From 1959 to 1965, nine major dams of the world failed (of an estimated 7800). The worldwide failure rate was about 2×10−4 per dam-year.
The probability of failure of a particular dam is difficult to estimate, but can be strongly dependent on geological setting, surface faulting and seismicity, as well as on the type of construction, size, and other factors. Damage will depend on such factors as topography, storage volume, the cause and mode of failure, the size of the population at risk (some thousands to more than 100,000), and on mitigating factors, such as evacuation.
Nuclear Power Reactor accidents are discussed in chapter 5 from a physical and engineering point of view. A major conclusion of that discussion is that the more serious the accident, the less probable its occurrence. Two important implications follow. First, the relatively frequent minor accidents provide opportunities to improve design and lessen overall risk. Second, there is a “dominant risk,” defined by maximum cumulative damage, that may be used to characterize and compare particular reactors. The dominant risk is chosen on the basis of a maximum value for the expression: (probability of incident)×(damage per incident).
It should be noted that even large accidents in light water reactors are not analogous to the explosion of an atomic bomb. There is no explosion or release of neutrons; buildings outside the reactor plant are not damaged physically, nor are fires ignited. Property, ground, and water damage are due to radioactive contamination. The danger to human beings is the inhalation or ingestion of radioactive substances or gases, or irradiation from such substances released to the atmosphere or deposited on the ground. (The occupational hazard within the plant could at some time involve exposure to neutrons.)
Three Mile Island In the case of light water reactors, accidents with dire consequences have not occurred in some 400 reactor-years of operation, but an accident that severely damaged the reactor core did occur at the Three Mile Island plant in Pennsylvania on March 28, 1979. It involved a partial failure of the cooling system. The analysis of what happened is being conducted by a presidential commission, whose report is due in October 1979, and other groups.
Of particular interest is the report of the Ad Hoc Population Dose Assessment Group, staffed by technical experts from three government agencies.132 During the period March 28 to April 7, it was necessary to vent quantities of radioactive gas. These gases constituted an atmospheric risk to approximately 2 million people who reside within a 50-mile radius of the plant. The principal emissions were xenon-133 (133Xe) and, to a