Keeping AAR in check4 October 2002
At Orlík dam in the Czech Republic, investigations have been carried out to assess the impact of alkali-aggregate reaction
ORLÍK dam is situated on the Vltava river southwards from Prague, the capital of the Czech Republic, and forms a vital link in the Vltava cascade. Its main functions include the generation of peak load electric energy, while its large reservoir accumulation capacity facilitates efficient water management activities downstream on the river and protection against floods. It is also significant from a recreational and water sports point of view.
Construction of the straight gravity concrete dam took place from 1956-66. With a crest length of 450m, its height in the deepest point of the river valley is 90m. The volume of deposited concrete is about 1M m3 and the retention capacity of the reservoir is 720M m3, while the backwater goes 68km upstream on the Vltava river and 22km and 8km upstream further on two significant tributaries.
Flood waters can be discharged over three 15m wide chute spillways equipped with 8m high steel radial gates, situated on the right side of the dam. In blocks underneath there are two 4m diameter bottom outlets. Flow capacities of these discharge devices are 3x728m3 and 2x185m3 respectively. On the left side of the dam there is a peak load water power plant with four Kaplan turbines of 4x150m3/sec capacity and 4x91MW power output. On the right bank there is a ship transport equipment and a road crosses the top of the dam. Povodí Vltavy is the state enterprise responsible for the Orlík waterwork, its reservoir and the whole Vltava river stream including tributaries and waterworks. CEZ is the owner and operator of the hydroelectric equipment.
Concrete composition for large waterwork structures is a long term development. For the Orlík dam suitable aggregates were chosen far in advance from carefully selected mining locations. Attention was also given to reducing heat of hydratation by using special cements with low hydratation heat generation. Unsatisfactory test results by the end of 1950 led to the abandonment of this type of concrete production. A new solution was found in 1958 when a part of cement was substituted for fly ash from fossil power plants. The research work into this problem came to an end in the spring of 1959, at which time over 140,000m3 of concrete with full cement mix content had already been deposited. Over the following years the yearly volume of deposited concrete reached 380,000m3. On this site three types of concrete were used, namely the B80 core concrete with fly ash addition having iron Portland cement and fly ash content of 130kg and 50kg respectively; the B170 face concrete with 250kg cement content; and B170 face concrete with fly ash addition having cement and fly ash content of 200kg and 50kg respectively. While the 28-day strength of the B80 concrete with fly ash addition oscillated closely above a 110MPa level, the cube strength of these concretes increased after one year to 23.4MPa. Similarly, average values of cube strength of face concrete with fly ash addition after one year increased to a 33MPa level.
Over the next 40 years, extensive testing of core bore samples was performed, showing that the actual strength of face concrete varies in an interval between 40 and 50MPa. This substantiates a commonly accepted assumption of a long term maturing hydration of blended cements, moreover with fly ash addition.
The iron Portland cement was produced in the Dvur Králové cement mill and its binding power at the time of the Orlík waterwork construction varied at about 40MPa.
The prospecting and selection activities aimed at finding aggregates that would suit both the concrete mixture formula and availability requirements proved to be quite complicated. At first, attention focused on utilisation of the local rock but some of this crushed gravel and gravel sands turned out to be of unsuitable quality or not available in sufficient quantity. It was therefore decided to transport quality natural aggregates from more distant Labe deposits. The gravel sand delivered was in five fractions 0/3mm, 3/10mm, 10/25mm, 25/50mm and 50/100mm from nine deposits, and had a negative influence on the concrete mixture homogenity. Because a fine fraction of constant composition could not be provided in time and in sufficient volume, irregular grain size variations in fractions complicated the work of technologists.
Crucial volumes of concrete were deposited during 1959 and 1960. By the end of 1961 the dam was finished up to the roadway. The concrete casting phase of the construction was accomplished without major problems and the gravity dam body performed its functions without any problems in relation to watertightness and other parameter concerns over the next few decades.
Technical and safety supervisors have continuously performed standard control measurements for the whole set of parameters. In 1997 and 1998 - after 35 years of dam operation - a wide range of additional measurements and monitoring was initiated in order to clear up the cause of acoustical phenomena, experienced by operators of this waterwork since 1994.
Various working hypotheses of temperature dilatation of possible uneven settlement were examined. A plausible explanation was the expansion of certain aggregate components due to a reaction between them and alkali-aggregates in the cement. This hypothesis is supported by others with experience from the US and other countries, where a number of concrete dams have a problem with major alkali swelling, especially those built in the 1930s.
Although it generally holds true that the used gravel sands from Labe deposits should be risk-free with regard to alkali-aggregate reactions, there was not enough unambiguous experimental evidence to support this general prediction because, for the last 40 years, this problem was not systematically investigated in the Czech Republic due to the fact that probability of alkali-aggregate reaction for most rock kinds in this region was considered very low.
To investigate any potential problem, 12 pieces of core bore specimens were taken from inspection gallery No. 4 and 5 and grouting gallery No. 1. These were intended to be mainly of face concrete, having a higher cement content, and were tested by standard methods. At first they were dressed on a diamond disk saw, then the surfaces were cut and fitted with special stainless steel contact targets, attached in pre-bored holes with special two-component adhesive. During testing, the bodies were put into a special measuring stand provided with digital deformeter, and every measurement was verified on a standard.
Upon initial measurement the specimens were put into a fog chamber at a constant temperature of +40°C and relative humidity of 100%. The plot of specimen volume changes versus time clearly showed that - after a six-month period - volume changes of eight out of eleven measured bodies stabilised on a level of 0.015 to 0.020%. After this time, only three specimens from grouting gallery No. 1 and inspection gallery No. 4 show progressive longitudinal expansion. When interpreting the measured values it is necessary to compare the acquired data with generally accepted limit values that make it possible to characterise whether the tested concrete is at risk or is risk-free in relation to alkali-aggregate reaction. The commonly accepted limit value of change in length is 0.06 or 0.05%. Of course, this value cannot be directly applied to the measured data. These tests assess residual length changes - if alkali-aggregate reaction took place in the past its considerable part is already over. To estimate the change in length proportional to 40 years of waterwork existence is very difficult. Values from literature, stating that six months exposure in fog chamber at +40°C is to 20 years of deposition in a humid environment, can be applied only as a rough measure. Hence it would follow that - for a rough estimation - to get an approximate value of the total change in length due to alkali-aggregate reaction the measured values should be multiplied by three. For bodies with 0.015 to 0.020% measured change in length the resulting value would be 0.045 to 0.60%. These concretes could be with high probability classified as problem-free with regards to alkali-aggregate reaction concerns. In the case of the remaining three test bodies, whose measured values are 0.028, 0.033 and 0.043%, this speculative value would significantly exceed the 0.060% limit level. These concretes could be considered as susceptible to show alkali-aggregate reaction effects.
Based on this pilot test it can be concluded that at Orlík dam, different sections are susceptible to change of length, with high probability due to so called slow type alkali-aggregate reaction, possibly inducing mutual movements of some blocks accompanied by acoustical effects, that were discussed years before.
From the presented results it can be found that:
• The results confirm preliminary findings made in the past on basis of simple colorimetric tests.
• Assessed differential changes in length could explain acoustical effects, experienced in the past at the Orlík dam.
• Based on a speculative estimation of total expansion due to alkali-aggregate reaction it is very probable that in some sections of the Orlík waterwork there are local concrete blocks showing mild forms of alkali-aggregate reaction.
• The fact that first signals of alkali reaction development appeared 35 to 40 years after putting the work into operation and the measured values for the most part are still in an acceptable range, it is possible to classify these alkali reaction effects as mild, in no case endangering the waterwork in the future. Nevertheless, detected facts show that such important engineering work as a gravity dam, even after a long time, can produce phenomena that, from standard engineering practice point of view, are absolutely atypical or unusual for the region. It is necessary for the operator to further monitor this problem, and to also pay due attention to the other concrete dams.
Technical and safety surveillance
Due to its parameters and significance, the Orlík dam is classified in the I category of technical and safety surveillance. The content and scope of surveillance activities are regulated in general legislature and specified according to a programme of technical and safety supervision for this waterwork. Key information is acquired by measurement of various dam behaviour characteristics at various time intervals.
Most of the measurement points have already been realised during the construction and form part of the project. Vertical and horizontal displacements are monitored geodetically. Relative movements in dilatation joints are assessed on more than a hundred bases. Dam inclinations and deflections are measured by pendulum meters. Hydrostatical pressures in dam body are monitored in 70 bores in front of and behind the grout curtain. The main seepage rate is measured on three points in the grouting gallery and foundation drain, partial seepages are measured on various adapted places in the staircases and shafts.
In the course of 1995 to 1997, significant uprating took place. The Orlík dam was equipped with an installation of automatic sensors by Swiss firm huggenberger. This installation included:
• Dam seepage measurement on three points in the shaft. Two partial seepage rates are monitored by FDU 80 ultrasonic probes and an ultrasonic transducer, the total seepage rate by conductometric sensor of seepage water limit levels in the observation well.
• Dam inclination measurements in two main directions are realised by three Telelot automatic sensing devices and two pendulum meters with Koordiskop manual reading bases on three levels. Two holes were bored in dam profile into the full block to house thermometers. Six temperature sensors were installed at various distances from the dam fac .
• The Telejoinmeter triaxial transducer, with 0.01 mm accuracy, is used to sense relative movements in ten selected points of dilatation joints. Forty-eight bore holes for static pressure measurement in the inspection gallery were fitted with the Telepressmeter string meters. Separately a geosys monitoring net to monitor dynamic seismic events was installed. It consists of six sensors, of which five are inside and one outside the dam.
The system further comprises three dataloggers and master computer with basic SW for communication with dataloggers. All values can also be measured manually by original methods. The acquired data together with visual observations by the dam operator are verified on place and compared with set limit values. They are then sent to the central office to be further processed for assessment of long term trends in individual variables behaviour.
The history of building artificial reservoirs for economic utilisation in the Czech Republic dates back to the Middle Ages. A great number of ponds were established during the 13th to 16th centuries, particularly in the South Bohemian region.
At present the international-commission-on-large-dams lists 118 dams in the Czech Republic, of which 77 are fill dams and 41 concrete and masonry dams. Seven dams (three concrete and four fill dams) exceed a height of 60m. There are also 43 dams over 30m high (23 concrete and masonry dams and 20 fill dams). The highest dam in the Czech Republic is the Dalesice dam which is a 100m high rockfill dam with an earthern core.
All reservoirs in the Czech Republic are multi-functional. Although some of these waterworks were planned for a single purpose, additional functions have been developed so that these reservoirs became more effective in service.
In the Czech Republic there are 41 concrete and masonry dams. Most of these are gravity dams, with the only arch dam being the 40m high single curved arch of Vrchlice dam. The 56m high Fláje waterwork is the only buttress dam in the country. Of the remaining gravity dams, 13 are masonry dams which were built at the beginning of the 20th century. Twenty-six classical gravity dams were constructed from the 1930s to 1992. The Vltava cascade was primarily built for hydroelectricity generation and flow compensation for waterworks in Prague. It also provides limited flood protection for the city of Prague, water sports recreation and ship traffic functions.