2024 – Concrete Gravity Dam and Other Spillway Assets – Automated Data Driven Condition Assessment using Unmanned Aerial Vehicle (UAV)

Harsh Rathod, Hetal Parmar

This paper discusses the use of Artificial intelligence (AI) and Machine Learning (ML) for Dam safety – Surveillance and inspection. The methodology includes collection of different datasets such as including visual/optical imagery models, thermal imagery datasets, and a unique acoustic sounding datasets mainly using an Unmanned Aerial Vehicle (UAV) equipped with a patented sensor pre programmed to replicate a traditional hammer sounding. Although the use of technology for dam safety is becoming increasingly popular, the question still remains on how to best utilize the datasets for dam safety management and maintenance workflows. There are over 170 large concrete dams and spillways in Australia. These assets are aging and that poses a significant challenge for dam safety professionals to maintain the assets in a predictive way instead of reactive.

This case study presents a methodology on collecting various datasets, processing the datasets, and developing the framework for the data analysis using different platforms, techniques, and processes. In this case study, the authors attempt to demonstrate how the analytics are being used to aid dam safety inspections program of some large dams within WaterNSW’s dams portfolio in Australia. The analytics/condition survey outcomes, accessible within a collaborative Geographical Information System (GIS) framework, is intended to assist dam safety engineers at WaterNSW to conduct concrete condition analyses utilizing digitized defect data in both two-dimensional (2D) and three-dimensional (3D) domains. The acquired raw data and processed analytics insights contribute to support informed decision-making processes related with respect to inspection, monitoring, and maintenance, thereby facilitating a comprehensive approach to asset management. Finally, the establishment of a high-resolution digital baseline further facilitates precise change detection in subsequent surveys, substantiating through the development of deterioration models. The technology presented has been peer reviewed and has been deployed on Niricson’s over 50 completed projects around the world, including 10 within Australia on dams and bridges. The data and analytics produced provide dam safety engineers, asset managers and other professionals and fostering an advanced automated condition assessment for proactive decision-making for prudent dam safety management and capital investment.