Modeling And Optimization Of Multi-line Ethylene Plant

Ethylene is one of the most important raw material in petrochemical industry. Optimization of operation for large-scale devices such as ethylene process is important. Shanghai petrochemical ethylene plant has two production lines, the old one and new one. Different process types are used in the two lines, ABB/LUMMUS process for the old one, and SW process fot the new one. These two lines are independent, so they can not only produce ethylene independently, but also can redistribute load through the cross line device between them.This paper establishes a neural network model to predict the products of cracking furnaces, and linear model for prediction of compressor suction pressure, these models are well fited with factory data. In addition, Aspen Plus was used to establishe the model of cryogenic separation of pyrolysis gas process, analyzed the relationship between operating variables and ethylene loss. Based on the actual production and with the help of models, the use of rich ethylene gas was optimized.As there are coupling effects between cracking furnace and compressor, and cross line between these two lines. This paper will combine cracking furnace and compressor, solving the problem jointly with the consideration of energy minimization, and find the relationship between the flow rate of cross line and the overall benefit.