Forecasting of Hydrate Formation Pressure for Natural Gas Using Artificial Neural Network
Advancing the chemical engineering fundamentals
Thermodynamics (T2-1P)
Keywords: Pressure Forecasting, Artificial Neural Network, Hydrate, Natural gas
Knowledge and prediction of natural gas hydrate formation conditions is one of the most important problems in oil and gas engineering. Although some mathematical models are available to predict temperature of hydrate formation, but there is no model for pressure. In this paper the Artificial Neural Network (ANN) technique have been applied for estimation of hydrate formation pressure. The neural network training and test data are 149 and 18 respectively and data are in the range of 32-74 Fo for temperature, 50-4200 Pisa for pressure and 0.554-1 for specific gravity. The standard feed forward back propagation algorithm is used for training neural networks. The sub optimal neural network structure is 10-6-1 which is obtained using trail and error method. In comparison of performance analysis of ANN, the relative error (RE) was studied and maximum error is obtained 3.051 % and R-Square value is 0.9988. The ANN structure for pressure prediction is greater than temperature (7-5-1) and it’s R-Square is better than temperature (0.9941). Thus it can be concluded that the proposed approach provides a good method for prediction of pressure and temperature of hydrates formation conditions.
Presented Monday 17, 13:30 to 15:00, in session Thermodynamics (T2-1P).