Research Paper
Year: 2021 | Month: January | Volume: 8 | Issue: 1 | Pages: 183-188
Development of Welding Fumes Velocity Predictive Models Using Artificial Neural Network
Usman .F1, Achebo .J .I2, Ozigagun A3
1Department of Production Engineering, Faculty of Engineering, University of Benin, P.M.B 1154, Benin City, Edo State
2Department of Production Engineering, Faculty of Engineering, University of Benin, P.M.B 1154, Benin City, Edo State
3Department of Production Engineering, Faculty of Engineering, University of Benin, P.M.B 1154, Benin City, Edo State, Nigeria
Corresponding Author: Usman .F
ABSTRACT
This research study concentrates on predictive models for welding fumes velocity using artificial neural network. The experiment was optimized using an expert design software, which produced thirty runs of tungsten inert gas mild steel weld specimen, thereafter measuring the welding fumes velocity response. The fume velocity data collected was trained, validated and tested. Analysis of variance ANOVA and model summary statistics helped to check for the models accuracy and reliability. The result obtained from the model indicates that the network developed can predict the fumes velocity adequately with a very low mean square error.
Keywords: development, welding, fume velocity, models artificial neural network, predictive.
[PDF Full Text]