SiCp/Al Metal matrix composite (MMCs) materials are known as hard to machine materials despite of their demanding application in aerospace, automobiles and heavy-duty industries. More specifically, high-volume silicon carbide particles reinforce aluminum-based metal matrix composite (SiCp/Al 50% vol:) are difficult to cut materials due to their exceptionally high strength ratio, stiffness, high thermal co-efficient, tough and harsh composite. However, machining SiCp/Al MMCs is a perplexing task for achieving required dimensional accuracy, surface quality, cutting force and tool life. This study focuses on the 3 factors, 4 levels of machining parameters selecting (orthogonal array <InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="12008_2025_2282_Article_IEq1.gif" Format="GIF" Height="23" Rendition="HTML" Resolution="72" Type="Linedraw" Width="58" /> </InlineMediaObject> <EquationSource Format="TEX">${L}_{16}\left({3}^{4}\right)$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi>L</mi> <mn>16</mn> </msub> <mfenced close=")" open="("> <msup> <mrow> <mn>3</mn> </mrow> <mn>4</mn> </msup> </mfenced> </mrow> </math></EquationSource> </InlineEquation>), a total of 16 lathe machining experiments using DOE method are used to obtain experimental data. Experiments were conducted to enhance the machinability of an exceptionally high-volume fraction of 50% SiCp/Al MMCs in CNC lathe machining operation. The optimization and modeling of the multiple input variables and key output variables including response factors such as tool life = TL, surface roughness = Ra and cutting forces = Fx, Fy, and Fz. Using multiple input variable parameters such as, cutting speed = cv, federate = f, and depth of cut = ap, are performed using the multiple regression method (MRA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) Modeling to predict cutting forces, surface roughness and tool life. The maximum and minimum error percentage for each output and optimal combination of cutting parameters are given in the conclusion section. The ANFIS modeling and multiple regression prediction model are seen accurate, precise and in close agreement with experimental results.

Parametric modeling and optimization for machinability performance enhancement of difficult-to-cut SiCp/Al (50%) MMCs using ANFIS and MRA

期刊论文
发表日期：

2025年4月23日

查看全文

Rashid Ali Laghari,
Ahmed Aly Diaa Sarhan

摘要

SiCp/Al Metal matrix composite (MMCs) materials are known as hard to machine materials despite of their demanding application in aerospace, automobiles and heavy-duty industries. More specifically, high-volume silicon carbide particles reinforce aluminum-based metal matrix composite (SiCp/Al 50% vol:) are difficult to cut materials due to their exceptionally high strength ratio, stiffness, high thermal co-efficient, tough and harsh composite. However, machining SiCp/Al MMCs is a perplexing task for achieving required dimensional accuracy, surface quality, cutting force and tool life. This study focuses on the 3 factors, 4 levels of machining parameters selecting (orthogonal array ${L}_{16}\left({3}^{4}\right)$ $L_{16} (3^{4})$ ), a total of 16 lathe machining experiments using DOE method are used to obtain experimental data. Experiments were conducted to enhance the machinability of an exceptionally high-volume fraction of 50% SiCp/Al MMCs in CNC lathe machining operation. The optimization and modeling of the multiple input variables and key output variables including response factors such as tool life = TL, surface roughness = Ra and cutting forces = Fx, Fy, and Fz. Using multiple input variable parameters such as, cutting speed = c_v, federate = f, and depth of cut = a_p, are performed using the multiple regression method (MRA) and Adaptive Neuro-Fuzzy Inference System (ANFIS) Modeling to predict cutting forces, surface roughness and tool life. The maximum and minimum error percentage for each output and optimal combination of cutting parameters are given in the conclusion section. The ANFIS modeling and multiple regression prediction model are seen accurate, precise and in close agreement with experimental results.