Internship time series analysis

Internship

Time series analysis

The fiscal metering of natural gas is often performed using a flow meter and a gas chromatograph.
The flow meter is used to measure the volume (𝑉) of the gas passing the metering station and the gas
chromatograph is used to measure the composition, from which the calorific value (𝐻) is determined.
The energy is calculated as 𝐸 = ΣNₜ=1 𝐻ₜVₜ
The uncertainty of the energy delivered (or received) is
an important parameter for operating the gas grid, ensuring that suppliers can feed the grid and users
can draw gas from the grid. Current methods to estimate the uncertainty associated to the energy do
not take into account temporal correlations among subsequent measurements. The aim of this
internship is to validate the proposed methodology to determine temporal correlations between
measurements using time series analysis. Specifically, the following steps will be performed:

1. Study relevant literature
2. Generate a synthetic dataset reproducing some of the characteristics typical in volume flow and composition measurements in a gas grid
3. Validate and improve the proposed methodology on the synthetic dataset
4. Further analyze the available measurement data