ABSTRACT
The squeeze casting process blends the features of casting, forging, and also helps in achieving better manufacturing abilities; attain smooth uniform surface, high productivity, superior mechanical properties, and refined microstructure. The precise control of process variables is the most feasible solution to obtain a defect-free casting. In the squeeze casting process, surface roughness, hardness, ultimate tensile strength, and yield strength are influenced by process variables. Hence, this paper deals with the new intelligent-based squeeze casting process using LM-20 alloy. Aluminium is commonly employed in automobile and aerospace applications. The proposed casting process considers the input parameters such as squeeze pressure, pressure duration, pouring temperature, and Die preheats temperature. The output variables such as Yield strength, hardness, Ultimate tensile strength, and Surface roughness is artificially computed using the regression equations obtained for different responses using the non-linear regression models based on Central Composite Design (CCD). Initially, the considered process variables are trained in a Neural Network (NN). Further, the employed parameters are optimised to attain the best solution within the concerned limit using an improved meta-heuristic algorithm called Modified Coefficient based DHOA called MC-DHOA. The fitness evaluation of parameter optimisation depends on NN for prediction, and the objective function intends to minimise the error between optimised and target value. The analysis proves that the implemental model helps to select the most influential process parameters in the squeeze casting process within less duration. The overall performance of the proposed MC-DHOA is 16.5% better than PSO, 1% better than FF, 11% better than GWO, 4.6% better than WOA, and 2.15% better than DHOA.
Nomenclature
Abbreviations | = | Descriptions |
ANOVA | = | Analysis Of VAriance |
BP | = | Back Propagation |
BPNN | = | Back Propagation tuned Neural Network |
CCD | = | Central Composite Design |
DHOA | = | Deer Hunting Optimization Algorithm |
DoE | = | Design of Experiments |
FF | = | FireFly algorithm |
GA | = | Genetic Algorithm |
GA-NN | = | Genetic Algorithm tuned Neural Network |
GWO | = | Grey Wolf Optimization |
Mpa | = | Ultimate Tensile Strength |
NN | = | Neural Network |
OES | = | Optical Emission Spectrometer |
PSO | = | Particle Swarm Optimization |
RNN | = | recurrent neural networks |
RSM | = | Response Surface Methodology |
SEM | = | Scanning Electronic Microscope |
SiC | = | Silicon Carbide |
WOA | = | Whales Optimization Algorithm |
Disclosure statement
No potential conflict of interest was reported by the authors.
Additional information
Notes on contributors
Pratheesh G Panicker
Pratheesh G Paricker had Secured B-Tech in Mechanical Engineering from CUSAT and M-Tech in Mechanical (Production) Engineering from Calicut University. Also secured MBA in Finance and HR from MG University. Pursuing PhD.in MG University. I have got 7 years of teaching experience and 3 year industrial experience. Currently working as factory manager in public sector. I have got membership in International Association Of Engineers(Membership No.108403)and Indian Science Congress Association(Membership No.1556).Currently doing research in squeeze casting.
Shajan Kuriakose
Shajan Kuriakose was the Head of Department of Mechanical Engineering ,M A college ,Kothamangalam and was the co-ordinator of research, National Conference on Manufacturing and Management, chairman of International conference on Design and Manufacturing. He was the member of Research Council of Kerala technological university and also expert in the Aicte inspection team.He is a member of Institution of Engineers and Indian society for technical Education. He had secured his B-Tech in mechanical Engineering from MA college, M-Tech from IIT Madras and Ph.d in Manufacturing from IIT Madras.