Modelling simulation and optimization of dehydration and desalting process

Abstract

The trigger for this research was to develop process model for Dehydration and Desalting Process (DDP) that may be used at oilfield plant to simulate various operating scenarios in order to support decisions for optimizing plant’s operation. For building such model, I have utilized the concept of Artificial Neural Network (ANN). This technique allows to correlate performance of a process with various process parameters of interest, in a generalized manner, through black-box modelling, which otherwise could be complex and specific to a particular equipment design if first-principles model is built. Further, to reap the benefit of ANN model’s generalization capability, to extend its application beyond modelling DDP, and to maximize its versatility and outreach, I implemented the ANN model in MS Excel. Thus, I achieved a versatile process model (named VP Model), that is, a modelling framework which may be populated, trained, tested and used as a model-based decision support tool at any plant where large amount of data is present.