 In tape one of this series, Dr. Ralph Peck noted how failures of dams have occurred due to seepage, which developed into piping. The many examples he cited demonstrate that these types of failures happen with dismaying frequency. New and apparently well-constructed projects have failed when first exposed to water loads. Dams that have operated successfully for years have suddenly failed. Often these failures have been without apparent warning, and the results have been catastrophic. And when structures are designed properly to control seepage and prevent erosion, other factors like undiscovered geologic features, or subtle events which occurred during construction have been shown to be weak links that eventually lead to failure. This makes it absolutely necessary for dam owners to maintain active programs of surveillance through instrumentation and visual monitoring of the structures. In many instances, timely observation and evaluation of performance has aided in avoiding catastrophic failure, allowing a reservoir to be drawn down to remove the water loads and permit repairs. To improve safety, dam owners and operators need operating procedures which include dam monitoring, inspection, and emergency response plans. Modern structures operated under proactive, comprehensive dam safety programs pose generally acceptable risk to the public. Once the program is developed and put into place, it must be aggressively pursued to maintain a high level of protection. The owner-operator must be willing and prepared to make conservative decisions and take immediate action in the interest of public safety. Because irrigation, water supply, power, and recreation benefits may be lost, that decision may have to be made with full knowledge that tough economic and political consequences may result. A quick decision may be required with no time for gathering additional data or second guessing. But in circumstances where significant hazard is posed to the public, any error of judgment or decision must be on the side of public safety. In this video, Dr. Peck discusses dam surveillance and features for the control of seepage and piping that should be designed into new dams. The same features can be designed and used to repair older dams that have seepage and piping potential. The absence of proper design elements in most older and some newer dams makes them especially vulnerable to failure by piping. Generally, dams designed and constructed before the 1950s may be conservatively assumed to lack these control features, and they require close attention to the continuing adequacy of their performance. Dr. Peck's involvement in soil mechanics and foundation engineering coincides with the development of soil mechanics as an engineering discipline, beginning when he worked on the Chicago subway project in 1939 as an assistant to Dr. Carl Trissage, the father of soil mechanics. He has since served as a consultant for many governmental organizations, including the Bureau of Reclamation and the Corps of Engineers. The San Francisco Bay Area Rapid Transit Project and the James Bay Hydro Project in Canada are two of many significant projects that he has been associated with, as well as a large number of major dams around the world. Dr. Peck was awarded the National Medal of Science by President Ford for his contributions in the United States, currently as Professor Emeritus at the University of Illinois. Dr. Peck. In our first tape, we saw some of the serious consequences of piping, and we learned something about the mechanism of the process. We now want to turn our attention to ways of preventing piping, to ways of detecting it if it begins to develop even after a dam has been constructed and filled, and ways of remedying the problem if the piping is detected. Almost all modern dams are designed with a clay or at least an impervious core to hold the water back and shells of granular material or rock that supports the core. Since most core materials are erodable or if cracks develop through them, they can be washed out, the cores are always protected by filters. So there is a filter or maybe a series of filters between the core and the shells, particularly on the downstream side. In addition, in most dams there is some sort of a cutoff below the core that prevents water from flowing through any permeable materials that may be beneath the dam, at least permeable soil materials. And so we see cross sections that ideally look something like these. There is a core, and if an impermeable surface is not too great a depth, the core may simply be widened to make a cutoff out of the same core material. Between the core and the shells you see the filters, one behind the core and usually one underneath the downstream shell. Even with the cores and the cutoffs, the assumption is made that there is likely to be some leakage, and so downstream there is a second line of defense, a drainage system that picks up any moisture, any water that may get through or that may bypass the core, usually filter wells, wells into which flow can take place readily, but into which fine materials cannot pass. There is a problem with filters however, that is that they are difficult to place without segregation, particularly if the filter material is obtained from borrow pits in which the materials are rather broadly graded, that is the sizes may range from fine sand up to even coarse gravels. And most borrow pits contain materials of several sizes, so that you don't find in nature very often a borrow pit that has materials of just the size one would like for a filter. If people try to bake filters out of the mixed grain material, it is very difficult to place them without having the coarse material accumulate along the edge. As you see in this picture, there is coarse material along the edge of the windrow that has been placed, and that coarse material is in contact with the material that is supposed to be protected. That's unfortunate because the finest material to be protected can move easily into the coarsest material along the edge of the windrow. And even though broadly graded filter material is still used in some instances, it's very preferable to screen the material first from the borrow pit to divide it into sections of finer and coarser materials, and to make several filters, each one protected against the next one, to make these filters of a more uniform size material, then segregation is easily avoided or doesn't occur. Unfortunately in a great many dams that already exist, segregation may have developed and so there is the possibility for internal migration some long time after the dam has been completed. When a dam is on rock, we can do several things to reduce the likelihood of piping. And piping in the core of a dam can really take two forms. The material can move from the core into coarser materials, that is it can try to get into the filter or go into some place where there isn't a filter, or the core may crack because there are tendencies for extension or tensile strains that occur under some circumstances. And when that happens, and water can pass through the crack in the core, the core material can be eroded, and that erosion material must also be stopped by the filter. So when we have a rock contour as we might in a deep gorge across which a dam is constructed, we try to shape the gorge to some extent at least to avoid cracks in the core. And it's necessary to remove overhangs in the rock because you can't compact against the underside of an overhang very well. And of course where there are joints and weathered materials, they have to be treated by one means or another. We see here a cross section of a gorge beneath Lower Notch Dam in Ontario. The dam itself is only a couple hundred feet high, but the gorge that is beneath part of it is another couple hundred feet deep. So that actually this dam of modest size, two hundred feet high, if you count the depth of the gorge is about the fourth highest dam in Canada. And the profile of the gorge originally had very sharp angles as you can see on the edge of the valley wall here. These were trimmed back in order to reduce the concentration of stresses in the area near the top of the gorge proper. The amount to reduce the amount to make the curvature is a matter of judgment, but in this case, some very early finite element calculations were made to determine the best balance between the improvement that could be achieved in reducing the extension strains that might occur and in the cost and difficulty of doing the trimming. And the compromise gave a economical solution that considerably reduced the pressures. The gorge itself, which you see in this picture, was quite impressive. You can see that it had very steep side walls and you can also see a considerable overhang and as the fill was placed in the gorge above the location of the top of the fill, the overhang was filled with concrete which was bolted to the gorge walls in order to make a surface against which compaction could take place and without overhangs. This is an extreme case to be sure. We don't very often build dams over deep gorges, but there are some and they have needed treatment of this sort. More often, we see situations like this. In the stratified rocks, such as we had at Jordan El Dam, there are a lot of there are bedding planes which are usually nearly horizontal but not always and there are vertical joints quite often at right angles to each other so that the material alongside the deep portion of the canyon, the material alongside the valley walls, constitutes a series of blocks and quite obviously some of these blocks should be removed. It would be easier to take a block like this away than to grout beneath it and to do concrete infill behind it. That may be obvious, but it's by no means obvious how far back one should carry this treatment because these joints extend many hundred feet back from the dam and certainly one can't continue to remove blocks and replace with concrete or replace with fill. So some judgment has to be exercised to determine at what point one stops removing material and begins to depend on grouting to fill the more or less horizontal joints and grouting and infill surface treatment to take care of the vertical joints. This is a matter of judgment again and I doubt if there are any rules or specifications that can be written in advance that would cover all the situations that arise during construction. The best way probably to handle this problem is to assemble a committee at the time the job is approaching this stage consisting of the engineer or some of the engineers who design the dam, who know what the requirements are, the field superintendent or resident engineer and some of his staff, chief inspector and so on and preferably the contractors, foremen and people who will be actually doing the work. And to go from point to point looking for particular situations discussing and describing and agreeing on how to make the treatment. This seems to work quite well. Everybody understands what the objective is and what should be done and it's a flexible treatment because as the work progresses it may turn out that there are better things to be done and this same group can get together and review the situation. And a record ought to be kept of what is done so people later on can see what the situation was. At Jordan L again we see in the center part of this picture a surface has been treated, you can see it is now quite smooth. The objective is to get a surface against which the core can be compacted without having re-entrant angles and places where the fill can't get access. And sometimes when there are steep slopes and perhaps overhangs the material can be blasted off to make it smoother or concrete and fill can be placed and here you see some form work behind which concrete can be placed to make such a smooth surface. I think it should be apparent that every dam site is different in this respect and therefore the treatment that would be optimum would also be different. This is the left abutment of WAC Bennett Dam, a 600 foot dam in British Columbia that was built about a quarter of a century ago. The abutment materials here are sandstones and shales with occasional coal seams and you can see that these seams have re-entrant angles in them, have grooves that go from upstream to downstream so that water could get into such a groove on the upstream side through the shells go along the side of the core if something weren't done about it erode material out of the groove or pipe the material, erode the material from the shell from the core that was up against the groove area and thereby start a process of piping. On this particular dam the abutment was covered with shotcrete. This is one of the early applications of shotcrete to abutment treatment in dams as a matter of fact and on this dam the shotcrete was placed against wire mesh that probably would be preferable not to use mesh because it tends to hold the shotcrete a little bit away from the rock. But the principle is sound because even if water does get behind this protective blanket and erode something out of the rock between the more resistant layers it is protected, it can't attack the core from the back side and therefore begin to cause a process of piping. Where there are igneous rocks and metamorphic rocks that are strong such as in the James Bay region where there are great many dams and dikes there are several hundred dams in the James Bay area on the project. There are certain situations that appear over and over again and it was possible to give to the field forces after a little experience sketches such as these that would give them guidance. You can't see them in detail I'm sure but for example in an overhang followed by a sort of a depression there are some dimensions suggested for shooting off the upper part of the overhang to make a smooth surface and instead of shooting off a lot of rock to fill in below the smooth part with dental concrete. Other combinations are possible and these kinds of guidance are very helpful for field forces. On the other hand we have to be careful not to try to apply criteria that are developed for one particular geology, one particular location to some other geology to which they might not be applicable. So on each particular geology in each situation studies can be made quite profitably to develop criteria of this sort they must be used with judgment and they mustn't be transferred to other sites for which they're not applicable. This is the finished product of one of the dikes, dike D7 on the James Bay project and you can see infill concrete filling up the grooves that the glacier made in the rock at the bottom of the valley. You can see some form work behind which concrete is to be placed to smooth out the surfaces. You can see some smooth surfaces already prepared and irregular but nevertheless smooth surface against which a moist plastic clay core material can readily be compacted and make a tight contact. If the rock is weathered, heavily weathered behind the clay core or alongside the clay core then the situation can be quite different. Weathering might be from the action of climatic factors such as you have very often in the tropics or in areas where tropical climate used to prevail such as in some parts of California. It also may be due to hydrothermal alteration which has so much the same effect. This is a photograph of a portion of the left wing dam of Eteshetesia Dam in Zambia where there was very deep tropical weathering of a granite, a very ancient granite weathered long ago and when the granite was formed it tried to incorporate some more ancient marbles and limestones which later were dissolved and so in the granite which itself is fairly friable and soft you find holes like this and these sometimes go to considerable distance and obviously the dam can't be compacted against those holes and expected to stay in place. This situation can be very difficult to contend with sometimes. At Eteshetesia for example there was a very deep cut off wall put under the fairly shallow wing dam, concrete slabs in the bottom, concrete facing on the downstream side, some grouting could be done behind that concrete face and the cut off itself protected by these concrete works. This again is unusual but is a situation that can develop and when it does develop it has to be handled by very heroic measures on some occasions. More often the weathering takes the form of reducing a rock which might otherwise be sound to something that's either soft and friable or even sometimes granular. This was the case in the right abutment of Jordan L Dam for example and you see some of the more friable granular material here. This material under the core needs to be removed for a reasonable depth and replaced with the concrete dental treatment and downstream of the core underneath the filter materials it is advisable to cover the granular material with another filter of its own, generally a filter that is fine enough to keep these weathered materials from moving should there be some water trying to come through the foundation beneath the dam itself. There are great many dams of course that are not founded on rock or founded partly on rock and when a dam is on soil or it has abutments that consist of soil then there are two treatments that are used and that are sometimes both necessary. Where possible there is a positive cutoff. We have seen some of these cutoffs in earlier slides they can consist of diaphragm walls of concrete for example. Used to be that steel sheet piles were the common choice but steel sheet piles are not really impermeable water can get through the interlocks and fine materials can but more likely the interlocks aren't interlocked and one doesn't know that so steel sheet piles themselves are rather undependable cutoffs and are not used much more for that purpose. But concrete walls slurry walls walls constructed by jet grouting by a number of techniques that are available today for making even very deep walls are feasible and economical. But in addition to the positive cutoff there should always be drainage downstream of the cutoff. There need to be filters and blankets and relief wells. The idea is the belt and suspender system. We cut off the flow with diaphragms with cutoffs as completely as possible and then we assume they don't work and put in drainage measures to make sure that any leakage is properly taken care of. And even if there is a great deal of attention paid to these cutoffs there is sometimes leakage. Not all leakage that appears downstream of a dam or on the face of a dam is necessarily fatal. You've probably all seen sand boils like this one where clear water bubbles up and where a little dike is built around the point of emergence and if you watch carefully you can see grains of sand rising and falling back into place. And you realize if you watch long enough that material is not escaping. That the material is simply being circulated at the bottom of the little depression and what has happened is that nature has built a little dike to produce a back pressure which is helpful. And very often when floods occur along big rivers like the Mississippi for example many of these sand boils occur in the downstream side of levees and they're built up as an additional precaution against piping. However ordinarily we don't like to see water escaping in the downstream toe of a dam partly because the material around the little springs is softened and that reduces the strength so that even such sand boils are generally taken care of by putting in drainage and putting in blankets over the top. So treatment after the reservoir comes up and one discovers where the treatment is needed usually consists of placing drainage blankets and relief wells. On dike D20 one of the long sand dikes about 120 feet high on the James Bay project dikes that rest on sandy glacial tills that go to considerable depth of bedrock. The design was to put in a cut off and here you see the top of a cut off wall a slurry panel wall that had been built first being exposed to make the connection with the dike that is yet to be built above it. Even when this cut off wall was being built it was realized that there were nests of boulders that were giving trouble, a slurry was being lost and it was beginning to be questionable whether the cut off wall would really be impermeable. And indeed when the dike was filled and the water came up in the reservoir the first time there were sand boils and springs and soft spots downstream where the cut off was supplemented by relief wells. Here you see some of the relief wells and you see that the downstream portion of the ground below the dike is nice and dry and stable which it wouldn't be in which it wasn't before the drainage was put in. You also notice that the drainage wells have rather unusual outlets here in this subarctic climate you have to provide discharge facilities before the drainage wells so that they will not freeze up during the wintertime. These are things that we can do during construction to prevent piping and they are measures that we can take after construction if piping is detected. We can put in cut off walls in dams that exist. Sometimes some very deep ones have been put in. A very deep cut off wall was placed throughout the entire length of Mud Mountain Dam a Corps of Engineers project near Mountaineer. We can use relief wells. We can use drains. There are great many procedures available to reduce the effects of piping or to prevent its development. But the most important thing when we have a dam that we wish to keep our eyes on is to keep it under close surveillance and every dam requires careful surveillance. What do we do in order to detect and see whether piping might be a problem? Probably the first thing we try to do is to see if there is a history of the construction of the dam. If the dam was built by an engineering organization or designed by one, very likely there is a considerable body of literature in the files somewhere about how the dam was built. Records in photographs, all these things should be dug up, looked at and studied to find out in effect whether the dam was designed with these modern principles and made as safe as possible against piping in the first place or whether one can see in the design certain defects which might lead to piping. It's very helpful to know how the dam was put together. Sometimes we don't have this information. All too often in older dams we don't have this information and then we have to explore to try to find out if we can what the dam is really like inside. One of the most important things to look for is whether it is possible for erosion products to escape without being detected. As we saw, for example, in the tail race at Walter Bolden Dam, piping could go on a long time under the tail race with nobody ever seeing it because the tail race water was always muddy. If this situation exists, if that is a possibility, then that is a warning flag to those doing surveillance that there are things that could be going on that might be missed by simply walking around and inspecting the site. And one looks during surveillance for such things as springs and boils and seeps. One collects the water from these seepages in ditches and passes them over weirs or other measuring devices so that the quantities can be determined. And it is important to keep track of this on a closely spaced basis weekly or daily perhaps. It's not a matter of whether there are wet spots or seepages. It's whether they have changed that counts. Whether surveillance has to be over the long haul, one has to be able to observe things throughout the year and from year to year to detect whether there may be subtle changes that are not associated with precipitation or snow melt. New wet spots, spots where green grass grows for the first time, for example, are a tip-off that something is changing and change in the body of a dam is something to be guarded against. And then, of course, in surveillance, one looks to see whether there are depressions. The beginning of sinkholes, for example. And in most surveillance programs for dams, one finds instrumentation today. There are piezometers installed downstream of the toe and sometimes within the body of the dam itself. But piezometers are something of a mixed blessing. If they indicate that something is changing, that in itself is important. That's a sign that there may be erosion going on. But it would be a matter of pure luck if piezometers would be installed at a place where they would be affected by an advancing erosion tunnel. One doesn't put in enough piezometers to catch all the possibilities for a place that an erosion tunnel might go. And so, because piezometric observations may not show anything suspicious, is not a good reason for believing that erosion tunnels cannot be developing. And so we see that even with careful surveillance, there are surprises that nature can have in store for us. Nature has the upper hand in that she knows what she's doing in that dam, and we don't. We have to try to read what signs she gives, but we also have to realize that there may be some signs that are not going to be apparent. If we do detect something, we have seen that we do have defenses. We can put in cutoffs. We can put in drains. We can do a number of things to improve the situation. But first, we have to know that something is going on, and that requires very careful vigilance. It requires individuals who are dedicated to the job, and it requires people who will stay with the job from week to week, even from year to year. And in this day and age, that is difficult to find, to find a person who is interested in such a job, because in one sense it is a routine job, and many people would consider it a dead-end job, and yet it is one of vital importance to the safety of a dam. That is a human problem that has to be considered in the surveillance programs for dams. Trotsage remarked once that nobody knows a piece of ground like the poor peasant who owns it and who has to scratch a living from it. He knows every little thing about his little plot of ground, senses all the changes that may occur, and that is the type of person that one needs to look after the surveillance of a particular dam or group of dams. The Miami Conservancy District, probably without realizing it, found a fairly happy solution for their five major hydraulic fill dams. There is a caretaker for each dam. He lives in a house at the dam site, sometimes just on the downstream side of the dam. He does a great many things around the dam site, including mows, he mows the grass on the downstream slope. If there are wet spots that develop, he is pretty sure to find them. It's a job in which the caretaker takes considerable pride, and in some instances, fathers have been succeeded by their sons in that job and in those houses. Those are dedicated people, and that is one way in which an important series of dams has been kept under surveillance by people who take the job seriously. And of course, it takes a dedicated owner to realize that somebody whose main job is looking for trouble and who is not producing dollars and cents on the income side of the ledger, it takes an owner with considerable foresight to realize that this is an important obligation, that the owner has to keep track of the surveillance, has to be willing to pay for it, otherwise he risks the possibility of being surprised by a serious failure. There isn't any known procedure of surveillance that can guarantee that any dam will be safe, that something could not happen, but that doesn't mean that one should relax just because it may not be possible to save every dam in which a defect develops. Sir Adam Beck II is a very good example of that, where we saw that the watchman on the holiday decided to make his rounds anyhow. He might have decided to stay home that day thinking that nothing had ever happened in the five years before, but it did happen that day and he was there at the right time and it was possible to avoid a complete failure of that dam just because the surveillance was conscientiously done. So we do all we can to detect the possibility of failure and to take remedial measures. We do the very best we can to reduce the risk, but we have to remember that we cannot reduce that risk to zero.