سال انتشار: ۱۳۹۳

محل انتشار: دومین کنفرانس بین المللی نفت، گاز و پتروشیمی

تعداد صفحات: ۳

نویسنده(ها):

b Bagheritabar – Azad University, Mahshahr,
g Moradi – Faculty of Chemical Engineering, Razi University

چکیده:

In characterization of heptane plus fractions in addition to bulk properties such as molecular weight and specific gravity, properties distribution is also required. Bulk properties can be measured easily but determination of properties distribution is more costly and time consuming. So the characterization methods are used for determining the properties distribution, which need at least some distribution data for calculating their constant coefficients. The main objective of this paper is to useartificial intelligence as a characterization method to estimate heptane plus fractions. In this work an artificial neural network has been trained and test with 62 samples of gas condensate from an Iranian field. Inputs of ANN are specific gravity (SG), molecular weight (Mw) and true boiling point (Tb). The results show that the estimated weight fractions distribution is very close to the experimental results.