SIMULATION OF UNSTEADY STATE FLOW OF NATURAL GAS IN PIPELINES USING FINITE VOLUME METHOD IN 2D CYLINDRICAL Coordinates

ABSTRACT             

Transient compressible natural gas flow through a pipeline was studied by the use of a finite volume method in 2D axisymmetric cylindrical coordinates. To account for turbulence within the pipeline system, the standard turbulence model was simulated together with the Navier Stokes System of equations via the Reynolds-Averaged method. The equation of state employed was the SoaveRedlich-Kwong equation. Implicit discretization was used for the temporal terms, whereas the central differencing scheme and the upwind differencing scheme were used in the discretization of the spatial diffusion terms and the spatial convection terms respectively. The Pressure Implicit with Splitting of Operators (PISO) algorithm was then used for calculating the pressure and velocities on a staggered grid. Computer simulation was carried out to determine variations in pressure, density, velocity and temperature within the pipeline system. Profiles for turbulence viscosity, turbulence kinetic energy and turbulence eddy dissipation along the pipeline were also obtained. To validate the model, data obtained from the following companies were used – Shell Petroleum Development Company, Port-Harcourt, Nigeria, Eroton Production and Exploration Company Limited, Nigeria and published data from the National Iranian Gas Company. Pressure validation was carried out with output pressure values obtained from the stated companies. Parametric analyses were also carried out in the work. This entailed ascertaining the effect of varied inlet temperature on some gas flow parameters. The temperature range considered was 290K – 330K with a temperature difference of 10K. In addition, a rupture that was assumed to have occurred at the middle of the pipe (18,000m long) was also analysed to determine pressure profile. For the purpose of comparison, simulations were carried out with the turbulence model and the – turbulence model using the Shell data. The shapes of the profiles obtained from the results were in agreement with others obtained in validated literature results. Grid independence was also investigated. It was discovered that grid independence occurred at a value of 182,000 cells. This means that there was no change in the results obtained after the use of 182,000 cells. Output pressure values from the industries were: 3.19917MPa, 5.55MPa and 6MPa respectively for Shell, Eroton and the Iranian Companies respectively, while the simulated values were 3.153MPa, 5.5124MPa and 6.016MPa for the Shell, Eroton and Iranian Companies respectively. Percentage error between the Industry and simulated values gave, 1.44% for Shell data, 0.27% for Eroton data and 0.77% for National Iranian Gas Company data. These results prove accuracy of the steady state model. Percentage error for transient validation performed with the Iranian data was 0.3067%. There was also no significant difference between the pressure profiles obtained from the two turbulence models. The present method, therefore, can become an invaluable tool and template for use by the oil and gas industry for mitigating the consequences of pipeline perturbations such as vandalism, explosions and ruptures by its application in leak detection systems.