On steady state modelling for optimization of natural gas pipeline networks

Current approaches to the fuel cost minimization problem (FCMP) emphasize the tractability of the pipeline model but compromise its validity. In this paper, the model simplifications and assumptions commonly used in the FCMP research are assessed quantitatively. Firstly, a rigorous pipeline model is developed. Secondly, outputs from the rigorous model are compared to outputs from simpler models used in the literature. It is found that there is a significant prediction error in these simpler models. The pipeline model is also extended to address a pipeline network, for which the mixing effect at the pipeline junctions is addressed. Lastly, a FCMP problem is solved for a pipeline network, and the obtained mass flow rates are used to assess the performance of different models. It is found that consideration of temperature variation and rigorous calculation of critical physical quantities are crucial for a more accurate model that better represents reality.