Project overview
Purpose: Hydropower and irrigation
Client: Sri Lankan Ministry of Irrigation and Water Resources Management
Contractor: Farab Company (Iran)
Subcontractor: Nimrokh Company (Iran head office)
Designer: Mahab Ghodss (Iran) and Poyry Energy Limited (Switzerland) joint venture
Methodology: Drill-and-blast, Atlas Copco 352 twin boom jumbo

The Uma Oya Multipurpose Development Project lies in the south-eastern part of the central highland region of Sri Lanka (Figure 1). For this project, the Ministry of Irrigation and Water Resources Management of Sri Lanka is the client. Farab Company from Iran is the principal contractor. A joint venture of Mahab Ghodss Consulting Engineering Company (Iran), and Poyry Energy Limited (Switzerland) is responsible for the design and supervision of the project for the principal contractor. The hydroelectric project includes dams, tunnels, roads and an underground power station, all constructed by various subcontractors.

The Uma Oya Multipurpose Development Project is intended to convey water from the Uma Oya River to the Kirindi Oya basin for irrigation and hydroelectric energy generation. The project scope includes:

  • Puhulpola Roller-Compacted-Concrete (RCC) Dam, with a height of 35m, constructed across the Uma Oya River
  • Dyraaba RCC Dam, with a height of 50m, across the Mahatotilla Oya River
  • Link Tunnel between the two reservoirs of the dams, with a length of 3.7km
  • Headrace Tunnel, 15.5km in length
  • Vertical pressure shaft, approximately 600m deep
  • Surge shaft upstream of the vertical pressure shaft, approximately 200m deep
  • ound powerhouse including powerhouse cavern (Length: 70m, width: 20m, height: 35m); transformer cavern (Length: 37m, width: 14m, height: 13m) and associated tunnels (Figure 3)
  • Tailrace Tunnel, 3.6km in length
  • Main Access Tunnel to powerhouse cavern, 2km in length.

This case study is for the excavation of the main access tunnel (MAT) which was constructed between November 2011 and February 2013, to provide access to the powerhouse for both construction and during operation. The main characteristics of the MAT are:

  • Length: about 2km
  • Gradient: 11.61 per cent (downwards)
  • Tunnel size (width by height): 8 x 7m.

Geology

The project lies entirely in the Precambrian Highland Complex, the central of the three main tectonic complexes of the island of Sri Lanka. The high grade and multiphase metamorphic bedrock derives from supra-crustal (former sediments) and intrusive (magmatic) material.

Three main geological units with uniaxial compressive strength of 30 to 150 MPa encountered in the project area are:

  • Gneiss including:
    • Garnet Quartzofeldspathic Gneiss
    • Undifferentiated Charnockite Biotite Gneiss
    • Biotite-hornblende Gneiss
    • net-sillimanite-biotite Gneiss
  • Marble
  • Quartzite

Usually, two main joint systems were found in the tunnel, one parallel (the dip is sub-vertical) and the other perpendicular (the dip is about 50 to 60 degrees from horizontal) to the tunnel direction. Occasionally other random joints or fault planes were found but generally the intersections of the discontinuities made a block rather than a wedge.

Excavation and Rock Support

Four excavation classes were planned, with rock support consisting of rock bolts and shotcrete to varying extents for excavation classes I, II and III. For excavation class IV, rock support is provided using steel ribs for more adverse ground conditions (Figure 5).

During construction, as more accurate information became available, some modifications were made on the rock support classes, adjusting these to be representative of the as encountered ground conditions. Seventy-five percent of the tunnel length was constructed in good to excellent rock, corresponding to excavation classes I and II. The remainder of the tunnel was constructed in mostly moderately good rock, requiring excavation class III.

The method statement for the MAT was initially based on a heading & bench excavation using drill and blast methods. The heading and bench parts had areas of 37.12m2 and 12m2 respectively. Drilling was carried out with a twin-boom drilling jumbo (Atlas Copco-352) with a drilling length of 3 to 3.5m. Later, full-face excavation of the tunnel was adopted and proved to be more productive. Also, as will be explained later, the length of the holes was increased to 3.9m.

On completion of the learning curve for the tunnel, with all site establishment and logistics now in place, the team began an in-depth study into ensuring the optimum drilling and blasting performance.

Benefits of Full-Face Excavation over Top Heading and Bench

Due to the good quality rock encountered in the MAT, in terms of rock stability, full-face excavation was feasible. Also, in view of the steep slope of the tunnel (11% downwards), water ingress in the tunnel (between 15 to 20 litres/sec) could cause delays during the bench excavation.

Excavation of the bench was time consuming and the pump pipe line, electricity cables and ventilation duct had to be removed prior to blasting. They had then to be re-installed after blasting, during which time the tunnel would be flooded by water inflow from the rock mass, resulting in several hours’ delay whilst the tunnel was dewatered. With full-face excavation, these problems were never faced.

Also, tunnels with a larger cross section usually have a lower powder factor (quantity of explosive required per cubic meter of ground excavated) compared to those with a smaller cross section in the same ground condition.

Therefore, changing the top heading and bench to full-face excavation helped to reduce the powder factor.

Design of the Drilling and Blasting Pattern

The critical factor in blasting in tunnels is the lack of an available free surface towards which breakage can occur (only the tunnel heading itself). The principle behind tunnel blasting is to first create an opening by means of a cut, usually at the centre, which provides a breaking surface for the surrounding holes.

Holes were charged with “Superpower 90” emulsion cartridges (40mm diameter, 390gm) from Solar Industries (India) Limited. To suit the 40mm emulsion, which was readily available to the project, the hole diameter was reduced from 52mm to 45mm.

Jimeno et al, in the 1995 book “Drilling and Blasting in Rocks” describe how the Powder Factor (quantity of explosive required per cubic meter of ground excavated) is given by:

As can be seen, moving the hole diameter closer to the explosive diameter increases the Coupling Factor (η1), and therefore reduces the Powder Factor (C). Changing the drilling diameter from 52mm to 45mm reduced the powder factor, producing real cost savings. Moreover, drilling a smaller diameter hole, reduces the drilling time.

The most critical holes to ensuring a successful blast are the ‘cut’ holes (usually at the centre), which, once blasted, provide a free breaking surface for the remainder of the holes. Parallel holes with four-section cuts were selected as they proved to be easier to drill and control, and do not require a change in the feed angle of the jumbo and therefore the advances are not constrained by the width of the tunnel, as happens with angled (or ‘wedge’) cuts.

In four section cuts, to minimise the powder factor, the optimum depth of the blast holes can be estimated by the following equation (Jimeno, et al., 1995):

When cuts of NB empty holes are used instead of only one large diameter drill hole, the former equation is still valid with:

As this formula shows, to achieve a minimum powder factor, the depth of the hole is related to the diameter of the empty hole. The diameter of the empty hole was increased to 128mm by drilling four 64mm holes separately (64*√4 =128). As a result, this enabled increasing the depth of the blast holes from 3.0/3.5m to 3.9m, which was the maximum possible depth of the drilling for the jumbo with one rod. This therefore improved the drilling efficiency.

Blasting pattern based on a four-section cut 

The four-section cut, which is close to the Swedish model, is based on the parallel hole cut. This model was first described by Langefors and Kihlstrom (1963) and has been further developed since. Holmberg published the complete blast design model in 1982 (Holmberg, 1982) which was subsequently updated (Persson, et al., 2001). These studies suggest the experimental equations listed in Table 1. In this Table, E and X are drilling error and length of each quadrangle side respectively. Due to the relative ease and precision in the drilling of parallel cut holes, E is taken as zero. The value of E for stopping and perimeter holes can be calculated by equation 5 (Konya, 1995):

In this model, the holes in the face are divided into separate sections as cutting holes, stopping holes, perimeter (roof, floor and wall) holes. The four-section cut method includes an empty hole in the centre as shown in Figure 8. The optimum lineal charge concentration is calculated from equation 7 (Jimeno, et al., 1995) :

For the cut holes, the Table 1 equations were used to calculate the optimum hole spacings. However, for the other holes the equations do not cover the controlled blasting which was required for the tunnel to minimise damage to the surrounding rock as the tunnel had minimum rock support with no in situ concrete lining. Therefore, the equations were used to give an initial indication only of the spacing and burden. Then controlled blasting requirements should be considered.

Controlled blasting involves, but is not limited to, a closer spacing of contour holes which are loaded more lightly than the other holes.

A common misconception is that the only step required to control blasting damage is to introduce pre-splitting or smooth blasting techniques. However, controlling blasting damage starts long before the introduction of pre-splitting or smooth blasting.

A poorly designed blast can induce cracks several metres behind the last row of blast holes (Hoek, 2006).

The more recent approach to controlled blast design is based on adding a new row of holes known as buffer holes between the contour and stopping holes with a reduced quantity of explosives in the stopping, buffer and contour holes. Older control blast designs are based on closely spaced perimeter holes using the blast hole diameter to determine the perimeter spacing and burden, and are not based on the explosive quantity or type for the perimeter holes, and do not consider the effect of damage by the other holes in the tunnel face.

As a rule of thumb, in controlled blasting, burden and spacing of the buffer and contour rows should be about 75 per cent and 40 per cent respectively of that of production rows (Sharma, 2009). The designed pattern was tested in the tunnel and some adjustments were then made due to the practical result after blasting. The final pattern is shown in Figure 9.

Implementing The Designed Pattern In The Tunnel
Drilling
Design targets cannot be achieved without accurate execution of the drill pattern in the tunnel.

This is a critical part of the process and requires specific attention. Drilling and blasting should be closely monitored in the tunnel to make sure it is being carried out in accordance with the design.

The results should then be reviewed and, based on the rock reaction and practical results in the tunnel, the pattern should be reviewed again. Reviewing the pattern should be repeated whenever the condition of the ground (including geology, Rock Mass Rating (RMR), discontinuities, etc.) changes.

While drilling, the accurate position of the holes is important as well as the angle and the length of all holes. Cleaning the holes by air pressure was found to be of paramount importance to ensure that the holes were cleaned of the water that they gained during drilling (due to the 11 per cent downward slope of the tunnel).

The holes were cleaned by inserting a steel pipe with a diameter of 1 inch and connecting it to a compressed air hose. Due to the quantity of water there was no dust issue using this method. As well as removing the water and cleaning, this helped to better compact the explosives inside the holes.

Charging
The detonators used were non-electric detonator. It is common practice to use millisecond (MS) detonators for parallel cut holes as the cut holes are close to each other and blasting one hole can cause some cracks in the surrounding rock and, if there is a long delay, this may allow the explosion gas to escape, dissipating the power of the blast. By using the MS detonators (instead of half second [HS] detonators), there is not enough time for the explosion gas to escape from these cracks.

Based on surveys of the early blasting and monitoring the team during the charging stage, it was realised that the explosive compaction needed some improvement. There should not be any gap between the explosive cartridges in a hole.

Stemming of the blast holes is the other important item which needs to be carried out properly.

Some major improvements were carried out in the process of stemming, including the length of stemming, quality and thickness of clay used, and compaction method (using a 2m round wooden stick for stemming instead of the 4m semi-flexible plastic stick that was used for the compaction of the explosives). In this drilling pattern, with empty holes close to the cut holes, a small connection between these holes (empty) and cut holes (charged) could provide a route for the explosion gases to escape.

As a precautionary measure, the first few metres of the empty holes were stemmed with clay to prevent gas from escaping. These holes remained fit for the purpose, as just the entrance was blocked.

Summary

With the re-designed pattern and careful control of the drilling and blasting methodology, the powder factor was reduced significantly to around 1.9kg/m3.

This improvement in the pattern then led to the following advantages for the project:

  • Reducing the explosive usage and consequently decreasing the costs.
  • By reducing the explosive usage, the negative effects on the rock mass outside the tunnel were reduced; and consequently the required support system was also reduced.
  • Reducing the time required for scaling, reducing the excavation time, and therefore increasing the production rate.
  • The blasted rock was of a more consistent size, with far less powder (caused by using extra explosives). The extracted rock was graded to the appropriate sizes and utilised in various places in the project such as road substructure, rock fills, embankment material for roads and to produce aggregates of different sizes for use in the concrete works in the project.
  • Reducing the negative environmental impacts due to a reduction in the explosive usage.

The principles outlined in this study have been taken forward into the remainder of the project with significant benefits.