A New Method for the Prediction of the Gasoline Yield of the MIP Process

摘要

Themaximizing iso-paraffins (MIP) process, a family craft of fluid catalytic cracking process, is designed to reduce olefin content and increase isoparaffin and aromatic content in gasoline. The authors propose a method used to predict the gasoline yield of MIP process. The prediction model with 17 input variables is based on generalized regression neural network and adaptive boosting (adaboost) algorithm. The mean square error between predictive gasoline yield and actual yield is 2.46.