This is a report on the activities of the International Committee on Large Dams (ICOLD) Technical Committee on Embankment Dams at the ICOLD annual meeting in Bali. It includes a contribution to the workshop on the new ICOLD bulletin on small dams.
1. ICOLD annual meeting
The meeting was well attended: over 1500 delegates, a record, enjoyed excellent hospitality on a compact site in gorgeous tropical surroundings next to the sea, with delightful Balinese music and dancing at the opening ceremony and as well as other functions. It was a very lively meeting, with much interaction among colleagues. A great deal of work was done, with many technical committees, up to six concurrent sessions at the symposium, and several knowledge transfer workshops – an innovation – to share information from newly completed International Committee on Large Dams (ICOLD) guidance bulletins with colleagues. There were 1-day technical tours around the island, with stops at dams, including Titab dam, under construction. There were also pre- and post-meeting technical tours to dams on many of the Indonesian islands.
This report is about the work of the ICOLD Embankment Dams Committee. Reports on the ICOLD meeting, the executive meeting and all the technical committee meetings (and free copies of bulletins, including preprints of new ones) are available in the members' area of the ICOLD website (www.icold-cigb.org). You will need the members' area password. British Dam Society (BDS) members can get this from Tim Fuller (tim.fuller@ice.org.uk).
2. ICOLD Technical Committee on Embankment Dams
The Technical Committee on Embankment Dams's activities started with the first of the new ICOLD knowledge transfer workshops. The workshop was on ICOLD bulletin 164, vol. 1, on internal erosion in existing dams. Jean-Pierre Tournier (Canada), chairman of the Embankment Dams Committee, gave an introduction acknowledging the outstanding work of Dr Jean-Jacques Fry (France), Professor Robin Fell (Australia) and the ICOLD European working group on internal erosion in advancing knowledge of internal erosion and preparing the bulletin.
Dr Jean-Jacques Fry (France) and Rodney Bridle (UK) made presentations dealing with the fundamentals and detection and monitoring of internal erosion in existing dams. The fundamental point is that internal erosion is a mechanical process. It initiates when the hydraulic loads imposed by water flowing through dams exceeds the ability of the materials in the dam and its foundation to resist them. Understanding the mechanics provides a basis for investigations, remediation, if necessary, and monitoring. The audience were pleased that the bulletin makes a currently poorly understood phenomenon more comprehensible. It also provides guidance on how to carry out engineering investigations and analysis to check if remediation is necessary to protect dams against internal erosion, and to design appropriate monitoring systems.
At the knowledge transfer workshop on small dams (bulletin 157), the relevance of the internal erosion bulletin to all dams – small and large – was emphasised. One of the reasons for writing the small dams bulletin and the internal erosion bulletin had been the failure in 2009 of the 15 m high Situ Gintung dam near Jakarta in Indonesia. The failure had happened suddenly early in the morning, releasing about 1 million m3 in 10 s of water through a breach at the spillway position. It caused more than 100 fatalities. As Rod Bridle explained in a paper presented at the symposium, the failure was not caused by overtopping: it was caused by internal erosion, probably disruption of the overflow structure by uplift through water pressure in concentrated leaks – cracks – in the embankment under it. ICOLD bulletins 159 and 164 are a lasting memorial to the victims, providing guidance for engineers to prevent similar failures and fatalities in future.
At the Technical Committee on Embankment Dams committee meeting, Jean-Pierre Tournier announced that at the congress in Stavanger, Norway, in 2015 he will be joint chairman with Harvey MacLeod, chairman of the Tailings Dam Committee, of the sessions on the Embankment Dams Committee's question 98, on five topics: high rockfill dams, internal erosion, foundation risks, interfaces between concrete and earthfill structures, and tailings dams. The general reporter will be Zeiping Xu from China. Responses (papers) will come from the national committees as usual; responses from the committees have not yet been introduced.
Volume 1 of bulletin 164 on internal erosion had been updated in English to respond to national committee comments received last year. Dr Jean-Jacques Fry had completed the translation into French, and he and Rod Bridle were making the final review and would be putting the ultimate preprint on the website shortly (available free at www.icold-cigb.org). A printed and bound version would be available in due course. Progress on volume 2 had been slow; we had concentrated on disseminating the important fundamental information from volume 1, with several workshops, papers and presentations over the past year. However, the intention is to complete the draft of volume 2 over the coming months. Meanwhile, the latest draft will be available from the website.
Thibaut Mallet, the technical director of the drainage authority responsible for the flood embankments protecting the Camargue in France, gave an interesting presentation on his paper with Dr J-J Fry (also presented at the symposium). It describes a risk assessment process using the internal erosion bulletin to establish the locations and priorities for maintenance of 200 km of dikes, parts of which had failed in 1994 and 2003, particularly by concentrated leak erosion through badger setts (13 breaches out of a total of 16). Only a few setts had been completely excavated out and refilled. At most, the previous private owners had only plugged the entrances, thereby hiding the location of the remainder of the tunnels in the setts. One of the findings of the risk assessment was that sections of the embankments with gravel access tracks on the crest were at low risk of failure. This was because the tracks made it possible to find and fix potential breaches promptly before breach occurred. Another finding was a marked increase in risk of failure where soils of low plasticity were used.
Helge Saxegaard (Norway) reported that the new bulletin on asphaltic core dams should be complete in draft for the congress in Stavanger in June 2015.
Danie Badenhorst and Kelvin Legge (South Africa) led discussions on the update of the 1986 bulletin on geotextiles in dams. They wanted to make progress in the use of geotextiles in dams. Geotextiles could now be manufactured to meet requirements exactly, unlike granular materials, which could not be precisely processed, and were subject to segregation, settlement and cracking, cementing, decomposition and chemical attack, and faced increasing objections because of the environmental damage caused by sand pits and quarries. Badenhorst and Legge proposed, and it was agreed, that the new bulletin should deal with geotextiles in all functions in dams, not just filters and transitions.
How to address the issue of the alleged lack of durability of geotextiles was discussed. Kelvin Legge explained that the anti-oxidant content of geotextiles was the major component affecting durability. The antioxidant content is only about 0·25% in most geotextiles, but it accounts for about half the cost. Doubling the antioxidant content would increase the cost by 50%, but increase the durability enormously. We should specify what we want, not take what we are given by the manufacturers, and the new bulletin will give typical specifications. To overcome doubts about durability, it was proposed to recommend dual filters (geotextile and granular) in critical zones to ensure good early performance from the geotextile when the granular one may be subject to cracking on settlement on first filling. The granular filter would take over in the long term, when any cracks would have filled on saturation, and the geotextile may have perished.
The question of clogging of geotextile filters was discussed. Why does it seem to occur only in geotextile filters? It may be because geotextile filters are unable to fracture briefly and relieve uplift if it develops, or because insufficient drainage capacity is provided, or because it is laid on smeared surfaces or slurry, which prevent drainage. If more extensive use of geotextile filters is recommended, advice on precautions against clogging will have to be included.
The plan is to complete the geotextile bulletin for the ICOLD meeting in Johannesburg in 2016. Danie Badenhorst appealed for case histories, good and bad, of the use of geotextiles in dams. Some UK case studies have been supplied by Malcolm Eddleston, but as many as possible from all countries would be most welcome.
Antonio Soriano (Spain) reported that the Spanish report, which would provide the basis of the new bulletin on filters, transitions and slope protection, was progressing, but had not yet received full approval. This is expected in the coming months, and a draft of the new bulletin may be available for ICOLD in Stavanger in 2015.
Juha Laasonen (Finland) reported on the meeting of the European working group on the management of dam incidents – a matter of particular interest to the Dam Safety Committee. He presented his paper on the subject at the symposium.
Dave Paul (USA) gave a presentation on the geo-impulse programme, developed in the Netherlands to encourage the better application of geotechnical engineering and knowledge to reduce the number of failures and incidents. It had been successful, and was being taken up in other countries. ICOLD bulletins played a part in this by disseminating geotechnical knowledge, and also benefitted from it. For example, the new understanding of backward erosion presented in the ICOLD internal erosion bulletin came largely from work done in the Netherlands.
Ingvar Ekström (Sweden) made a presentation on the first secant pile wall embankment repair in Sweden. This was to replace old sheet piles, which had formed an incomplete cut-off and allowed what was thought to be contact erosion to occur in glacial deposits in the foundation.
Former ICOLD president Cassio Viotti (Brazil) gave presentations on the failures of three dams. One had failed by backward erosion in residual soil above sound gneiss below the dam. Another had failed because of unsatisfactory details at the connection between the dam and the spillway. At the third, on weakly cemented sandstone foundations with open joints, leakage had caused uplift below the spillway structure and it had been disrupted and transported downstream.
The symposium included presentations of papers on seven themes: social and environmental aspects; engineering issues in dam development; tailings dams; water quality management; catchment area management; safety policy and implementation; and climate change. As Rod Bridle explained when presenting his paper at the symposium, a benefit of the improved understanding of the mechanisms of internal erosion is that the water level at which it initiates can be estimated. The probability of occurrence of that water level can be estimated from the flood hydrology. If the consequences of failure are estimated by dambreak studies, the risk equation (risk = probability × consequences) and risk diagrams can be deployed to make rational and equitable policy decisions on dam safety improvements, to deal with increasing populations and climate change, for example. In response to a question, it was pointed out that the probabilities and the consequences are estimated and that risk studies are not exact, but make use of many ‘conventions’, conservative assumptions, to arrive at responsible decisions on the safety status of a dam.
3. Contribution to the workshop on the ICOLD bulletin on small dams
The ICOLD Technical Committee on Embankment Dams has prepared ICOLD bulletin 164 (available from www.icold-cigb.org) to give guidance on investigating and protecting existing dams against internal erosion. The bulletin applies to all dams – small and large. The bulletin will provide a lasting memorial to the more than 100 victims of the failure of Situ Gintung dam near Jakarta on 27 March 2009.
More was said about the Situ Gintung failure at the symposium (Bridle, 2014). Note that the failure mode was unpredictable, even with current knowledge.
The bulletin and experience from safety investigations of many small (and large) existing dams provide some simple engineering criteria in relation to the safety of small dams, as follows.
3.1 In relation to the consequences of failure of small dams
The minimum height of dam H that may cause loss of life is about 3 m. This can be said because if a dambreak flood has depth D (m) and velocity V (m) and D × V > 2, fatalities can be expected among people in the floodway. D, depth, close to the dam is often about two-thirds of H, or 2 m.
3.2 In relation to probability of failure of small dams
3.2.1 Probability of failure by overtopping
Small dams often retain small volumes of water in the reservoir. If a small dam retaining a small reservoir fails by overtopping, the volume of water that escapes from the reservoir may be very much less than the volume of the ‘natural’ flood water. In such circumstances, any fatalities or damage would be the consequence of the ‘natural’ flood, and the dambreak flood would have had little additional impact. In the UK, dams retaining ‘small’ volumes of water in the reservoir are defined as those containing less than 10% of the total volume of the 1-in-1000-year flood. If no fatalities are expected, only damage to the dam, then it is recommended that the spillway capacity of such dams is sufficient to pass the 1-in-150-year flood safely. In these circumstances, the annual probability of failure is approximately 1 in 150 years or 6·7 × 10−3. An alternative is to estimate the depth of flow over the crest that would cause failure by surface erosion, and estimate the probability of that water level from the flood hydrology.
3.2.2 Probability of failure by internal erosion
Bulletin 164 on internal erosion enables engineers to estimate the water level at which internal erosion initiates and may lead to failure if the dam cannot arrest it. The probability of occurrence of this water level can be estimated from the flood hydrology.
3.3 Use of the risk equation (risk = probability × consequences)
Whether small (and large) reservoirs impose high risk, moderate risk or low risk on people downstream can be estimated using the risk equation (risk = probability × consequences) as explained in the symposium paper and in the presentation. The risk equation can be used to make rational policy decisions to strike an equitable (fair and just) balance between the benefits of dams that are enjoyed by everyone and the risk of loss of life imposed on those living in the dambreak floodways downstream of dams.
