MATHEMATICAL MODELING AND OPTIMIZATION OF REFINERY HYDROGEN NETWORK

Abstract

In the modern day refineries, there is a burgeoning need of hydrogen because of heavy crudes and the advent of draconian environmental norms. This has led to an increased demand of clean fuels having low sulphur content. In this backdrop, hydrogen management in refinery presents itself as an effective solution to maximize the profits accrued by the refinery along with maintaining the environment regulations. The present investigation deals with the prediction of an optimum network to manage hydrogen in refinery based on the models developed. Five models namely Model-0, Model-1, Model-2, Model-3 and Model-4 have been developed in the present work. Model-0 and Model-1 are NLP models while the remaining three are MINLP models, which takes into account both binary variables and continuous variables. The objective function of all the models is the total operating cost, which has to be minimized. Model-0 is a base case optimization with an aim to validate the existing network. Model-1 considers changes in piping. It also considers the effect of recycle and pressure constraints. The consideration of recycle will reduce the requirement of fresh hydrogen, as the recycle stream will, to a certain extent, aid in fulfilling the demand of hydrogen in the network. According to pressure constraints, hydrogen flows from a source to sink if the pressure of the former is greater or equal to the pressure of the latter. Model-2 is MINLP model which considers changes in piping and pressure constraints. Model-3 is developed considering the addition of new compressor in Model-2. Model-4 is MINLP model accounting the incorporation of new compressor and new purifier in the existing network. The constraints common to all models are source flow rate balances, sink flow rate balances, sink purity constraint, compressor flow rate balance, compressor hydrogen flow rate balance and compressor capacity limit. The present investigation considers different capital cost functions for compressors taken from Dekker, 2003. Based on the Total Annual Cost (TAC), the screw compressor has been chosen as the most economical compressor. The present work also takes into account piping cost and export cost of unused hydrogen to calculate the TAC.