Tunneling problems for 1020 MW Tala Hydroelectric Project in Bhutan Himalayas

D.P. Goyal Tala Hydroelectric Project Authority R.N. Khazanchi Tala Hydroelectric Project Authority

ABSTRACT: Tunneling in Himalayas has always been a challenging task. Thi s paper highlights various tunnelling problems encountered and solutions being implemented in excavation of a 22.9km long 7.6m excavated dia. Hydraulic Tunnel for Tala Hydroelectric Project which is located in southwestern part of Royal Kingdom of Bhutan in eastern Himalayas. As of end October 2001, 15.2km of tunnel has already been excavated through 6 adits totalling 2.23km in length and many challenging situations have been tackled peculiar to tunnelling in Himalayas. Fractured nature of rock mass and d eep permeation has created pockets of ground water which has posed problems during tunneling. Barton's rock mass quality index 'Q' is being used for evaluation of support system in the Norwegian Tunneling Method. The difficult situations are being managed by 'DRESS' (Drainage, Reinforcement, Excavate, Support, Solution) technique. SFRS thickness up to 150mm in a single application by using alkali free accelerator, Meyco SA 160 of MBT, is being provided

INTRODUCTION been taken up from Oct. 1997 starting from developing the infrastructural facilities and is scheduled to be commissioned by Dec. 2004. The head for power genera tion is available through a 22.9km long hydraulic tunnel, from intake to powerhouse complex. (refer Figure 1, Plan of the Project & Fig. 2, L -Section through Head Race Tunnel).

The Tala Hydroelectric Project is a run -of-river development immediately downstream of the existing 336 MW Chukha Hydroelectric Project on river Wangchu in South -West Bhutan in eastern Himalayas. It envisages construction of a 92m high concrete gravity dam at Wangkha; three intakes and thr ee underground desilting chambers each 250m x 13.9 m x 18.5 m; a 6.8 m modified horse shoe 22.9km long head race tunnel to conduct 142.5 cumecs design discharge and terminating into a 15m dia and 178m high surge shaft; two inclined pressure shafts 4m in di a and 1.1km long each trifurcating into penstocks of 2.3 m dia; and underground power house complex at Tala to generate 6x170 MW of power through six pelton turbine based generating units and a 7.75 m horse shoe 3.1km long tail race tunnel to discharge the water back into river Wangchu. (refer Fig.1, Plan of the Project). The Project execution has

Figure 1,

Plan of the Project

Figure 2,

L-Sectio n through Head race tunnel

REGIONAL GEOLOGY

comprising quartzites, phyllites, with carbonate bands and basic rocks. The area is essentially occupied by the folded sequence of metamorphic rocks comprising gneiss, augen gneiss, granite gneiss, high grade schist, chlorite and sericite schist, quartzite, calc silicates and marble with acidic and basic intrusives. The rocks are highly folded, jointed, sheared an d faulted in nature.

The project area falls within the central crystalline belt of Thimphu Formation and metasedimentaries of the Paro Formation. These formations are underlain by younger Shumar Formation. The rocks in the Bhut an Himalayas have suffered at least three cycles of deformation, as a result of which major thrust zones have developed. The most prominent of these is the Main Central Thrust (MCT) along which the older crystalline rocks of Thimphu Formation comprising v ariant of gneiss, biotite schist, marble and quartzite and granitoid bodies are thrusted over the younger metasedimentaries of Shumar Formation

GEOLOGY ALONG HEAD RACE TUNNEL

The head race tunnel (HRT) aligned on the right bank of river Wangchu has six kinks due to the prevailing topographical features. It crosses seven

deeply incised major drainages, all tributaries to Wa ngchu. These cross drainages necessitated hillward shift and kinks in the tunnel alignment so as to ensure sufficient rock cover, maintain the tunnel grade and locate convenient sites of intermediate adits for speedy excavation of the tunnel. The HRT has a vertical cover of 95m to 1100m covered largely on the surface with dense forest. The 22.885m long and 6.8m dia horseshoe shaped HRT is being excavated from 11 faces with six construction adits and an additional face from the inlet end. The locations o f the adits are so fixed that the maximum length of tunnel from any face is not more than 3 km. The tunnel alignment passes through bedrocks comprising biotite gneiss, augen gneiss, gneiss with bands of quartzite, biotite schist, muscovite schist, quartz mica schist, calc silicate, quartzite and gneiss with subordinate chlorite schist and sericite schist with acid and basic bands. The tunnel is mostly parallel or sub parallel to the foliation trend of the formation. The rocks along the tunnel alignment a re generally folded into open synforms and antiforms. A number of cross faults and shear zones are intermittently present. Tunneling through the anticlinal axial zones results into wedge formations and overbreaks due to the fractured nature of rock mass. The Project area gets fairly high annual precipitation as rainfall of upto 5000mm. Low rock cover and deep weathering posed problems during tunneling. Fractured nature of rockmass and deep permeation has created pockets of ground water. Soft bands of mus covite schist and biotite schist, in saturated conditions, create flowing conditions and tunneling problems.

TUNNELING METHOD

The tunnel work of Tala Hydroelectric Project is being carried out in accordance with the Norwegian Tunneling Method. Tunnel e xcavation is by conventional drill and blast. Rock support is

mainly by rock bolts and plain/steel fibre reinforced shotcrete (SFRS). In case of extremely poor rock conditions, positive supports are also being provided by structural steel ribs. To check late dilation of rock and sagging tendency, excavation with rock bolt and SFRS support to about one excavation span length followed by 3m length of steel rib support alternately is used. We have achieved desired thickness of SFRS upto 150mm in a single a pplication by using alkali free accelerator, Meyco SA 160 of MBT. Tunnel work is being carried out by three Indian Construction Companies, namely M/s. Hindustan Construction Co. (HCC) - Mumbai (13.5km); M/s. Jai Prakash Industries Ltd. (JIL) - New Delhi (5 .0 km) and M/s. Larsen & Toubro (L&T) - Chennai (4.4 km). These Company's have generally deployed Atlas Copco 352H 2 -boom drill Jumbos; Atlas Copco 435H Boltec (for rock bolting of 3.5m length); CIFA Spritz wet shotcrete machine (30 m 3/hr); MBT Meyco wet shotcrete machine (30 m 3 /hr); Atlas Copco Wagner MT 420 dumpers (18 T); Ajay - Fiori CM -4000E Transit mixers (4m 3 ); Deici dumpers (12.5 T); Deici Transit mixers(5m 3 ); L&T CK90 electrical hydraulic excavator (1.2m 3 ); Schwing pumps(30m 3 /hr) etc. as main exc avation equipment. N Barton's rock quality index 'Q' is being used for excavation of support system in Norwegian Tunneling Method. Q = RQD x Jr x Jw Jn Ja SRF It is based on numerical assessment of rock mass quality using six different parameters, given below: ROD - Rock Quality Designation Jn - Joint Set Number Jr - Joint Roughness Number Ja - Joint Alteration Number Jw - Joint Water Reduction Factor SRF - Stress Reduction Factor These parameters are combined in pai rs and are found to be crude measures of: RQD - Relative block size (represents overall Jn structure of rock mass) Jr - Inter -block shear strength (represents Ja roughness and degree of alteration

Jw SRF

of the joint walls or filling materials) - Active stress (consists of two stress parameters; Jw is a measure of water pressure; SRF is a measure of

(i)loosening load (ii) rock stress and (iii) squeezin g or swelling loads in plastic incompetent rock. It can be regarded as a total stress parameter.

Excavation Classes - Support Measures (Based on Barton's 'Q' classification) Value of 'Q' 40 and above 10 to 40 4 to 10 1 to 4 Rock Class I (very good) II (Good) III (Fair) IV(Poor) Support measures Excavated section remains basically unsupported. Spot rockbolts or shotcreting in localised areas may be required Support system consists of individual rock bolts and shotcrete Support system consists of pattern rock bolting and shotcrete (Plain or SFRS) in one or more layers. Support system consists of steel ribs with or without invert bracing with concrete/steel lagging backfilled with concrete. The steel ribs may be complemented wit h some rock bolts and shotcrete (Plain or SFRS). Few forepoles, if required. Support system consists of steel ribs and lagging placed continuously at the heading face and backfilled with concrete immediately after excavating whole o r a part of the cross section. If fibre reinforced shotcrete is used, outer flange of the steel ribs may be encased in concrete. Forepoles may also be required. Multiple drifting may be required in some stretches. Water -cement admixture grouting may be carried out in and around the future advance followed by careful excavation at small steps and immediate installation of supports at the face and in the excavated space. In event of squeezing ground, excavated secti on shall be restored to its theoretical dimensions. Multiple drifting may be required in some stretches. Chemical grouting and ground freezing etc. to be used to continue with the advance

0.1 to 1

V (Very Poor)

0.01 to 0.1

VI (Squeezing Conditions)

Adverse Geological Occurrence The appearance is unexpected and there is no possibility of further advance by construction methods including forepolin g and multiple drifting even by using best qualified tunneling crew.

TUNNELING PROBLEMS

Although a number of situations have been encountered in the excavation of adit s and HRT wherein occurrence of cavities due to formation of wedges in the crown ranging from 1m to more than 5m in depth at times combined with seepage water resulting in flow of rock mass have been tackled so far, some of the serious situations are highl ighted hereunder which indicate the difficult tunneling conditions in Himalayas. 5.1 Cavity at Ch. 118 m of Padechu adit At Ch. 118m of Padechu adit, a sudden collapse at crown was noticed at 13.00 hrs on 7 th May 1999 through rock fall with initiation of wat er flow of about 40 - 50 litres/minute. Within half an hour flow of rock mass started heavily compounded with high discharge of water, measured to about 1200 litres/minute at 16.00 hrs of the same day. Flow of rock mass covered floor of adit back to adit Ch. 100 m. The excavation indicated pressure of as foliation shear and a cross shear zone with fairly thick fracture zone which resulted in collapse. Initially, an attempt was made to remove the rockmass to make crown free of any such loose rock and to pr oceed for further advancement. About 900 m 3 of rock mass was removed upto 11 th May. But further flow of rockmass was continuing with removal of fallen muck. However, flow of rockmass reduced when the same occupied the floor of tunnel from Ch. 118 m to C h. 95 m. Discharge finally decreased to 600 litres/minute and remained almost constant. With this flowing condition of rockmass tending to collapse of crown, steel supporting of crown from Ch. 100m was taken up. A ramp was made on muck for movement of equipment etc. A bulkhead at the crown by sandbags was made on the muck at about 10m behind the face to prevent sliding of muck and by sandbags placed at the face near the crown. Steel ribs for the arch portion from Ch. 111m to Ch. 112.50m @ 500 mm c/c wa s erected, in pieces, supported on horizontal beams, placed on the muck. These ribs were also propped with

steel pieces and also to arrest movement of the bulkhead for arresting flow of muck. Backfill concrete behind the ribs was done and at the same time pressure relief holes were driven at places to take care of increase in water pressure around the crown. Steel sections of channels/boxes of 100mm x 75mm were driven through the bulkhead to arrest flow of boulders/muck inside the cavity. Concrete place r pipes and drainage pipes were duly placed inside the cavity for concreting and channelling flowing water to drain outside the cavity (refer Fig. 3). Since it was observed that location of unstable rock mass above spring level was continuing, a repeat pr ocedure of making a bulkhead, driving forepoles of heavier steel section was made in stretches and steel ribs were erected in parts, and placed on the fallen muck. About half a metre of space was cleared for accommodating the rib and a further bulkhead wa s made for advancement. Care was always taken to channelise gushing water coming from face to the tune of 1200 litres/minute. Any attempt to open space of more than 0.5m led to hazards of flow of rockmass along with water. Once channelisation of water was made possible, attempt was also made for injection of cement : sand (1:1) grout following concreting. In some places about 800 bags (50kg/bag) of cement were injected. In all advance operations through placement of ribs in parts above spring level, excavation was done manually without any blasting. During construction, the seepage flow also increased suddenly from 800 litres/min. to 2000 litres/min. for about 4 hours before it came back to 800 litres/min. probably because of puncture of a ground wat er pocket. With advancement, muck from the rear end i.e. Ch. 100 m was slowly removed and extension of columns to support erected ribs on muck was carried out. Extended columns were duly anchored in side rock.

During whole process, it was observed that w ith advancement of face, discharge was also found to be advancing at face with tendency of face collapse. Full face was opened at Ch. 132 m with some improvement in rock strata and the face

could take full blast. It took four months to handle this situati on between Ch. 118m - Ch.132m in a conventional method.

Figure 3,

Tackling Cavity at CH. 118 m of Padechu Adit

5.2 HRT junction with Mirchingchu Adit The Adit to the HRT at Mirchingchu was originally planned for a length of 280m in fair tunneling media. However, the length of 280m was mostly done in poor/very poor rock mass conditions and it was considered to increase the length of the adit, so as to locate the junction with the HRT in fair rock conditions. But even after taking the adit to a length of 320m, there was no improvement in the rock conditions and it was accordingly decided to provide a steel ribbed junction with the HRT in very poor rock. The Adit was D -shaped having an excavated diameter of 7.85m and was to be enlar ged to a section of 10.02m (w) x 9.593m (h) to accommodate the HRT in that portion. Enlargement of adit section to the required section at the junction was achieved by providing ribs in transition from Ch. 317 m to Ch. 320m. Length of

the adit from Ch.32 0m to Ch.329m in the enlarged section provided the required junction with the HRT, which was fully supported with steel sections. Support system in junction comprised four ISMB 600 x 210 columns; two ISMB 450 x 150 beams placed across for the HRT; and ISM B arches supporting the roof resting on two beams. The arrangement is indicated in Fig. - 4 & Fig. -5. In order to excavate beyond Ch. 320m upto Ch. 329m, it was decided to support the crown with the crown arches resting on the temporary columns of sectio n ISMB 350 x 140 which were placed outside the main columns in such a manner that these columns would not obstruct the erection of beams. Each crown arch was welded on both sides with ISMB 250 sections with a plate at the bottom just above the central lin e of the beam in such a way that when the beam was placed in position,

the plates welded to the ISMBs would be resting on top flange of the beam thereby transferring the load from the roof through the arch crowns to the beams and then to the columns. Exca vation upto Ch.329m was carried out by supporting crown arches on temporary columns provided on both sides. The remaining two main columns were placed when excavation reached upto Ch. 329m. Temporary columns were backfilled with concrete using form work, so that these columns did not get disturbed when excavation was advancing. Excavation was also carried out by advancing in

0.75m lengths and chipping with the boomer to avoid blasting as far as possible seeing very poor nature of the rock mass in which j unction was being made. After erecting the beams, crown arches were made to rest on these beams through ISMB 250 provided with a plate on each side of the crown. Any gap between the plates and the top flange of the beams was filled with steel pieces for uniform transfer of the load. All four columns were strengthened by embedding in concrete. Before embedding, columns in concrete, steel brackets connecting beams with columns were also provided to reduce the span of the beam.

Figure 5, Section A -A Figure 4, Adit Junction with HRT at Marichongchu the beams was found to be within the permissible limit. The rock mass at the junction of the adit with the HRT encountered comprised biolite gneiss, biolite schist and amphibolite schist bands. A foliation shear of 30cm thickness was observed in the middle part at hea ding, dipping at 20 o due west. Active flow of water was coming out from the right upper crown. The rockmass was moderately weathered, sheared and shattered. Foliation joints had clay infilling. On the upstream and downstream side of the HRT, squeezing rock conditions were encountered.

After the beams and the columns were in position and duly strengthened, temporary supports in upstream face of the HRT were first removed one by one so as to accommodate the steel rib for the HRT section. This rib was connected with the beam and columns at the junction to provide the transition and the excavation in the HRT continued for a length of 10m by installing ribs at 0.5m centre to centre before opening the downstream face. The ribs were further strengthened with steel sections in the form of stiffeners and runners to check squeezing behaviour of rock mass and pressure coming on to the ribs. The deflection in

5.3 HRT junction with Padechu Adit Junction of the HRT with Padechu adit has also been in very poor and squeezing rock conditions with steel ribbed sections used in a similar manner. At Padechu adit junction, thinly foli ated quartzite with biolite schist bands, dipping at a low angle with water seepage was present. The presence of thick shear seams at the crown resulted in 'Cave -in' of the order of 6m to 7m. The rock was saturated and separated out along the foliation. Squeezing conditions had been observed at the junction and upstream and downstream side of the junction. In the HRT on the downstream side, thinly foliated gneiss with quartzite was present with a foliation dip of 25 o S -20 o E. A cross shear was observed below the spring line trending almost horizontally having 50 cm thick crushed material and clay gouge.

collapse in the areas where cracks had formed, steel supports were provided in the Ch. 817m 820m and Ch.824m - 827m making safe the HRT in the rear portion. (refer Fig.6) The fallen muck was pushed toward the face and face was closed by providing a plug comprising steel plates in the bottom portion a nd gunny bags filled with muck in the upper portion. After sealing the face, GI pipes in lengths of about 6m - 9m were pushed into the face at different locations into the cavity for grouting. Cement : sand grout (1:1) was injected to consolidate the flow ing mass. This was attempted so that the rock mass above the crown in the cavity portion was consolidated which would enable further advance by providing rib supports below this. Water seepage of 100 -150 liters per minute was observed in the crown and a round. By grouting, water was brought towards the rear zone which was required for further advance otherwise it would further lubricate flowing mass and accentuate the cavity. With cement : sand grouting and strengthening of the crown by providing forepol es using steel sections such as ISMBs, pipes, angles, channels etc., it was possible to progress in heading by providing steel ribs in segments. It was experienced that every time after grouting when an advance was made by installing a segmental rib, at a distance of 0.5m or less, muck in the heading in cavity portion would slide down making water once again flow through the cavity portion unless rock mass was grouted once again. The cavity, which finally extended from Ch. 835m - Ch.841m, could be overco me by sealing the face; grouting with cement sand mix; supporting with steel forepoles; channelizing the water; proceeding in heading; providing segmental rib 2m -3m long at a time and back packing with concrete. The size of cavity could be appreciated fr om the fact that 2706 m 3 of loose rock fall was removed from the tunnel. It took 2 months to advance through the blind shear and the affected zone from Ch. 835m Ch. 841m.

5.4 Cavity In HRT D/S from Padechu Adit A major shear zone was encountered along with seepage between Ch. 838m - Ch.841m. The shear had a dip of 75 o - 80o /N 23 o S and was arculately disposed within this reach. It had an effective width of + 1.75m with crushed and fractured gneiss with multiple thin clayey gauge seams. Due to its steep faceward dip, it made a sudden appearance leading to flowing condit ion. The Head Race Tunnel downstream from the adit at Padechu at Ch. 835m was progressing in class III by providing design support of rock bolt and shotcrete. A blast at Ch. 835m was undertaken to reach the face at Ch. 838m. While mucking was in progres s, the rock mass started flowing from the crown forming a cavity extending backwards. Flow of rock mass continued along with the flow of water. The flowing muck covered the full tunnel face forming a heap and was allowed to remain as it was to check furt her flow from the cavity. While flow of rockmass at Ch. 838m was forming a cavity extending backwards, cracks were seen in shotcrete in the crown portion behind Ch. 838m. In order to check any possible roof

Figure 6,

X-Section at Ch. 841 m of HRT D/s of Padechu Adit

5.5 Shear Zone In HRT U/S from Kalikhola Adit The HRT between Ch. 662m - 700m upstream from Kalikhola adit had been advancing in very poor rock conditions having an unstable crown, formation of large overbreaks and cavities and flowing of rock mass at many locat ions. In the stretches where the crown was unstable causing large overbreaks but where rock mass was not flowing, it was possible to advance by putting steel supports back filled with concrete in smaller pulls of 1m -2m. In stretches where the rock mass was flowing, the face had to be plugged to consolidate the crown with cement grout for advancing further. Cement grouting also helped in channelising the seepage. Alternate bands of quartzite, amphibolite and biotite schist with 5cm to 20cm thick shear seams were encountered from Ch. 662m dipping 20 o t o 2 5 o into the upstream face causing roof collapses. The rockmass was highly folded and crushed along the axial plane of fold and weathered in nature. The strata was highly charged with seepage water. At C h. + 699m, a 10m to 12m thick cross shear with 1.5m to 2m thick clayey gouge and 5m to 6m thick pulverised quartzite was encountered at the face. This shear acted as barrier for the underground seepage and resulted in flowing and squeezing conditions. At Ch. + 700 m where the shear dipping at 20 o to 25 o met the sub -vertical shear, tunnel advance became

very difficult. The shear mass comprised of clayey material mixed with quartzite and amphibolite having extremely poor permeability. The conditions as enc ountered at Ch. 700m made the rockmass to flow heavily when it was tried to advance further. In view of flowing rockmass condition, the face was plugged with gunny bags filled with muck. It was tried to consolidate rock mass with cement grout but because o f impermeable rock mass, grout intake was almost negligible. Efforts were made to isolate water by injecting polyurethane grout 2m - 3m above the crown with the purpose that once water was isolated the rock mass in the crown, which was mostly clay, woul d become dry giving time for installing positive supports for further advance. While attempting polyurethane grout due care was taken that water above the crown was properly drained out from rear portion without allowing to develop pore pressure. Through the polyurethane grout helped in making crown comparatively dry and gain some strength, the standup time was not enough to support the crown in heading with steel ribs. It was also not possible to proceed in small segments because of the pressures and sq ueezing conditions on all sides. An umbrella with Self -drilling Hollow Core Anchors (MAI -Austria) 51mm dia, 12m - 16m

long and spaced at 300mm centre to center was also attempted to provide a shield so that advance could be achieved by putting steel supp orts one by one. Along with providing of umbrella, long pipes up to even 20m were also pushed into the face to have free drainage. The drainage was further tried by suction by putting 60HP vacuumatic pump. Because of the impermeable nature of the rockmas s no effective drainage could be achieved and seepage water to maximum of 30 liters per minute could only be collected from the face (refer Fig. 7 & Fig. 8 ). The heading face was consolidated by pushing bamboos into the face. This did help checking rock mass from flowing from the heading and helped to remove the plug in the heading portion without any disturbance to face which was seen standing even vertical. With installation of the umbrella and the drainage, as detailed above, when advance was tried, rockmass once again started to flow from the space between MAI anchors thereby forcing closer of the heading once again. A small drift 1.8m x 2.1 m is also being attempted through the face to know conditions ahead of face

and to release pressure and also to drain out water behind the shear. Progress in the drift also has been very slow in view of high stress conditions. We now plan to deploy 'PUNTEL' fore -poling machine for installing 15m - 20m long 110mm pipes into the face which would be closely spaced to form an umbrella for advancing further. This methodology known as 'DRESS' (Drainage Reinforcement - Excavate - Support - Solution) technique was successfully followed in similar situation in excavation of the 28km long HRT of 1500 MW Nathpa Jhakri Hy droelectric Project in Western Himalayas in India. Besides installation of longer forepoles, consolidating rock mass with Jet -grouting at high pressure is also planned so that rockmass does not flow as it happened in case of umbrella installed with 'MAI' S elf -drilling Hollow Core Anchors. Ch. at 663m was reached in Oct. 2000 and a length of 40m only has been done up to the end of Oct 2001. Tunnelling in such rock mass conditions is a real challenge and is an indication of tunneling difficulties in Himalay as.

tackling such surprises cannot be always available before han d for unpredictable situations. Whereas

Figure 7,

Schematic L -Section of Kalikhola Shear Zone

Figure 8, Section A -A some of the difficult tunneling conditions can even be tackled by conventional rock support measures, yet some serious situations need to be consid ered for tackling by jet grouting, DRESS technique etc.

CONCLUSIONS

There are surprises during tunneling in Himalayan geologic formations. State -of-art equipment for