AI-equipped robots could inspect and repair power plant boilersDavid Wagman | February 25, 2020
The Energy Department's National Energy Technology Laboratory said it is partnering with the Colorado School of Mines to develop artificial intelligence-enabled robots capable of inspecting and repairing power plant boilers.
Robotic crawlers already exist, NETL said. Unlike the AI-enabled robots of this project, however, those machines are typically incapable of repair, are not fully autonomous and are not equipped with AI to enable smart autonomy and predictive analysis.
The AI algorithms developed in the School of Mines project are intended to enable the robot to perform 3D mapping and information fusion as well as spatio-temporal crack tracking, map updating and smart damage analysis by robot learning.
Boilers are one of the most important components of a power plant, responsible for superheating water to create the steam that drives energy-producing turbines. However, it is difficult for humans to perform critical inspection and repair of these components.
Successful robotic inspection could limit or eliminate the need to send inspectors to assess difficult-to-access or hazardous areas, said one of the researchers. The robots will also allow automated live inspection, reduce risk to human operators during maintenance or unplanned outages and enable smart collection of well-organized data.
Boilers often include large sections of vertical structures, like wall surfaces, which must be maintained to ensure proper operation. Instead of constructing scaffolding to inspect and repair this area, the AI-enabled robots use magnetic treads to scale the high, steep walls.
As the robots crawl along, they can employ a variety of nondestructive evaluation sensors to perform real-time inspection of boiler furnace walls. If they find a crack, they can then operate repair devices to make an immediate repair, all while using AI to enable smart data analysis and autonomy.