TY - JOUR
T1 - A geoelectrical pre-study of Älvkarleby test embankment dam
T2 - 3D forward modelling and effects of structural constraints on the 3D inversion model of zoned embankment dams
AU - Norooz, Reyhaneh
AU - Olsson, Per-Ivar
AU - Dahlin, Torleif
AU - Günther, Thomas
AU - Bernstone, Christian
PY - 2021
Y1 - 2021
N2 - Electrical resistivity tomography has potential as a complementary long-term monitoring method in embankment dams; however, the 3D character of the geometry including the shape of the embankment, its internal zoned construction and the reservoir water make interpretation challenging. To tackle this problem, a qualified inversion model considering the 3D environment is necessary. In this paper, prior information about the resistivity of different parts of a test embankment dam was used as constraints in order to increase the capability of defect detection in a complex 3D context. Five small defects were incorporated into the core of the model. Laboratory measurements were made on samples of the materials intended to be used for the construction of a test embankment dam, and resistivity values provided from the laboratory measurements were used in the forward modelling. A measurement sequence of around 8000 synthetic data points using extended gradient, crossline bipole-bipole and corner arrays between horizontal-horizontal, vertical-vertical, and vertical-horizontal lines were modelled and inverted all at once. The structural constraints were applied to increase the accuracy of inversion, using the L1 and L2 methods. Different mesh qualities with different boundaries for each region and 3D geometric factor calculation were applied for the inversion to evaluate the effects of region control incorporated in the inversion process. The results showed that L1 and L2 norm inversions combined with region control can determine the location of very small defects and finding the defects located near each other, respectively. Removing the region control from the inversion caused unrealistic resistivity prediction for some regions and the inability to discover the dam defects. Therefore, the proposed methodology can decrease non-uniqueness in the inversion and make time-lapse ERT a valuable monitoring tool that complements other instrumentation techniques and based on these results it was concluded promising to continue with the construction of the test dam using the same type of defects and electrode set-up, for verification under field-conditions.
AB - Electrical resistivity tomography has potential as a complementary long-term monitoring method in embankment dams; however, the 3D character of the geometry including the shape of the embankment, its internal zoned construction and the reservoir water make interpretation challenging. To tackle this problem, a qualified inversion model considering the 3D environment is necessary. In this paper, prior information about the resistivity of different parts of a test embankment dam was used as constraints in order to increase the capability of defect detection in a complex 3D context. Five small defects were incorporated into the core of the model. Laboratory measurements were made on samples of the materials intended to be used for the construction of a test embankment dam, and resistivity values provided from the laboratory measurements were used in the forward modelling. A measurement sequence of around 8000 synthetic data points using extended gradient, crossline bipole-bipole and corner arrays between horizontal-horizontal, vertical-vertical, and vertical-horizontal lines were modelled and inverted all at once. The structural constraints were applied to increase the accuracy of inversion, using the L1 and L2 methods. Different mesh qualities with different boundaries for each region and 3D geometric factor calculation were applied for the inversion to evaluate the effects of region control incorporated in the inversion process. The results showed that L1 and L2 norm inversions combined with region control can determine the location of very small defects and finding the defects located near each other, respectively. Removing the region control from the inversion caused unrealistic resistivity prediction for some regions and the inability to discover the dam defects. Therefore, the proposed methodology can decrease non-uniqueness in the inversion and make time-lapse ERT a valuable monitoring tool that complements other instrumentation techniques and based on these results it was concluded promising to continue with the construction of the test dam using the same type of defects and electrode set-up, for verification under field-conditions.
KW - Electrical resistivity tomography
KW - Geoelectrical monitoring
KW - Zoned embankment dam
KW - Finite element method
KW - Internal erosion
KW - Structural constraints
U2 - 10.1016/j.jappgeo.2021.104355
DO - 10.1016/j.jappgeo.2021.104355
M3 - Article
SN - 0926-9851
VL - 191
JO - Journal of Applied Geophysics
JF - Journal of Applied Geophysics
M1 - 104355
ER -