Can rockbolt design lives of 100 years or more be justified? The evidence so far is that it can, says David Baxter, providing that bolts are rigorously corrosion protected and installed, but this has been the case on few projects.
Rock tunnel projects increasingly adopt single shell linings. The primary support, rockbolts and shotcrete, must provide both immediate ground support and permanent support.
Although the criticality of bolts varies with the shape of the tunnel and the ground competency, their longevity is a prime issue since maintenance of rockbolts is impossible. If they are major supporting elements appropriate design lives of the bolts may be 100 years or more, depending on the practicality of replacement and the criticality of the system’s operations.
Durability issues for a 100-year design life for rockbolts are problematical, because the records of such materials used as permanent support are short.
There appear to be no worldwide standards for rockbolting or rockbolt corrosion protection. Guidance can be obtained from national standards, particularly for ground anchors.
One of the most rigorous is BS 8081, British standard code of practice for ground anchorages. The standard is for highly stressed anchorages which are primary supporting elements, but it does refer to rockbolts. The principals of corrosion and corrosion protection are the same.
Although raising the questions of safety versus cost of protection, BS 8081 recommends that as a general rule permanent anchorages (design life over two years) should be protected because there is no way of predicting corrosion rates precisely. Current (UK) practice ranges from double protection, implying two physical barriers against corrosion in aggressive permeable soils, to simple grout cover for low permeability non-aggressive rock or where low capacity bolts are secondary reinforcement.
Single protection implies one physical barrier before installation, while double protection provides a second barrier to ensure an undamaged inner barrier during handling and placement. For high capacity anchorages in low permeability rock at least one physical barrier should be used. The code notes that in some cases providing an alkaline environment has given satisfactory performance.
Protection is a once-only treatment, since it can be neither replaced nor monitored. Total enclosure to exclude a moist gaseous atmosphere is the usual solution for ground anchorages, combined with coatings and materials introduced as fluids.
The code specifically excludes the use of grout injected in-situ to bond the tendon to the ground as part of a protective system because the quality and integrity of the grout cannot be assured. In most underground projects the only corrosion protection system for rockbolts is the use of in-situ injected grout which bonds the bolt to the grout.
Under-stressing the anchorages creates cracks in the grout which may lower the alkalinity of the environment. The code suggests that corrosion may be inhibited, provided the upper limit of crack width is 0.1mm. This is for cement grouts only, since inert resin provides no protection. The code reports a scarcity of technical information on epoxy or resin based grouts for anchorages.
Corrosion occurs at pits or surface irregularities, so using thicker metal gives no effective protection. Bonded metallic coatings can be used, but sacrificial coatings such as galvanising should not be used where load is transferred by interfacial bond. The code does not recommend mechanical bolts for permanent anchorage because of likely damage to coatings.
The draft Eurocode prEn 1537: 1996 — Execution of special geotechnical work: ground anchors requires anchor coatings to last the design life of the anchor. This requirement would severely limit the use of coatings such as galvanising or epoxy which might substitute for sheathing.
Stillborg reports that permanent bolts are generally specified as being doubly protected against corrosion and that in most countries this means galvanising the bolt and fully encasing in resin or cement grout. Stillborg goes on to state that around 50% of all bolts installed have reduced or insufficient grouting (presumably from Scandinavian field research). Thus automatic protection from cement or resin grouts alone cannot be assumed. He concludes that a permanent bolt must be coated with an elastic corrosion resistant coating, so the grout provides the anchoring mechanism and prevents water flow along the bolt.
Atlas Copco reinforces the view that grouted rebars are not ideal for fractured rock since the grout disappears into cracks and encapsulation is not complete. Atlas Copco also refers to reports that galvanised fully encapsulated resin or cement grout bolts are fraught with problems. For instance, galvanising gives minimal protection for pH<6.5-7.0. The integrity of resin or cement grout encapsulation is unpredictable because of the uncheckable variables involved in the quality of installation of grouted bolts.
Hydro histories: Canada
Ontario Hydro’s Niagara River Hydroelectric Development (NRHD) required an exploratory audit and trial enlargement programme. Primary rock support consisted of (temporary) resin-grouted untensioned 2m long 25mm diameter rockbolts with welded wire. Spalling, 0.5-1.0m deep, controlled by a bedding plane within 2m of the crown, occurred three months after excavation. All local bolts were exposed and just 25% of bolt shafts showed hardened resin.
hydro-quebec has experienced bolt failures principally caused by problematic installation of resin bolts under common but imperfect conditions, or water inflow. Its experience demonstrates that:
•Quality control of resin anchored bolts is difficult in large permanent installations.
•Long resin-anchored bolts should not be used as primary support in large permanent excavations.
•Resin anchored bolts should not be longer than 3m. The above project used bolts 5-7 m long.
The doubts concerning resin grouting expressed above have been experienced in projects in Australia. In terms of corrosion protection, BS 8081 would not recognise resin grout placed in-situ to bond the bolt as a protective barrier.
Hydro Quebec’s current practice for permanent bolts (eg the 882MW SM-3 project) is to use mechanically anchored hollow groutable bolts. Grouting is through the hollow when horizontal or downward dipping and the reverse when upward dipping. Denso tape and mastic behind the face plate act as a plug during grouting and provide long term corrosion protection to the ungrouted section. Dywidag or equivalent anchors are specified with double corrosion protection if higher strength bolts are required. Most Hydro-Quebec installations are in igneous rocks and ground water has not been aggressive.
In the Delivery Tunnel South of the Lesotho Highlands Water Project the rock was stable: no concrete secondary lining was warranted. Rockbolt assemblies were designed from the start to provide longevity and to be upgradable ‘off line’.
The design was based on experience from Norway, which has had unlined hydro tunnels in operation for over 20 years. Because of the aggressive water and its proximity, the end hardware was hot-dip galvanised (river water pH6.8) and shotcrete was applied over the bolt head off-line. To counter incomplete filling of the bolt hole with grout the bolt shaft was hot-dip galvanised to 85µm thickness. Galvanising is sacrificial, but it provided cheap rugged protection.
On Australia’s Snowy Mountains Scheme bolts were expansion shells on 19mm diameter cement-grouted bolts, later superceded by hollow cement-grouted bolts with mechanical anchors, galvanised plates and fittings and surface mortar corrosion protection. Few problems have occurred and there have been few failures in 40-odd years.
The Eucumbene–Snowy Tunnel exhibited some spalling up to ten years after construction. Local damage to the gunite sometimes left rock bolts exposed, but corrosion did not occur. Some rock falls were large enough to bring down rockbolts, bent but in good condition. Note that grouting and collar protection were rigidly enforced by the client’s inspectors observing installation and grouting of every bolt. This would have been the practice adopted in many projects of this era, regrettably now replaced by Quality Assurance. The ground support worked well with very few failures, but the granite was competent and possibly many bolts were redundant. Groundwater and ground conditions are benign.
Preferred current practice by the Snowy Mountains Engineering Corporation is to use hollow groutable bolts with mechanical anchors with grout return via the hollow, not the reverse. If anchors do not hold, a resin anchor is used instead.
Tasmania’s Hydro-Electric Commission (HEC) operates many hydroelectric projects. At Poatina, completed in 1965, the cavern spans 13.9m supported by systematic rockbolts and shotcrete. The rock reinforcement consists of 4.3m and 3.7m long slot and wedge type bolts with attached tubes for cement grouting.
Currently passive grouted dowels are used for rock support installed at the face. Centralisers have been used in the walls but not the roof because they get stuck and have to be hammered home. Galvanising and Molybond barrier coatings have been used for some situations. In acidic conditions, the practice is to drill the bolt hole, grout to seal it and then redrill to install the bolt.
For cement grouting, a correct w/c ratio is considered the key element. The hole is filled with a thick grout to the surface (miners otherwise underfill the hole to avoid getting covered with grout). The bolt is then pushed in. Resin capsule anchors are only used for temporary bolts: HEC’s experience is that sometimes only 10% of the bolt is encapsulated with resin.
HEC has caverns up to 25 years old, and bolt inspections indicate sound conditions.
In Sweden, grouted dowels have traditionally been used. The hydro plants owned by Vatterfall involve about 1.5M m2 of tunnel roof, 15% belonging to plants built before 1970. Three types of bolt have been used : tube injected; Perfo; and grouted dowels (pushed into the hole filled with a stiff grout). The commonest are 25mm diameter ribbed bars. Simple grouted dowels without corrosion protection have been used since 1958. Inspections have shown only minor rock falls due to malfunctioning bolts.
Water tunnels in Swedish hydro schemes are unlined and shotcrete is only used to a small extent in power houses and other underground installations. However, inspections and systematic maintenance on a 4-5 year basis involve scaling and complementary support such as selective bolting, replacement of malfunctioning bolts and shotcreting. Grouted dowels seem to be more reliable than the types mentioned above.
For the Drakensberg pumped storage scheme power house fast-set resin anchorages proved to be best suited to the weak bedded strata. The 50t capacity 26mm diameter Dwyidag threadbar primary bolts were double corrosion protected and the 30t capacity secondary bolts were epoxy coated and full column grouted.
In theory cement grout offers excellent corrosion protection, but Scandinavian work suggests installation problems may reduce its effectiveness. Grout defects and unpredictable integrity lead to corrosion.
The many corrosion mechanisms indicate that bolts will eventually rust. From standards for ground anchors the minimum corrosion protection required for a design life greater than 50 years is one physical barrier. Double corrosion protection will assure a design life approaching 100 years. Grout that bonds to the rock is not acceptable as a physical barrier because of the inevitable defects.
Thicker metal sections give no effective protection because of the accelerated effects of pitting corrosion. Galvanising is not recommended because of its sacrificial nature, possible effect on bond and vulnerability to damage.
Practice in tunnelling varies, at base treating all bolts as temporary and relying on a secondary concrete lining to ensure a long design life. Problems with resin anchoring appear to be widespread.
A key element in ensuring quality is close scrutiny by an inspector. Reliance solely on testing and audits of paperwork is no guarantee of quality. Hydro projects mostly adopt simple grouted bolts, sometimes galvanised. Civil projects use a more rigorous approach to corrosion protection.
| At the Laboratory of Mining Engineering, Helsinki University studies were carried out including pullout tests on experimental bolts and overcoring of in-situ bolts up to seven years old in five mines. The studies showed that: •Using the wrong type of bolt caused many defects.
•Cement and resin grouted bolts both had cracks, voids and major lack of material.
•There were large pieces of resin cartridge cover between the grout and borehole wall.
•Cement grouting seemed the best protection against corrosion.
•Corrosion damage occurred only at the proximal end of the bolt where the cement grout was often very poor.
•Cracks and empty spaces deeper in the cement grout had not led to any corrosion.
•Free parts on resin-grouted rebar were very badly corroded within two years.
Swedish Rock Engineering Research Foundation (SveBeFo) carried out boltmeter measurements in 1987 of 52 operational bolts and over coring of 16 bolts in the Centralgruvan mine. The bolts included cement grouted bolts (up to 20 years old), resin grouted bolts (5-6 years old) and uncoated Swellex bolts (seven years old).
•Around 50% of cement grouted bolts had insufficient grouting quality.
•Severe pitting took place in one bolt within the grouted section, outside mainly general surface corrosion.
•Ungrouted bolts showed less surface corrosion.
•Resin grouted bolts indicated increasing rust formation with age.
•Cement grouted bolts had none or poor grouting at the distal end whereas resin grouted bolts showed the same defects at the proximal end.
In Norway, Orsta Stahlindustri (manufacturers of the Combi coated bolt and the CT bolt), used an accelerated test method done in a salt fog chamber to ASTM B117, which cannot be used to predict corrosion rates but provides comparative testing and locates possible sites for severe corrosion attack. Three groups of 20mm rebars 300mm long were grouted into granite cylinders as follows: poorly grouted eccentrically placed rebars, no cracks in the rock and one open end; poorly grouted rebars, eccentrically placed, two cracks in the rock; and well grouted centrally placed rebar, two cracks in the rock. Exposure time was 183-371 days. Reference rock bolts were also exposed for 371 days – an untreated bolt, a galvanised bolt and a galvanised epoxy powder coated bolt (Combi coated bolt).
Reference bolts tests indicated weight losses of 142g with local pitting for the uncoated bolt and of 19g with no pitting for the galvanised bolt. The galvanised epoxy powder (Combi) coated bolt showed no rust on the coats surface. Small blisters showed zinc corrosion products or rust.
Sundholm and Forsén at Helsinki University of Technology split, photographed and then chemically cleaned the samples of corrosion products. In the first group the poorly grouted open ends were all corroded, with minor corrosion on the ribs. The poorly grouted rebars in the second group showed extensive corrosion. In the crack and near the upper end of the bar, the ribs had locally corroded away. The well-grouted samples of the third group were still in good condition but early stages of corrosion were evident in the vicinity of the cracks.
Sundholm and Forsén also reported on overcoring of 41 installed rock bolts of different ages from six Finnish mines. Most of the measurable damage appeared along the first 300-350mm of the bolts. The Pyhasalmi mine, with ground water pH2.8-4.2, had bolts up to 7 years old showing little damage where the grouting was intact. The corrosion damage of 20 year old bolts from the Vihanti mine was quite low (reduction of bolt thickness 1-3.5mm) even though the bolts were installed without grouting. Beyond the first 300mm of bolt only 25% of the remaining length was adequately grouted.
Sundholm and Forsén conclude that cement grout is a protective layer against corrosion for rebar if it is intact. The protective properties are reduced considerably if there are defects in the grout which coincide with fractures in the rock serving as flow channels for corrosive media. The corrosion is worst at the end of the rebar which is closest to open space. Local attack of the rebar is very serious. After a relatively long initiation period the damage can take place at an accelerated rate.