Workreport 2011-88

Semi-Automated Fracture Classification Procedure Based on Geophysical Drillhole Logging Data

Heikkinen, E., Heinonen, S. & Ravimo, I.

English

120

Posiva Oy is responsible for preparation of final disposal of spent nuclear fuel in Olkiluoto. The knowledge about existing network of fractures is important for the safety and feasibility of the final repository. The bedrock properties essential for safety case are analysed in investigations of Rock Suitability Criteria (RSC). One subtask in RSC is avoidance of large (long) fractures adjacent to disposal holes. The long fractures have been defined in tunnel mapping to indicate tunnel cross-cutting features (TCF) or full perimeter intersections (FPI).

This work has assessed possibilities to detect and classify the significant fractures from the geophysical data of a pilot drillhole, drilled on the centre line of the tunnel. The data from the pilot drillholes ONK-PH08…PH14 and ONK-PH16 – PH17 were used. Anomaly value and corresponding local background value was picked from logging profile at the fracture location. Picking used method specific windows. Also average and standard deviation was picked, and a gradient was computed. Standard deviation was used as threshold value in selection, and gradient in rejection of anomalies caused by other reasons than a fracture.

Anomaly values have been normalised within each method at range 0 – 100% and the contributions of several logging methods was added together. The applicability of different methods and their combinations in detection and classification was analysed. The match between anomalies and tunnel mapped long fractures (TCF) projected to core intersection, have been examined in the analysis.

Considering the fractures reported from core, some of these produce geophysical anomalies, and many others will not. A part of the logging methods gives a weak response for majority of fractures, but may be prominent for few fractures (like in density). The results based on electrical conductivity and full waveform sonic data match the examined TCF fractures fairly well. The sum of anomalies is describing the significant fracture better than the index from single methods. Anomalies can be found at fractures which are not significant in core or tunnel wall observation.

The TCF fracture locations show in several cases an indication in summary geophysical anomaly. Often these indications are found at small groups of several fractures and even at brittle fault indications. All of the fractures at narrow zone will gain high scores in these cases. At some of TCF fractures there is no anomaly at any of the methods.

The developed anomaly picking method which is based on windowing, threshold values and normalisation, can be performed immediately after the drill core logging fracture locations are available and geophysical data are depth adjusted. The technique will allow all fractures of the investigation target to be compared on basis of the results. Method is giving tentative information on targets where are likely to exist significant fractures, serving as supplementary information for geological logging.

Spent nuclear fuel disposal, geophysical drillhole logging, fracture mapping, rock suitability criteria, classification, large fractures.

Semi-Automated Fracture Classification Procedure Based on Geophysical Drillhole Logging Data (pdf) (3 MB)