During excavation of the power house cavern at Sardar Sarovar Project, India, distress problems were encountered due to the limited amount of cover and the presence of shear zones. After cracks were observed on the cavern walls, further excavation was suspended and additional treatments to the walls were provided. A construction ramp on the downstream wall was to be excavated by drilling and blasting without causing further damage.
The excavation of the ramp was critical as it was providing restraint to the movement of the downstream wall, which was intersected by large openings for six draft tube tunnels below the ramp. Three alternative blast designs were considered for the excavation and the safest method was adopted. The selected method of excavation divided the ramp into two parts – ‘main’ and ‘bark’. The ramp was excavated in 2m layers to minimise the unsupported area and was followed by support and reinforcement. The damage to the downstream wall was assessed by taking multipoint extensometer readings and by observing cracks in the glass plate installed at a critical location, by vibration monitoring and observing drillhole impressions on the wall.
The Sardar Sarovar Project is an inter-state project in which the states of Gujarat, Madhya Pradesh, Maharashtra and Rajasthan are participating. One of the largest water resources development projects in India, it envisages construction of a concrete gravity dam, powerhouses, one underground and another surface. The proposed underground powerhouse is 23m wide, 56.6m high and 212m long.
The area for this comprises lava flows of sub-horizontal basalt, separated by hard conglomerate and intruded by two dolerite dykes. The rocks are hard but intensively jointed. The main shear zone at the contact of dolerite dyke and basalt, is running across the cavern with a dip of 600-650o towards the river side. This shear zone is 1 to 2m wide, consisting of rock fragments with little clay. Two shear zones ‘A’ and ‘B’ dip to the north west forming a wedge. These are 0.3 to 0.5m thick and consist of highly chloritised, calcified rock fragments with little clay. Three sub horizontal minor shear zones ‘X’, ‘Y’ and ‘Z’ are also observed along the downstream wall.
The excavation of the cavern commenced in November 1987. Support consisted of pattern rock bolting together with 38mm layers of shotcrete with a welded wire mesh in between. Unfortunately, the 6 to 7.5m long rock bolts for the sidewalls were too short and could not provide adequate restraint to prevent the development of cracks (Hoek, 1995).
The first crack was observed along the upstream wall in the area of shear zones in March 1991. Subsequently, cracks were also observed on the downstream wall and bus galleries (Dasgupta et al, 1995). As remedial measures, additional treatment for the upstream and downstream walls, bus galleries and for the junction of pressure shafts, bus galleries and draft tube tunnels with the power house walls were developed (Mittal et al, 1999) and treatment was completed in 1994.
A construction ramp was left along the downstream wall from elevation 20m at the service bay to the floor level of the cavern at elevation -2m on the river side, as shown in Figures 1 and 2. This ramp was the only access to the floor level of the cavern until draft tube tunnel 1 was broken through. The ramp width was approximately 8m on the top or nearly 12m at the bottom. The estimated volume to be excavated was 15000m³.
The overall stability of the powerhouse cavern was a matter of great concern because the ramp, temporarily supporting the downstream wall, was to be finally excavated. The rock mass in the downstream wall was weakened by the intersection of several large openings including six for draft tube tunnels and three for bus galleries. Although the magnitude of the deformation recorded after the treatment was small and no alarming signals were discernible, excavation of the ramp by drilling and blasting was considered risky. Therefore Sardar Sarovar Narmada Nigam Ltd (SSNNL) requested the National Institute of Rock Mechanics (NIRM) to recommend and execute a blasting design for safe excavation of the ramp. The removal of the ramp by controlled blasting started in May 1999 under NIRM supervision and was completed in March 2000.
Earlier excavation
The blast design for the earlier excavation of the cavern were examined in arriving at a suitable final design. Horizontal holes were drilled to a depth of 2.4m with jack hammers. The maximum charge per delay was restricted to 5.46kg and the total charge was about 50kg with the specific charge of 0.54kg/m³. Holes were also drilled up to 3m in the same pattern but the maximum charge per delay was increased up to 7.54kg.
As progress with jack hammer holes was slow, blasts with 76mm diameter, 4m deep holes were tried. It was understood that this design was discontinued after some trials because it caused more damage to the rock mass. Both the designs control blasting damage to the wall by reducing the charges in graduation from the production holes to the perimeter holes.
It was felt that a conservative blast design would delay the excavation process and increase costs. A liberal blast design, on the other hand, should not be permitted for the safety reasons. Therefore an optimum controlled blast design had to be evolved. The controlled blasting should be followed with immediate support and reinforcement of the exposed wall to minimise rock mass movement.
Ramp blasting had an additional free face compared to the blasts conducted earlier which had only two free faces. The ramp blasting for the same amount of charge per delay was expected to produce lower vibration and hence less damage to the adjoining rock. For safety the ramp was divided into two parts. The major part was blasted as bench blasting while controlling ground vibration. Smooth blasting was also required for the bark portion of the ramp to reduce the damage on the downstream wall of the cavern. Therefore two blast designs were made: one for main (Stage 1) blasting, another for bark (Stage 2) blasting. Stage 1 was advanced by about 6m before taking Stage 2.
Ramp blasting in two stages (leaving the bark of 1.5 – 2m) was expected to give better results compared to blasting in a single stage (full width of the ramp). Though the loading and hauling equipment were to be deployed under congested conditions, this option was safer.
The blasted muck was removed through draft tube tunnel 1, loop and access tunnels. Movement of rock through this route was possible throughout the year. As the ramp was divided into layers of 2m, 11 layers had to be removed. The ramp blasting progressed from the river side towards the service bay by vertical benching leaving a bark 1.5m – 2m wide. After removing the third layer, the muck was also removed through draft tube tunnel 6 after breakthrough with the downstream wall of the cavern in January 2000.
Superdyne, a cap-sensitive small diameter aluminised slurry explosive, manufactured by IDL Industries was used. Each cartridge of explosive was 25mm dia, 200mm long, and weighed 0.125kg. The density of the explosive was 1.15-1.25gm/cc and velocity of detonation 3400-4000m/s.
Two types of IDL Industries electric detonator were used. Short delay electric detonators had delay numbers from zero to 10, with a nominal time interval of 25ms between successive delay numbers from 1 to 6, 50ms for 7 and 8, and 75ms for 9 and 10. Long delay electric detonators had delay numbers from zero to 10, with a nominal time interval of 500ms between each successive delay number. The delay periods of either short or long delay detonators were not sufficient. Hence, a combination of short and long delay series was used for the blast design to restrict the maximum charge per delay to 4.5 kg.
Blast design for main
The hole diameter deployed for drilling was 51mm using Ingersoll-Rand CM 341 mounted on EVL 130 wagon drills. The same wagon drills were used for rock bolting and cable anchoring. It was important that excavation and support were concurrent. Therefore, all necessary arrangements for treatment of the downstream wall after removal of bark were made ready to provide treatment without delay. Vertical holes were drilled to a depth of about 2.2m on a pattern of 1m x 1m and charged with Superdyne. The charge per hole was 0.750kg. The maximum charge per delay was restricted to 4.5kg to avoid damage to the rock mass. This was less than the 7.54kg used earlier for horizontal bench blasting and much less than 10.1kg used in vertical benching with 76mm diameter holes. The suggested drilling and initiation pattern for removal of the main portion of the ramp shows a V-cut or diagonal cut, depending on the desired direction of throw and other site conditions.
The design specific charge for the ramp blasting was 0.30 to 0.35kg/m³, which was lower compared to those used earlier because of availability of an additional free face (Adhikari et al, 1999). The actual specific charge calculated from the consumption of explosives and the volume of mucking was 0.23kg/m³.
The 51mm dia drill hole and the 25mm dia explosive cartridges were not matching. Six cartridges of explosives loaded in the hole settled at the bottom, leading to an increased stemming length of about 1.5m. Because of this, fragmentation was not uniform but the longer stemming column controlled the occurrence of flyrock reducing the risk of equipment damage.
A rock slide occurred in the ramp between chainage 1575 – 1594 after the blasts on June 22 and 23, 1999. The sliding occurred along the joint plane that was dipping towards the upstream side. Another area prone to sliding was then drilled and blasted in a controlled manner to avoid sudden collapse of the rock.
Blast design for bark
With the drilling and charging pattern for bark blasting the bark portion of the ramp was removed by drilling jack hammer holes 2m deep. The suggested design consisted of production holes and perimeter holes. The latter were drilled at a spacing of 0.2m and the production holes were 0.5m apart from the perimeter holes at the spacing of 0.8m.
The production holes were charged with two to three cartridges per hole while the perimeter holes were charged with two cartridges using PVC spacers. However, in the shear zones the charge was reduced to two cartridges in production holes and one cartridge in perimeter holes. Trials were made leaving two dummy holes and one dummy hole in between the charged holes. The result was better leaving only one dummy hole.
Monitoring blast vibration
During excavation of the ramp, blast vibrations were monitored at different locations in the powerhouse. The vibration levels were compared with the computed values from Equation 1, which was derived for excavation of the same cavern (Gupta, et al, 1987):
V= 64.23(D/√Q)-0.80, r = -0.86 where
V = peak particle velocity (mm/s),
D = radial distance from blast to monitoring station (m),
Q = maximum charge per delay (kg) and
r = correlation coefficient.
The measured peak particle velocities were very much less than those computed from Equation 1. As the ground vibrations generated by ramp blasting were lower, the damage to the surrounding rock mass was controlled.
Monitoring movements
Single and multipoint borehole extensometers were installed at locations around the cavern to regularly monitor the behaviour of the surrounding rock mass in the cavern. The instrumentation readings did not indicate any considerable change in the rock mass response during and after removal of the ramp. This shows that damage to the surrounding rock mass was controlled.
A glass plate was fixed on the river side wall of the bus gallery 1 before commencing the removal of the ramp. It was at this location where the distress was maximum. After each blast, the glass plate was observed for crack formation. No cracks showed before completion of the excavation, confirming that blasting damage was controlled.
During removal of the ramp, even though the over break was negligible on the downstream wall of the cavern, the drill hole impressions obtained were not satisfactory. This may be due to the unfavourable joint orientations on the downstream wall. However, the profile of the downstream wall, as confirmed by the survey, was satisfactory and there was no noticeable damage to the wall rock.
Conclusions
The presence of shear zones and distress problems in the cavern at the Sardar Sarovar Project complicated the excavation of the ramp. The following conclusions were drawn:
Related Files
Fig 3. Drilling and Blasting pattern
Fig 4. Drilling and Blasting pattern – vertical
Fig 4. Key
Fig 1
Fig 3. Key
