D. B. Edwards, B.H. Jackson & R. H. Wright
Ground anchorages are installed to support structures such as dams, slopes and tunnels. Failure of anchorages could be serious.
The condition of these critical supports is currently assessed by monitoring the load in the anchorages by either load cells or lift-off testing (jacking). Both methods are expensive and testing may damage the corrosion protection beneath the anchorage head.
A non-destructive testing method for ground anchorages needed developing and the UK Universities of Aberdeen and Bradford developed a testing system called GRANIT with patent applications on the system filed world-wide.
Full scale measurements were conducted during the construction of Penmaenbach and Pen y Clip Tunnels on the UK’s A55, where rock support was provided by prestressed rock anchorages. In all 9000 records of anchorage response were analysed.
A major finding from the research was that the response of the anchorages to the dynamic impulse motion produced by the blast loading depended on how the anchorage had been constructed and on the nature of the surrounding rock mass. If the prestress load in the anchorage was changed, or the free length increased, a noticeable change was observed in the response ‘signature’ as monitored by an accelerometer located at the anchorage head.
Applying a known impulse load to the anchorage head immediately after construction and measuring the response, provides a datum response signature for the intact anchorage. If the anchorage was to deteriorate in any way, eg loss of prestress, this should be noticeable on subsequent response signatures. This approach is the basis of the GRANIT system.
A short programme of anchor calibration testing for bolts was conducted in Hawkesbury sandstone in Sydney during March 1998 and developments in Australia and UK are proceeding.
— OR —
D. C. Green
The disaggregation of public water supply bodies in recent years has seen the functions of ownership, design and operation transferred to separate bodies. Consequently , issues of risk management associated with legal liability which previously could be ignored because all risks were absorbed in -house must now be faced and addressed in a more formal way.
This paper looks firstly at the general principles of legal liability for dam performance associated with construction and design, ownership of an existing dam and monitoring of its performance. Liability under several different areas of the law is discussed. Special issues associated with “design and construct” contracts are then highlighted, and warnings are given for project sponsors who control the letting of contracts and the briefing of consultants.
Richard I Herweynen
For concrete gravity dams, when the foundation’s value of cohesion is low, it is very difficult to meet the sliding criteria proposed by ANCOLD. Low cohesion is generally associated with serious foundation defects. This was the case for Meadowbank Dam, with a foundation having persistent horizontal seams containing material of a clayey silt size classification. By adopting the ANCOLD strength reduction factors, it was found that a large number of ground anchors would be required to meet the ANCOLD sliding criteria. During original design, extensive laboratory and insitu testing was performed on the seam material. This paper proposes a methodology for arriving at less severe strength reduction factors based upon a statistical analysis of the strength parameters measured in the Meadowbank Dam foundation.
Additionally, a probabilistic approach using a Monte Carlo simulation is used to give further weight to this argument. This paper concludes that the probability of Meadowbank Dam failing due to sliding is very low and within acceptable limits.
Leonard A McDonald and Chi Fai Wan
A risk assessment has been undertaken as part of a comprehensive review of the safety of Hume Dam. Use of risk assessment techniques, to assist in evaluating the safety of existing dams, is a relatively recent trend. Hume Dam was a particularly challenging subject for the application of risk assessment techniques at their present stage of development. The challenge lay in the number and diversity of dam elements to be analysed, in the number and complexity of the potential failure modes and in the fact that there were significant safety issues under normal operating conditions.
This paper outlines some of the key lessons learned from that phase of the risk assessment that was concerned with estimating the chance of dam failure. Some of the issues discussed have not previously been addressed in the literature and some demonstrate a clear need for improved analysis procedures.
Kurt Douglas, Matt Spannagle and Robin Fell
This paper describes a method for estimating the probability of failure of concrete and masonry gravity dams through the dam or the foundation. The method is based on the research and analysis of historic failures and accidents performed at The University of New South Wales over the last two years. The method accounts for dam type; age; foundation; height/width ratio; dam performance observations; and monitoring and surveillance.
Trevor Daniell, David Kemp and Jenny Dickins
Early February 1997 saw the occurrence of heavy rainfalls over a wide area of South Australia’s north. One of the worst hit areas was near Olary, in eastern South Australia, where over a three day period, rainfall totals up to 320 mm were recorded. Within this period, localised, short duration intense rain occurred. In one four hour period on 7 February, about 200 mm fell.
The rain produced floods that washed away large sections of the main Sydney to Perth railway and inundated long sections of the Barrier Highway. Repair costs were of the order of $6 m for the railway and $1.5m for the road. Damage to rural infrastructure in the region was substantial. Flows within the catchment would have been sufficient to wash away most stream gauging stations.
The airmass over much of South Australia was of tropical origin, contained a high amount of moisture and was unstable. Thunderstorms were the main rain producer, consequently the event was characterised by localised, very intense rain episodes. This contrasts with the March 1989 floods, where it rained at a fairly steady rate over large areas for durations up to 24 hours, as a monsoon low tracked across the state.
Analysis of the depth-area relationship for the Olary storm indicates that the relationship to be used for design purposes should be the humid area relationship of Australian Rainfall and Runoff, not the arid area. This is reinforced when it is considered that the 1997 rainfall was localised, not general rain as in 1989.
Investigation of the event indicates that the Olary Creek catchment experienced overland flow, resulting in much higher peak flows than would occur with more frequently occurring “normal” processes. It is possible that any catchment may change its behaviour with extreme rainfall, and produce flows well in excess of those predicted with currently available runoff routing models, or flood frequency analysis of “normal” events.