Utility Week

UTILITY WEEK | 21ST - 27TH FEBRUARY 2020 | 27 Drones and utilities so you can't put it in an explosive atmosphere. e older version of the drone had a fully rotational cage that would spin if you hit something so the cage would appear in the video footage, but the latest model has a window in the cage to prevent that from happening." Drones can now be adapted to carry bespoke pay- loads and equipment, not just conventional cameras. LiDAR remote sensing is a potentially more effective method to pinpoint vegetation that needs to be cut back to prevent damage to infrastructure leading to fires or outages. New York Power Authority plans to use the system to manage vegetation along its transmission lines and a similar use is being investigated by business and IT consulting firm CGI for a utility client. Rich Hampshire, vice president of digital utilities at CGI, tells Flex: "Earth observation satellite imagery could identify potential areas of vegetation growth and that 'crude' picture can be used to target specific problem areas using other imaging techniques, such as LiDAR mounted on a drone to capture much higher resolution imagery." Modern UAVs can even be hardwired to "smell" the environment. Canadian company Scentroid sells a Sample Drone fitted with sensors and equipment designed to measure and sample chemicals like ammonia, sulphur dioxide, VOCs (volatile organic compounds) and other pollutants. e samples are brought down and analysed to help wastewater treatment plants keep tabs on levels of odour and air pollution. Flying machines aren't all about snazzy hardware; advances in software processing allow utilities to do much more intelligent things with captured footage. Connected Drone, developed by Norway's eSmart Systems, uses AI to automatically scan thousands of drone-captured images of assets, like towers or poles, for signs of damage or faults. e software was used by a major transmission system operator in Denmark to detect rust, in a data- set of 7,500 pictures the issue was identified with 96 per cent accuracy. According to eSmart, the ability to track rust development over time has enabled the operator to more accurately forecast maintenance requirements. A utility in Norway that used the system to scan infrastructure was able to avoid as much as 25 per cent of outage-related costs caused by defects that previously went undetected. e system needs human training to reach its full potential, says Phillip Yoxall at eSmart Systems: "Connected Drone automatically analyses all images and links them to the related asset. When unknown components and defects are not detected, a human

• Cover