Assessment and role of risk in dam building

Abstract

for commercial and drinking water, and the creation of zones for recreation and possible navigation in previously inaccessible sections of the riw,rs [12]. In addition, the existence of reservoir-impounding dams. together with the above-indicated benefits will, by itself, generate different kinds of risks, which are apprecial)ly lower in magnitude than the expected benefit. and by their own nature, are probabilistic; of these risks, the most familiar in terms of negative consequences are social, material, constructive (operation in the design mode), hydroh)gical, geodynamic, ecological, etc. Here, risk is understood, in tire broad sense, to be tile incapability of the project to provide the expected benefit over a given period of time. About 3.000 reservoirs, each with a volume of 1 million m a and greater, have been impounded within tile territory of the Russian Federation. As analysis of the catastrophic failures of the Kiselevsk (1993) and Tirlyansk (1994) dams h~ indicated, study of the causes and htws governing different risks, and their consideration and regulation assume practical significanc'e. Provision for reliability and safety has always been a primary condition of dam building. Statistical data derived from emergency situations are being used to assess the actual magnitude of the risk of failure, especially in recent times. Dam Reliability. Figure I shows the distribution of tile risk of emergency situations (P is equal to tile ratio of the number of failures to tile number of dams in service: the figure is presented in logarithmic scale) at various types of darns as their number increased from 1,200 at the turn of the century to 44,000 at the present tinle. According to these data, the risk of emergency situations for all types of dams, which had reached 18% in the period from I920 to 1930, had been reduced by more than a factor of three by 1971-1980. For concrete dams, a similar reduction in failure risk amounted to approximately five times from 1981 through 1990. The risk of failure for all types of concrete dams, which had reached 4% in the 1920s and 1930s, had also been reduced by up to 0.7% from 1971 through 1980. Here, the failure risk of concrete dams had been lowered by a factor of approximately 30 (from 3% in 1911-1920 to 0.1% in 1981-1990). Social risk over the entire period has served as a principal criterion of reliability assessment - the safety of existing types of dams. Catastrophes have been timely beacons for tile inspection of existing design criteria and selection of more effective methods of monitoring dam safety, such that use of the social-risk criterion reached maturity long ago in design practice. The introduction of government monitoring of tile condition of dams has been associated with large-scale dam catastrophes: after tile Bois disaster in France in 189.5, after the disaster at the Saint Francis Dam in California in 1928, after the disaster at the Eidzhi Darn and Dollarog hydroelectric plant in England in I935, and, finally, after the disasters at the Kiselevsk and Tirlyansk Dams in Russia in 1997. Considering the increa.sed attention that society has given to the safety question in analyzing problems of social risk, we can utilize the most (:autious estimates. Figure 2. which is based on data derived from various publications [1, 2, 3, 5, 6], shows the distribution of the xmmber of human victims in the largest dam disasters during the period from 1900 through 1995, according to which an increase in tile number of human victims from 1,074 to 4,165, i.e., 3.9 times, was observed in the periods between 1920-1940 and 1940-1960, with a subsequent increase to 4,576 persons in the period from 1960 through 1980. A reduction in the number of human victims to 3.858 after 1980 is explained by the incompleteness of this period; the number of victims during this period w~, nevertheless, higher than during the entire 19th century'. It is possible to criticize tile method used to calculate the number of victims;