Advancing AI-Based Sewer Inspection at West Boise Sewer District

The West Boise Sewer District (WBSD) faced challenges in maintaining efficient and accurate sewer inspections, limited by conventional methods that were often less accurate, time consuming and labor-intensive. In response, WBSD collaborated with Jacobs to integrate DragonflySM , a cloud-based Artificial Intelligence (AI) sewer defect coding solution developed in partnership with Hitachi (Figure 1 illustrates the AI tools three-step process). This initiative aimed to assess whether AI could offer a more precise, reliable, and cost-effective approach compared to traditional manual inspection methods. The integration of this AI technology was a key part of WBSD's broader strategy, targeting resource optimization, cost reduction, and enhanced sustainability in sewer management processes. The methodology of the project was comprehensive, involving a pilot program that processed over 25,000 linear feet of closed-circuit television (CCTV) sewer footage. AI algorithms were developed to perform tasks traditionally reliant on human effort, such as identifying, categorizing, and coding various defects within the sewer system. (Figure 2 illustrates how the AI identifies defects with a boundary polygon). Manual defect coding, despite being performed by experienced experts, often faced challenges such as human error and inconsistencies. The pilot study's findings were significant. The AI technology demonstrated a marked improvement in detecting sewer pipe defects, especially those previously overlooked in manual inspections (refer to Figure 3 for defect snapshots from Dragonfly). This improvement was not just in terms of defect detection accuracy but also in operational efficiency. The time saved in sewer inspections and the reduction in the need for extensive manual labor led to a considerable decrease in operational costs. This shift in methodology reflects a larger trend towards adopting innovative technologies in infrastructure management, aiming for better, more efficient, and more sustainable outcomes. Furthermore, the pilot introduced WBSD to an integrated asset management module, Argon, which enhanced the analysis and cataloging of inspection data. Argon not only streamlines this process (as depicted in Figure 4) but also consolidates up-to-date data on the system's overall health in a single, accessible platform. This integration brings together video footage, detailed inspection data, a comprehensive catalog of sewer assets, along with maintenance and improvement plans, which were previously spread across multiple platforms. Figure 5 illustrates the defect family score, showcasing Argon's efficiency in managing condition assessments. The modules capabilities allowed WBSD to develop comprehensive maintenance plans, including, detailed clean schedules, reinspection timelines (highlighted in Figure 6), and prioritization strategies for asset rehabilitation and replacement. The introduction of Argon represented a significant step in enabling the transformation of inspection data into actionable strategies, thereby offering the potential to improve the overall efficiency of the maintenance process. Following the positive outcomes of the pilot, WBSD is actively in the process of integrating AI across their entire 65-mile collection system, an initiative still underway. This strategic decision, influenced by the potential for substantial cost savings, is projected to reduce expenses by approximately 75% compared to the originally budgeted amount for a network-wide implementation. It is pertinent to mention that these savings are based on a comparison with costs that would have been incurred if the work was outsourced to an independent contractor, rather than conducted internally by WBSD staff. This expansion, though not yet complete, stands as a testament to WBSD's commitment to adopting innovative solutions to enhance the efficiency and cost-effectiveness of their operations. The adoption of AI technology in sewer management at WBSD represents a significant advancement in municipal infrastructure management. It underscores the potential of AI to enhance operational efficiency, accuracy, and predictive capabilities in sewer system maintenance. The outcomes included improved inspection accuracy, notable cost reductions, better resource utilization, and a shift toward a proactive, data-driven maintenance approach. This change aligns with the broader trend in infrastructure management of leveraging technological innovations to achieve improved results. WBSD's integration of AI technology in sewer inspection is more than just a successful application case; it represents a significant shift in municipal operations. This experience provides valuable insights into the challenges and potential solutions in modern infrastructure management, especially in contexts where resources are limited and demands are growing. It's an instructive example for utilities worldwide, demonstrating how embracing innovative, technology-driven solutions can lead to improved operational efficiency and sustainability. This approach is crucial in addressing complex challenges in utility management, paving the way for more efficient, accurate, and sustainable practices in the sector, and setting new standards for technological integration in infrastructure management.