The tie-in works comprised the excavation by drill and blast methods of two new 6m diameter access adits from a surface platform to 8m diameter tie-in chambers approximately 80m below ground level.

The tie-in chambers were excavated to leave a 4m thick pillar of which approximately 1m comprised prepacked concrete backfill between the tie-in chamber and existing 1950s pressurised penstocks to allow excavation of two new 4m diameter penstock connection tunnels from the tie-in chambers to a new guard valve chamber located approximately 100 to 60m away.

Through further design analysis when the geological conditions were better understood the pillar width was subsequently reduced to 3m to assist the construction schedule while the penstocks were still pressurised. The 3m pillar was then removed once the rest of the tunnelling had been completed and the penstock was depressurised.

The design needed to consider the geometrically complex proposed tie-ins as well as passing below an existing access adit, with a maximum of 5m middling. Coupled with limited information on the condition of the 1950s penstock pipe and thickness and surrounding pre-packed concrete backfill, added to the complexity on the type and nature of the constraints and the development of practical blasting criteria that could be successfully implemented by a contractor.

The rock mass typically comprised strong (125MPa) blocky locally very block and seamy Granodiorite where intersected by dykes and altered shear zones. An average rock mass rating, RMR89, value of 65 and typical Geological strength index, GSI, values ranging from 60 to 80 were recorded at the Tie-In Chamber locations.

Development of the drill and blast criteria that would allow a contractor to successfully excavate and support the tie-in chambers, while the penstocks were still pressurised, and subsequently remove a pillar of rock and packed concrete backfill to expose a depressurised penstock to allow the tie-in works to be undertaken, within a very tight construction window, without damaging the penstock pipes, was a key challenge. Consideration was given to blasting effects, predicting the vibrations and the amount of strain on the pipe and the amount of control that could reasonably be placed on the drill and blast excavation methodology to reduce the risk.

Standard industry damage criteria and ‘safe levels’ of ground motion are generally based on peak particle velocity (PPV) and frequency of motion. General safe PPV limits for pipes generally range from 76 to 127mm/s. Since blasting needed to accomplish this work would occur within a metre of the pipes, calculations of PPV for minimum practical charges indicated levels would exceed 2,500mm/s. This motion seems extremely high and at first glance pipe failure would be expected.

Particle velocity values should not be confused with ground displacement. For instance, if a measured peak particle velocity for very near-field blasting is 2,000mm/s, the ground has not moved 2,000mm because ground particles disturbed by blast vibration waves will oscillate back and forth many times in a second. This is why frequency of motion is important because, unlike earthquakes where frequency of motion is quite low, cycles of ground particle shaking (frequency) caused by careful blasting in hard rock at distances of 1 to 5m usually occurs between 100 to 1,000Hz. Since the ground particles are shaking back and forth or up and down so quickly, similar to running in place, they do not move very far. At any given PPV, the amount of elastic particle displacement is inversely proportional to frequency (f). With the sinusoidal motion created by blast-induced vibration, displacement can be approximated by the formula below where ‘f’ is motion frequency.

Peak Displacement (mm) = PPV/(2pf) [1]

For example, a seemingly high PPV of 2,000mm/s occurring at a frequency of 500Hz would cause a temporary elastic particle displacement of only 0.64mm [2000 / (2 x 3.14 x 500)]. To convince other engineers reviewing the risk of the work that blasting could be accomplished safely while working outside typical PPV limits, microstrain in the pipe was modeled based on blast-induced ground displacement. While considering the unknown conditions of the penstocks and potential for adverse rock mass conditions, a reasonable factor of safety was established if peak ground displacement at the pipe did not exceed 1mm. Since vibration monitoring equipment cannot directly measure displacement, and velocity transducers could not handle the high frequency and PPV ranges, accelerometers with a 500g and 1Hz to 3kHz range were used to perform primary measurements of ground motion. Displacements were estimated using sinusoidal motion relationships.

When explosive charges detonate in rock, they are designed so that most of the energy is used in breaking and displacing the rock mass. However, some of the energy can also be released in the form of transient stress waves, which in turn cause temporary ground vibration. Detonating charges also create rock movement and release of high-pressure gas, which in turn induce air-overpressure (noise), airborne dust and audible blast noise. In the very-near zone, if holes are filled with bulk explosives or tightly tamped charges crushing usually occurs in the rock around the charge. The extent of this compressive and shear failure zone is usually limited to one or two charge radii (half the diameter of the charge).

Beyond the plastic crushing zone, the rock or ground is temporarily deformed by elastic strain waves. For some distance, tangential strain intensity exceeds the rock’s strength and new fractures are created. The magnitude of dynamic strain and particle motion decrease as distance from the charge increases. Radial cracks are created in rock around detonating charges as a result of induced strain that exceeds the rock’s tensile strength. These cracks for fully-charged holes generally do not extend farther than 13 charge-diameters. For instance, if the diameter of the charge is 38mm, radial cracks might extend 494mm into adjacent rock. Direct rupturing or overbreak of rock beyond the desired limits of a blast area might also occur if ground is weak or jointed and or poor perimeter control methods are used for blasting. For this work, much less radial cracking was expected in the packed concrete due to the discontinuities at the boundaries and varying properties of aggregate surrounded by weaker cement. This result was later confirmed when exposed concrete was very cohesive and showed little crack damage.

Early on, it was decided that the expected concrete with a thickness of 1m (+/-) around pipes would be excavated by mechanical methods. Hence, the blasting done near the pipes had to be designed to assure blast- induced pressure would not cause direct rupturing of the penstocks.

Charge decoupling and relief
With the understanding that direct rupturing of ground or high gas pressures occurring near the pipes would cause damage, all charges placed near the pipes were designed with excellent relief so blasted rock can easily move toward open rock faces and away from the existing pipes. When holes are fully-charged with explosives that fill or nearly-fill the hole, rock or concrete around the charges is crushed and fractured because the explosion generates pressures as great as 400,000 psi (2.75 x 106kPa), which greatly exceeds the strength of rock and concrete. The physical properties of concrete are generally consistent so blasting results are more predictable.

Since the borehole pressure of fully-coupled charges exceeds the dynamic strength of rock or concrete, the size of charges used in holes located at and near excavation perimeters and the penstocks were reduced to control damage. Similar to controlled blasting methods like Pre-splitting and Smoothwall blasting, borehole pressures were reduced by using extremely decoupled charges. For the very close-in blasts, a series of decoupled charges were used to lightly slash rock away from surfaces where clean break was desired. Fast millisecond delay timing was also used to create a cooperative cleaving effect within each row of very lightly-charged holes.

Blasting limitations
At the outset of the work, very specific limitations regarding minimum scaled distances, maximum charge weights, hole-diameters, charge diameters, PPV and maximum calculated displacement were specified at locations where blasting took place within 20m, 3m to 5m and 1m to 3m of existing structures. These limitations are summarised in Table 1.

In addition, in the 1m to 3m zone burden distance for all charges was set to not exceed 0.5m and the delays set to assure best charge relief. Holes were to be orientated on vectors away from or tangential to the penstock pipe. These limitations would typically be developed by a contractor to comply with PPV and frequency limitations as part of their working methodology. Specifying these limitations together with the input of a blasting specialist in the development of the project provides more control and assurance to owners when blasting close to their existing assets and assist in the managing the risks at the design phase.

Compliance monitoring
Triaxial accelerometers were grouted into the toe of boreholes drilled to locations centered about the tie-in excavations about 1.5m above each penstock. Since all blasts would have varying proximity to the closest section of pipe and the measurement points, extrapolation calculations based on site-data-curves were used to predict the intensity of motions at the closest part of the pipe. Conventional PPV measurements were also done at other more distant points of concern.

As blasting progressed towards the pipes, regression curves were continuously updated to establish site-specific relationships between the dependent variables of distance and charge-per-delay; and frequency of motion, PPV, acceleration and predicted displacement. Upper 95 per cent confidence curves were used to establish the compliance of all blasts with the 1mm displacement limit at the pipes and with other PPV limits for existing openings and other facilities.

High frequency geophones were also used as a back up to verify the blast data from the triaxial accelerometers.

The high frequency geophones were mounted on the walls to as close as possible to the blasting with protective plates. Measurements of wall-mounted geophones typically recorded higher motions in directions normal to the wall surfaces. This was attributed to a greater degree of freedom due to the open wall surface. The location of each instrument was surveyed so the distance to the blast and the nearest structure or penstock could be determined before each blast.

The raw blast data was analysed using advanced software to identify the dominant frequency, which may or may not have been related to the peak acceleration or velocity and to discount any erroneous readings such as impacts of fly rock to wall-mounted velocity transducers. The results of each blast were reviewed by the project team and the results were compiled into a working spreadsheet continuously updating curves based on respective acceleration, PPV and predominant frequency measurements. With further information being made available from geological face mapping of the advancing excavations and laboratory testing of core taken at the location of the tie-in chambers to confirm the ground conditions it was possible to review the blasting limitations and progressively relax some of the limitations based on measurements and results. Agreements by all parties to the work were confirmed before implementing changes to specifications or relaxing blasting limitations. The first relaxation was given to reduce the scaled distance for zones 1 to 3m and 3m to 5m as well as allowing drilling orientations to be directed in the general direction of the penstock pipe, with a number of quality control conditions such as establishing an appropriate survey control.

Control was implemented on site by setting out each hole by survey and orientation and depth guides were surveyed and demarcated using paint and ribboning to allow the jumbo operator and miners sight in the drilling of each hole. In addition, holes were then checked upon completion of drilling to ensure there were no deviations and allow the contractor to finalise their blast plan. It was critical that the drill hole-depths were carefully controlled to avoid any unnecessary over excavation that could have exacerbated the depth of potential geological over break in the pillar face while the penstock was pressurised. Given the orientation of the main discontinuities, geological over break in the form of face wedges were foreseen through kinematic analysis and were going to be difficult to control.

This was considered by the design when considering the safe pillar width and the over break that occurred in localised wedges up to 0.6m deep were within the permissible tolerances set out by the design and did not lead to any destabilisation of the pillar while the penstock was still pressurised. By comparison, with similar jointing but generally better ground, less over break was witnessed on the second tie-in chamber pillar. These better results occurred despite allowing reduced scaled distances and increased charge weights.

As works progressed in the 3m to 5m zone the results from the blast monitoring indicated that the charge weight per delay and scaled distance could be further relaxed for the 1m to 3m zone that were predicted not to adversely affect the penstock pipe. Advance planning and discussion also meant that the contractor was also prepared for such relaxations and had procured appropriate materials to allow them to blast four rows at once and help control vibration. This allowed a significant saving to the construction schedule to be afforded. Such a relaxation remained conditional on the importance of quality assurance and quality control on the drill depths, amount of burden and survey data that were required to be put in place.

Due to the complex geometry of the tie-in locations, some blasting was done above and below the penstock pipes. To create needed relief, the contractor chose to use a Brokk excavator equipped with a hoe-ram to carefully excavate rock and concrete located within a metre of the pipes. Vibration by the Brokk excavator was very low and caused no damage.

The unique challenge of excavating by drill and blast methods in very close proximity to the penstocks, with a very tight excavation schedule in hard rock required the use of special methods, measurements and controls. The applied approach dealt with the risk, made the project viable, and allowed significant reductions in the excavation schedule.