Wisely Managing Vegetation at the Distribution Level

Ever since the 2003 blackout, vegetation management has improved, largely due to FERC regulation. Still, improvement has been mostly in transmission. The “last mile” of distribution has not achieved maturity. The good news is that analytics infrastructure – monitoring, data access, analytics and operationalization – could bring more wisdom to managing distribution line vegetation.

Vegetation Management for Distribution Needs Help

In the United States, there are over 200,000 miles of transmission lines and about 5.5 million miles of distribution lines according to Stephen Cieslewicz, Independent Consultant at UVMI. News stories about trees falling onto distribution lines during storms and high winds abound. Still, there are few publicly available studies on vegetation-related distribution outages. That’s because, according to arborist and utility vegetation management (UVM) consultant William Porter, utilities have been slow to develop uniform codes for capturing vegetation-related outages.

Unlike transmission lines, which fall under federal regulation, distribution vegetation is regulated by the states. The most adopted vegetation requirement is NESC Rule 218. This rule does not mandate active management or prescribe measurable metrics.  More importantly, there is no guidance on factors essential to proper management – line voltage, species’ growth rates, failure characteristics, proximity to conductors and conductor sag. Plus, regulations differ on who is responsible for keeping lines free of trees. In some service territories, the customer is responsible for trimming around service drops.

Most utility operators outsource vegetation management and specify a cycle length to complete the system. Arborists inspect and control vegetation within a defined distance based on expected cyclical growth characteristics. Enhanced vegetation management (EVM), a new approach, calls for increasing that distance, to prevent “danger” trees from hitting lines and sparking fires, ala California.  A new buzz phrase is condition-based vegetation management, which targets tree trimming at hot spots where trees are might cause outages.

Bringing Analytics to Vegetation Management

Traditionally, vegetation management systems track vegetation inspections and work performed by utility personnel or contractors. Utilities have also used systems like ArcGIS to provide visibility into geo-referenced vegetation associated with circuitry. Data is static in these systems. Yet, vegetation management is dynamic; conditions change frequently.

Advances in LIDAR and aerial imaging have improved vegetation management for transmission. However, these technologies might not always be right for distribution. High-resolution LIDAR data is too expensive to refresh frequently. Processing the data is complicated. For example, one utility found that handling and storage LIDAR – which is a 3-D point cloud – requires an entirely different architecture from processing traditional aerial imagery.

That is where analytics come in. Multiple low-cost information sources, such as open source satellite and agricultural data can be married with grid data, to train models to identify trees in relation to circuitry. One startup, Live-EO, has established a high degree of accuracy by applying AI to satellite data AI to identify the top tree species in Europe. [1] On the near horizon are analytic techniques for determining growth rates and tree health.

Today, some utilities are testing out the use of analytics to enhance planning, assess risk and reduce the cost of vegetation management at the distribution level.  Some examples of pilots covering distribution:

• Commonwealth Edison – using LiDAR and hyperspectral imagery to determine ways to restructure vegetation management contracts.
• Oncor – trial and validation conducted with The Weather Company (IBM), using AI and a geo-temporal cloud-based platform to predict hot spots.
• PPL – leveraging machine learning to direct vegetation crews to high risk areas.

Where to Go from Here

Even if a utility contracts out vegetation management, “The one who has the information can manage the outcome.  Whether building new lines or dealing with the regulators, utilities need to have the best information,” according to Porter.

Vegetation management strategy needs to take local conditions into account. Utilities do not need to have the same approach. What works in one area, may not be prudent in another. Soil condition, tree species, soil moisture and predominate weather patterns vary across micro-climates.

Here are a few words to the wise:

• Look at in-house data that could improve vegetation management. Start with the most obvious – your vegetation management application. Nina Cohn, utility vegetation analyst, recommends considering structured or unstructured customer data in the CIS/CRM, such as vegetation related inquiries or even merging these two datasets to further define scope of work.
• Investigate using low or no-cost sources, albeit less frequent, vegetation data such as Sentinel 2 open source satellite data, the NAIP aerial imagery, the NDVI vegetation index. You can use that data to identify hot spots and then move to higher resolution data sources for those areas.
• Put an agile analytics team together to experiment with machine learning to address one or two use cases. From there, you will be in a better position to assess whether to build or buy a platform for analytics-enhanced vegetation management.

Looking to the Future – Collaboration with Customers and Communities

Looking to the future, utilities will need to move beyond removal and pruning to a more holistic approach to nurturing the environment. Trees provide vital benefits.  They act as a carbon sink and provide natural cooling with shade.

Plus, utility customers like having a welcoming landscape. Communities are thinking about protecting the tree canopy and having the right tree in the right place.  Municipals like CPS Energy offer their customers an energy tree guide. Expect more sharing of data and information between utilities and communities. Successful vegetation management will depend on it.

Want to learn more about vegetation management?  Attend the Utility Analytics Summit. Ankush Agarwal, Director, Grid T&D Analytics, Exelon and Steve Russell, Chair, Transmission Technology Board, Duke Energy will present Advanced Vegetation Management at Duke Energy and Exelon Utilities on April 16.

[1] Live-EO.com has projects with E.ON, EPRI and AEP.