Abstract
This case study focuses on the use of an off-the-shelf autonomous robotic platform to inspect an overland conveyor at an open-pit mine in Chile. Manual inspections for the conveyor were labor-intensive, operator-dependent, error-prone, and exposed workers to various risks such as harmful substances and environmental hazards. The robotic platform carried a payload of moving sensor arrays to detect thermal, audio, and visual information, without requiring fixed sensors to be installed on the conveyor. Unstructured data from the moving sensors was then converted into a structured format, with anomaly detection algorithms and dashboarding applied. This presentation shows the benefits for maintenance teams to leverage mobile robotics and AI for collecting, extracting, analyzing, and visualizing data to avoid costly conveyor breakdowns.
The autonomous robots worked with two levels of operational intelligence for mobilization through the plant:
1) High-level behavior to navigate around a 3d digital twin of the site, which allow the robots to work in GPS-denied environments.
2) Independent low-level programming provides safe movement between waypoints.
Those robots were equipped with sensor arrays to capture multi-variable field data (audio, images, thermography). Although the data gathering was done with commercially available technology, the data-analysis software layer defined how value was generated for stakeholders. That software layer was created using artificial intelligence, signal analysis, data-clustering, and expert interviews to generate objective, comparable, numeric descriptors for data, specific for the application on hand. These models were then used by on-board AI computers to generate real-time automated cloud-based dashboards providing detailed, objective data for strategic decision-making.
The case study’s results highlight the robots’ ability to perform consistent, focused inspections at a speed comparable to humans but with greater effectiveness and reduced risk. This approach showcases the significant potential of adopting autonomous software-defined-robotics as a strategy for inspection of operations in variable industrial settings, offering improved safety, accuracy, and efficiency.