Operating experiences - BDS conference report28 November 2006
The British Dam Society’s biennial conference, held in Durham, UK, offered delegates the chance to learn more about improvements in reservoir construction, operation and maintenance. Carrieann Davies reports
This year’s British Dam Society (BDS) biennial conference – with its theme ‘Improvements in reservoir construction, operation and maintenance’ – appeared to be very timely, as recent events illustrate the importance of efficient O&M practices. For example, in October 2006, AmerenUE, a subsidiary of Ameren Corporation, agreed to pay the US Federal Energy Regulatory Commission (FERC) a record US$10M civil penalty following the breach of Taum Sauk pumped storage project in Missouri on 14 December 2005. In proceedings following the event, FERC staff alleged that the company committed 15 violations of various FERC regulations and licence conditions, including failure to notify FERC of conditions affecting the safety of the project and failure to use sound and prudent engineering practices. In entering into the settlement, AmerenUE neither admitted nor denied the alleged violations. On 7 April, the company submitted a report prepared by an independent consultant it retained under a directive from FERC, whose dam safety staff then conducted an exhaustive forensic investigation of the breach. The report found that substandard construction and specification and maintenance of instrumentation and control systems were partly responsible for the breach, which released 6.3B litres of water, causing extensive environmental and property damage.
Although such disastrous incidents are rare, it is important that lessons are learned from such events. The BDS conference – held in September at Durham University, UK – offered delegates just such an opportunity to learn from previous rehabilitation schemes, while addressing many of the issues that face dam engineers today.
One particularly interesting paper presented during the event looked at plans by the Environment Agency (EA) for a system of recording ‘near-miss’ incidents at UK dams, a move designed to help reduce the risk of flooding from reservoirs across England and Wales.
In his presentation, Ian Hope, the EA’s Technical Manager for Reservoir Safety, announced details of a Post-Incident Reporting System (PIRS) to help inform the reservoir industry about what incidents have happened and how they were dealt with.
Hope says: ‘Since legislation was introduced in 1930, the UK safety record for dams has been good. But learning from the experiences of those handling accidents and emergencies has been shared on a purely ad hoc basis.’ He urged reservoir owners (undertakers) and panel engineers (specialist engineers) to comment on the proposed system ahead of its introduction in January 2007.
The EA has been working with the Department for Environment, Food and Rural Affairs (Defra) and the Welsh Assembly Government (WAG) to develop PIRS.
‘Appreciating that other potentially hazardous industries have formal reporting systems we, Defra and WAG have recognised the need for something similar in the reservoir industry and we want to hear what people think,’ says Hope. ‘We will then use the comments to fine tune the system before it comes on line.’
Although the number of incidents on UK dams is comparatively low, reservoir structures can move with the seasons, deteriorate with age and be affected by natural forces such as earth tremors. This can lead to various consequences from minor leaks to major flooding.
From January 2007, undertakers can call the EA’s Reservoir Safety Team to volunteer details of an incident. Panel engineers will also be encouraged to support the system by providing information as part of their regular reservoir inspections.
Hope adds: ‘No one will be legally required to give information because we would rather develop a successful system where people feel able to report freely.’
In straightforward incidents, details will be taken down in a 10 to 15-minute phone call but, in more complex cases, the EA may offer to send a panel engineer to the site to discuss events in more depth.
Details about incidents and the lessons that have been learned will be published through a regular bulletin to undertakers and panel engineers, direct communication with the industry and via the EA’s website at www.environment-agency.gov.uk.
To avoid a project appearing on such incident systems, it is important to ensure that efficient refurbishment, construction and maintenance techniques are carried out. The paper, ‘Refurbishing and upgrading old spillway gate installations’ by J Lewin, G M Ballard and P To looked at how many old spillway gate installations systems are approaching the limit of operating life in their present form. The majority were designed and constructed to be robust, with little or no redundancy, and incorporate many single point and common cause failures. Design deficiency, age and degradation present a reliability risk to the dam system and downstream population.
There have been failures of spillway gate installations, resulting in cases of dam collapse. The paper mentions the Machhu II project in India, when gate malfunction during a catastrophic flood in 1979 caused overtopping and washout, with an estimated 2000 people killed. In 1982, the Tous dam in Spain overtopped during an extreme flood event, leading to 16 deaths. In fact it has been suggested that about 13% of dam failures are associated with a spillway gate (Foster et al, 2000).
As dam safety has become an ever more important subject, there has been a greater realisation that spillway gate installations are part of the dam system, and as such they need to be reliable. The paper identifies issues with spillway gate installations which require investigation and resolution, including: provision of reliable standby electrical power arrangements; installation of robust electrical power distribution with redundancy for gate operating equipment; adequate segregation and protection of standby and operating equipment; improved protection against hazards such as lightning and fire; comprehensive assessment of seismic resistance; maintenance of gate roller bushes at vertical lift gates; maintenance of seals; reliable reservoir water level instrumentation; formal staff training; planning of major maintenance and replacement operations; and analysis of operational and maintenance records.
A reliability assessment will identify a range of deficiencies in a spillway gate installation which will have a detrimental effect on the reliability of flood discharge, says the paper. Not all issues will have the same significance, and the paper admits that there is no single simple answer to questions of relative priority.
An evaluation of the relative importance (or effect on reliability) of remedial works and improvements can act as a guide in the decision making process and form the basis of a programme of work. This can take the form of a simple reliability model of the dam flood discharge facilities, such as that constructed for BC Hydro after an extensive survey of spillway gate installations.
The paper gives as an example the reliability assessment of three 50-year old spillway gate installations which identified the following significant contributors to potential failure:
•Failure of the gate electric motor or associated equipment - 30%.
•Failure to raise a gate because motors are overloaded due to seized rollers (bearings or debris) - 20%.
•Failure of electric power supply to the gate - 15%.
•Failure of gate mechanical drive - 15%.
•Failure of gate control equipment - 10%.
All this shows that unless regular refurbishment is carried out over the years, the scope of work and costs involved in upgrading an old spillway gate system to meet modern reliability standards can be very significant. As part of the asset management process, the dam owner may have to evaluate the need to invest money in gate refurbishment against other priorities, and consider the potential consequences of deferring or failing to take action.
The paper points out a number of factors that need to be considered when prioritising improvements. These include: function of the gate system within the dam system; consequence of gates failing to operate on demand; reservoir and operational characteristics; and interim control. As the paper says, improving the reliability of a gate installation is not a one off investment. Dam owners must recognise that a recurrent investment of resources will be required to upkeep an improved standard of reliability in the long term.
Another interesting paper that came from the session ‘refurbishment, construction and maintenance’ looked at developments in grouting processes, and how inefficient grouting procedures can affect dam safety. The paper ‘Application of modern grouting technology to remedial works on dams’ by A K Hughes and C T Kettle gave a historical background to the grouting process and identified the need for change. It also outlined the GIN process and illustrated the rapid changes in the grouting process based on computer packages.
The paper says that historically although much excellent work was carried out by experienced and professional specialist contractors using traditional methods, there was clearly potential for many errors of execution or interpretation. Some of the key areas of risk include:
•At the mixing station, grout batching could be very inaccurate, requiring ‘reconciliation’ at the close of each shift between the dry weight of material delivered and used, the volume of grout mixed, and the injected volume measured from observation of the agitator tank.
•Errors in mixing grouts of differing water/cement ratios, errors in measuring the volumes prepared in the mixer, errors in measuring the volumes pumped from the agitator, all compounded to produce large discrepancies.
•Poor mixing practice, inefficient mixers, and the practice of starting with ‘dirty water’ style mixes of water cement ratios as high as 10:1, even 15 or 20:1, all combined as factors in allowing the injection of thin, unstable mixes.
•Poor control of pressure at the point of injection arose due to inadequately damped pressure and flow from reciprocating pumps, non-continuous observation of the manometer and/or control of the return line valve, and inaccurate, poorly calibrated, manometers. As a result, limiting hydro fracture pressures were often exceeded, and readings of oscillating pressures at the gauge were inaccurate. The resulting analysis on the basis of the final injection pressure achieved took no account of the time/pressure relationship, and often did not identify or highlight premature blocking off of the fissure or borehole.
•A related problem was that of poor control of hydro-fracture in the formation, or breakout of the grout at the surface, due to poor observation of the fall in grout pressure under these conditions. This led to a potential weakening of the formation, an excessive volume being injected for the given location, and to the execution of subsequent supplementary injections due to an apparently high absorption.
•Frequent changes from one mix to another added hugely to the confusion on site and the difficulty of analysing results.
This is just a brief summary of some of the major problems associated with controlling the injection process. Added to problems in controlling drilling practice and accuracy of drilling, says the paper, and the whole field of water testing control and design, it can be seen that there was much the industry needed to do to ensure good design, best practice and adequate control.
Over the years many incremental practical improvements have been made to the manual control which have helped significantly in improving the quality of work. These include: introduction of automated batching plants; pressure damping devices; improved pumps; high speed grout mixers of improved efficiency; improved quality and reliability of manometers; introduction of pressure recording rotary graphs; introduction of flow meters and pressure relief valves; reduction in water/cement ratios; improved grout materials and additives; better quality, more controlled regulating valves; better understanding of the theory and practise of grouting; and the introduction of routine Lugeon testing.
Not withstanding these improvements the end result was dependent entirely on the skill, observation, diligence, and honesty of field personnel, and the quality of management and planning of the site operations by the contractor.
The paper goes on to explain that in the late 1970s and early 1980s a new concept in rock grouting developed by Lombardi, Deere et al was introduced. The GIN system involves the application of a constant factor P x V (pressure v cumulative volume) for each injection, allowing for a more accurate and dynamic treatment of fissured rock and a better engineering evaluation of the injection process. The method sets out a limiting pressure envelope which decreases according to the volume injected, until the limiting volume (Vmax) is reached. This procedure is very useful in formations where grout absorptions are expected to vary significantly, such as Karzstic limestone, in weak formations susceptible to hydro-fracture, and in very open formations where excessive grout travel is likely.
This process, says the paper, although recognised immediately as a major breakthrough, was difficult to implement for a number of years. The boundary curve had to be modelled by instructing a series of decreasing steps of constant injection pressure, each linked to a limiting volume, creating a series of injection domains which were touching the boundary curve. All this had to be controlled with great diligence on site by very experienced operators and supervisors.
The availability of the personal computer however helped solve this difficulty. The next step, says the paper, was the development of reliable software to collect injection data, record, analyse and present results in a clear visual format. This was followed by the use of the new computing technology to dynamically control the injection pumps in such a way that the pressure and injection rate can be automatically adjusted throughout the injection according to the design parameters.
Major specialist contractors have either used or developed their own equipment to exploit the advances in theory, technology, and practice. The paper gives the example of a software package developed by Bachy Soletanche which consists of a suite of independent but compatible interactive programmes.
•ENPASOL is a package of sensors, read-out device, and software mounted on the drill rig to collect data such as torque, rate of advance, bit pressure, flush pressure, hydraulic feed pressure etc in real time as drilling progresses. The data is presented as a continuous graphical profile of these parameters plotted against depth. These plots, which are visible to the drill operator during drilling, can be later combined by the software in different ways to highlight variations in ground density and cohesion, giving a measure of the competence and consistency of the ground. These plots are used in planning the injection programme.
•CASTAUR is a design tool which creates a 3D model of the site; optimises the location of grout holes and produces graphical plots and tabulated instructions for the setting out and drilling crews, defines and calculates the theoretical volume of ground to be treated based on the soil and rock characteristics, and established a 3D structure for the SPHINX database within which the grout injection data can be analysed and presented spatially.
•SPHINX organises information into a database which processes data on all aspects of the grouting works and provides all the formats for day-to-day management of this huge amount of data.
•SPICE is the software which physically controls all grouting operations from a PC computer installed at the grouting plant, and is linked to a specialised programme which servo controls the electro hydraulic grouting pump.
•SCAN 3-D is a highly configurable software which produces sectional CAD drawings which display the grouting results and boreholes in the correct spatial position for easy analysis.
This system has been employed on a large number of dam grouting projects, including the Piedra del Aguila dam in Argentina which employed over 100 grout pumps and six computerised grouting plants, with 18 PCs for technical supervision.
According to the paper, the main advantages of such a system include: pre-determined grouting procedures and parameters can be pre-programmed into the software; local, semi-skilled personnel can be quickly trained to an operational level; computerisation enables optimisation of design, control, quality control, and production efficiency; and reporting is fast, automatic and accurate.
In the end, says the paper, the desired effect must be to free the specialist engineer from hours of routine data processing, allowing them to employ analytical tools rapidly, giving quality engineering time to apply skills and experience to interpret the data and refine the grouting parameters and programme for maximum effect and efficiency.
A number of examples were presented during the conference of successful rehabilitation projects. For example the paper ‘Emergency Underwater Rehabilitation of the Poti Main Diversion Weir, Georgia’ by Ljiljana Spasic-Gril reports on a dam safety project implemented by engineering firm Jacobs from August 2003 on the main diversion weir on the river Rioni in the town of Poti, the third largest town in Georgia.
The diversion structure, which discharges flows of 4000m3/sec through 10 identical openings, was constructed in the 1950s to protect the town of Poti from frequent flooding.
As part of the Irrigation and Drainage Community Development Project financed by the World Bank, Jacobs carried out a safety evaluation of the main structure from August 2003 to July 2004. An underwater survey was undertaken as part of the safety evaluation and the results showed that the foundation condition had deteriorated rapidly since the last inspection in 2003. Emergency rehabilitation works were recommended and the implementation of the works was carried out from August 2004 to March 2006.
An underwater survey by divers was conducted from January to March 2004, firstly when the gates were closed and then after the gates had been opened and sediments were flushed. It was found that the downstream erosion line was highly dependent on the operational regime of the gates and the flow of sediments, related to the volume of the discharged water. Scour was also taking place at the edge of the main structural foundation in the area of the central sections and pier No 5 when 3-4m deep voids were encountered. Other findings were that stilling slabs had been completely demolished in the central sections and partly in the remaining sections, the downstream apron had been completely damaged and had largely disappeared, and the upstream apron had been damaged in the central part.
Based on the results of the investigations, three main areas of concern were identified: Can the structure withstand the hydraulic forces? Can the structure withstand the seepage forces? Can the structure cope with the potential loss of bearing capacity due to the 3-4m deep voids under the foundation and increased hydraulic gradients?
All these areas of concern were analysed and discussed, and as a result emergency rehabilitation works were recommended. As it was not possible to divert the river away from the weir, works had to be undertaken underwater, during a complete closure of the 10 gates.
The following actions were recommended:
•Prohibit the use of the diversion structure by heavy traffic until emergency rehabilitation works were implemented.
•Rehabilitate hydromechanical and electrical equipment on the main regulating structure and the left bak regulator.
•Clear the Poti canal allowing discharge of flows of up to 4000m3/sec.
•Backfill the voids (approximately 500m3) close to the structural foundation, in the areas of the central sections and the pier no 5, that are associated with erosive processes and are likely to have impact on the structural stability. This was carried out by placing sand-cement-clay mix underwater through tremies installed through voids identified in the broken stilling slabs, close to the structure. Prior to the backfilling a site trial was carried out to check the exact mix and the pumping pressures. Quality of backfilling works was checked by drilling and sampling of the backfilling.
•Reinstatement of the stilling slabs and the downstream apron in the areas where the slabs and the apron are damaged. This was carried out using Maccaferri gabion baskets which were assembled on the river bank, dropped into position from barges and interconnected underwater by divers. A geofabric was placed along the outer sides of the gabion baskets to prevent upward migration of the fines from the foundation soils into the gabions. Prior to placement of the gabions, a bedding layer of sand and gravel was placed over the scoured foundation to level the area. In places where the gabions were to be placed over broken slabs close to the main structure, smaller and more flexible cylindrical gabions were used, 1m long and 0.65m in diameter. They were also used at the ends of the ‘gabion structure’ to provide a transition between the gabions and the natural river bed. Divers were employed to ensure that the gabions were properly placed and interconnected.
Detailed design of the emergency works was carried out by GWP under guidance and review by Jacobs. Works started in September 2004 with rehabilitation of the hydromechanical and electrical equipment on the main regulating structure and the left bank regulator, and clearing of the Poti canal. These works were delayed and had an impact on the programme of placement of the gabions. Between February and March 2005 the works were undertaken with full closure of the gates, except for a short period at the end of February when the flood flow was about 1400m3/sec and gates 1,8, 9 and 10 were opened. After passing the flow a bathymetric survey was carried out, where it was found that the bedding layer and gabions installed and not been affected.
The works had to be temporarily halted on 15 April 2005 due to high flows in the river Rioni. By 15 April, backfilling of the voids, placement of the bedding layer and installation of gabions were completed in the central sections up to an elevation of -6m. The flood period lasted till mid August, with the maximum flood of 3000m3/sec being discharge through the weir at the end of April 2005. Underwater surveys carried out after the flood showed that the gabions were unaffected by the flood releases, and that the river bed in the areas downstream of the placed gabions was not subjected to additional scouring. The works were then resumed and successfully completed in February 2006.
The papers mentioned above were just a selection of the many informative and interesting presentations given which stressed the importance of efficient construction, operation and maintenance programmes. These papers together demonstrate that if dam projects are not efficiently managed, serious consequences can occur.
For more information on the British Dam Society’s 14th Biennial conference, please visit www.britishdams.org