Using Time Series Analysis in the Evaluation of Measurement Uncertainty in Gas Metering

Dynamic effects in gas grids influence the measurement of the quantity and energy supplied to and withdrawn from the grid. These dynamics arise in gas grids due to the supply and withdrawal of gas from the grid. The magnitude of these effects increase due to the injection of, e.g., biomethane and hydrogen. Said dynamic effects add to measurement uncertainty associated with the quantities used in the metering of gas. In this work, it is argued that generally, the measurements in gas grids can be modelled as time series. The data in such time series may or may not show autocorrelation, which can be assessed by time series analysis. In this work, it is shown how the outputs from a time series analysis can be combined with the uncertainty budgets of the measured quantities to comprehensively describe the dependencies between the measurement results used for obtaining the mass, volume or energy supplied to or withdrawn from the grid. Enriching the measurement model used for propagating the measurement uncertainty is the key to integrating the various uncertainty contributions. For temperature and volume measurement, it is demonstrated how measurement models can be constructed to deal with autocorrelation and instrumental uncertainty. From the model development, it follows that the dependencies between the data in the time series have a different cause than the dependencies arising from, e.g. using the same instrumentation, so they should be combined rather than perceived as double counting.