Predictive modelling of machining parameters of S45C mild steel

Abbas, Adnan Jameel (2016) Predictive modelling of machining parameters of S45C mild steel. Doctoral thesis, Universiti Teknikal Malaysia Melaka.

	Text Predictive Modelling Of Machining Parameters Of S45C Mild Steel 24 Pages.pdf - Submitted Version Download (391kB)
	Text (Full Text) Predictive modelling of machining parameters of S45C mild steel.pdf - Submitted Version Restricted to Registered users only Download (2MB)

Abstract

The determination of the ideal parameters and performance are among the most crucial and complex factors in the process planning and economics of metal cutting operations. Minimization of undesired parameters in production operations is very necessary to increase the productivity and reduce the costs. Turning process is one of complicated operations to control its cutting parameters because it depends upon several conflicting cutting parameters that must be adjusted at the same time accurately. In this research, minimization of cutting temperature, work piece surface roughness, cutting time and cutting tool flank wear are achieved in CNC turning operation. A mild steel material type JIS S45C and a tungsten carbide insert type SPG-422 Grade E30 are used as workpiece and cutting tool materials via dry machining respectively. The temperature of primary plastic deformation zone which called shearing zone, and secondary deformation zone which called chip slides on the rake face zone are measured. This research adopts the utilization of three types of heurestic algorithms to achieve the minimization operation; Genetic Algorithm (GA). Particle Swarm Optimization (PSO) and Artificial Immune System (AIS). Four objective functions are used as input for the intellegent algorithms for minimization purpose, two objective functions for temperature minimization and one for surface roughness minimization and one for cutting time minimization. The outputs of huerestics algorithms are; minimum temperature, minimum surface finish, minimum cutting time. This research includes simulation and experimental work results. The simulation operation is executed by PSO, AIS and GA to find the ideal results, then the these results are tested by CNC turning experimental work to find the accuracy percentage of algorithms and seleceting the ideal one. The simulation results of GA, PSO and AIS showed that the GA1 algorithm which used the first main temperature objective function gives the best temperature value (35. 7 0C) compared with other algorithms, followed by PSO1 (70.2 0C), then AIS1 (112.8 0C). The PSO1 algorithm which used first main temperature objective function gives the best roughness value (0.52 μm) compared with other algorithms, followed by the AIS2 and PSO2 that give (0.86 μm). In cutting time estimation, it is shown that the results of the second main objective functions estimations are better than the first main objective function results. The AIS2 algorithm gives the best time value (3.22 min) compared with the other algorithms, followed by AIS1 (5.05 min), then PSO2 (5.16 min). The experimental results indicate that the best value of cutting temperature which ranged between (150.2-175.3 ͦC) can be obtained with the combination of input parameters- cutting speed (40 m / min), feed rate (0.05 mm / rev) and depth of cut (0.6 mm). In addition, the best value of surface roughness which ranged between (0.26-1.63 μm) can be obtained with the combination of input parameters-cutting speed (140 m / min), feed rate (0.05 mm/rev) and depth of cut (0.9 mm). Also, the best value of flank wear which ranged between (0.07-0.16mm) can be obtained with the combination of input parameters-cutting speed (40m/min), feed rate (0.05mm/rev) and depth of cut (0.6mm). The artificial neural network type Network Fitting Tool (NFTOOL) is used as a modeling technique for manipulating the ideal algorithm parameters. The results of NFTOOL indicates that (9-6-3) network is the ideal type because it gives lower testing (MSE) equal to (3.97214 *10-12). The effects of cutting parameters on performance characteristics are studied using the signal-to-noise (S/N) ratio method. Finally, selection the better algorithm that gives the best and ideal results of temperature, roughness and cutting time is selected as an ideal network for prediction the ideal cutting performance for future works.

Item Type:	Thesis (Doctoral)
Uncontrolled Keywords:	Machine-tools, Machining Automation, Metal-cutting tools, Surface roughness
Subjects:	T Technology > T Technology (General) T Technology > TJ Mechanical engineering and machinery
Divisions:	Library > Tesis > FKP
Depositing User:	Nor Aini Md. Jali
Date Deposited:	01 Jun 2017 05:33
Last Modified:	13 Jun 2022 15:47
URI:	http://eprints.utem.edu.my/id/eprint/18559
Statistic Details:	View Download Statistic

Actions (login required)

View Item