Prediction of the product yield from catalytic cracking (MIP) process by an 8-lump kinetic model combined with neural network

Abstract

ABSTRACTBased on the characteristics of maximizing iso-paraffins (MIP) process and industrial data, an 8-lump kinetic model for MIP process is developed. And the 47 kinetic parameters of the model are calculated by Runge-Kutta method and genetic algorithm. It is seen that kinetic parameters show good consistence with the reaction mechanism of catalytic cracking. The average relative errors between calculated values and real values of product distribution are all below 5%. Then the model is modified by 14-7-5 type of back propagation (BP) neural network. As a result, the product distribution can be predicted more accurately by the hybrid model. Therefore, the combination of lump model and neural network can provide a new direction for simulation and optimization for heavy oil catalytic cracking.