Structural health monitoring of ageing off-river gravity dams: a framework and case study

This research focuses on off-river gravity dams, a subject not well covered by the existing literature. Dams play a pivotal role in modern society, serving as essential infrastructure whose failures can have catastrophic consequences. Ageing concrete dams in Australia and around the world, particularly those constructed in the early 1900s under less stringent safety standards than are currently enforced, present significant challenges regarding integrity, monitoring, and risk management. These challenges stem from the increased risk of failure as these structures age, with many nearing or surpassing their design lifespan. This concern is further compounded by the limited precedent for managing ageing concrete dams and the lack of understanding surrounding concrete dam failure progression, including the complex interplay of structural and geotechnical mechanisms. Several factors contribute to the persistence of the knowledge gap, one of which is the relative infrequency of concrete dam failures to date. Monitoring an off-river gravity dam involves addressing three questions: What parameters should be measured? Where should these measurements be taken? And which values should trigger an alarm for stakeholders? This thesis provides valuable insight into these questions by exploring material investigations, collecting performance data with subsequent statistical analysis, and evaluating stability using both numerical and analytical methods. Moreover, it presents a general framework for the structural health monitoring of off-river gravity dams. This framework is organised into phases, beginning with a preliminary understanding of the structure, followed by empirical data collection of material properties and performance metrics. Statistical and numerical analyses then guide long-term monitoring strategies and the establishment of alarm thresholds. A case study involving an eight-metre-tall off-river gravity dam was used as a benchmark to supplement the framework. The dam’s serviceability crest deflection was primarily affected by annual temperature fluctuations, leading to variations of 1-2 mm. Conversely, deflections due to changes in headwater levels were minimal, with a variation from zero headwater to full serviceability loading resulting in approximately 0.1 mm of crest deflection. To statistically model this interaction effectively, at least two years of data collection is recommended, most notably to parse the time variable adequately. Numerical modelling revealed that multiaxial loading, or confinement, significantly influences the system’s equilibrium path by enhancing ductility and strength, which is essential information for establishing alarm thresholds. Moreover, the most likely failure progression involves overturning, culminating in tensile failure at the concrete-rock interface.