Harris has applied deep learning technology to automate a wide range of utility asset management problems, and has deployed its advanced deep learning technology, MEGA, in various utility scenarios including asset location and identification, anomaly detection, and monitoring change over time. In addition to standard libraries designed to detect common anomalies on T&D infrastructure, new classifiers can be developed with specific data collected by the utility and deployed to provide tailored inspection analytics. These analytics can be automated to run on data ingest in real-time, or used interactively by analysts to review and improve the deep learning models.
Our deep learning technology can detect anomalies on assets such as: missing or damaged components, pole split/rot, bird’s nests or other animal infestation, lightning strikes, corrosion, or rust. Using infrared data, analyses of temperature profiles on equipment can be automatically generated to provide insights on power flow and predict potential equipment failures.