Monitoring to support the integrity of structures immersed in water has been a challenge for dam engineers and other construction technicians for decades. In addition, it is known that the planning and execution of these activities are often linked to high costs and risk assessment. Non-destructive methods can act in these tasks with the production of useful information for a more optimized decision making. Ultrasound (US) is a wellknown tool that works with the emission and reception of wave signals. US stands out mainly for providing low radiation to operators, high resolution and flexibility in defining sensor arrays. Data acquisition can be performed even in immersed and tilted geometry environments. This article presents benchtop measurements from an approach focused on the microscale study of upstream slopes of earth dams. The model was developed to simulate the approximate geometry and composition of a dam slope. Traditional ultrasonic acquisition techniques are applied using immersion transducers. The investigations carried out in a model with dimensions of 30 cm, 30 cm and 100 cm, rip rap under 2H:1V inclination, central frequency of the transducer of 2.25 MHz and depth of water for measurement of up to 20 cm, demonstrated the ability to classify the ultrasonic reflection in contact between water and the different types of interfaces that mimic typical dam materials. Lessons learned from this microscale development should be tested in a controlled reservoir model before moving to field applications. The perspective of this study is to generate information capable of composing a database for machine learning and subsequently assist in decision-making for engineering solutions that act in maintenance, safety and predictive interventions in dam bodies.
Read full abstract