Loading...

Application of ANFIS-PSO as a novel method to estimate effect of inhibitors on asphaltene precipitation

Malmir, P ; Sharif University of Technology | 2018

798 Viewed
  1. Type of Document: Article
  2. DOI: 10.1080/10916466.2018.1437637
  3. Publisher: Taylor and Francis Inc , 2018
  4. Abstract:
  5. Asphaltene precipitation in petroleum industries is known as major problems. To solve problems there are approaches for inhibition of asphaltene precipitation, Asphaltene inhibitors are known effective and economical approach for inhibition and prevention of asphaltene precipitation. In the present study Adaptive neuro-fuzzy inference system (ANFIS) was coupled with Particle swarm optimization (PSO) to create a novel approach to predict effect of inhibitors on asphaltene precipitation as function of crude oil properties and concentration and structure of asphaltene inhibitors.in order to training and testing the algorithm, a total number of 75 experimental data was gathered from the literature. The results of this model showed that average absolute relative deviation (AARD), the coefficient of determination (R2) and root mean square error (RMSE) for the dataset of the algorithm are 2.5058, 0.99342 and 0.64238 respectively. According to the graphical and statistical reports, the proposed ANFIS-PSO has acceptable potential for investigation of effect on asphaltene inhibitors. © 2018 Taylor & Francis Group, LLC
  6. Keywords:
  7. ANFIS-PSO ; Asphaltene ; Inhibitors ; Precipitation ; Predicting model ; Corrosion inhibitors ; Crude oil ; Fuzzy inference ; Fuzzy neural networks ; Fuzzy systems ; Inference engines ; Mean square error ; Optimization ; Particle swarm optimization (PSO) ; Precipitation (chemical) ; Adaptive neuro fuzzy inference systems (ANFIS) ; Asphaltene precipitation ; Average absolute relative deviations ; Coefficient of determination ; Predicting models ; Root mean square errors ; Training and testing ; Asphaltenes
  8. Source: Petroleum Science and Technology ; Volume 36, Issue 8 , 2018 , Pages 597-603 ; 10916466 (ISSN)
  9. URL: https://www.tandfonline.com/doi/abs/10.1080/10916466.2018.1437637