Some people call alkali-aggregate reaction (AAR) ‘concrete cancer’, and you can see why — it does behave rather like a disease. The first symptoms are usually cracking on the surface of the material. These get wider and more numerous and the concrete starts to expand. High stresses develop in the concrete structure with often alarming results, including distortion and spalling of the concrete.

‘AAR was only diagnosed in the UK for the first time in the mid-1970s, and since then, several hundred structures have been diagnosed as having it,’ says Professor Jonathan Wood, director of British firm Structural Studies and Design and a specialist on AAR in concrete. He points out that most of these structures — including buildings, bridges and car parks as well as some dams — are still in service. But he warns that the incidence of AAR in a concrete dam is a more serious problem than finding the same damage in, say, a bridge or a building.

‘Dams are sensitive to very small expansions,’ explains Wood. He points out that, because they are such large structures, the movements caused by expansion due to AAR is compounded in three dimensions. Dams are also massive — in comparison to structures like bridges — and seldom contain much steel reinforcement.

‘While an expansion of 1.0mm/m may only cause marginal damage in a well reinforced structure, expansion of less than 0.5mm/m can cause serious operational and serviceability difficulties and significantly reduce overall safety margins in dams,’ says Wood. ‘Dams are basically big lumps of concrete with turbines and generators buried inside them,’ he adds. ‘Fairly small expansions of the concrete can cause enough distortion to jam or otherwise damage these machines.’

There is plenty of evidence to bear this out. Wood cites two structures on the St Lawrence River in North America — the Sanders and Beauharnois dams — both of which have suffered from AAR to such an extent that machinery was rendered inoperable due to distortion of the concrete in which the components were bedded.

Wood believes that dams might be exposed to a higher risk of damage from AAR than other structures because current design standards are not appropriate to dam construction.

‘If you take the normal precautions recommended to guard against AAR in most concrete structures, they might not be sufficient for dams,’ he says. He cites the standard concrete prism test required in the UK which deems concrete acceptable if a sample shows less than 1% linear expansion after one year immersed in water. But he points out that in Canada — where AAR has long been recognised as a threat to dams — the standard demands no more than 0.4% expansion after two years. ‘A dam is a long-term structure; they are meant to last for generations, so you have to be more rigorous with their design,’ says Wood.

Experience from Kariba

Among the dams to suffer from AAR is Kariba in Zimbabwe. Built about 40 years ago, the dam is a 620m long, 130m high concrete arch impounding some 160,000M m3 of water in one of the largest man-made reservoirs in the world. The dam is suffering from AAR which causing swelling of the concrete — of particular concern around the six sluice gates.

Monitoring the expansion is British consulting engineer Gibb. Principal engineer Andrew Dodd says that the appropriate remedy for this particular case of AAR is repair on an ‘as-needed’ basis. Dodd points out that AAR does not necessarily threaten the strength of the concrete itself in the same way that, for example, chloride attack of steel reinforcement can. Swelling of concrete can increase the compressive forces within the concrete and, as concrete performs well in compression, it will not automatically lead to a weakening of the structure.

‘The arch dam is extremely robust and is not threatened by the current level of swelling,’ explains Dodd. The structure itself is quite sound, with only minor cracking. Where it has caused problems is at the interface between concrete and machinery, where the swelling has caused jamming of the sluice gates at the lintel seal and jamming of the stoplogs in their slots.

Andrew Dodd’s ‘as-needed’ repair has so far extended to breaking out and repairing the concrete at the sluice gates to give clearance, grinding and adjusting the stop beams and adding an epoxy coating to the sluice waterways to reduce water ingress.

Dodd — who is currently working on a MPhil thesis on the treatment of deteriorated concrete dams — stresses that there is no simple remedy to the problem. The solution to AAR in any dam will depend upon a combination of factors — how old the dam is, its location, the extent of the damage, the design of the structure and, of course, economics.

Reto Baumann, an engineer with Swiss firm Kraftwerk Brusio, is currently working on three cases of AAR in concrete dams. One is the dam at Illsee, near Geneva; the other two are the North and South dams at Lago Bianco, 20km from St Moritz.

‘At Lago Bianco we identified AAR about five years ago,’ says Baumann. ‘On the North dam there appeared horizontal cracks on the upstream face which we first thought were caused by frost damage.’ Experiments carried out by Swiss testing house TFB confirmed the presence of AAR in the concrete. ‘We feel this is unusual,’ says Baumann. ‘Five years ago, people were saying you did not get AAR at low temperatures.’ But the Lago Bianco dams are 2200m up in the Alps and temperatures seldom exceed 5°C.

Damage to the North dam is serious, says Baumann, with cracks worsened by the effects of frost. Repair work, which is starting in April, will involve breaking out the worst-affected concrete and replacing it. Work will take three years and can only be carried out during the summer months. The Lago Bianco South dam is less badly affected than its companion and will be treated with a synthetic polymer membrane on the upstream face which will reduce water ingress and slow down the AAR reaction. This is the solution previously chosen for the Illsee dam, work on which has recently finished. In both cases, the affected concrete is left intact.

Andrew Dodds’ work at engineering consultants Gibb has involved several dams affected by AAR, and nearly every one has required a different solution. In a recent paper, written with colleague John Sawyer, senior executive engineer at Gibb, Dodds compared the remedies chosen for several AAR-affected concrete dams in different parts of the world. These range from the minor repairs carried out on Zimbabwe’s Kariba dam, through the use of galvanised spiral strand cable to stabilise and support the 23m high Dinas dam in Wales, to complete replacement of the Maentwrog dam, also in Wales.

Dodds concludes that each case of AAR should be properly assessed and the likely consequences of the problem addressed, determining what functional and safety issues are compromised. Most importantly, Dodds asserts that each case should be considered individually. ‘Appropriate solutions,’ he says, ‘can range from monitoring and observation to complete replacement.’

Prevention is, of course, better than cure. When it comes to new dams, common sense should help designers avoid the problems associated with AAR, according to Wood. ‘If you start out saying “there is a risk”, you can set about minimising it. With a big project like a dam, you have to use the aggregates nature gives you, but you can still quarry selectively to avoid any seams of highly reactive rock. Then you must look at your cement and see if you can use less cement, perhaps by incorporating pulverised fuel ash or another low-alkali substitute, or by using a leaner mix.

‘And if you’ve still got a risk of AAR, you must look at the consequences: are you designing a dam which will be sensitive to expansion? You can do a lot in the detailing to reduce the impact.’

In the end, like Jonathan Wood, Andrew Dodd believes that the fear of AAR can be more worrying than the problem itself. And, he assures us, ‘there is life after AAR’.