Authors | Sandeep1, Deepak Chhabra1, R.K. Gupta2 |
Affiliations |
1Department of Mechanical Engineering, UIET, Maharshi Dayanand University, Rohtak Haryana, India 2Department of Mechanical Engineering, Manipal University Jaipur, Jaipur, Rajasthan, India |
Е-mail | deepak.chhabra@mdurohtak.ac.in |
Issue | Volume 13, Year 2021, Number 2 |
Dates | Received 12 January 2021; revised manuscript received 25 March 2021; published online 09 April 2021 |
Citation | Sandeep, Deepak Chhabra, R.K. Gupta, J. Nano- Electron. Phys. 13 No 2, 02004 (2021) |
DOI | https://doi.org/10.21272/jnep.13(2).02004 |
PACS Number(s) | 68.35.Ct, 62.23.Pq |
Keywords | Fused deposition modeling, Artificial neural network integrated with genetic algorithm, Surface roughness (2) , Experimental design matrix, Optimization (14) . |
Annotation |
The surface characteristics of components fabricated by additive manufacturing techniques are greatly affected by the input parameters. In this work, momentous input factors (layer thickness (LT), temperature (T), printing speed (S), outer wall speed (OWS), raster angle (RA), orientation (Or.), outer wall line width (OWLW), infill overlap (IO), infill line width (ILW) of fused deposition modeling (FDM) printer are modeled and optimized for getting the better surface roughness (SR) of carbon based nylon (PA-CF) composite material fabricated parts. To develop input experimental matrix, central composite design method has been utilized and on these input parameters, surface roughness of each run has been measured using Mitutoyo Talysurf surface roughness measuring instrument. A total number of 61 specimens have been fabricated on different input parameters and their surface roughnesses are tested. The minimum surface roughness value of test specimens with experimental design matrix was recorded as 6.331 mm. The modeling and optimization of experimental design matrix has been carried out using evolutionary algorithm i.e. artificial neural network integrated with genetic algorithm (ANNGA). The minimum value obtained using ANNGA for roughness is 5.01788 mm, corresponding to various optimum input factors as LT = 0.1776 mm, T = 236.0609 ºC, S = 40.7369 mm/s, OWS = 20.0676 mm/s, RA = 43.9177º, OWLW = 0.3445 mm, Or. = 0.00180, IO = 56.6295 %, ILW = 0.3488 mm. At these optimized input factors value one end-use part is also fabricated and the developed hybrid model is validated. The artificial neural network integrated with genetic algorithm could be anticipated for better prophecy, factors optimization and outcomes for any engineering application tribulations. |
List of References |