Toward an intelligent approach for H2S content and vapor pressure of sour condensate of south pars natural gas processing plant

In this study, artificial neural network is employed to develop a model to predict process output variables of an industrial condensate stabilization plant. The developed model is evaluated by process operating data of south pars natural gas processing plant located Asaluyeh/Iran. A large dataset of 4 variables consisting of temperature and pressure of the stabilization column in addition to Ried Vapor Pressure (RVP) and H2S content of the processed condensate is utilized to train the network. In order to determine the optimized topology and decision parameters of the network, the values of Mean Square Error (MSE), Mean Absolute Error (MAE) and the coefficient of determination (R2) are minimized by the method of trial and error. Since precision of ANN model is dependent on the amount of training data used, the extensive set of samples applied in this work can offer accurate reliable predictions. Model output is compared to actual data of the plant and the values of Average Absolute Deviation percent (ADD%) are reported as 1.6 for RVP and 3.8 for H2S concentration.