Salt body geophysics III

BMBF funds the joint project “Universal integrated geophysical measuring and evaluation instruments to characterise problem zones in salt bodies (salt body geophysics III)” of which the subprojects - electromagnetics, georadar and the quantitative characterisation of problem zones - are undertaken by the GGA institute. The geoelectric logging side of the project is the responsibility of the University of Leipzig, seismic is handled by KUTEC Sondershausen, and sonar by the Fraunhofer Institute for Non-destructive Testing in Dresden.

In the two preceding projects, work led to the preliminary development of a complex measurement and evaluation toolkit for detecting and characterising problem zones in the geological barriers of geologic repositories and underground disposal sites. The current project focuses on turning the previously qualitative, i.e. primarily visual evaluation of the results of each method, into a means of characterising problem zones in salt bodies quantitatively. It is hoped that this approach will reveal phenomenon patterns from the data generated by the various measuring techniques, with the ultimate aim of identifying, evaluating and dealing with geological problem zones, particularly brine occurrences in salt rocks.

This involved the use of geoelectrics, georadar, electromagnetics, seismic and sonar at the new test sites in Sollstedt and Teutschenthal.

The EM logs cannot be used to derive any geologically relevant conclusions because data from both of the test sites was strongly distorted by metallic objects within the underground workings. The EM logs will therefore not be used in the combined evaluation.

Models indicating the presence of possible problem zones at some localities were elaborated for both test sites on the basis of the georadar results. Geological boundary horizons and possible zones of deconsolidation were also interpreted from the data. Figure 1 shows a change in reflection properties in the immediate proximity of a brine seep.

Commercial GSSI antennae with a range of frequencies were mainly used for the georadar surveys. Figure 2 shows the survey results at several frequencies from the Sollstedt test site.

In addition, specially made resistance-suppressed dipoles, “Wu-King antennae” were used. These give shorter separate signals and thus higher resolutions.

The results from the geoelectric, seismic, sonar and geological work carried out in parallel were combined with the georadar results. The multivariant statistical learning methods SVM (support vector machine) and ANN (artificial neuronal network) were used on all of the geophysical survey results with the aim of quantitatively characterising problem zones (especially brine occurrences in salt rock). This produced better results than the previously used univariant methods. Overall, pattern recognition methods improve the identification of anomalies which could either not be clearly interpreted from the results of any one method, or were only hinted at (cf. Fig. 3). The data from the separate surveys (overlapping area approx. 150 m) reveals a few possible but not unequivocally interpretable anomalies. At the - 50 m line co-ordinate, there were indications of the presence of an anomaly, but this only became clearer by integrating all of the data.