Modeling Water-in-Oil Emulsion Formation Using Fuzzy Logic


Yetilmezsoy K., Fingas M., Fieldhouse B.

JOURNAL OF MULTIPLE-VALUED LOGIC AND SOFT COMPUTING, cilt.18, ss.329-353, 2012 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 18
  • Basım Tarihi: 2012
  • Dergi Adı: JOURNAL OF MULTIPLE-VALUED LOGIC AND SOFT COMPUTING
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.329-353
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

Several compositional factors including density, viscosity, asphaltene, aromatic, saturate and resin contents play an important role to compute a class index (Stability C) which yields either an unstable or entrained water-in-oil state or a meso-stable or stable emulsion. Considering the complex structure and tedious determination procedures of water-in-oil emulsions-based problems, old regression models are not able to capture the nonlinear relationships existing between variables in a complex water-in-oil emulsion system. To undertake these tasks, derivation of a motivation for developing a robust and reliable model has become a particular field of investigation to predict a proper stability index due the involved uncertainties and their poor generalization performance. Recently, it has become apparent that alternative artificial intelligence-based methods, such as fuzzy logic methodology, have been successfully used to deal with subjects having ambiguities and uncertainties. In this study, a MISO (multiple inputs and single output) fuzzy-logic-based model was proposed as a new numerical modeling scheme for the prediction of water-in-oil emulsions formation. The fuzzy-logic predictions were compared to the actual data from some common oils and against a 15-term old regression model. Statistical results clearly indicated that, compared to the regression approach, the proposed MISO fuzzy-logic-based model showed a superior predictive performance on forecasting of water-in-oil emulsions stability with a satisfactory determination coefficient over 0.98.