1) Overview of the MIU-2 investigations^40),78)

(1) Concept of the investigations

With the main aim of verifying the validity of the above-mentioned rock mechanics conceptual model, amounts and positions of tests are set carried out in the MIU-2.

(2) Contents of the investigations

Tests for physical/mechanical properties, AE/DRA tests and hydraulic fracturing tests are originally planned to be carried out at intervals of 100m, 100m and 50m, respectively, in the MIU-2. These purpose are understanding the distribution of physical properties in the above-mentioned three zones and verifying the investigation results obtained in the AN-1 and MIU-1. However, the MIU-2 crosses the Tsukiyoshi Fault, of which the footwall has too many fractures to undergo hydraulic fracturing tests. Thus, the hydraulic fracturing tests are carried out only in the hanging wall of the fault. Details of tests for physical/mechanical properties are shown in Tab.4.30. The depths for initial stress measurements are shown in Tab.4.31.

^* : Sampling depth for AE/DRA
^** : Middle point depth of test interval for Hydraulic fracturing

(3) Results of the investigations

Results of physical/mechanical property tests in the MIU-2 are shown in Tab.4.32 and Fig.4.61. Vertical initial stresses obtained by AE/DRA tests of cores from the MIU-2 and initial stress states on horizontal planes obtained by the hydraulic fracturing tests are shown in Figs.4.62 and 4.63, respectively. The hydraulic fracturing tests result in an echelon-type of fractures at depths of 604.0m, 651.0m, 682.0m and 698.5m⁷⁹⁾, where it is probable that one of the principal stresses is not vertical.

	Item	Unit	Result
Physical property	Apparent density	-	2.512.65
	Effective porosity	%	0.72.0
	Water ratio	%	0.240.50
	Seismic wave velocity (Vp)	km/sec	4.06.0
	Seismic wave velocity (Vs)	km/sec	2.03.0
Mechanical property	Uniaxial compression test	MPa	130240
	Young's modulus (E50)	GPa	3263
	Poisson's ratio	-	0.300.46
	Tensile strength (by Brazilian test)	MPa	410
	Cohesion	MPa	1325
	Internal friction angle	Degree	5563

(4) Evaluation of the results

Both physical/mechanical properties of the rock mass in the MIU-2 show uneven distributions like in the AN-1 and MIU-1. Comprehensive examinations on the distribution of physical/mechanical properties obtained in the MIU-2 indicate that the rock mass at the MIU-2 is divided into the following four zones with different trends of the properties: 0400m, 400600m, 600900m in depth (the depth of the Tsukiyoshi Fault) and deeper than 900m. These zones have the following characteristics.

Zone 1: ground surface400m in depth
Apparent density is small, but water ratio and effective porosity are large. Physical properties tend to vary a little with the depth. Young's modulus, uniaxial compressive strength and tensile strength (by Brazilian test) slightly increase with an increase in depth.

Zone 2: 400m600 m in depth
Variations in average physical properties with the depth are relatively small. Physical properties abruptly change around the upper and lower boundaries of this zone. All the physical properties but Poisson's ratio is considerably lower than those in the overlying and underlying zones.

Zone 3: 600m900m in depth (the Tsukiyoshi Fault)
Physical properties tend to change with an increase in depth. Apparent density and seismic wave velocity increase with the depth, whereas water ratio and effective porosity decrease with it. Young's modulus and Poisson's ratio tend to increase with the depth.

Zone 4: deeper than 900m
Physical properties discontinuously change from those in the overlying zones. While Young's modulus and tensile strength (by Brazilian test) abruptly drop, all the other properties little change.

The MIU-2 crosses the Tsukiyoshi Fault at 890 to 915m in depth. It is known that the hanging wall and footwall differ in lithofacies from each other (hanging wall: Biotite granite, footwall: Felsic granite). Nevertheless, no meaningful correlation between the facies change and the difference in physical properties is found yet. Consequently, it is presumable that a discontinuous change in physical/mechanical properties deeper than 900m is originated in damages associated with the formation of the fault.

Results of AE tests of the MIU-2 vary more widely than those of DRA tests. While sensors for AE tests are usually set in the central part of a test specimen, the sensors are set at the both ends of the specimen in this measurement. Accordingly, it is presumable that this results in the different count number of AE. Therefore, the distribution of vertical initial stresses is assessed mainly by the results of DRA tests. The values of vertical stress obtained by DRA tests are approximately equal to the estimated overburden pressure from unit weight and overburden between the ground surface and 800m in depth. However, they are a little smaller than the estimated overburden pressures below 900m in depth.

The initial stress states in horizontal planes are assessed only in the hanging wall of the fault by the hydraulic fracturing test because the hydraulic fracturing test cannot be carried out below 900m in depth. The results indicate that the stress values generally tend to increase with the depth, but the maximum principal stress drops at about 300m and 600m in depth. Assuming that vertical stress is equal to overburden pressure, 3-D stress states of reverse-fault-type (_H>_h>_v), transitional-type (_H>_hv), and lateral-fault-type (_H>_v>_h) are expected at 0300m, 300600m and 600900m in depth, respectively. The direction of the maximum principal stress tends to rotate in the three sections ranging in depth from 0400m, 400700m and 700900m. Specifically, the azimuth of the maximum principal stress rotates about 60 from N-S to WNW-ESE in the two sections ranging in depth 0400m and 400700m. Furthermore, it rotates from N-S to WNW-ESE and returns to N-S again in the section between 700900m in depth.

Accordingly, the initial stress states in the hanging wall of the Tsukiyoshi Fault around the MIU-2 are divided into three sections ranging in 0300/400m, 300/400m600/700m and 600/700m900m.

Distribution histograms of fractures counted by BTV investigations in the MIU-2 are shown in Fig.4.64. These histograms indicate that not only the fracture numbers in the MIU-2 discontinuously changes at 400m and 700m in depth but the fluctuating trend of the fracture numbers is also nearly identical with that of the MIU-1.

Thus, physical/mechanical properties, initial stress states, and fracture distributions abruptly change at 300400m, 600700m and some 900m in depth. It indicates that the hanging wall of the Tsukiyoshi Fault is composed of three zones with different physical/mechanical properties and stress states. Also, it is probable that the footwall of the Tsukiyoshi Fault has different properties from the hanging wall.

2) Comparison between the results of the MIU-2 investigations and the rock mechanics conceptual model (See Chapter 4.4.2)

The mean values of the results obtained by physical/mechanical property tests carried out in the AN-1, MIU-1 and MIU-2 are shown in Tab.4.33.

Average values of physical properties in the MIU-2 are thought to be nearly identical with the results obtained by the previous investigations. They are close to the values obtained in the AN-1 but tend to be slightly lower than those obtained in the MIU-1 as a whole. Specifically, values of physical properties, such as unit weight, effective porosity, water ratio and sesmic wave velocity are smaller than those of the MIU-1. All other value of mechanical properties but internal friction angle and Poisson's ratio of the MIU-2 are smaller than those of the MIU-1. The variation of physical properties with an increase in depth tends to increase from the AN-1 to the MIU-2 (from south to north). Though the vertical distribution of Young's modulus shows a similar tendency, none of the rest shows a clear tendency. Based on these facts, it is presumed that stronger mechanical damage has exerted during the fault formation in the hanging wall of the Tsukiyoshi Fault and it increases towards the fault.

It is presumable that the values of vertical initial stress in the hanging wall of the Tsukiyoshi Fault are nearly equal to the estimated overburden pressures. The values of principal stresses in horizontal planes tend to increase with the depth in the sections between 0300m and deeper than 600/700m in the AN-1. However, the MIU-2 shows no vertical variation in principal stress values. Though the AN-1 and the MIU-2 are similar to each other in distribution tendency of stresses in the section 300600/700m in depth, stresses in the MIU-2 are larger than those in the AN-1. The azimuth of principal stress shows rotation of about 45 at some 300m in depth in the AN-1. On the other hand, it rotates from N-S to WNW-ESE with an increase in depth in the sections 0400m and 400700m in depth in the MIU-2. However, it changes complicatedly below 700m in depth in the MIU-2. These facts suggest that the existence of the fault might exert a great influence on the in-situ stress field.

3) Revision of the rock mechanics conceptual model⁸⁰⁾

Based on the investigation results by the three 1,000m-class boreholes (AN-1, MIU-1 and MIU-2), it is proven that the rock mass on the hanging wall side of the fault could be divided into three zones with different geological and mechanical properties down to 1,000m in depth. Specifically, the first, second and third zones range 0300/400, 300/400600/700 and deeper than 600/700m in depth, respectively. Based on the investigation results of the three boreholes, the rock mechanics conceptual model of the rock mass on the hanging wall side of the fault in the Shobasama site are constructed as is shown in Fig.4.65. The rock mechanics conceptual model is constructed on the following assumptions.
The vertical distribution of initial stress values in the MIU-2 is complicated due to the effects of the fault. Thus, attention is paid mainly to the azimuth variation of principal stresses, which well corresponds with the variation in vertical distribution features of fractures.
The maximum principal stresses in the AN-1 located farthest from the fault show a NW-SE trend constantly below 300m in depth, which is coincident with the regional compressive axis⁸¹⁾. Consequently, it is assumed that stress axes at this location is little affected by the fault.
The results of BTV investigations in the AN-1 show almost the same variation trend of the fracture numbers as those in the MIU-1 and MIU-2. Therefore, the structural zones divided in the rock mass are extended nearly horizontally from the MIU-2 to the AN-1. However, it is assumed for simplification that stress values at locations far from the fault linearly increase with the depth in the each structural zones.

This rock mechanics conceptual model images that the rock mass on the hanging wall side of the Tsukiyoshi Fault is obliquely pressed against the large-scale tectonic discontinuity plane by the regional tectonic stress field to be resultantly pushed up along the fault plane.

This model seems to give an explanation to the phenomenan that vertical stress in the footwall of the fault is smaller than the estimated overburden pressure. Also, a phenomenan that the azimuths of principal stresses rotate in the individual structural zones could be explained. It is thought that the validity of this hypothesis could be indirectly verified by understanding the stress states on the footwall side of the fault in the MIU-3.