ABSTRACT

Blasting of a 76cm diameter shaft was required to create an artificial entrance into Carroll Cave. A 6-hole and an 8-hole blast design were produced around an existing 23cm pilot hole. The burden was too large on the 6-hole design leading to poor pull and fragmentation (blockages of pilot hole). The 8 hole design caused too many shelves to be developed and poor pull because drill cuttings flushed into the drillholes before loading. An optimised design was developed using 6 holes being charged as two separate 3-hole blasts. This significantly increased pull to consistently over 80% and reduced the number of pilot hole blockages.

INTRODUCTION

Carroll cave is located in central Missouri, approximately 220km southwest of St. Louis. First discovered in 1957, the limestone karst feature is thought to have over 32km of passages. It is home to many geologically and biologically significant features, including a bats graveyard, schools of blind fish and types of limestone structures that are yet to be classified by science (Weaver 1980). Future studies have been planned for the cave, however, the only known natural entrance to the cave lies on private property, onto which access has not been granted. Consequently an alternative entrance is required.

CARROLL CAVE CONSERVANCY

The Carroll Cave Conservancy (CCC) is dedicated to the conservation and 'perpetual protection' of Carroll Cave and its associated structures as well as construction of a new entrance to the cave. It was established in 1998 by a group of cavers who had been trying since 1995 to find an alternative entrance into the cave. Previous attempts to gain access to the cave centred on mechanical excavation of a sinkhole (Hines 2002).

AIM

The aim of the project was to create an artificial entrance from the surface into Carroll Cave at a location known as the T-Junction. As the project was being carried out by the CCC, whose role it is to protect the cave, any disturbance to the geology and biology of the cave was to be minimised, if not prevented. Consequently, airblast, ground vibrations and flyrock entering the cave due to blasting were minimised. The shaft was to be 33m deep with a minimum diameter of 76cm. CCC had already sunk a 23cm pilot hole at the blast location in previous attempts to jackhammer through the rock.

GEOLOGY

Carroll Cave forms part of a large structural dome known as the Ozark Plateau; which forms the major geological feature of Missouri, northern Arkansas and eastern Oklahoma. Located within the Gasconade Dolomite, Carroll Cave is immediately below the basal sandstone member of the Roubidoux Formation, believed to be the controlling structure in the cave's formation. There are three zones of cave formation within the local stratigraphy, all of which fall immediately below such sandstone units. The stratigraphy overlying the cave consists of flat-lying, interbedded dolomites, cherts, sandstones and clay seams, which are thought to be Upper Cambrian to Early Ordovician in age (Orndorff, Weary & Sebela 2002).

INITIAL BLAST DESIGN

It was assumed that the 23cm pilot hole would be used as both a conduit for the extraction of the broken material from the shaft and the initial free face for the burn-cut style blasts. Two blast designs were initially produced, a 6-hole design and an 8-hole design, both of which were designed according to method suggested by Langefors & Kihlstrom (1978). This method uses Figure 1 to estimate the required burden for the blast design.

Figure 1 Burdens required for a burn cut (adapted from Langefors & Kihlstrom 1978)

The 6-hole design (Figure 2) provides the maximum allowable burden for the free face provided by the pilot hole. It has less drillholes, so less explosive is required and cycle times are reduced. As can be seen the calculated burdens (1.5 times the pilot hole diameter) are too small for the required shaft diameter, so the pattern is expanded to 76cm in diameter. This forces the burden into the ‘breakage’ region rather than the preferred ‘clean blasted’ region (Figure 1).

Figure 2 6-hole blast design

The 8-hole design lowers the burden and spacing, increasing fragmentation (Figure 3). It is blasted on two separate delays, reducing the likelihood of hangups in the pilot hole. The reduced burden and spacing also allows for inaccurate drilling.

Figure 3 8-hole blast design

TESTING

A number of tests were carried out to provide information on the strength parameters of the rock as well as the degree of influence each of the blasting variables has on the ground vibrations produced by the blast. The first such test was the point load test to estimate the Uniaxial Compressive Strength of the rock. The results displayed a lognormal distribution with a mode of 80.27MPa, a mean of 99.44MPa and standard distribution of 36.74 MPa.

Brazil tests were carried out to estimate the tensile strength of the rock. Tensile strength is the most important rock property with respect to blasting as rocks break under tension, not compression (Worsey 2001). Results were separated into 3 categories to show the heterogeneity of the rock mass.

1. Clean fracture in fresh rock - 12.7MPa

2. Clean fracture in perforated rock due to the leaching of carbonates - 8.55MPa

3. Fracture along existing discontinuities - 3.99MPa

The initial designs assumed that the natural scatter produced by a 6.4sec, #12 detonator would be sufficient to produce a millisecond delay. As all burdens were the same, the order of initiation was not considered important. This assumption was tested to determine its viability. Results showed a normal distribution with a much higher delay than anticipated of approximately 7.5s (mode – 7.48s, mean – 7.52s). This higher delay is thought to be due to excessive time spent in the magazines. The standard deviation of 89.8ms suggests the delays produced are sufficient for a successful blast.

The final tests to be carried out were vibration tests on some of the blast parameters that could be adjusted to improve the blast performance. These tests aimed to identify which parameters had the maximum impact on the reduction of ground vibrations, in order to optimise the blast design in a more efficient manner. The parameters tested included powder factor, stemming, decking, drillhole diameter, burden and spacing. Geophones were placed at 7.6m and 15.2m away from the blast to represent the nearest limestone structures to the pilot hole.

Results from the vibration tests (Table 1) show that the powder factor had the greatest influence on the production of ground vibrations. Increasing the powder factor by as little as 20% is seen to increase the ground vibrations as much as 150%. It should also be noted that introducing a cast booster as a primer also reduced the level of ground vibrations. Adjusting the stemming, burden and spacing also showed a considerable reduction in PPV. The only parameter that didn’t show a significant change in round vibrations was adjusting the hole diameter. A number of drillholes were set up using decks to lower the charge per delay. Though all decks had sufficient stemming according to the literature, all of the holes sympathetically detonated. It is suggested that the rules of thumb regarding the stemming length of decking drillholes breaks down as the drillhole diameter is reduced.

Table 1 Summary of vibration test results

			7.6m		15.2m
Blasthole	Powder Factor		PPV (mm/s)	% Change	PPV (mm/s)	% Change
1	3 water gel chubs		-11.18	46.67	-15.24	150.00
2	2.5 water gel chubs		-7.62	0.00	-6.10	0.00
3	2 water gel chubs		-5.59	-26.67	-6.10	0.00
4	2 water gel chubs + 1 dynamite stick		-15.24	100.00	-9.65	58.33
5	2 water gel chubs + 1 cast booster		-5.08	-33.33	-4.57	-25.00

Blasthole	Stemming		PPV (mm/s)	% Change	PPV (mm/s)	% Change
2	33cm dirt		-11.18	0.00	-6.10	0.00
6	18cm dirt		-8.64	-22.73	-6.60	8.33
7	33cm air		-12.70	13.64	-9.65	58.33
8	33cm water		-8.64	-22.73	-8.64	41.67
9	33cm fine gravel		-8.13	-27.27	-7.62	25.00
10	33cm coarse gravel		-8.64	-22.73	8.13	33.33

Drillhole	Hole Diameter		PPV (mm/s)	% Change	PPV (mm/s)	% Change
13	38.1mm		-3.05	0.00	-4.57	0.00
16	44.5mm		3.56	16.67	-4.57	0.00

Drillhole	Burden	Spacing	Max	% Change	Max	% Change
3	46cm	61cm	-5.59	0.00	-6.10	0.00
13	91cm	122cm	-3.05	-45.45	-4.57	-25.00

BLAST OPTIMISATION

In order to optimise the design, a number of performance parameters were established to measure the success of the blast design. They included ground vibrations, pull, fragmentation, backbreak, cycle time, simplicity of design, cost and safety. Safety was deemed the most important parameter due to the inexperience of the workers undertaking the project.

The optimisation process was carried out simultaneous with the excavation of the shaft on a trial and error basis. The individual variables were altered one at a time to assess their impact on the blast performance. There were two main problems with the blasting process, poor pull and blockages of the pilot hole. As the blasted material was being flushed into the cave through the pilot hole, blockages would stop the excavation process until the pilot hole could be unblocked. The main causes of blockages were too much material trying to flow through the pilot hole at once and poor fragmentation. Damage to the structures within the cave also had to be monitored.

There were a number of problems associated with the 6-hole design. Having all holes on the same delays saw too much material flow through the pilot hole simultaneously, creating blockages. Introducing separate delays eased the problem, however, the burden and spacing of the design was still too great to create sufficient fragmentation. This problem was exacerbated as the drillholes needed to be angled to keep a constant shaft diameter. Good pull could not be achieved because the burden at the toe was too great to fragment the rock (Figure 4).

Figure 4 Increased burden at toe results in poor pull

While the 8-hole design solved many of the problems associated with the 6-hole design, it introduced a series of new issues that had to be considered. The reduced burdens and spacings increased fragmentation, however, the addition holes increased the drilling time significantly, leading to two main problems. The additional drill cuttings tended to flush into the adjacent drillholes before loading of explosives. Pull was therefore unpredictable as the loaded length of each drillhole was different. This lead to many shelves being produced in the shaft as well as an uneven floor. By this time the shaft was well below the water table. The additional drilling time and abundant water reduced the driller's comfort, leading to inaccurate drilling, which affected the blast performance.

The optimised blast reverted back to a 6-hole design, however, it was drilled and charged as two separate 3-hole blasts. Because two blasts were required, any shelves produced by the first blast could be taken out with the second blast. The first 3-holes had a slightly smaller diameter to keep the burden below 1.5 times the pilot hole diameter to ensure sufficient fragmentation (Figure 5). It significantly increased pull to consistently over 80%, often gaining 100%. Though the two blasts increased the cycle time of the design, the additional gains in pull meant the advance rate was actually quicker than previous designs.

Figure 5 Optimised blast design

Drillholes for the final blasts were 107cm deep and 38mm in diameter. The three inner holes could be drilled vertically, which are more accurate than angled holes. Reduced drilling times improved the quality of drilling, further improving the accuracy of drilling. Each drillhole was charged with 4 sticks of extra-gelatin dynamite with shelves charged at one stick per 30cm. Drillholes were initiated with a separate LP delay to control the amount of fluid passing through pilot hole at any time. Drill cuttings were used for stemming, though groundwater was always present, increasing coupling.

BREAKTHROUGH

The last 1.2m was excavated using mechanical methods. Though this method is much slower, it was expected to reduce ground vibration, airblast and flyrock that could potentially damage the limestone structures within the cave. It was also expected to reduce the backbreak in the cave roof, improving the structural integrity and working life of the shaft.

RECONCILIATION

Inspection of the cave after the construction of the shaft showed no visible signs of damage to any of the caves structures. Monitoring of the ground vibrations created by the optimised blast design showed that the maximum induced PPV was 21.34mm/s. This was 50% higher than the suggested destruction limit for a cave formation of 15.2mm/s (Kirkbride, Worsey & Rupert 1995), suggesting more research is required in this field. The shaft walls and cave roof immediately surrounding the shaft have proven to be structurally sound, though long-term damage is still being monitored. The air blast values recorded were surprisingly low, all below 100dB. It is thought that the separate delays caused the mullock flowing through the pilot hole acted as a muffler for the flowing blasts.

CONCLUSION

In conclusion, it is suggested that Langefors and Kihlstrom's method for confined blast design be revised for different host rocks. Their testing was carried out in Swedish Granites, which has different blasting properties to Missouri Limestones, leading to many of the problems associated with the initial blast designs. The geometry and not the powder factor was deemed to be the most important parameter in improving the blast design as geometry is what distinguishes between fragmentation rather than fracture. Figure 6 shows just how little impact the powder factor had on the fragmentation of the rock. Finally, separate LP delays should be used for each of the drillholes, as too much material flowing through the pilot hole at once will cause blockages.

[1] Student, School of Geosciences, Minerals and Civil Engineering, University of South Australia, Mawson Lakes Campus, Adelaide S.A. 5095, Australia