Adaptive neuro fuzzy inference system to predict average asphaltene particle diameter case study: in thermal de-asphalting process

In thermal de-asphalting procedure, where asphaltene particles are aggregated through heating crude oil at an elevated temperature, predicting the asphaltene particles' diameter is of major importance. In this study, thermal de-asphalting process is modeled using adaptive neuro fuzzy inference system by considering isolating temperature, crude oil API and asphaltene content as inputs to predict asphaltene particles' average diameter. The experimental data measured in a custom built thermal de-asphalting set-up, are applied to construct neuro fuzzy model structure. The average relative error of proposed model, for all data points is 3.06%, indicating an excellent agreement between model predictions and experimental values.