Enhanced investigations and modeling of surface roughness of epoxy/ Alfa fber biocomposites using optimized neural network architecture with genetic algorithms
Loading...
Date
2023
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Currently, there is a notable attraction within the industry towards biocomposites, driven by the increasing fascination
with natural fber-reinforced composites (NFRCs). These NFRCs ofer remarkable benefts, including cost-efectiveness,
biodegradability, eco-friendliness, and favorable mechanical properties. As a result, the manufacturing processes of natural
fber reinforced polymer (NFRP) composites have garnered attention from both industrial professionals and scientists. The
emergence of these eco-friendly materials in the automotive and aerospace industries has sparked interest in understanding
their production techniques. However, the machining processes of NFRP composites pose signifcant challenges due to the
complex structure of natural fbers, necessitating thorough studies to address these issues efectively. This research paper
presents a comprehensive investigation on surface roughness during the milling process of Alfa/epoxy biocomposites. A
set of 100 experimental trials was conducted to test the surface roughness, and analysis of variance (ANOVA) was used to
assess the impact of cutting parameters and chemical treatment on surface quality.
To develop a predictive model for surface roughness, a hybrid approach called ANN-GA (artifcial neural networks-genetic
algorithms) is proposed in this research. This approach combines ANN and GA to determine an optimal neural network archi tecture. The performance of the ANN-GA model is compared to the Levenberg–Marquardt backpropagation (LM) algorithm.
ANOVA results show that the feed per revolution have a signifcant infuence on surface roughness, followed by the chemi cal treatment of fbers, while machining direction has a smaller efect. The ANN-GA model demonstrates good accuracy in
surface roughness prediction compared to the LM algorithm
Description
Keywords
Biocomposite · Alfa fbers · Surface roughness · Optimization · ANN · GA