In city gate stations (CGS) experimental facilities are currently employed to specify natural gas compositions and subsequently calculate its density. Then, natural gas mass flow rate is calculated by employing an Ultrasonic Flow Meter (UFM), and a volume corrector to convert the given value to its equivalent in standard conditions. As online measurement of natural gas compositions is a costly and difficult task, this study presents an innovative method for natural gas mass flow metering in CGSs. In this method, a novel correlation, which is a functional of only three simple measurable variables i.e. temperature, pressure and sound speed, is presented for calculating natural gas molecular weight, and then, employing the authentic equations of state (EOS) of AGA8 for natural gas, its density could be calculated. The correlation was developed based on curve fitting and data mining approaches on a large database associated with four different natural gas fields of Iran and its accuracy was validated with available experimental data for seven other Iran's natural gas fields and two more gases with sample compositions. Each database in correlation development and validation stage (13 databases altogether) consists of 17,000 sound speed values in all possible temperature and pressure ranges in CGSs after expansion process. Mean absolute error (MAE) and mean absolute percentage error (MAPE) methods are used to validate the correlation accuracy and compare its performance with the correlation presented previously for the same objective. The evaluation results prove high prediction accuracy for the presented correlation and its superiority comparing to the previous work.

Keywords