No abstract

A n arti®cial neural network (ANN) model for determining the steady-state behaviour of an industrial Fluid Catalytic Cracking (FCC) unit is presented in this paper. Industrial data from a Greek petroleum re®nery were used to develop, train and check the model. FCC is one of the most important oil re®nery processes. Due to its complexity the modelling of the FCC poses a great challenge. The proposed model is capable of predicting the volume percent of conversion based on six input variables. This work is focused on determining the optimum architecture of the ANN, in order to gain good generalization properties. The results show that the ANN is able to accurately predict the measured data. The prediction errors in both training and validation data sets are almost the same, indicating the capabilities of the model to accurately generalize when presented with unseen data. The neural model developed is also compared to an existing non-linear statistical model. The comparison shows that the neural model is superior to the statistical model.

No abstract

A computer-aided tool for the simulation, optimization and analysis of the combined operation of the hydrodesulphurization (HDS) and fluid catalytic cracking (FCC) processes in an oil refinery is presented. The optimization of these processes is an important yet difficult engineering task, because of the complexity in the integration of the two units, the large number of interacting variables, the product quality specifications and the financial benefits associated. The proposed tool is developed in a user-friendly Visual Basic environment and operates in two different modes: the modelling-prediction mode and the optimization-sensitivity analysis mode. The modelling of the processes is based on 'short form' models, which were created following statistical and neural network approaches. This kind of model usually has short computing time requirements, which is critical for the optimization mode. The optimization algorithm is based on a financial objective function with a flexible form, which gives the user the option to explore a variety of scenarios. Industrial runs have verified the modelling accuracy of the tool. The optimization scenarios examined include the contemporary needs of modern refineries for LPG and gasoline maximization, subject to strict quality specifications. The demonstration of this tool aims to give an insight into the system dependencies and add knowledge on the possibility of a more profitable operation of such a complex process.

No abstract