The Pointe du Bois Generating Station (GS) is the oldest hydroelectric plant operating in Manitoba. The station was constructed between 1909 and 1926 with the first unit in service in 1911. Pointe du Bois was acquired by Manitoba Hydro in 2002. The existing GS is situated on the Winnipeg River approximately 170km by road northeast of Winnipeg and 43km east of Lac du Bonnet (Figure 1).

The pre-project spillway facilities consist of 97 spillway/sluiceway bays, all of which are manually operated, with the exception of six remotely-controlled bays used to manage minor inflow variations.

Despite extensive ongoing maintenance and upgrades over the years, the preproject facilities at the Pointe du Bois GS required major repair or replacement for the following reasons:

  • to meet current dam safety standards,
  • to provide a safe work environment for staff,
  • to ensure reliability of power production.

New primary spillway
In the fall of 2012, Manitoba Hydro initiated construction of a new sevenbay gated spillway to replace the existing spill facilities. The new concrete structure is approximately 125m long and consists of concrete piers (3m wide), rollways, bridge structures and dedicated steel water control gates and towers. The new spillway is capable of passing a total safe flow of 5,040m3/s at a maximum forebay elevation of 299.7m. The approach and discharge channels for the new spillway have a curved shape, measuring 690m in length on the outer curve and 430m on the inner curve. The maximum depth of excavation is approximately 23m. The approach and discharge channels serve to direct the water flow into and out of the spillway.

Structure and site safety
The existing structures remained in-service to control the river during construction, and required protection during blasting operations for excavation of the new spillway and channels. These include the existing sluiceway structure located 100m distance from the nearest blast; existing spillway and east gravity dam between the sluiceway and the powerhouse, located 600m from the nearest blast. In addition, an existing residential and recreational community is also situated in close proximity. The existing site and facilities are shown in Figure 2.

In order to preserve the integrity and safety of the more than 100-years-old structures, the ground vibration criteria was initially set at 10mm/sec; this was gradually increased several times on a case-by-case basis as the blasting operations progressed and monitoring results were reviewed. In addition, flyrock control measures were enforced to ensure safety of the construction site, workers and adjacent community.

Contractual arrangement
KGS Group Consulting Engineers (KGS) from Winnipeg, Manitoba, and Peter Kiewit Infrastructure Group (PKI), were the prime consultant and contractor to Manitoba Hydro for the spillway replacement project. KGS was responsible for overall project design and provided quality assurance assistance to the client during construction. PKI was responsible for all construction activities, as well as for the contractual arrangements with their subcontractors. To facilitate final design and implementation of construction, Manitoba Hydro adopted an "early contractor involvement" (ECI) model. The ECI contract delivery model is where an owner, engineer and contractor work in a process of open communication and collaboration on a project to develop the scope of work, final design, cost estimates and schedules. The key benefits of this model are a focus on the best value for the project, mutual sharing of risks and balanced contract development.

Bedrock structure
The majority of the spillway excavations encountered hard rock of granitic to grano-dioritic origin, which were generally considered to be of "very good" quality. The rock is brittle in nature, with a network of discontinuities (joints). Amphibolites and narrow coarse-grained granite pegmatite dykes were encountered within the bedrock excavations. The majority of the joints have a dip ranging from 70 to 90 degrees. To a large extent the network of joints controlled the stability of the spillway channel walls by creating an assembly of "partially" or "fully" formed blocks, requiring a pattern of rockbolts (6m-long William’s hollow-core tensioned to 130-146kN); while the more massive portion of the bedrock (free from joints) remained stable without any support.

The upper portion of the bedrock structure exhibited horizontal exfoliation joints, requiring a large amount of scaling and loose block removal after excavation, with vertical, un-tensioned rock dowels to stabilise potentially unstable large blocks that could not be feasibly removed.

Excavation plan

Volume of rock to be removed
The blast program entailed a total of approximately 682,000m3 of solid bedrock excavation; the excavations consisted of three main segments: Main Channels and Structure Footprint, in-the-dry (560,000m3); Approach Channel in-the-wet (62,000m3) and Upstream Rock Plug (60,000m3). The specifications, collaboratively developed for the project, required the contractor to perform a series of trial blasts to optimize the drill-hole pattern, hole depth, spacing, burden, explosive type, explosive quantity, blasting sequence and delay pattern. Benching operations were specified not to exceed 9m height.

Excavation model
For the main channel excavations completed in the dry (approach and discharge channels), the excavation plan comprised two primary benches broken up into 46 shots for the upper bench, and 56 shots for the lower bench (Figures 3 and 4).

Through the new spillway structure area, three production benches and a buffer section were used on each side of the spillway to facilitate better control for the structure footprint.

As illustrated in Figure 5, the top bench comprised blasting the center of the primary structure (from original ground surface down to elevation 294m) with no sub-drill. The middle bench consisted of blasting the center of the primary structure channel (between elevations 294m and 287m). The drilling pattern for both benches was set at 2.4m x 2.4m, using 102mm diameter drill holes, loaded with bulk emulsion explosives supplied by Austin Powder, and was generally consistent with the main production benches for the rest of the channel. The bottom bench comprised the final grade of the spillway structure. Excavation for this bench was accomplished by a combination of line drilling, pre-splitting and production blasting, using 150mm sub drill. Pre-splitting on each side of the buffer within the top and middle benches was done in advance of production blasting, both for the spillway structure and for the main channel excavations, to maintain a high tolerance in wall control.

Drilling rates
Drilling for production blasting was performed with several different drill models, including: Tamrock Ranger-DX800; CATMD5090; Ingersoll Rand ECM590; PowerROC T35; Furukawa HCR1500 and Agassiz Drilling DX800. Typical penetration rates in hard granite ranged from 0.3m/min to 0.7m/min; averaging 0.5m/min. As an example, a 102mm diameter hole, 23m deep, was completed in 11 minutes on February 13, 2013 in the cold winter weather (temperature minus 25oC). The 0.5m/min average is considered reasonably good progress, considering that a large portion of drilling was performed during sub-zero temperature conditions.

Perimeter blasting

Drilling and blasting criteria
Technical specifications required that the maximum acceptable perimeter hole deviation remain within 0.15m of the vertical alignment of the drill hole. The overall averaged final grade tolerance for the spillway walls and the floor was required to protrude no more than 50mm inside of the neat-line grades.

In addition, any localized individual protrusion was required not to extend more than 100mm inside of the neat-line for the concrete structures, and no more than 150mm for the channels (outside of the concrete structure area).

Field optimisation of drilling and blasting pattern
As part of the contractor’s trial blasts, variable hole-diameters and spacings were initially experimented with to determine the optimum configuration and achieve the best results while minimizing the cost. The initial test section for pre-splitting was approximately 60m long with holespacing varying from 500 to 600mm using 70mm diameter holes, and from 700 to 800mm using 89mm diameter holes. All holes were drilled and blasted to the full depth of the channel (including 0.9m of over-drill).

Visually inspected rock faces following the trial blasts indicated the pre-splitting performance was better with the larger diameter holes.

The best results were achieved when using 89mm diameter holes at 800mm spacing, with the corresponding shear factor of 0.58kg/m2. Within the 70mm diameter hole section, at 500 and 600mm hole spacing, pre-splitting performance appeared less effective; the rock faces were generally rougher and more irregular compared to the 700mm and 800mm spacing. This is largely attributable to the much lower shear factor in these sections where the holes were loaded with Primaflex detonation cord as opposed to Emuline packaged emulsion. In addition; the drill-hole survey showed that misalignment was also more prevalent in the 70mm diameter holes.

The drilling alignment in 89mm diameter holes was generally within the specified deviation criteria of 0.15m, which is partly attributed to the stiffer steel rods used for drilling. Table 1 summarises the results of perimeter blasting experiments.

In-Water Blasting
The upstream portion of the approach channel required rock removal in a submerged condition. This required approximately 1,980 blast holes. The area was divided into three separate blasts; each approximately 700 holes; 480 holes and 800 holes, respectively.

Staging and drilling for blast-holes was done through a rockfill mattress, which entailed pre-placement of approximately 68,000m3 of excavated rock to facilitate access to underwater bedrock, which required removal. The field procedure included installation of temporary steel through the mattress portion (down to the bedrock surface); installation of PVC pipes to the bottom of the blast hole; withdrawal of steel casing and loading of explosives within the bedrock portion, only. Blast holes for each blast were 100mm in diameter, drilled on a 2.1m by 2.1m pattern. To ensure there were no tights or high spots in the floor of the channel, a sub-drill of 2.1m (equalling the burden) was used. The blasting agent was Subtek Intense bulk emulsion, supplied by Orica. With recommendation by Orica; the programmable i-kon detonators were used to initiate each blast.

In order to meet the environmental overpressure criteria for protection of aquatic life of 50kPa (max), the Contractor was initially required to maintain a 40m setback distance for the exclusion zone.

This distance was later reduced to 20m after determining the pressure pulse would be decreased by 50 per cent as it passed through the rock mattress (determined through monitoring conducted by Explotech, May 13, 2013). Figures 6 and 7 illustrate the concept of mattress blasting.

Rock plug removal
A final upstream rock plug of approximately 60,000m3 in volume was blasted in a single shot. The plug was located upstream of the main spillway structure at about 130m distance measured along the west channel wall and 275m measured along the east channel wall. The plug measured about 14m in height; 30m in width and about 130m in length. The rock within the plug was generally competent granite, containing bands of amphibolite and horizontal jointing in the upper 2-3m.

Prior to blasting the rock plug, the approach channel between the plug and the new spillway was flooded to the forebay FSL of 299.1m, to minimize the chance of a water surge flowing towards the spillway after the blast. The water overpressure and vibration limits for the new gates and the spillway structure were set at 50kPa and 75mm/sec, respectively. Mitigation of water overpressures in relation to the spillway gates was done by installation of a double bubble curtain. Assuming the line was successfully weighted down to the channel floor, it was estimated that the double bubble curtain would yield a total overall reduction of overpressure levels in the order of 60 per cent (Explotech, August 3, 2013).

The plug was successfully removed in a single shot with no damage to the new spillway from pressure waves, vibrations or flyrock. Subtek Intense was again used as the explosive agent for the rock plug shot. Drilling was done on a 2.1m by 2.1m pattern, using 100mm hole diameter. All holes were sub-drilled 2.1m below the design grade of the channel. Just prior to the plug blast, a line of pre-shear holes was established on each side along the channel walls.

A total of 800 holes were required for plug removal, resulting in 100kg of explosive per delay. The i-kon programmable detonators, supplied by Orica, were chosen on the basis of accuracy and timing flexibility; assigning one hole per five millisecond delay interval.

Conclusions
Manitoba Hydro, KGS Group and Peter Kiewit Infrastructure successfully completed the excavation phase of the spillway replacement project.

The key benefits of the ECI model adopted at this project were the timely planning of field operations, logistics and work schedule, along with management of the risk which lead to relatively close proximity blasting and removal of 682,000m3 of solid bedrock with no damage to any of the existing or new structures.

The remainder of the work on the project entails construction of new earthfill dams to connect the new spillway with the existing powerhouse, replacing the water retaining structures, scheduled for completion in 2015.