LRRT

 

Home

FRACTURE MAPPING AND ANALYSIS FOR TUNNEL CONSTRUCTION

 

When carrying out fracture mapping on a 3-D model of a tunnel round, the 3-D model must:

 

  • Include face, walls and crown. However, all too often only the face is modelled, and this may result in severely underestimation of the fracture sets, as explained below.
  • Correctly represent the entire wall and crown geometry, otherwise fractures will be artificially obliterated.
  • Exactly attribute pixels to the underlying 3-D geometry, otherwise fracture traces will generate erroneous fracture orientations. This can never be accomplished with a Laser scanner (LiDAR), where there is always an offset between pictures and point cloud. Tonon USA, on the other hand, creates the point cloud from the pixels, hence the exact pixel-geometry overlap.
  • Correctly represent the rock color in order to appreciate the rock types, intrusions, veins, water seepage, etc...

Tonon USA developed special techniques to achieve all of these results without infringing into the construction schedule.

The figures above illustrate an example of a three-dimensional model of the entire round, and of the face, walls, and crown textured with high resolution pictures (click on a figure to enlarge). Notice that the models of the walls correctly represent the tunnel geometry all the way to the tunnel face. Fracture traces and planes were digitized on this model: colors refer to the different fracture sets identified in the stereonet (click to enlarge), where a white color indicates fractures that could not be assigned to any fracture set. In a 6-m wide tunnel, 260-280 fractures were digitized per round, only 40 of which were digitized at the face. This is due to the different way in which the rock breaks at the walls (where smooth blasting is used) and at the face (where smooth blasting can not be used and fractures are smeared). In many cases, were the survey be restricted to the face, important fracture sets not parallel to the face would not have been identified at all.

The number of measured fractures per round is in line with the ISRM (International Society for Rock Mechanics) recommendations for a reliable statistical characterization of fractures on an outcrop that would then lead to a rational rock engineering design (ISRM (1978). "Suggested methods for the quantitative description of discontinuities in rock masses". Int. Journal Rock Mechanics, Mining Sciences & Geomechanical Abstr. 15: 319–368). It would be very difficult to acquire such a large amount of information by using traditional means (geologist that moves about on a cherry picker to take measurements with a compass) because this operation would require several hours of work at the face under unprotected ground, and would unacceptably impact advance rates. In many countries, it is even forbidden to step closer than 2 rounds from the tunnel face: in these cases, traditional fracture mapping is simply impossible.

In the following, some uses of the obtained information are exemplified; although not all of them may be of interest in all cases, the objective here is to highlight possible uses of the three-dimensional model textured with high resolution pictures.

Interaction between fractures and blasting

Fractures in the orange set in the figures above appear to be much more frequent at the tip of the crown blastholes: when blasting the i-th round, the waves generated by the blast encounter the free surface of the (i-1)-th round and propagate the excavation along the orange fractures, thus creating over excavation along the tunnel axis. It is highly unlikely that this detail and its consequences be documented, understood, and defensibly proved without the use of a 3-D model textured with high resolution pictures.

Fracture clustering and intensity

Here is a typical example of a fracture cluster: the scale is about 1 m. The trace-digitized fractures are in red, the plane-digitized fractures are in cyan.

Typically, water ingress concentrates at fracture clusters. By documenting these clusters one knows exactly where these clusters were encountered during construction, e.g., if in the future an anomalous water ingress is observed through the lining. Cluster characterization (including geostatistics and Dershowitz’s PIJ system) ,allows one to refine the fracture modeling (e.g., by using stochastic, geostatistical, and fractal models) prepared at the design stage. Such an improved model can then be used for groundwater inflow, fluid flow, or contaminant migration predictions. Indeed, any type of scan line may be traced on the rock walls, including circular scanlines that provide unbiased data.

Fracture coalescence

Here is a typical case where two fractures coalesced to form a stepped fracture, which may explain in part the orientation variability within a fracture set, and limits the reliability of rock block-size prediction based on natural fracture lengths. Indeed these detailed three-dimensional models revealed that oftentimes fallen blocks were formed by fractures (in the same set) that coalesced because of the blasting and of the stress redistribution induced by the excavation.

Accuracy of fracture traces

This example shows a typical case in which two fractures in the same fracture set were digitized either by using a plane or a trace. The resulting generated planes have the same orientation.

 

This is a consequence of the fact that Tonon USA exactly attributes pixels to the three-dimensional surface. As a result, the lengths of the fracture traces may be accurately determined; Tonon USA has developed techniques to obtain the fracture size distribution from the fracture trace probability distribution (F. Tonon, and Chen, S. Closed-form and numerical solutions for the probability distribution function of fracture diameters. Int. J. of Rock Mech. and Mining Sci. 44 (3), 2007, 332-350. DOI: 10.1016/j.ijrmms.2006.07.013). Finally, the ability to rotate and zoom the three-dimensional model textured with high resolution pictures allows one to easily identify relationships between structures and/or fractures that otherwise would be difficult to identify in situ, especially in the short time allowed to the geologist at the face

 

2028 E Ben White BLVD #240-2660• Austin, TX 78741

 

Phone: +1-512-200-3051

 

E-mail: info@tononeng.com