droplets during flushing, exposing fresh surfaces. The aluminum alloy's groove width is larger compared to the composites. Thus, it can be observed that the wear tracks on the base Al6061 alloy are larger and deeper in comparison to the Al- 15 wt% TiB2 surface composites due to the presence of hard TiB2 particles. Also, the recast layer may be advantageous due to its wear-resistant properties. TiB2 has importance in an industry where components slide with each other it is used in disc brakes in automobiles. Compared to industry standards, single-objective optimization considerably enhanced the outcomes: MRR increased by 10.61%, TWR enriched by 25.91%, and Ra increased by 17.60%. According to an ANOVA, peak current contributed 87.88%, 98.18%, and 95.80% respectively to MRR, TWR, and Ra. It is requisite to investigate multi-objective optimization for EDM of Al6061-15 wt% TiB2 composites' owing to the diverse optimal factor combinations for responses. Ra was slightly increased by 1.14% due to simultaneous optimization utilizing hybrid techniques with Equal and Entropy weights, which diminished TWR by 11.74% and increased MRR by 11.66%. ML, particularly DT, improves machining performance by assisting productivity, durability, and cost-efficiency under changeable circumstances and raising forthcoming study possibilities.

Keywords

• Electrical discharge machining
• Aluminium Alloy-6061/TiB2 composite material
• Squeeze casting
• Metal matrix
• Machine learning
• TOPSIS-AHP-Entropy weights