Gasum Now!: Method of calculation

There are 9 stations in Finland continuously measuring the geomagnetic field (as a part of the IMAGE magnetometer network), of which the Nurmijärvi Geophysical Observatory is located close to the pipeline system. Magnetic recordings are saved at each second, but in this project we use 10 second averages. Data files are transfered to the server at regular intervals for the calculation of the electric field.

Calculation of the electric field from the magnetic field requires a model of the earth's conductivity. As a starting point, we have used recent models, of which the most recent ones are provided by the BEAR project on the Fennoscandian Shield. The earth's conductivity in Finland varies very much from place to place. However, it is lucky that concerning GIC, small-scale variations are not important. Because the Finnish pipeline covers a relatively small region, we assume a single layered earth model for the whole region, which is also computationally simple. With some experimentation, we have noticed that even a two-layer model yields reasonable results.

When we know the electric field, GIC in the pipeline is determined using an advanced model (DSTL = distributed source transmission line method), which is applicable for any buried pipeline network. Necessary pipeline input parameters are its conductivity and topology. A simplifying fact is that geomagnetic variations are slow so that GIC is nearly DC. Further, it is quite accurate to assume that the electric field is spatially uniform in the area of the pipeline system. Then GIC at each point in the pipeline is a linear function of the two horizontal components of the electric field. The DSTL method allows for using a spatially varying electric field too.

The method has been validated with comparisons to measured values since November 1998 at Mäntsälä. The median error of calculated near-real-time values is about 30%, when GIC of at least 1 A are considered. One error source is that the electric field is not exactly spatially uniform near Mäntsälä. Furthermore, the active cathodic protection system of the pipeline causes some bias to the measurements too. A detailed description of the GIC phenomenon and its modelling is given by Antti Pulkkinen's PhD thesis, whose introduction is publicly available as a PDF file on the e-thesis server of the University of Helsinki.

Example

Comparison of modelled and measured values

Error distribution in November 1998 to May 2003:

Examples of single events