Growing old gracefully

8 October 1999



The integrity of concrete dams not only depends on a good design but on the quality of concrete used for construction. Tim Dolen from the US Bureau of Reclamation gives a US perspective on ageing concrete


The US Bureau of Reclamation’s 348 dams and reservoirs are an important component of the utility’s concrete infrastructure. But with an average age of 50 years, and with the oldest concrete structures approaching 100 years of age, time is not on Reclamation’s side. What is more, over half of its concrete dams were constructed before advances in concrete materials technology were identified to withstand ageing from sulphate attack and alkali-aggregate reaction, as well as to resist freezing and thawing.

The material properties of ageing dams may degrade with time. Although these dams may meet our safety requirements today, they may not in the future if deterioration is allowed to continue unabated.

A focal point of Reclamation’s work at the present time is a research project called the Decision Support System for Ageing Concrete. This will gather information from a historic database of concrete materials properties, from the Dam Safety Information System and from facilities’ operations and maintenance records. The decision support programme is a database system that will access information on concrete materials used to construct dams, power plants, pumping plants and water conveyance facilities. It will identify those most susceptible to ageing.

One of our goals will be to place our concrete structures on a time line for ageing concrete. This will help identify the structures constructed before major advances in concrete materials technology and improved construction practices. By constructing a time line for projects we can identify the most likely ageing mechanisms for individual structures; ultimately determining those in need of rehabilitation.

Some of the main causes of ageing and factors which affect the rate of deterioration of Reclamation’s concrete dams include:

•Quality of the original concrete construction.

•Amount and severity of repeated freezing and thawing cycles (in the presence of moisture).

•Exposure to soil and groundwater containing sulphate-bearing minerals or acids.

•Quality and mineral composition of the aggregates and cement.

•Unanticipated events — floods, earthquakes, or overloading.

•Number of wetting and drying cycles.

Quality of original construction

The quality of original concrete at the time of construction greatly influences the ageing process. As the state-of-the-art in concrete technology and construction methods improved, so did the resistance to deterioration and ageing. Early twentieth century concretes were not as well controlled as they are today. Generally, the quantity of cement and water used in the paste determined the strength and relative durability of the resulting hardened concrete. If too much water was added, the paste would be weaker and more porous, leading to poor durability. Because early cements were expensive and hard to obtain, some attempts to economise the mixture by reducing the cement content and adding water for fluidity led to poorer quality concrete.

Pre-World War II concretes were far less durable than the ‘modern’ concretes of today. Significant advances in concrete materials technology occurred between 1932, when construction of Hoover dam took place, and World War II. Low heat, sulphate resisting and low alkali cements and air-entrained concretes improved the durability more than five-fold. Ageing concretes are most often those early twentieth century concretes constructed before the advent of new methods and materials.

Freezing and thawing cycles

The number of freezing and thawing cycles experienced by concrete when saturated directly affects its durability. Most pre-World War II concretes cannot withstand repeated saturated freezing and thawing (FT) cycles. These concretes can only withstand about 100 to 200 cycles before failure of the cement paste matrix. Concretes in Arizona that only experience one or two cycles of FT may be durable for many years. However, most of Reclamation’s projects are constructed in climates that experience as many as 50 to 100 cycles of FT each year.

Large dam structures have considerable mass and experience surficial FT deterioration without jeopardising safety for years. However, spillways, canals and other appurtenant structures can suffer from serious deterioration within a decade after construction. Thin arch and buttress dams constructed up to the 1930s have more problems due to freezing and thawing deterioration.

The discovery just before World War II that a small percent of air bubbles can absorb FT expansion resulted in a five-fold increase in FT durability. Modern, air-entrained concrete can resist 500 to 1000 cycles of FT without failure.

Soil and groundwater sulphates

The harsh western climate of the US causes accumulations of sulphate-bearing minerals to concentrate in the soil and groundwater. These can be seen as the white ‘alkali flats’ across the west. Sulphates attack the reaction products formed with tri-calcium aluminate hydrate (C3A) in hardened cement paste. This results in the formation of ettringite, followed by expansion and cracking of the concrete. Cracks allow even more infiltration and accelerate the process.

Early twentieth century cements did not limit the percent of C3A and were very susceptible to sulphate attack. Some canals and pavements failed in two or three years due to severe attacks. The process was aggravated by more porous mixtures and repeated wetting and drying.

Initial investigations into sulphate attack occurred during laboratory investigations for construction of the Hoover dam. Researchers found that low heat cements formulated with C3A content also had greater sulphate resistance than standard Portland cements. Cements purposely formulated to reduce sulphate attack by limiting the C3A to less than 5% first appeared in the late 1930s. The introduction of Type II and Type V cements, as well as making more impermeable mixtures with higher cement and lower water contents, essentially eliminated sulphate attack in Reclamation structures by about 1940.

Aggregate-cement composition and reactions

Most sands and gravels are sufficiently dense and durable to withstand physical and chemical attack. Certain aggregates can be dissolved by chemical reactions with alkali compounds in cements. Although most cements are composed of less than 2% alkalis, the chemical reactions between certain primarily siliceous minerals such as cherts, opals, andesites, rhyolites, and water form an expansive gel.

The dissolution of the aggregate and volumetric expansion of the gel cause map cracking of the concrete called alkali-aggregate reaction (AAR), or in the Western US alkali-silica reaction (ASR). AAR only affects concretes containing certain percentages of reactive aggregates in the presence of high alkali cement and suitable temperature and moisture. Internal expansion within the concrete causes cracking, which allows moisture to enter and continue the reaction or promote other forms of deterioration such as freezing and thawing. Expansion and movement of the mass concrete itself may cause mis-alignment of hydraulic machinery, closing of joints, and binding of spillway and outlet gates. Stewart Mountain dam in Arizona expanded and moved almost 30cm upstream and 15cm upward. The upper portion of Friant dam, California, expanded until spillway gates were inoperable and had to be replaced. In the most extreme case American Falls dam, constructed in 1927, was deemed unsafe primarily due to AAR and was replaced in 1978. (See IWP&DC, April 1999 pp28-29).

A similar type of reaction can occur between alkalis and certain carbonate aggregates and is called alkali-carbonate reaction. Alkali-carbonate reaction is not a significant problem in the Western US but is a major problem in some eastern parts of North America. Limiting the alkali content of cements to 0.6% or less in 1942, being selective in aggregate choices, and Reclamation’s historical use of good quality pozzolans essentially eliminated AAR as a major durability concern by 1950.

Unanticipated events

Unanticipated events can be considered those caused by acts of nature (floods and earthquakes) and those caused by acts of man (errors in design or construction, overloading or operational changes). These events are normally rare and single occurrences that must be corrected by remedial measures. One common event is abrasion/erosion damage in spillways and outlet works caused by high water releases, either man-made or flood induced, combined with sands and gravels in the river channels.

Cavitation of concrete due to high velocity flows can cause severe deterioration in spillways and outlet works. These forms of deterioration often require structural modifications to rectify the problem. Cavitation damage at Glen Canyon dam spillways required substantial repairs and construction of air slots in 1984.

Wetting and drying

Wetting and drying cycles cause slight volumetric expansions and contractions in the cement paste. Reclamation’s water conveyance facilities experience frequent fluctuating water surface levels depending on irrigation or power demands. Most concretes do not experience deterioration from wetting and drying alone, however, it can accelerate other deterioration mechanisms such as sulphate attack or FT deterioration. Frequently, the fluctuating water zone or spray zone is the first concrete to suffer damage.

Growing old gracefully

Ageing in general can be considered the accumulation of all of the above deterioration mechanisms over time. The weaker the concrete, the faster damage occurs, and the faster the structure ages. Many durable concretes were constructed before modern advances in concrete technology. These naturally durable concretes had favourable environments that were not destructive or were composed of cements and aggregates that happened to be chemically resistant when manufactured. Ageing concretes are those less fortunate in either their environmental exposure and/or their chemical composition.

The effects of ageing normally begin where moisture is present, adjacent to construction or contraction joints or cracks, and at exposed surfaces such as the tops of dams, parapets, or spillway and outlet walls. Leaking lift lines and contraction joints in older dams are often the first areas to show freezing and thawing deterioration. If the outer, damaged layer is removed by rain or wind, freezing and thawing deterioration progresses inward and shows lenses and cracking parallel to the exposed surface.

Repairing the physical effects of ageing may vary from initiating a monitoring programme and not taking any further action, to repairing cracks and replacing damaged concrete, to major rehabilitation efforts.

Major dam safety modifications are being initiated using risk assessment techniques first to fix those structures with the greatest risk. Concrete repairs often are only a part of a major dam safety modification project.

Some of the major dam modifications carried out by the Bureau of Reclamation due to ageing include:

•Installing vertical tendons through Stewart Mountain dam (to repair alkali-aggregate reaction and dis-bonded lift lines), Arizona. The dam was constructed in 1936 and modified in 1992.

•Adding a downstream concrete buttress to Santa Cruz dam, (to counter FT), New Mexico. The dam was constructed in 1929 and modified in 1990.

•Rebuilding the spillways at Cold Springs, Ochoco, and Tieton dams (to counter FT), Oregon and Washington. These were constructed between 1908-25 and modified between 1995-9.

Reclamation engineers perform annual inspections of our dams, including a comprehensive facilities review of each structure every six years. Evidence of cracking, disintegration and movement are noted throughout a dam’s inspection history. Concrete coring and testing programmes may be conducted to determine the overall condition of the concrete and identify any specific ongoing deterioration mechanisms. The test programme includes determining material properties for static and dynamic strength and elastic properties; lift line bond strength in tension and shear; and petrographic examination. Concrete coring programmes may be designed to give an unbiased sampling for determination of the overall material properties or a biased sampling to compare areas of good quality concrete versus deteriorating concrete. Updated material properties are used for performing a revised structural analysis for current static, hydrologic (flood) and seismic loadings.

Mistakes

Most of the fundamental developments in concrete dam and materials technology were accomplished in the first half of the twentieth century. The first generation of concrete technologists left us with an abundance of knowledge of the fundamentals of ‘modern’ concrete required to withstand ageing. Methods to improve concrete quality such as utilising the appropriate water to cementitious materials ratio; using special cements and pozzolans to resist sulphate attack; identifying potentially reactive aggregates and using low-alkali cement; and utilising air-entraining admixtures were all identified and published before the end of World War II.

Our primary mistakes since then have been due to poor availability of suitable concrete-making materials or a lack of education and training within our profession. Technical societies and organisations such as icold, American Society of Testing Materials, Portland Cement Association and American Concrete Institute have been at the forefront of both the dissemination of information and development of education and training programmes for most of this century. However, education concerning the fundamental concepts of durable concrete construction is a continuing need for our profession.

Concrete advances

Some of the developments that resulted in improved concrete dam performance in the US include: Identification of the fundamentals of concrete mixture proportioning for strength and quality related to the ratio of water to cementitious materials (Abrams, 1918). Elemental analysis of cement compounds and the development of low heat cements for mass concrete (Bogue, 1929; Boulder Canyon Studies, 1933). Development of petrographic examination techniques for analysing concrete and aggregates (Mielenz and White, 1936 to 1941). Identification of the cause of sulphate attack and limiting the C3A content of cements (Thordvaldson, 1927; Tuthill, 1936). Utilisation of pozzolans to improve concrete quality (Davis, 1935). ASTM specifications for five types of cement based on chemical composition (1940). Identification of the cause of alkali-aggregate reaction leading to the development of low-alkali cements (Stanton, 1942). Identification of the cause of freezing and thawing deterioration and development of air-entrained cements and admixtures (1938 to 1945). Studies on effect of pozzolans (Class N, F, and C) on alkali-aggregate reaction and sulphate attack (Kalousek, 1948-1972). Reclamation and other federal agencies require use of 15 to 25% pozzolan (ASTM C 618 Class N, F, or C) to improve resistance to alkali-aggregate reaction and sulphate attack (Central Arizona Project, 1970). Advances in concrete dam construction have also improved dam performance, for example: Mechanised concrete manufacturing equipment improved rate of concrete placement and reduced the number of cold joints (1900 to 1920s). Development of weigh-batching plants and separation of aggregates into individual size gradings (1920s). Utilisation of contraction joints to reduce thermal cracking in concrete dams (1920s). Improved cleaning and treatment of horizontal lift lines between placements (late 1920s). Mechanical vibration for concrete consolidation (1933). Process quality control techniques applied to concrete manufacturing (Hoover dam, 1936). Development of block construction techniques and post-cooling of mass concrete (Owyhee, 1932 and Hoover dams, 1936). Improved designs and development of techniques to minimise cavitation damage in high velocity spillways (Yellowtail dam, 1968).


Quality assurance worldwide

Members of the hydro power and dams industry give their thoughts on concrete quality assurance. Chris Oosthuizen, chief engineer of civil design (dam safety), Department of Water Affairs and Forestry in South Africa 'Our department owns approximately 190 ICOLD listed dams, of which 112 can be classified as concrete dams or dams which have large associated concrete works. I have been involved in dam safety since its inception in our department in 1977 and have inspected most of these dams. 'The main cause of ageing/deterioration is carbonation and forms of AAR. In tunnels aggressive water is the main cause of deterioration. Our experience is that these conditions are more susceptible to pore water than cyclic warming (winter/summer). We have only had one case where superficial deterioration of concrete has been ascribed to extreme cold. Heat has not been a problem for us as yet.' Talking about repairs to ageing concrete Oosthuizen says that only minor repairs have been necessary for South Africa's dams. In most cases AAR has stabilised with time but as its long-term effect is unknown it is monitored extensively for behavioural patterns. Oosthuizen added that due to the scarceness of water water in South Africa it is very unlikely that a reservoir would ever be drawn down or emptied to carry out repairs. The quality of concrete for dam construction from the outset is also vital. 'When we had identified AAR and associated problems we realised that cement being produced in the southern part of South Africa was the main contributor to the problem,' Oosthuizen explains. 'Problems with batching plants, where scales for weighing materials are not calibrated regularly can and have led to localised areas of poor quality concrete as a result of an incorrect mix of proportions. 'The average ambient temperature of packing large volumes of concrete is also important as this could lead to cracking. This is particularly pertinent in regions where high temperatures occur, and would also apply to countries which experience very low temperatures.' BJ Parmar, secretary of Narmada Water Resources and Water Supply Department, India Parmar believes that there is no fine dividing line between good quality and deteriorating concrete in dams. 'Concrete is manufactured with natural ingredients and there may be a vast variation in properties,' he said. 'But what needs to be established is whether the concrete still possesses the properties actually needed to ensure the integrity of the dam.' He went on to explain that cracked concrete does not transmit loads properly and so the continuity of the concrete mass needs to be re-established. A large scale rehabilitation was preformed at El. Atazar dam in 1977 to rectify this and since then 20 dams have been repaired worldwide using the grouting technique. Reflecting on AAR, Parmar gave the example of Chambon gravity dam in France. This was suffering from AAR and to limit leakage through the upstream face of the dam a 2.5mm thick watertight membrane was placed on the surface to slow the rate of the reaction. Armando Camelo, Department of Materials for Construction, Hidrorumo, Portugal Camelo believes that the quality of concrete used for dam construction is very important. 'In fact,' he says, 'It is vitally important for dam owners to have a quality policy, such as ISO 9002, which should be included in tender document specifications for quality management and control. The audits for measuring quality are vital.'




Privacy Policy
We have updated our privacy policy. In the latest update it explains what cookies are and how we use them on our site. To learn more about cookies and their benefits, please view our privacy policy. Please be aware that parts of this site will not function correctly if you disable cookies. By continuing to use this site, you consent to our use of cookies in accordance with our privacy policy unless you have disabled them.