Comparison of the artificial neural network model prediction and the experimental results for cutting region temperature and surface roughness in laser cutting of AL6061T6 alloy
Abstract
In this study, a function approximation procedure is used, which called artificial neural network (ANN), according to the experimental results of the temperature of cutting region and surface roughness in cutting by the laser of AL6061T6 alloy. The cutting speed, laser power, sheet thickness, and assistant gas pressure as the inputs parameters and the surface roughness and cutting temperature as the target attributes are considered. The novelty of this study is shown by preparing 30 unalike ANN procedures to propose suitable architectures and training algorithms for them. The results of ANN are compared with the experimental results, and the error percent between them is derived. According to the comparison, the error percent between the experimental data and ANN is in a reasonable range, and this numerical method can be applied with low times and costs. For the cutting temperature, the mean value of the error percentages between the numerical results and experimental data is 0.66%. Also, for the surface roughness, the mean value of the error percentages is 5.79%.
- Publication:
-
Infrared Physics and Technology
- Pub Date:
- August 2020
- DOI:
- 10.1016/j.infrared.2020.103364
- Bibcode:
- 2020InPhT.10803364Y
- Keywords:
-
- Optimization procedure;
- Artificial neural network;
- Surface roughness;
- Training algorithms;
- Experimental results;
- Cutting temperature