Modeling and optimization of a catalytic naphtha reformer

Date

1996-05

Journal Title

Journal ISSN

Volume Title

Publisher

Texas Tech University

Abstract

Catalytic naphtha reforming is one of the key processes in petroleum refining, converting gasoline boiling range low-octane hydrocarbons to high-octane compounds which can be blended into gasoline. Other valuable byproducts include hydrogen and cracked light gases.

Modeling of a typical semi-regenerative catalytic reformer has been carried out involving most its key constituent units. Kinetic modeling of the reactions occurring in the fixed bed reactors connected in series formed the most significant part of the overall simulation effort. A reaction scheme involving 36 pseudocomponents connected together by a network of 35 reactions for components in the C5-C10 range has been modeled. The Hougen-Watson Langmuir-Hinshelwood type reaction rate expressions are used to represent rate of each reaction. Deactivation of the catalyst was modeled by including the corresponding equations for coking kinetics.

The kinetic model was parameterized by bench marking the kinetic model against plant data. A feed characterization procedure was developed to infer the composition of chemical species in the feed and reformate from the given ASTM distillation data. A nonlinear regression procedure was carried out to calculate the rate parameters that provided the best match between the model and the plant data.

The key to most optimum reformer operation lies in choosing the four catalyst bed inlet temperatures, and recycle ratio. Sometimes, in practice, four beds are operated at same inlet temperatures and varying influence of each bed behavior on the outlet properties is not taken into account. This leads to a sub-optimal operation. It is also important to consider the objective functions over the entire catalyst run-length, which has been predetermined for optimization analysis. The optimization analysis was conducted in two stages. In the first stage, the decision variables are optimized but held constant throughout the life of the catalyst. This time-invariant mode of operation showed a significant improvement in objective function over the base case. In the second stage of the analysis, the problem was expanded to optimize the path of the decision variables over the run-length. This time-optimal problem also showed a substantial improvement in profits over the time-invariant case. Finally, sensitivity of objective functions to uncertainties in model parameters was examined.

Description

Citation