Natural Gas Combustion Modeling using Artificial Neural Networks- A Case Study of Bidboland Gas Refinery

There is a variety of pollutants, Due to the combustion of natural gas-fired gas.A number of pollutants in the combustion of natural gas, depend on the quality and composition of the gas components. Major changes in the composition of natural gas causes changes in furnace performance and emissions levels.The analysis of Refinery production of gas composition changes on furnace combustion products Combustion is required to solve complex relationships Increased error rate at the time it analyzes and also makes it more time. This cause the use of modeling methods that have predictive power and high accuracy are required. the goal of the paper is to investigate the effect of gas composition on combustion products. In this  article, using experimental data Bidboland  gas refinery, Changes in composition and Convection of refinery gas-fired combustion products of a refinery was studied in the period of time And model by  neural network. So increasing the amount of 5% methane, NOX, increases 39 percent, the amount of CO increases , 28% and CO2 increase 15 percent. Increasing 30% ethane, NOX pollutant, increased 23 percent, CO2 , increases 8%, CO got 2\4 time more. Increasing 30% propane , CO  increase 21%in the amount , CO2 has increased seven percent and NOX increases 39%. Generally , increasing 5% in hydrocarbon gas, the carbon content is increased to an average of 11 percent. Increasing 40% H2S , amount of SO2, increase on average 1\44 times. Increasing 30% N2 , amount of NOX ,  increase 19%.  Keywords : Combustion of natural gas , Gas compounds , Combustion products , furnace, Artificial Neural Networks, Bidboland  gas refinery, compositions change