Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
1 Introduction The seismic safety of dams is an issue that has been receiving increasing attention in many parts of the world during recent years. It has become a major factor in the planning for new dams proposed to be built in seismic regions and in the safety evaluations of existing dams in those areas. There are two main reasons for this increasing level of concern: (1) The risk of a major disaster due to dam collapse increases each year because the popula- tion is increasing downstream of nearly all large dams. Consequently, large population centers are now at risk in locations where only a few people may have lived at the time when the dams were built. (2) Knowledge of the complex nature of the earthquake motions that may attack a dam is increasing rapidly as more earthquake records are obtained all over the world, and it is becoming apparent that the seismic safety precautions taken in the past did not fully recognize the hazard. The NRC Panel on Earthquake Engineering for Concrete Dams was ap- pointed to evaluate the present status of seismic safety evaluation of concrete dams and to prepare a report summarizing its findings. Specific objectives of the report were stated to be as follows: 1. To indicate to research workers the present state of knowledge of the earthquake problem as it relates to concrete dams, and the areas where knowledge is lacking and where further research is needed. 2. To bring the earthquake problems of concrete dams to the attention of government agencies and other organizations that may initiate, direct, or fund research and to provide them with information helpful in planning such activities. This evaluation has been limited to concrete dams because analysis proce- dures and research needs for embankment dams are quite different. 9
10 THE HAZARD OF DAMS Dams probably were among the earliest major structures to be created by humans; the reservoirs retained by dams were key elements in water supply systems that dated back at least 5,000 years. The first dams undoubtedly were small earthen embankments that were designed by trial and error; failures of these dams must have occurred frequently but with very little consequence. However, with expanding needs for water and with confi- dence derived from previous successful dam construction ventures, the sizes of dams grew, leading to increasing potential for disaster. Understanding of the changes in behavior associated with bigger dams tended to lag behind the increasing boldness of the planners, with the result that failures contin- ued to occur with increasingly tragic consequences. Ultimately, this history of dam tragedies led to the formation of agencies responsible for dam safety. A notable example of a dam failure that was responsible for imposition of controls on dam safety was the collapse of the St. Francis Dam, located about 45 mi north of Los Angeles, California. This concrete gravity dam, 205 ft high by 700 ft long, and impounding a 35,000-acre-foot reservoir, is shown before and after the collapse in Figure 1-1. The collapse occurred almost instantaneously just before midnight on 12 March 1928, creating a surge of water 125 ft high and an estimated discharge rate of 500,000 cfs. The resulting extensive property damage, and especially the loss of over 400 lives, clearly demonstrated the hazard presented by major dams and led directly to the organization of an agency of the State of California responsible for the safety of dams. This agency, now known as the Division of Safety of Dams of the California Department of Water Resources, has jurisdictional authority over nearly 1,200 dams. Its operation during the past 60 years has served as a model for the organization of similar agencies in many other states and in other countries. The hazard posed by large dams has been demonstrated by failures of many dams since 1928 dams of many types located all over the world. Two recent examples of dam disasters occurred with concrete thin arch structures Malpasset in France and Vaiont in Italy. The Malpasset Dam failed in a sudden burst on the evening of 2 December 1959, creating a huge wave that destroyed towns and villages downstream with tremendous loss of life. As was the case with the St. Francis Dam, the failure occurred not in the concrete structure but rather in a weak seam in the foundation rock supporting the dam. The situation at the Vaiont disaster was entirely different in that the dam did not fail; it remained intact despite being severely overloaded when it was overtopped on the night of 9 October 1963 by a great wave reaching more than 300 ft over the crest. This wave was generated by a tremendous landslide that dropped 350 million cubic yards of rock and earth into the reservoir from a cliff high on the side of the valley. However, the overtopping
: : :: - ala ^. - ~ e ~ O O O ~ O ^ o _ ~ . O O O ~ em O .- a a O ~ O a a .E O ~ ~ a O ~ ~ O ~ ~ O ~ .- ~ O -a e ~~ e .5 ~ 8 o is: o ~ ~ o o .b . o o > o o o o o ~ dig ~ ~ ^ ~ o ~ ~ ~ . o ._ ~ ~ 2.8 ~ ~ o ~~ e ~ 0 ~ ~ ~ O . ~ ~ _ ~ O =.s ~ ~ ~ O rig - _ A. 0 ~ _
12 had the same effect on property and people downstream as if the dam had been breached; over 2,000 fatalities resulted. The failure of Teton Dam, built in Idaho by the Bureau of Reclamation? demonstrated that embankment dams also present a threat to populations downstream, even when they are built by modern construction techniques. The failure occurred on 5 June 1976, when the water level was 3.3 ft below the spillway during the first filling of the reservoir. It was initiated by leakage in the contact zone between the impervious core material and the foundation rock; the leak quickly led to erosion of the embankment and caused total release of the reservoir within a few hours. The resulting flood produced property damage estimated at over one-half billion dollars, but fortunately flood warnings were issued early enough that only 14 lives were lost. None of the aforementioned catastrophes was related to earthquake activ- ity, but several other events during the same time span have demonstrated the additional risk that applies to dams located in seismic regions. Two significant examples of earthquake damage to concrete dams were incidents observed at Hsinfengkiang Dam in China and at Koyna Dam in India. These are discussed more fully in subsequent chapters of this report, but are men- tioned here to emphasize that earthquakes truly are a major hazard to concrete dams. Hsinfengkiang Dam is located in Kwangdong Province; it is 1,444 ft long and 344 ft high and was completed in 1959. As originally designed, it had 19 buttress blocks in the central segment, with gravity sections located on either side. The original design was intended to resist ground shaking of only intensity 6 on the Modified Mercalli Scale, but when numerous earth- quakes were experienced in the vicinity during and shortly after the filling of the reservoir, the dam was strengthened to withstand earthquakes up to intensity 8. On 19 March 1962 the dam was severely shaken by ground motions considerably above intensity 8, caused by an earthquake of magni- tude 6.1 located very close to the dam. No instrumentation had been installed in time to record this event, but strong-motion seismographs were installed in the dam to record the effects of aftershocks, and a magnitude 4.5 aftershock produced a peak horizontal acceleration of 0.54 g recorded at the dam crest. During the even more intense motion that occurred in the main shock, the dam developed horizontal cracks at a change of section in the nonoverflow buttress blocks on each side of the spillway, but no leakage occurred. Sub- sequent analysis demonstrated that cracking was to be expected at this location on the structure during an earthquake of this intensity, and the dam was then strengthened so that it could withstand even more intense ground mo- tions. Koyna Dam is a straight gravity structure built of rubble concrete in southwestern India. It is 2,800 ft long and 338 ft high at the tallest block; it has an unconventional change of slope on the downstream face resulting
13 from changes introduced while construction was in progress. Originally the dam was intended to be built in two stages, but while the first stage of construction was in progress it was decided to build the dam to its final height in one stage. This change of plans necessitated modification of the dam cross section. On 11 December 1967 a magnitude 6.5 earthquake occurred, with its epicenter at a distance of 8 mi but producing a fault trace that passed as close as 2 mi from the dam. Instrumentation within the dam recorded peak horizontal accelerations of 0.63 and 0.49 g in the cross- channel and the stream directions, respectively, and horizontal cracks caused by these motions appeared on both the upstream and downstream faces of the tallest nonoverflow blocks. They were located at the elevation where the slope of the downstream face changed, and subsequent dynamic finite element analysis demonstrated that such cracking was to be expected from an earthquake of the observed intensity. However, it is noteworthy that the reservoir was not released by this damage, even though minor leakage was observed. Thus, the cracking did not cause dam failure, and there was no flooding disaster downstream, although 180 people in the vicinity were killed by the earthquake ground vibration effect. Subsequently, the dam was strengthened by the addition of buttresses on the downstream face of the nonoverflow blocks. A notable example of earthquake damage to an embankment dam occurred at Van Norman Dam near Los Angeles, California, during the February 1971 San Fernando earthquake. This hydraulic fill structure was built in a series of stages starting in 1912, with the last stage completed in 1940. It retained the lower San Fernando reservoir, and by 1971 tens of thousands of people were living immediately downstream from it. During the earthquake a large slice of the upstream embankmentextending up to include the dam crest slid into the reservoir. Only the fact that the reservoir had been lowered significantly due to concerns about seismic safety (it was 35 It below the dam crest at the time of the earthquake) prevented the dam from being overtopped, with disastrous consequences to the population downstream. This near disaster was a major factor leading Congress to promulgate the National Program of Inspection of Dams in 1972. The subsequent nonseismic failures of Teton Dam in 1976 and Kelley Barnes Dam in Georgia in 1977 caused additional concern for dam safety on the part of Congress and resulted in the appropriation of funds to implement the inspection program. SEISMIC SAFETY OF CONCRETE DAMS The dam inspection program was the first formal step toward improve- ment of dam safety in the United States. It provided for talking a census of all dams in the country and for identifying those for which detailed studies would be needed to determine whether they met the required level of safety.
14 Possible failures due to all types of causes were considered for both con- crete and embankment dams. For dams located in seismic regions it generally was concluded that further investigations of seismic safety were needed, due to advances in knowledge of the earthquake input as well as to better understanding of the dynamic response mechanisms. Accordingly, major reevaluations have been made of the seismic safety of many existing dams during the past decade. In addition, performance predictions have been made for the relatively small number of new dams built or proposed to be built in seismic regions in this country during this time. In order to put the question of the seismic safety of concrete dams in proper perspective, it must be noted that there is no record of any concrete dam ever having been damaged due to earthquake excitation to the extent that the reservoir was released. This perfect record stands even though more than 100 concrete dams of all types have been subjected to measurable shaking due to earthquakes in many parts of the world. But this record does not justify complacency, because in very few of these numerous examples were measurements actually made of the intensity of shaking, and the severity of the tests generally is not known. In one of the few cases where severe shaking was recorded at Pacoima Dam, a concrete arch in the epicentral region of the 1971 San Fernando, California, earthquake the earthquake forces were greatly reduced because the reservoir surface at the time was at less than half the design water level. Hence, it is not possible to draw definitive conclusions concerning seismic safety from this history of perfor- mance. Consequently, even though the earthquake safety record of concrete dams is excellent, much more research must be done if the details of their seismic behavior are to be understood and predicted reliably. Until two or three decades ago, the only consideration given to earth- quakes in the design of concrete dams was to apply a static horizontal force specified as a fraction of the weight of the structure to represent the seismic design loading. In later stages of its use this equivalent static load design procedure often included an additional weight to represent the inertial resistance of the water behind the dam. In these analyses a gravity dam was modeled as a simple cantilever column of varying cross section, but with the development of digital computers and the finite element method of analysis, it became possible to model the geometric configuration of the dam in a realistic way. By using three-dimensional finite elements, even arch dams could be mod- eled reliably. Subsequently, the analysis procedure was extended to treat the true dynamic character of the earthquake motions and the dynamic interac- tion of the dam with the reservoir water and foundation rock; then the aforementioned use of equivalent static loads was applied only in preliminary analyses. Present studies that take advantage of these advances in analytical procedures give greatly improved estimates of the seismic safety of existing or proposed concrete dams. However, uncertainties still remain in nearly
15 every aspect of the analyses, and an intensive continuing research effort is needed to reach the point where full confidence in the predictions of seis- mic safety will be attained. SCOPE OF THE REPORT This report describes the present state of the art of earthquake response analysis for concrete dams and identifies major areas where additional research is needed. The subject matter that follows is divided into five chapters, each dealing with a different aspect of the analysis procedure. It is impor- tant to note that the critical issue of defining the design earthquake (i.e., the ground motions expected to act at the dam site) is not considered in this report. It is assumed that geologists and seismologists will have established the magnitude and epicentral position of the greatest earthquake that must be considered in the seismic performance evaluation and that this informa- tion will have been applied, together with knowledge of the foundation rock properties, to obtain an estimate of the most intense free-field ground motions that may be expected in the vicinity of the dam. Thus, this report deals only with procedures for evaluating the response of the concrete dam to these specified earthquake motions. Urgent research needs that have been identified in the various subject areas are listed and described at the end of each chapter.