The confluence of several technological innovations including drones, photogrammetry, and thermal imaging has enabled the development of a novel approach to defect mapping and monitoring for large dams. A pilot project trialling the methodology was completed at a rockfill embankment dam with a concrete spillway and is presented as a potential means of improving the accuracy and reliability of condition monitoring. The pilot project included two main objectives: digital inspection and mapping of defects within the concrete spillway; and drone-based photogrammetric survey of the rockfill embankment. Defect mapping of the concrete spillway utilised drone-based photography and Structure from Motion (SfM) photogrammetry to develop a high-fidelity 3D model, from which visual defects could be identified and mapped in a virtual environment. Thermal infrared (IR) imagery of the structure provided an indication of potential shallow subsurface defects in the concrete. Photogrammetric survey of the embankment structure utilised drone-based photography, SfM photogrammetry and a network of precisely surveyed ground control and verification points to develop a georeferenced point cloud, digital elevation model and elevation contours. The results of the project were delivered via a web-based digital twin which included georeferenced results from defect mapping, aerial survey and tools for visualisation, measurement, and reporting.
Gideon Steyl, Ralph Holding, Lis Boczek
A Monte Carlo method for assessing liner systems is applied with outcomes demonstrating the range of discharge that could occur over the liner interface. The Monte Carlo approach allows for variation of fill material over the liner system and includes the assessment of a second compacted zone either above or below the liner zone. In this paper clay liners were evaluated due to regulatory guidelines and it could be demonstrated that similar performance to a 1 m clay liner could be attained using compacted material to reduce discharge over the liner interface. The approach applied in this paper allows for at least a worst-case quantification of seepage risk which could be included in liner selection criteria or presenting liner options to regulators.
Shane McGrath, Mark Arnold, Josh Rankin, Gavan Hunter
Greenvale Dam is a critical storage for the supply of potable water to Melbourne. The dam had been upgraded through current risk management techniques, and an ALARP assessment completed at that time. However, it was decided that a more comprehensive demonstration of ALARP was warranted to satisfy the dam owner’s duty of care. Since there is no comprehensive guidance in the dams industry for owners and their advisors to reference, the safety case approach used extensively in other hazardous industries was adopted. Considering the approaches used by Victoria’s Worksafe, the Institution of Engineers Australia and the National Offshore Petroleum Safety and Environmental Management Authority (NOPSEMA), the key components of the safety case for Greenvale dam were identified then developed to provide a logical, structured and comprehensive argument for the safety of Greenvale Dam. This paper provides an overview of components of the safety case developed for Greenvale Dam, the use of safety cases for dams and where process improvements could be made.
Hench Wang, Peter Hill, Sam Banzi, Muhammad Hameed
Dam owners can often struggle to demonstrate the dam safety risk benefits that can be achieved through non-structural risk reduction measures, such as adoption of smart technological solutions that improve the timeliness and quality of decision making. WaterNSW collaborated with HARC to develop a novel way of demonstrating benefits from improved data management. This paper discusses the use of HEC-LifeSim to demonstrate the reduction in life safety risk from improved monitoring through DamGuard for a case study dam in Sydney. DamGuard is a real-time dam safety monitoring system implemented by WaterNSW. This case study was the first time in Australia where a simulation model such as HEC-LifeSim was applied to quantify the life safety risk benefits pre and post the implementation of DamGuard. The implementation of DamGuard to the sample dam was estimated to reduce the life safety risk by 15%.
Qian Gu, Joshua Chan
Tailings Storage Facilities (TSF) constructed using upstream methods may have static liquefaction risks due to the strain softening behaviour of contractive tailings. Conventional Limit Equilibrium Analyses (LEA) using either peak strength or residual strength fail to address the stress-strain compatibilities between materials at different stages of softening or hardening, resulting in over or underestimating embankment stabilities. Static numerical analyses (Finite Element or Difference) are unable to identify the threshold stability due to their inability to converge close to or beyond equilibrium conditions.
In this study the failure triggering process is modelled with dynamic Finite Element Analyses (FEA) with the stress-softening behaviour of contractive tailings simulated by Norsand Model. The embankment failures are identified by either non-zero residual velocities along downstream face, or a drop in average shear stress along potential failure surfaces under increasing disturbing surface pressure. Threshold disturbing surface pressure estimated using these two methods are in close agreements. Factor of Safety (FoS) values estimated from peak mobilised shear strength are found to be between those estimated using the peak and residual shear strength in LEA. q-p’ stress paths in tailings clearly show the stress ratio increasing towards and beyond instability ratio during undrained triggering process. The developments of zones of shear softening and p’ reduction with increasing undrained disturbances help visualise the failure triggering process.
Andrew Northfield, Peter Hill, Muhammad Hameed, Hench Wang, Sam Banzi
In 2018 WaterNSW undertook a Portfolio Risk Assessment (PRA) for 20 dams across the greater Sydney area.
This paper describes the estimation of consequences for this large and diverse portfolio of dams. For some dams the population at risk were greater than 100,000 people whereas for others there were no permanent PAR which required the careful consideration of itinerants. This diversity of the dams required that the approach for estimating the consequences be tailored to the specific characteristics. For example, the approaches for estimating the potential loss of life (PLL) varied from a detailed simulation model (HECLifeSim) to a simpler empirical approach (Reclamation Consequence Estimation Methodology (USBR, 2014) to bespoke consideration of itinerant campers and users of walking tracks. For some dams the economic costs were driven by direct infrastructure costs whereas for other the indirect costs dominated the total economic cost for failure.