Following the failure of Paloona Dam’s intake trashrack during the 2016 floods in northern Tasmania, a replacement trashrack and support structure was designed, manufactured and installed (by diver) within five months. This was a remarkable feat and hailed as a success at the time.
The euphoria, however, was short lived. A routine dive inspection in January 2018 revealed cracked
trashrack bars on one of the panels and this was after less than twelve months’ operation. This prompted a rigorous investigation where it was determined that the bars suffered fatigue due to flow induced vibration. Indeed it is possible that the bars cracked within a few weeks of returning to service.
The science of flow induced vibration is relatively mature, having been extensively researched over several decades. Its application to trashracks is well documented. However, this experience has shown that the common design approach overly simplifies the fluid-structure interaction. For Paloona, the result was a trashrack design which has proven to be inadequate, not having the resilience required for a dam outlet works component.
This paper revisits flow induced vibration theory as it pertains to trashracks, outlines the findings of vibration testing at Paloona, and suggests a design approach which will avoid similar issues. It is hoped that similar failures can be prevented and the design life expected of trashracks achieved.
Design floods for most dams and levees typically have an annual exceedance probability (AEP) of 1:100 (1E-2) or less frequent. In the U.S., high hazard dams are designed to pass the Probable Maximum Flood (PMF), which typically has an AEP of 1:10,000 (1E-4) or less frequent. In order to reduce epistemic uncertainties in the estimated AEP for extreme floods, such as the PMF, it is important to incorporate as much hydrologic information into the frequency analysis as reasonably possible. This paper presents a Bayesian analysis framework, originally profiled by Viglione et al. (2013), for combining at-site flood data with temporal information on historic and paleofloods, spatial information on precipitation-frequency, and causal information on the flood processes. This framework is used to evaluate the flood hazard for Lookout Point Dam, which is a high priority dam located in the Willamette River Basin, upstream of Portland, Oregon. Flood frequency results are compared with those from the Expected Moments Algorithm (EMA). Both analysis methods produce similar results for typical censored data, such as historical floods; however, unlike the Bayesian analysis framework, EMA is not capable of incorporating the causal rainfall-runoff information in a formal, probabilistic manner. Consequently, the Bayesian method considered herein provides higher confidence in the fitted flood frequency curves and resulting reservoir stage-frequency curves to be used in dam and levee safety risk assessments.
The Keepit Dam Safety Upgrade Project is being implemented to bring the 54m high concrete gravity dam in line with current guidelines for flood and earthquake loading. Stage 2A of the project involves the installation of two vertical 91 strand post-tensioned anchors on each monolith of the spillway section.
During coring of the anchor head blocks for the vertical anchors, deep cracks were observed across some monoliths, dipping diagonally in an upstream direction. In two of the monoliths the cracks were found to be continuous enough to possibly daylight at the upstream face and form freestanding blocks. If the freestanding blocks postulate is correct, the block stability could be currently relying on the friction of the cracked surface and on the engagement with shear keys of adjacent monoliths, which are provided in the vertical contraction joints.
This paper will explain the complex 3-D nonlinear Finite Element Analysis (FEA) conducted to replicate the conditions of the cracked spillway monoliths during the post-tensioned anchor installation. The nonlinearity captured the expected opening, closing and sliding of the crack, as well as its potential pressurisation, and the residual shear strength retention due to asperities of the crack surface. For the shear keys of the vertical contraction joints, the nonlinearity captured the force-deformation relationship of the plain concrete, up to a brittle failure condition if the shear strength threshold was reached.
The 3-D nonlinear FEA was also used to design the optimum number of Macalloy post-tensioned bars required to stitch the freestanding block to the monolith, so that the vertical anchors can be safely installed. In addition, the remedial design accounted for future extreme design flood and extreme earthquake loading conditions, the latter modelled with a time-history analysis.
Many quantified risk assessments finish the failure mode event tree at the estimated occurrence of an embankment breach leading to dam failure outflows and downstream consequences. In some situations, for dams with multiple embankments with potentially different consequences downstream of each embankment, the possibility for further breaches may be pertinent if there may potentially be higher consequences for a multiple breach scenario. The location of an initial breach and sequence of subsequent breaches could also result in different contributions to total risk.
This paper discusses a method applied to investigate the conditional probability of flood overtopping breaches for multiple earth-fill embankments with grass covered downstream slopes.
For the subject dam, preliminary modelling identified that for a flood overtopping breach of an embankment the breach’s development may not be sufficient to reduce the lake level and sustained overtopping flow over the remaining embankment crests could lead to further embankment breaches.
A Monte Carlo dam breach simulation modelling approach was used with a large number of flood events. The simulation modelling considered erosion initiation for a grass slope due to the combination of velocity and duration of flow, and erosion continuing to breach based on duration of flow after erosion initiation. Potential uncertainty of erosion initiation and erosion continuing to breach were represented with probability distributions in the Monte Carlo modelling.
The results from the large number of dam breach simulations were then analysed with post processing to derive conditional probabilities for single or multiple breaches and breach sequence.
Leslie Harrison Dam is located on Tingalpa Creek in the Redlands region, approximately 18 km southeast of Brisbane. It is classified as an extreme hazard category dam with a large population at risk only a short distance downstream.
The dam comprises a 25 m high zoned earthfill embankment, with a dry well concrete intake tower and an outlet conduit located at the base of the dam near the old river channel. The spillway has a 43 m wide concrete gravity ogee crest, with a concrete lined chute terminating in an energy dissipator structure.
Seqwater is undertaking a staged upgrade of Leslie Harrison Dam to address deficiencies identified during the Portfolio Risk Assessment (URS 2013) and Geotechnical Investigations (GHD 2016).
While the dam has met the water supply needs of the community for the past 50 years, the upgrade ensures local residents will be well served into the future. Additionally, the structure will meet the most up to date requirements of dam safety management and national industry standards.
Construction of the Stage 1 upgrade commenced in June 2018 and involved the removal and replacement of liquefiable material in the foundation, modernisation and extension of the outlet works, addition of a new downstream filter buttress to the embankment, and lastly, the installation of both active and passive anchors within the spillway ogee and lower chute floor.
As with any major project, the works involved a number of challenges that had to be addressed. This paper provides an insight into the key challenges encountered and how these were overcome by the design and construction teams using practical engineered solutions. The intent is to provide the reader with an account of the “lessons learned” during the construction phase, along with recommendations for future dam upgrades.
As part of the development of some dams and hydroelectric power schemes, deep infrastructure is often required which requires and understanding of the in situ stresses of the rock mass. Recent works completed in southern Australia and Europe have led to improved methodologies for conducting effective, reliable, and repeatable measurements of in situ strain and/or deformation, as well as the subsequent estimation of in situ stress.
In situ stress testing is generally an item that is specified as part of a geotechnical investigation, however it is not commonly well understood in terms of reliability, repeatability, or, in fact, what the result actually means and its implications to project design. Commonly, a handful of tests are completed, with variable results, which often generates more confusion than answers.
This paper provides a discussion of recent in situ stress testing completed for two deep Australian projects. It summarises the aim of the investigations, test selection process, laboratory testing, data review and model development. This is to illustrate how complex the estimation of in situ stress can be and some of the pitfalls that may be avoided whilst acquiring and assessing the data. It also examines several different testing methods available in the Australian and International industry and some of the analysis techniques available to dam and tunnel projects. Finally, the paper provides an update on topical developments provided at recent workshops in Europe.