Optimized processing power and trainability of neural network in numerical modeling of Al Matrix nano composites |
| |
Authors: | Ali Asghar Tofigh Mohammad Reza Rahimipour Mohsen Ostad Shabani Mehdi Alizadeh Fatemeh Heydari Ali Mazahery Mansour Razavi |
| |
Institution: | Materials and Energy Research Center (MERC), Tehran, Iran |
| |
Abstract: | In this research, an experimental study of reinforcing alumina nano-particles into the aluminum alloy matrix was implemented to verify the accuracy of modeling results obtained by feed forward neural networks. Artificial neural network combined with numerical technique were used to predict the various parameters of mechanical properties such as hardness, tensile and compressive yield stress, UTS and elongation percentage. Much experimentation were taken to discover a suitable number of hidden neurons, avoid detraction from the trainability and enable feed forward neural networks to solve more complex problems. The predictions were found to be consistent with experimental measurements. |
| |
Keywords: | Aluminum Metal matrix Nano composites Modeling |
本文献已被 ScienceDirect 等数据库收录! |