Publication: Aplication of artificial intelligent (ai) to predict co2 laser cut quality of 304 stainkess steel
| datacite.subject.fos | oecd::Engineering and technology::Mechanical engineering | |
| dc.contributor.author | Chong, Zhian Syn | |
| dc.date.accessioned | 2024-08-12T01:31:38Z | |
| dc.date.available | 2024-08-12T01:31:38Z | |
| dc.date.issued | 2008-04-01 | |
| dc.description.abstract | In this research, an empirical study is carried out to develop a neural network model to predict the effect of the CO2 laser cut quality based on the cutting parameter on 5mm 304-stainless steel which the data are obtained from the present students regarding on their final year project topics. The neural network model is developed using Levenberg-Marquardt approximation algorithm as it can converge the network in a much faster way and the error goal value is able to set to a much lower value. The architecture, established using this algorithm model for training phase consisted of three input (power, cutting speed, and assists gas pressure), three hidden layers, fifteen neurons in each hidden layer, 'tansig' as the transfer function for these layers and four output nodes (top kerf width, hardness, heat affected zone and surface roughness) using the 'purelin' transfer function. The chosen architecture model (3 - 15 - 15 - 15 - 4), is designated through a trial and error method. Furthermore, mathematical equations are also developed using regression analysis to describe the effect of the cutting parameters on the laser eut quality that can be used for comparison purpose with NN model. Both predicting models are done using MATLAB toolbox and the accuracy are evaluated and tested by a set of experimental data and both models are found to be in good agreement. From the prediction process that is performed, it is found that the average efficiency of both models using MATLAB program are 92.04% (Neural Network) and 92.07% (Regression) for four predicted outputs. The GUl form of expert system is also constructed using MATLAB based on the neural network and regression model for the prediction. | |
| dc.identifier.uri | https://erepo.usm.my/handle/123456789/20205 | |
| dc.language.iso | en | |
| dc.title | Aplication of artificial intelligent (ai) to predict co2 laser cut quality of 304 stainkess steel | |
| dc.type | Resource Types::text::report | |
| dspace.entity.type | Publication | |
| oairecerif.author.affiliation | Universiti Sains Malaysia |