Utilities need good quality data so that they can implement asset management, leverage predictive analytics, and more efficiently manage daily operations. AI can help identify and corroborate asset location, asset ID tags, equipment type, and attachments.
Utilities often find it difficult to accurately track their assets because of two reasons. First, utilities have widespread territories - often spanning thousands, tens of thousands, or even over a hundred thousand square miles. Second, the grid is dynamic - assets move, equipment is replaced or upgraded, vegetation grows into assets, and winds and storms can quickly change an asset’s condition. All of these factors impact internal asset databases, as employees struggle to keep the databases updated; and “data cleanup” projects are constantly needed.
However, it is more critical than ever to have good, clean data. Accurate asset information can enable utilities to make good investment decisions that mitigate asset failure risk and bring the most value to their customers in terms of safety, reliability and resiliency benefits. Good quality data will also support day to day operations - including work management, outage restoration, and storm impact prediction. As utilities proactively manage their assets and respond to daily operational challenges, they need to improve data quality.
Asset location is important because it enables utilities to proactively predict and plan for storm, wind, or fire events by estimating the predicted trajectory of the storm or fire event and how that will impact the assets located in the storm’s path. Utilities can leverage aerial or ground surveys, which will take images or video and ingest this media into the AI model. The AI can flag and identify assets and tag those assets with the longitude and latitude location, embedded within the metadata of the media files. This information can be used to corroborate existing datasets by confirming or flagging location issues.
AI can extract labels, identification tags, and other markers from images or video taken in the field and connect it to the location, as well as the internal information about the asset. For example, internal utility databases may contain large amounts of historical information about pole attributes (such as the type of pole, make, model, serial number, age, loading capacity), performance information (e.g., subject experience as well as actual data), historical records (when planned or unplanned work was last completed on the pole, historical actual costs) and future planned work (e.g., upcoming inspection or planned replacements). By associating real-time, current images and video with existing, historical datasets, utilities can make better decisions about how to prioritize the work on that asset.
AI can detect objects located on the pole, as well as distinguish between different pole materials. AI can help quickly identify crossarms and guy wires, as well as equipment that the pole is carrying. Besides carrying electric power lines (primaries, secondaries and neutrals), utility poles often also carry transformers, reclosers, insulators, cutouts, lightning arrestors, communications lines and antennas, and sometimes even streetlights. The more equipment mounted on to the pole, the heavier the loading on the pole, and more of a safety hazard it can become because strong winds can knock it over.
The ability to gather more data is easier than ever. However, the ability to manually analyze that data manually is limited - and often inefficient, time-consuming, and expensive. As utilities collect more images and video of their assets, AI can help them save time and O&M dollars by quickly analyzing that data. AI can locate assets, identify asset tags, equipment types, as well as component attachments. This extracted information can help a utility clean up their existing asset databases and enable them to have a single, up-to-date, database.