The Comparison of Artificial Intelligence and Traditional Approaches In FCCU Modeling


Creative Commons License

Zeydan M.

INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING-THEORY APPLICATIONS AND PRACTICE, vol.15, no.1, pp.1-15, 2008 (Journal Indexed in SCI) identifier

  • Publication Type: Article / Article
  • Volume: 15 Issue: 1
  • Publication Date: 2008
  • Title of Journal : INTERNATIONAL JOURNAL OF INDUSTRIAL ENGINEERING-THEORY APPLICATIONS AND PRACTICE
  • Page Numbers: pp.1-15

Abstract

FCCU (Fluid Catalytic Cracking Unit) is a part of oil refinery production process whereby valuable products such as gasoline, LPG (Liquid Petroluem Gas), diesel are manufactured in a short period of time. The objective of this paper is to find the most robust model by comparing the models of FCCU that are developed using different methodologies. The models of FCCU are developed by using Artificial Neural Network (ANN), Fuzzy Logic, Neuro-Fuzzy, and traditional methodology. In this paper, the criteria used for measuring the performance of different models is root mean squared error (RMSE). The models are applied to the real data obtained from TUPRAS (Turkish Petroleum Refineries Corporation)FCCU. Kurihara (1967) model is used as the traditional model for comparing with intelligence modeling techniques. Finally, the Fuzzy Neural Network (FNN) model was found as the model with the minimum RMSE. Qwicknet 2.23, MATLAB 6.5, and Neuro-solutions 4.1 softwares have been used for the construction of ANN, fuzzy, and neuro-fuzzy models, respectively.