... the construction cost. From this point a series of preliminary investigations were carried out on the geotechnical subjects concerning CAES with using unlined caverns in a sedimentary rock site near the sea in Japan. Among the subjects the prevention of the air leakage from the unlined caverns was a matter...

... embankment, 15m high, with associated problems such as the release of acidic water from the fill, poorly established vegetation, collapsing bank protection, surface gulleying and leakage. Canals retain water in artificial channels so any leakage can cause saturation of embankments, loss of strength...

...% of the water pressure, and with cement grouting for sealing. Leaks primarily occured at the concrete-rock interface, and often it was necessary to perform several rounds of post grouting to bring the leakages down to an acceptable level. During the last 10 years, the Norwegian hydropower industry has pushed...

... units were declared commercial in December 1985, and the plant has operated continuously since that date. Initial filling of one of the three power tunnels produced unacceptable amounts of leakage from the tunnel. In addition, seepage pressures buckled and ruptured a short Section of one of the steel...

... fundamental aspects: Structural stability during operation; avoid excessive leakage due to high permeability of the mass of rock, and avoid excessive 1eakage due to hydraulic fracturing when the stress level in the mass of rock is not high enough. The only alternative to guarantee that the phenomena...

... overburden stress water pressure evaluation reservoir geomechanics test chamber lining leakage tunnel permeability alignment internal pressure deformation interface prestress stress condition ROCK ENGINEERING ASPECTS OF THE CONCRETE LINED PRESSURE TUNNELS OF THE DRAKENSBERG PUMPED STORAGE...

... chart the valley Upstream Oil & Gas pressure tunnel Norway inclination surge chamber water pressure leakage rock mass Reservoir Characterization principal stress water head tunnel powerhouse operation unlined pressure shaft hydropower plant pressure shaft THE DEVELOPMENT...

... of filling the shafts and tunnels, leakage control is done. The tunnel system is filled in steps of 100–150 m in order to carry out the leakage measurements. Summary of design, supervision and leakage measurements is given for''6 hydropower plants with water pressure on unlined rock in the range of 25–59...

... called for extensive sealing and grouting works to be performed in these parts of the project in order to avoid excessive leakages. The main part of the unlined pressure shaft and pressure tunnel designed for a maximum 455 metre static head was, however, placed in moderately jointed rock. (Figure...

... design required a method to predict air leakage rates as a basis for selection of compressor capacity. The compressor size depends on the maximum time allowable for chamber filling, and the capacity necessary to compensate for air loss through the rock mass. In Norway, the State Power System, uses mainly...

... structural geology abutment bedrock Reservoir Characterization Lugeon value cement leakage softer rock Shirakawa Dam water level Thickness characteristic open crack water management foundation permeability injection pressure rock foundation soft rock foundation permissible...

... <Il welded) Platy slightly wide 0 Welded tuff Columnar.c narrow '" Wtl (poorly 1. 50-1. 90 100-300 welded) Platy narrow Pumiceous tuff Columnar rarely Shimamatsu (extra poorly present Ps 1.40-1. 68 150 underpumice welded or nonwelded) Platy absent Older terrace Gv Gravel deoisit - - - - Pre-Nopporo PNo Sandstone deposit Conglomerate - - - - overlies Wt almost horizontally. The rocks which raises problems from the point of the seepage control of foundation are Gv, Ps and Wt. Among them, Gv is most permeable. However, since firstly it exists at a depth of 40 meter from the riverbed at the location of dam axis, secondly Ps underlain by Gv is expected to function as a blanket for Gv, and thirdly the depth to Gv from the riverbed increases in the down- stream because the dip of Gv is slightly steeper than the inclination of the river- bed, Gv can be excluded from the kind of rocks which have engineering problems. Therefore leakage and piping of Ps and leakage through open joints of Wt became problems in design of dam, and investiga- tions of permeability were performed followingly. 3 PERMEABILITY INVESTIGATION FOR FOUNDATION 3.1 Water pressure tests in drill holes Water pressure tests were conducted 230 times in 33 drill holes by the packer method in which an air packer of high pressure of 100 cm length was used for the prevention of the leakage around' the packer. The relations between the injection pressure (P) and quantity of injection rate per length (Q) in drill holes are categorized into three patterns as follows, where P is corrected by subtracting the ground water table and head loss due to the viscosity in a drill hole from the sum of the pressure at the top of the drill hole and hydro- static pressure at testing location. Type I Q increases proportionally to P up to the maximum pressure (about 10 kg/cm2) Type II Q increases proportionally to P up to the critical pressure, bU~ if the injection pressure excee s the critical pressure, Q showS sudden increase Type III: Q is extremely large even for ot small pressure, so that P can n reach the maximum pressure due to lacking of pumping ability TlC II b I~ o 'l Figure 2. Relations between injection pressure (P) and injection quantity per unit length (Q) 1154 An extensive network of joints has rendered the otherwise dense welded tuff into a high- ly permeable aquifer. Void velocity of see- page flow was measured to investigate the behavior of seepage flow through joints. Saline water mixed with fluorescein was used as tracer. Tracer was placed into the drill holes and was detected at locations of abutments where the ground water carne out and at the springs of ground water in an exploratory adits. Figure 3 shows the test results accompa- nied with the distribution of the direction of joints obtained from the analysis of bore hole core which was drilled in an inclined plane. The main directions of the flow coin- cide with the predominant directions of joints, both heading for the downstream direction. Therefore opened main joints was supposed to form the paths for seepage flow. Table 2 Results of Laboratory high pressure permeability tests These three types of P-Q relations are illustrated in figure 2. Type II is further subdivided to Type II-a of irreversible relation and Type II-b of reversible relation. Type I prevails especially among the upper portions of Wtml and Wtl with no large open joints and Ps of partial welding. Seepage flow of Type I is assumed to be laminar flow and its Lugeon value ranges from 10 to 20. Type II is very popular in scarcely jointed Ps, and it sometimes exists in Wtp with hair crack and in fresh portions of Wtdl. Type II-a is a fracture type, in which Q increases abruptly at failure by piping around the packer or connection of joints and Q does not decrease when injection pressure is reduced. Type II-b is derived from the deformation of the rock which is associated with the upheaval of the ground, being featured by the reversible relation of P and Q. Lugeon value for Type II were calculated by extending the linear relation below the fracturing injection pressure. It was about 15, which is not low, but less permeable than expected. Type III is prevalent among the rocks with many wide-opened cracks, such as Wtm2, Wtml and Ps at the deepest. The deepest portion of Ps comprises a lot of lapilli with porous structure and constitutes a confined aquif- er. Ground water table sometimes cannot be found in these jointed rocks. These features indicate that Type III is significantly permeable, having the Lugeon value larger than several tens. 3.2 Pumping tests Water pressure tests in drill holes by Packer method were not applicable for some POrtions of Ps because of its low strength. For the deatailed survey of permeability for these portions, five pumping tests using a shaft and large size drill holes were per- formed. Test results sUfgest that the permeability ranges from lxlO- to 3xlO-4cm/sec and for confined aquifer it ranges from 5xlO-4 to ~XlO-3cm/sec, with a little variation, which ~s due to the difference of the test pro- cedures and analytical methods. Right abutment Concentrating degree of joints Velocity J Frequency - J of joints Figure 3. Joint airection in dense welded tuff and direction and velocity of ground water 3.3 Velocity measurements by tracer method Sample size Initial coefficient Critical hydraulic Critical velocity Flow direction Diameter length of permeability gradient (em/sed to bedd ing plane (em) (e~) (em/sed 15 3 1. 7 x 10-4 241. 5 0.019 oeroendicular 15 5 1. 7 x 10 200 0 0.0445 ner nend Lcu Lar 15 15 1. 7 x lO-q 126.3 0.072 oeroend icular 15 20-25 1. 7 x 10-Q 131. 3 0.038 oeroend lcular 15· 20-25 2.0 x 10 227.8 0.072 parellel 1155 3.4 Laboratory high pressure permeability tests Hydraulic fracturing tests was conducted for nonwelded pumiceous tuff as it is porous and permeable. Undisturbed sample was used for this test. The diameter of the specimen is l5cm and the length of the specimen varies from 3 to 25 cm. Two directions of perpendicular and parallel to the bedding plane were selected for the direction of the seepage flow. Test results are summariz- ed in table 2, in which critical hydraulic gradient and critical velocity were also shown. 3.5 Field hydraulic fracturing test Field hydraulic fracturing tests for pumice- ous tuff were carried out using an explora- tory adit and a horizontal drill hole which is parallel with the adit. Figure 4 shows the locations of adit, test drill hole, and installation of pore pressure gauges. Water was supplied into a borehole under gradually increasing pressure to cause the fracturing. 0 '0 o 0 0 Injection pressure 00 "!3 '.0III~ 0.'0.~ 0 0.' Dial gauge Test drill hole ¢55mm ~ec :J '"bO ~ ~~0. "5/c, NO.5 Adit Figure 4. Schematic view of the field hydraulic fracturing test Test results are shown in figure 5. pore pressure increases suddenly at about 10 kg/cm2. The rock around the packer fractured "0 '70 Time in minutes Time in minutes Figure 5. Result of the field hydraulic fracturing test I~ I no 180 'iIO 1156 at 11 kg/cm2 resulting in the whole injected water discharging from the inlet of borehole. The hydraulic gradient was 45/1 when the injection pressure was 10 kg/cm2 correspond- ing to the sudden increase of pore pressure, and it was 47/1 for 11 kg/cm2, between the borehole was and the pore pressure gauge No.1. The hydraulic gradient corresponding to the fracturing around the packer gave the value of 110/1, as the packer length is 100 cm. When the injection pressure was 11 kg/cm, the hydraulic gradient between the pore pressure gauge No.1 and the wall of the adit was 7/1 to 5/1, which was ten times of the maximum hydraulic gradient in dam foundation caused by impounding. The rock of the adit did not show any marked deformation nor movement of soil particles, but showed only slight leakage of water all OVer the wall. Laboratory and field tests indicate that the critical hydraulic gradient for pumice- Ous tuff ranges from 40/1 to 200/1. However such loose portions exist as sampling is impossible, and there are several examples that fracturing around the packer occurred at the injection pressure of 2 to 5 kg/cm2 during drill hole water pressure tests. In these portions the critical hydraulic gradient is 20/1 to 50/1, which indicates the minimum value for pumiceous tuff. 4 DESIGN OF SEEPAGE CONTROL AND GROUTING WORKS 4.1 Design concepts for seepage control As mentioned previously, the foundation rOcks of Izarigawa Dam raised two problems from the point of seepage control. One is open joints of welded tuff in abutments, and the other is porous pumiceous tuff in the foundation. Each required suitable COuntermeasures. 4.1.1 Design concepts for seepage control for welded tuff ~elded tuff in both abutments is extensively JOinted and the ground water table in this rOck is very deep. Although the ground water ~evel fluctuate with season and rainfalls, ~t is nearly horizontal and slightly affect- ed by the predominant direction of joints and topography. The unsaturated part between the reservoir water level and the stable ground water level is regarded as the main range where leakage might occur and con- ~eqUently grouting for seepage control ecomes very important. Grouting in both abutments was planned as fOllows: 1. Grouting tunnels were provided for effective curtain grouting works. The extent of the tunnels in the left and right abut- ment was 280 m and 200 m respectively. This enabled grouting curtain to cover all the unsaturated part between the reservoir water level and the ground water level. 2. The depth of the curtain grouting was 40 m which reach the pumiceous tuff overlain by welded tuff. 3. Consolidation grouting was provided for the foundation of the spillway structure. 4. Blanket grouting to reduce the shallow under seepage was performed for the foundation of impervious zone by six lines of grout holes parallel to the dam axis. The spacing of holes was 6 meters. 5. As a lot of leakages were observed at the walls of welded tuff, grouting to cut off these leakages was performed by using single line of grout holes spaced 2 meters. 4.1.2 Design concepts for seepage control for pumiceous tuff Pumiceous tuff is so porous because of poor welding that permeability of rock itself becomes problem. The results of grouting test suggested that it is difficult to improve its Lugeon value as much as the value for welded tuff. The purpose of grout- ing for pumiceous tuff was intended for the potentially existing small cracks, aquifer lense and disturbed rocks by excavation. 1. Curtain grouting was performed by a single line of grout holes to a depth of 30 m. The spacing of grout holes is 75 cm. Colloid cement was used for grout mix. The maximum injection pressure was 3 kg/cm2 2. Blanket grouting for the foundation underneath the impervious zone was performed by using horizontal grout holes in the foundation gallery after embankment of minimum 15 m thickness was placed. ("1m) 60 a ~ 40~~ 11 30 ~ "~ 20~ '0 Wtm2 Wtm, Wtl Ps Figure 6. Grout takes for each lithoface 1157 4.2 Grouting works 1. Grout takes Figure 6 shows the cement takes for each lithofacies. Wtm2 differs distinctively from Ps, and the intermediate quantity of grout takes were recorded for Wtml and Wtl. In Wtm2, cement takes in the primary and secondary holes are very large, decreasing very.much in the third and check holes. However, in Ps, cement takes are small in all holes and variation is small. The results of cement takes indicate that a large amout of cement was required for the dense welding portions because of open joints and cement injection was difficult for nonwelding portions, as was expected previously. 2. Degree of improvement of permeability Degree of improvement of permeability in each lithofacies is shown in figure 7 by using Lugeon value. Wtml and Wtm2 were improved effectively by sealing of open joints. Wtl and Ps were improved gradually, though it is not as effective as in densely welded parts. However the Lugeon value of Ps is about 8 even in the check holes, which is greater than the other litho- facies. This value was expected from earlier, and it is considered to be the ultimate value of improvement. As no unusual behavior was detected both in leakage and pore water pressure after impounding, the foundation treatment of pumiceous tuff was judged to be successful. 50 '" 40~ 30 c: o ~20 :Jl Pilot and ~ primary holes Secondary andx third holes x .0 Check holes 0 10 Wtrn, Wtl Ps Figure 7. Degree of permeability improve- ment 5.BEHAVIOR AFTER IMPOUNDING OF THE RESERVOIR Impoundment of the reservoir began in April, 1980 and the water level reached the maximum surcharge water level of EL. 176.5 m in November, 1980. Figure 8 shows the leakage collecting system of Izarigawa Dam. The reservoir EI.179.5" Impervious' zone I,' : ". Figure 8. Leakage collecting system water level and leakage are illustrated in figure 9. When the water surface reached maximum level in November, 1980 the total leakage from route 1 to 6 amounts 420 l/min, which is approximately coincide with the . result estimated from seepage flow analysks, ,. '00 .5 :: -, " . _lLJ 'M -- _ . l ht> 1980 1981 Figure 9. Relation between the water levei and leakage The ground water table in both left and right abutments is near-horizontal. At th~ farthest drill hole from the abutment wal , the ground water level exists at about EL. 165.0 m, w~ich fluctuation due to reservoir water level is very...

... of dam and Leakage measur- capacity of mud flow of foundation. plan and cross section are indicated 1 and Fig. 2, respectively. The dam located in the midst of a region in northern Japan which is covered with deep snow in winter. Accordingly, its con- struction was forced to be discontinued entirely...

... SUMMARY From December 1971 up to November 1974 an important and difficult grouting work was executed in the rock• foundation of the Canelles arch-dam. Leakages and percolations- of water through- the foundations, were around 6 m3 / sec. at the beginning of the works. Once completed the maip...