PLT - Building a Digital Defect Timeline for Concrete Dams and Spillways – A Pathway to Predictive Monitoring for Dam Safety Engineers

This study presents an innovative method for concrete dam health monitoring that utilizes a defect detection and quantification software coupled with photogrammetry to create mm-scale digital defect baselines of a critical concrete dam and spillway. The software utilizes optical imagery, thermal imagery, and acoustic sounding points from a novel Unmanned Aerial Vehicle (UAV)-based data collection technique and Machine Learning based image processing to detect and quantify surface and subsurface defects such as delamination, voids, and cracking.

In this case study, Niricson partners with a US dam owner to collect data on a run-of-river concrete dam to construct a high-resolution digital baseline model with mm-scale surficial defect maps of exposed concrete. The data processing resulted in a 3D model for the structure along with quantified defect maps at 1mm pixel resolution. Cracks, spalls, delamination along with efflorescence are viewed and analyzed in a georeferenced GIS environment. Quantified defects at high resolution allow for advanced analytics such as spatial density mapping, analysis of defect spatial correlations, along with change detection through time.

Results of these analyses provide advanced insight for inspection, monitoring, repair and maintenance decision making with the opportunity to use defect growth trends for predictive modeling. Furthermore, the condition-based life cycle assessment on a timescale allows owners to make cost-efficient business decisions to plan and execute the maintenance repair schedule across all assets, using the risk modelling methods to extend the lifespan of the structure rather than replacing the entire structure. This approach grants an accurate, date-driven environment for building predictive deterioration models using the historical performance which can be utilized for complete asset management.