1) Overview of the MIU-3 investigations^41),82)

(1) Concept of the investigations

With the main aim of verifying the validity of above-mentioned rock mechanics conceptual model and constructing the conceptual model of the rock mass on the footwall side of the Tsukiyoshi Fault, sampling locations for physical/mechanical property tests and depths to measure the initial stresses are set in the MIU-2.

(2) Contents of the investigations

Based on the above-mentioned concept, test positions are set to facilitate obtaining the data of mechanical properties mainly on the footwall side of the Tsukiyoshi Fault. From the hanging wall of the Tsukiyoshi Fault, a few samples in the individual structural zones set in the rock mechanics conceptual model (See Chapter 4.4.3) are planned to be obtained. From the footwall of the Tsukiyoshi Fault, as many samples as possible at an equal intervals (some 100m) are planned to be obtained. The MIU-3 crosses the Tsukiyoshi Fault at 693.2709m in depth. Details of physical/mechanical property tests, and measurement depths of initial stress measurements are shown in Tabs.4.34 and 4.35, respectively.

(3) Results of the investigations

Results of physical/mechanical property tests in the MIU-3 are shown in Tab.4.36 and Fig.4.61. Vertical initial stresses obtained by AE/DRA tests of cores from the MIU-3 and initial stress states on horizontal planes obtained by hydraulic fracturing tests are shown in Figs.4.62 and 4.63, respectively. However, hydraulic fracturing tests carried out at two depths (338m and 462m) in the MIU-3 produced no longitudinal fracture to satisfy a theoretical assumption. Therefore, these results are judged too unreliable to be used for assessment of stress states and to be shown in the figure.

	Item	Unit
Physical property	Apparent density	-	2.612.65
	Effective porosity	%	1.21.7
	Water ratio	%	0.450.65
	Seismic wave velocity (Vp)	km/sec	5.46.0
	Seismic wave velocity (Vs)	km/sec	3.03.5
Mechanical property	Uniaxial compression test	MPa	50200
	Young's modulus (E50)	GPa	2062
	Poisson's ratio	-	0.300.40
	Tensile strength (by Brazilian test)	MPa	38
	Cohesion	MPa	2050
	Internal friction angle	Degree	4860

(4) Evaluation of the results

It is known from the existing investigation results that Tsukiyoshi Fault is associated with some 100m-wide fractured zone on its both sides. Accordingly, it is assumed that 100m-wide areas on the both sides of the main part of the Tsukiyoshi Fault (693.2719.3m in depth) are affected by the fault. Comprehensive examinations on the distribution of physical/mechanical properties and this assumption suggest that the granite at the MIU-3 could be divided into three zones with different variation trends along the depth. That is, approximately 0300m, 300600m and 600800m in depth. Each of them has the following characteristics.

Zone 1: 0300m in depth
Despite a small amount of tests, water ratio and effective porosity tend to slightly increase with the depth. No change is found in apparent density and sesmic wave velocity. Young's modulus, uniaxial compressive strength and Poisson's ratio tend to slightly increase with the depth, whereas tensile strength (by Brazilian test) decreases with it.

Zone 2: 300600m in depth
Apparent density tends to increase with the depth, whereas efficient porosity and water ratio tend to decrease with it. Seismic wave velocity shows no clear change with the depth. Little change is recognized in Young's modulus, uniaxial compressive strength and Poisson's ratio, whereas only tensile strength (by Brazilian test) shows a slight increase with the depth.

Zone 3: 600800m in depth
Individual physical properties discontinuously vary compared with those in zones 1 and 2. As a whole, effective porosity and water ratio increase and seismic wave velocity decreases with the depth, but the variations are small. Values of mechanical properties show remarkable changes with the depth. Young's modulus, uniaxial compressive strength and Poisson's ratio increase, whereas tensile strength (by Brazalian test) decreases with the depth.

The results dispersion of AE tests is smaller than that of DRA tests at the same depths. Therefore, the distribution of vertical initial stresses is assessed mainly by the results of AE tests. Values of vertical stresses obtained by AE tests are a little larger than the overburden pressures estimated in the hanging wall of the Tsukiyoshi Fault but smaller than those in the footwall of the fault. They vary abruptly around the Tsukiyoshi Fault.

Initial stress states in horizontal planes obtained by hydraulic fracturing tests tend to increase with the depth. The values of the maximum stress drop at about 600m and below 700m in depth in the footwall of the Tsukiyoshi Fault. Assuming that vertical stress is equal to overburden pressure, 3-D stress states of reverse-fault-type (_H>_h>_v), lateral-fault-type (_H>_v>_h) and normal-fault-type (_v>_H>_h) are expected at 0�`550m, about 600m, and deeper than 700m in depth, respectively. The maximum principal stress trends N-S at about 100m in depth and NNW-SSE at deeper than 300m. Thus, the rock mass around the MIU-3 can be divided into three sections : 0550m, around 600m and deeper than 700m in depth.

Vertical variations in the number of fractures in the MIU-3 are shown in Fig.4.64. While they show a trend similar to the results obtained in the MIU-1 and MIU-2, the vertical contrast in the number of fractures is not as clear. However, the vertical distribution of the fracture numbers change abruptly at the similar depths to those of physical properties' changes in the MIU-3. These results are conformable with the structure predicted by the rock mechanics conceptual model based on the investigation results of the MIU-2.

2) Comparison between the results of the MIU-3 investigations and the rock mechanics conceptual model (See Chapter 4.4.3)

The mean values of results obtained by physical/mechanical property tests carried out in the AN-1 and MIU-13 are shown in Tab.4.37.

Average values of physical properties in the MIU-3 are close to the values of the MIU-1. Specifically, physical properties, such as apparent density, effective porosity, water ratio and seismic wave velocity, are all larger than those of the AN-1 and MIU-2. Mechanical properties, such as Young's modulus, uniaxial compressive strength, and tensile strength (by Brazilian test), are the smallest among the four boreholes. Values of cohesion, internal frictional angle and Poisson's ratio are almost the same as those of the MIU-1.

All of the vertical distributions of physical properties in the four boreholes (Fig.4.61) are uneven to greater or lesser degrees. Though variations in vertical distribution of physical properties tend to increase from the AN-1 toward the MIU-2 (from south to north), the variations in the MIU-3 are almost the same as those in the MIU-1. It indicates that there is no effect of the fault formation on the variations. Vertical distributions of mechanical properties are also uneven. The extent of variations in the MIU-2 and MIU-3 are larger than those in the AN-1 and MIU-1. While physical properties in the MIU-3 discontinuously change immediately above the Tsukiyoshi Fault, they start changing about 400m above the fault in the MIU-2. These results indicate that physical/mechanical properties of the rock mass in the Shobasama site are vertically uneven enough to allow dividing it into zones. The changes in physical/mechanical properties are probably generated by not only mechanical damages by the formation of the Tsukiyoshi Fault but other factors.

Vertical initial stresses in the hanging wall of the Tsukiyoshi Fault vary widely but are nearly equal to the estimated overburden pressures (Fig.4.62). On the other hand, they are smaller than the estimated overburden pressures in the footwall of the fault. It suggests that vertical stresses are released or there are conditions to decrease them around and below the fault.

Principal stresses on horizontal planes tend to increase with the depth (Fig.4.63) to allow dividing the rock mass down to 1,000m in depth into zones with different stress states. Based on the investigation results, the hanging wall of the Tsukiyoshi Fault is divided into three zones : 0300m, 300600/700m and deeper than 600/700m in depth. On the other hand, the footwall forms merely one zone. The azimuth of principal stress measures rotates about 45 at about 300m in depth in the AN-1, while it rotates from N-S to WNW-ESE with an increase in depth in the sections at 0400m and 400m700m in depth in the MIU-2. Here, the maximum compressive axis of the regional stress field in the Shobasama site also trends WNW-ESE. Therefore, the maximum principal stress is thought to generally trend WNW-ESE at 3001,000m in depth in the Shobasama site.

3) Revision of the previous rock mechanics conceptual model⁸³⁾

The validity of the previous rock mechanics conceptual model of the hanging wall of the Tsukiyoshi Fault is verified by the investigation results in the MIU-3. Also, following the acquisition of the mechanical data on the footwall side in the MIU-3, the rock mechanics conceptual model of the footwall is examined. However, the rock mechanical data of the rock mass on the footwall side is restricted to the depth exceeding 900m in the MIU-2 and 700m in the MIU-3. To make up this information shortage, the rock mechanics conceptual model of the footwall is examined using the data on the basement granite obtained by rock mechanical investigations^{83), 84)} in the Tono Mine and the measurement results⁶³⁾ in the DH-9 for the RHS Project⁸⁵⁾, in addition to the above-mentioned data. The DH-9 is a 1,000m-class borehole drilled about 1km north of the northern border of the Shobasama site and the whole hole is within the footwall of the Tsukiyoshi Fault. Results of mechanical tests of the rocks on the footwall side in the MIU-3 and the 99SE-02 (200m-class) are shown in Fig.4.66⁸⁴⁾. Though the about 50m-thick upper part of the basement granite in the 99SE-02 is too intensely weathered to collect cores, values of mechanical properties below the weathered part tend to increase with the depth. Accordingly, none of the mechanical properties obtained in the borehole probably represent the sound granite with the exception of the data in the deepest part. Rocks at the lowermost part of the borehole (207m in depth) have rock mechanical properties, such as apparent density of 2.62t/m³, E₅₀ of 50GPa, and Uniaxial compressive strength of about 150MPa. They are almost the same as the physical/mechanical properties on the footwall side in the MIU-3 and similar to the mean values obtained on the hanging wall side.

The information on the stress states on the footwall side of the fault is provided by hydraulic fracturing tests carried out deeper than 900m in the MIU-2, at deeper than 700m in the MIU-3, and in four boreholes drilled in the Tono Mine (TM-1, TM-2, 98SE-01 and 99SE-02). The measurement results are shown in Fig.4.67. The four boreholes in the Tono Mine cross the Tsukiyoshi Fault and the basement granite at different depths. Therefore, the measurement results of the individual boreholes are arranged from south (hanging wall side) to north (footwall side). Values of principal stresses on horizontal planes measured in the basement granite tend to clearly decrease from the hanging wall side to the footwall side. The minimum principal stresses are nearly equal to the overburden pressures in the northernmost 99SE-02 and the maximum principal stresses are 1.41.7 times the values of the minimum principal stresses. On the other hand, stresses on the footwall side in the MIU-3 are much lower than those on the hanging wall side. The maximum principal stresses in the Tono Mine trend NNW-SSE to NW-SE except near the rock boundary.

The investigation results⁶³⁾ in the DH-9 drilled for the RHS Project is used for geological structure investigations of the rock mass on the footwall side of the fault. Results of BTV investigations, seismic wave velocity loggings and density loggings in the individual boreholes (MIU-1, MIU-3 and DH-9) are shown in Fig.4.68. This figure indicates that the number of fractures in the DH-9 is extremely smaller than those in the rock mass on the hanging wall side in the MIU-2 and MIU-3. Seismic wave velocity and density drastically drop in fracture-concentrated parts in the MIU-2, MIU-3 and DH-9. As for the rest of the rock mass, seismic wave velocity changes a little where the fracture numbers abruptly change. Sections with the more fractures show the larger extent of velocity variation, and vice versa. There is little change in seismic wave velocity in the DH-9 except for the fracture-concentrated parts. Results of density loggings seem to show a similar trend though their trend is not as clear as the results of seismic wave velocity loggings.

Based on the investigation results of the MIU-3, the RHS Project and the studies in the Tono Mine, the rock mechanics conceptual model is revised. The revised models for the hanging wall and the footwall of the Tsukiyoshi Fault are shown in Figs.4.69(a) and (b), respectively. Furthermore, the following hypotheses could be proposed on mechanical properties and stress states of the rock mass in the footwall side.
According to the results of laboratory tests on cores obtained from the boreholes in the Tono Mine and the MIU-3, it is extremely improbable that mechanical properties of the matrix part in the "Moderately fracture zone" differ a great deal from those on the hanging wall side.
According to the results of initial stress measurement by hydraulic fracturing tests in the Tono Mine and on the footwall side in the MIU-3, stress in the footwall side is thought to be considerably lower than that in the hanging wall side. This is supported by the fact that there are a small number of fractures but a highly weathered fracture zone develops at deeper than 600m in d epth. In general, stress states are closely related to the opening width of fractures. It is presumable that a small number of fractures which develop under weak stresses in the footwall causes larger opening widths to facilitate infiltration of groundwater near the ground surface. Also, in the footwall side in the DH-9, quite a few discontinuities of vertical stress variations are expected to occur taking the development of considerably thick and fracture-concentrated zones into consideration. However, the extent of the discontinuities would not be as remarkable as in the hanging wall due to the weak stresses. The measurement data in the boreholes in the Tono Mine and the MIU-3 suggest that the maximum principal stresses in horizontal planes generally trend NNW-SSENW-SE like in the hanging wall of the fault though there is no data in the intermediate depth.

The rock mechanics conceptual model of the footwall of the Tsukiyoshi Fault has the following characteristics.
Physical/mechanical properties of the rock matrix in the footwall are almost the same as those in the hanging wall. Their vertical variations are not large.
Fractures are so few in the footwall that mechanical properties of the in-situ rock mass are better than those on the hanging wall side.
The minimum principal stresses are equal to the estimated overburden pressures or smaller. The difference between the maximum principal stresses and the minimum principal stresses is small. The maximum principal stresses trend WNW-ESE to NW-SE and don't vertically change a great deal.