Development of regression models to control product yields and properties of the fluid catalytic cracking process

Regression models for predicting product yields and properties for the Flui d Catalytic Cracking (FCC) process are presented. The feedstock properties are used as the primary correlation parameters. The models are based on dat a collected from various sources from the literature on pilot and commercia l plants. The predictions of these models are compared to previous models a nd also to real data. The proposed models give consistently good prediction s. These models can be used to evaluate different feeds to the FCC process and can also be integrated within general refinery mathematical programming (e.g. LP) software for planning and scheduling purposes.