I - EXCEPTIONAL CIRCUMSTANCES:
Risk assessment for exceptional circumstances is logically based on the probability and impact of such circumstances.
Ia - Floods
Floods have caused half of failures and their victims. 40% of existing reservoirs are larger than 10 x 106 m3 and have caused 75% of flood failures and 90% of the corresponding victims.
Risk analysis may be highly effective for floods, which are classified hereunder in 3 categories:
 During dam construction, despite the limited flood control, overtopping has caused no failure of concrete or masonry dams. There have been few failures of fill dams less than 30 m high-which are often built in a single dry season-but 0.5% of fill dams (or cofferdams) higher than 30 m have failed as a result of overtopping. This risk may be easily assessed and reduced by building more diversion tunnels and RCC cofferdams; human losses may be avoided by early-warning systems used for just the most critical months of the year: since 1950, 3 failures in highly-populated Asia have caused many victims, but 5 failures in the U.S. or Europe have caused no victims, and in Brazil practically all 100,000 people at risk were spared when Oros dam failed.
During operation, 0.5% of fill dams outside China and 2% in China (where very high specific flows have often been underestimated) have failed by flood overtopping: such fill dam failures usually occur for a 0.20 or 0.50 m overflow nappe on the crest of cohesive materials-earlier for non-cohesive materials, later for rockfill-and are initiated by downstream toe or slope erosion.
Few concrete dams but 1 % of masonry dams have failed due to floods. Failures were caused by sliding in the foundation or cracks in masonry and not by toe erosion. Failure may occur before overtopping and is more likely for long, low, gravity dams for which a small rise in the water level has a great impact on stability.
Most flood failures have been due to underdimensioned spillways: risk assessment for this aspect is quite easy and effective if it is focused on the real problem: which dams are most at risk and what is the probability of actual failure (corresponding to imminent failure flow (I.F.F.» and not merely of exceeding a regulatory high water level? Due to their freeboard, many ungated dams where the 500-year flood exceeds the regulatory high water level may weil withstand the PMF, but some large gated dams with a 1000-year design flood have little safety margin. The catchment area of the great majority of dams is less than 500 km² and simple regional flow formulæ will apply to most reservoirs in a given climatic area: for instance the 10,000-year peak discharge, Q, may be simply evaluated by the formula Q = K S0,75, S being the catchment area surface and K a regional coefficient. Some simple adjustments taking account of the shape of the catchment area and yearly local rainfall can be made. Comparing the imminent failure flow with this calculated 10,000-year discharge can help do two things: it can identify the dams most at risk, and it can estimate a range of failure probabilities. All the factors used are easily determined. The impact storage has on peak flow should be taken into account; it may be important if the reservoir area of ungated dams is more than 1 or 2% of the catchment area.
For gated dams an additional risk is overtopping due to gate jamming: the yearly probability of this occurring has been 0.5 or 1 x 10-4: exceptional rains causing large floods may disrupt electric power supply, transmissions, computers, and access, and may cause panic. There are two key points for evaluating this risk: the probability of the flood overtopping the dam when all the gates are closed, and the quality of maintenance, training of operators, and emergency measures: practical guidelines on how to reduce this risk have been given by a freely available report published in 1998 by the French Committee on Large Dams and based upon the operation of five hundred dams and related incidents.
After assessing the probability of failure, it is often possible, when necessary, to divide the probability by 5 or 10 by introducing low-cost structural measures: raising the top of fill dams by one or two metres, protecting downstream slopes for small specific flows, lowering the sill of ungated dams and installing fuse elements to retain storage capacity (flashboards, fuseplugs, fusegates), and preparing a possible failure place where the dam is less than 10 m high (for long fill dams), etc. The risk due to floating debris (trees, etc.) is often overlooked and may be serious if the gap between piers is less than 10 m. Many small spillways partly closed by sand bags, temporary wood or masonry walls which are supposed to be-but are not always-washed out by floods should be cleaned before the flood season.
Finally, emergency planning properly adapted to each case and early-waming systems may be very effective for floods: these days their effectiveness is further enhanced by weather forecasting giving advance waming of the possibility of heavy rains.
Large floods not endangering the dam but inundating large areas (100 or 1,000-year return period) may also be handled more effectively as a result of these latter measures. In future the global human risk associated with these floods may well be higher than the failure risk. This point is often overlooked as the risk is considered to be beyond the responsibility of dam owners, though this may not in fact be the case for three reasons:
The operation of many dams reduces or avoids small floods downstream and areas which appear safe for some years may be occupied by many people then exposed to larger floods, especially in densely populated Asia.
Gate operation (or failure) may cause sudden flow increases which may possibly be dangerous, even for rather small flows.
For ungated dams, a flood partly stored beneath the spillway level may be dangerous as the downstream hydrograph may subsequently be much steeper than for the incoming flood, and discharge may rise from nil to 500 m3/s in just one hour.
For all the above flood problems, risk analysis seems to be a very cost-effective solution if properly adapted to the real problems. It would appear that in a few years it will be possible to considerably reduce the corresponding human risk at low cost.
Ib - Upstream dam-break wave
There have been 6 failures of masonry or fill dams-the most prone to flood-induced failures-as a result of failure of a dam upstream. Against the recorded 800,000 dam-years this figure is low, but it is high compared to the number of dams situated downstream of a relatively large reservoir: for these dams, failure of the upstream dam may be one of the highest risks, if not the highest, with a yearly probability in the range of 10-4 or more.
Dams subject to sudden failures may considerably increase the risk for dams downstream by not giving time for gates to be operated normally.
Dam-break-wave risk assessment may be efficacious, but has to be tailored to each case and will be more complex than for floods.
Another risk is that of natural dams caused by landslides. These may fail hours, months or centuries after they are formed.
Ic - Earthquakes
Earthquakes have often caused cracks or settlement, but have caused few failures of dams more than 15 m high. Nevertheless, this risk should not be overlooked: hundreds of smaller dams have been destroyed in minutes, and the failure of Van Norman hydraulic fill dam (41 m high, 1971) was a near-disaster.
The consequences of earthquake may be:
Sudden failure, for instance by liquefaction of fine non-cohesive materials or structural failure of buttresses.
Cracks which may continue to extend hours or days later, particularly in the case of masonry dams or old fill dams with no filters or drainage.
Though the yearly failure probability of dams is lower than 10-6 in most cases, it may well be in the range of 10-3 for some dams in seismic areas.
Seismic risk assessment is less precise than for floods and emergency planning is less effective. More expensive structural measures may be necessary.
Id - War
Breaching the dikes of the Yellow River in China has been a military tool for over 1000 years. In 1938 it caused the loss of over 500,000 lives.
Outside China one per thousand dams higher than 30 m have been severely damaged by wars and at least 3 of them have failed completely (masonry or concrete). As the number of years when dams higher than 30 m have been involved in wars is small, the number of corresponding dam-years is less than 1000 and the corresponding yearly failure probability between 10-2 and 10-3. The dams at risk may well be those whose failure would have the worst consequences.
Although risk assessment can hardly determine probabilities of wartime failure, it may be very useful for estimating possible consequences and for preparing emergency planning and early-waming systems. Two similar German dams failed on the same night in 1943, with the same breach flow of 8,000 m3/s. Victims numbered 1,200 and 70, the alarm possibilities being different. Previous < 1 2 3 4 > Next
|