Deploying Advanced Drones and Artificial Intelligence in Sewer Condition Assessment

Drone inspections in storm and sanitary sewers can be beneficial as they can quickly and efficiently access hard-to-reach areas and can fly over and around obstacles such as sediment and pipe obstructions that can interfere with a traditional crawler. In addition, the cost for this technology is significantly decreasing, and many of the sensors that are deployed on traditional inspection equipment can be added to specialized drones including high-definition CCTV, Lidar, and thermal imaging. Although there are numerous benefits to drone inspection, several challenges exist which needed to be planned for. Underground sewer systems present challenges in obtaining GPS signals, making precise positioning and navigation more difficult for drones. Alternative positioning technologies can overcome this limitation and ensure accurate mapping of sewer infrastructure. Drones also have limited flight times due to battery capacity which affects the distance they can cover in a single operation. Efficient flight planning and battery management strategies are required to optimize the inspection range and duration. Advanced lighting systems integrated into drones, such as adjustable LED lights or infrared illumination, can help overcome lighting limitations and improve data quality. The weight of onboard sensors can impact drone maneuverability and endurance. Utilizing lightweight sensors without compromising data quality and analysis accuracy is essential. Drone inspections in sewer systems are subject to regulatory requirements, including permits, air space regulations, and licensed pilot requirements. Complying with these regulations is crucial to ensure safe and legal operations. Collaboration with relevant authorities and stakeholders is necessary to navigate regulatory challenges effectively. Combining drone inspection with artificial intelligence (AI) for automated defect coding and analysis streamlines the traditional approach to sewer condition assessment. Automated defect coding has become available through a variety of providers, but limitations inherent in the training data from which the AI algorithms are created must be acknowledged and understood when utilizing AI. Inherent bias that is built into the NASSCO PACP scoring system must be adjusted for when considering sewers for rehabilitation. This presentation will compare the results from several storm & sanitary sewer condition assessment projects. As part of the City of Columbus, Ohio Large Diameter Storm Sewer Condition Assessment Program, 100,000 LF of storm sewers between 36-inch diameter and 120-inch diameter were inspected using a variety of technologies including drone, CCTV crawler, and zoom camera. In addition, storms culverts 42-inch and larger were inspected in the City of Greensboro. The quality of the inspection data, productivity, and cost of these technologies were compared against traditional methods to determine the cost effectiveness of drones for future inspection work. In summary, utilities that are contemplating condition assessment of their large diameter sewers and tunnels will be able to apply the lessons learned from this case study to understand the state of drone technology and potential cost savings that can be achieved using this technology.