Hard Turning on JIS S45C Structural Steel: An Experimental, Modelling and Optimisation Approach

  • R. Kumar School of Mechanical Engineering, Kalinga Institute of Industrial Technology (KIIT), Deemed to be University, Bhubaneswar-24, Odisha, 751024, India, Tel: +91-0674-6540805
  • A. Modi School of Mechanical Engineering, Kalinga Institute of Industrial Technology (KIIT), Deemed to be University, Bhubaneswar-24, Odisha, 751024, India, Tel: +91-0674-6540805
  • A. Panda School of Mechanical Engineering, Kalinga Institute of Industrial Technology (KIIT), Deemed to be University, Bhubaneswar-24, Odisha, 751024, India, Tel: +91-0674-6540805
  • A. K. Sahoo School of Mechanical Engineering, Kalinga Institute of Industrial Technology (KIIT), Deemed to be University, Bhubaneswar-24, Odisha, 751024, India, Tel: +91-0674-6540805
  • A. Deep School of Mechanical Engineering, Kalinga Institute of Industrial Technology (KIIT), Deemed to be University, Bhubaneswar-24, Odisha, 751024, India, Tel: +91-0674-6540805
  • P. K. Behra Dyna Force Hydraulics Pvt. Ltd., Pune-11, Maharashtra, India
  • R. Tiwari Dyna Force Hydraulics Pvt. Ltd., Pune-11, Maharashtra, India
Keywords: JIS S45C Steel, hard turning, CVD-coated cutting tool, BNN, RNN, WPCA

Abstract

The present research is performed while turning of JIS S45C hardened structural steel with the multilayered (TiN-TiCN-Al2O3-TiN) CVD coated carbide insert by experimental, modelling and optimisation approach. Herein, cutting speed, feed rate, and depth of cut are regarded as input process factors whereas flank wear, surface roughness, chip morphology are considered to be measured responses. Abrasion and built up-edge are the more dominant mode of tool-wear at low and moderate cutting speed while the catastrophic failure of tool-tip is identified at higher cutting speed condition. Moreover, three different Modelling approaches namely regression, BNN, and RNN are implemented to predict the response variables. A Back-propagation neural network with a 3-8-1 network architecture model is more appropriate to predict the measured output responses compared to Elman recurrent neural network and regression model. The minimum mean absolute error for VBc, Ra and CRC is observed to be as 1.36% (BNN with 3- 8-1 structure), 1.11% (BNN with 3-8-1 structure) and 0.251 % (RNN with 3-8-1 structure). A multi-performance Optimisation approach is performed by employing the weighted principal component analysis. The optimal parametric combination is found as the depth of cut at level 2 (0.3 mm)-feed at level 1 (0.05 mm/rev) – cutting speed at level 2 (120 m/min) considered as favourable outcomes. The predicted results were validated through a confirmatory trial providing the process efficiency. The significant improvement for S/N ratio of CQL is observed to be 9.3586 indicating that the process is well suited to predict the machining performances. In conclusion, this analysis opens an avenue in the machining of medium carbon low alloy steel to enhance the machining performance of multi-layered coated carbide tool more effectively and efficiently.

Published
2019-12-30
How to Cite
Kumar, R., Modi, A., Panda, A., Sahoo, A. K., Deep, A., Behra, P. K., & Tiwari, R. (2019). Hard Turning on JIS S45C Structural Steel: An Experimental, Modelling and Optimisation Approach. International Journal of Automotive and Mechanical Engineering, 16(4), 7315-7340. https://doi.org/10.15282/ijame.16.4.2019.10.0544