Universiti Teknologi Malaysia Institutional Repository

Computational intelligence approach for prediction of hardness performance in coating process

Mohamad, Muhammad ‘Arif (2014) Computational intelligence approach for prediction of hardness performance in coating process. Masters thesis, Universiti Teknologi Malaysia, Faculty of Computing.

[img]
Preview
PDF
170kB

Abstract

Nowadays, coated materials are widely used due to their excellent properties especially for the hardness performance. The hardness of coated tools is determined by the coating process parameters. Traditionally, optimization to obtain the best coating performance of the parameters in a coating process was done by trial and error approach. However the traditional approach has raised issues with regards to cost and customization. In this research, these two issues were addressed by using a computational intelligence approach to develop a model for predicting the output responses in order to identify the optimal parameters used in coating process. Previous studies have shown that this approach was successfully adopted for optimization purpose in many types of domains. However, it was not yet applied in the coating process domain. Thus, two methods from computational intelligence approach were applied, namely Support Vector Machine (SVM) and Artificial Neural Network (ANN). The comparisons of the performances of the developed models were conducted based on predictive performance measurements such as percentage error, mean squared error (MSE), co-efficient determination (R2), and model accuracy and complexity. The results showed that, SVM obtained better predictive performances and less complicated in comparison to other prediction models. As a conclusion, SVM has demonstrated its capability in predicting the hardness performance of coating process and outperformed the other models. Besides that, the model is a promising alternative tool for coating process optimization as compared to the traditional approach.

Item Type:Thesis (Masters)
Additional Information:Thesis (Sarjana Sains (Sains Komputer)) - Universiti Teknologi Malaysia, 2014; Supervisors : Dr. Nor Azizah Ali, Prof. Dr. Habibollah Harun
Uncontrolled Keywords:support vector machine (SVM), artificial neural network (ANN)
Subjects:Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Divisions:Computing
ID Code:48024
Deposited By: Fazli Masari
Deposited On:13 Oct 2015 02:51
Last Modified:06 Aug 2017 09:54

Repository Staff Only: item control page