Predictive Vegetation Management and Grid Infrastructure

 

Aging grid infrastructure, dramatic increase in distributed energy resources and regulatory scrutiny has necessitated focus on lowering operation and maintenance cost while delivering enhanced customer experience and value.  But, innovation/investment on vegetation management has attracted least attention – not because it is cheap. It is learnt that California’s three investor-owned utilities seem to have spent more than $250 million annually on vegetation management on distribution lines alone.  Similarly, Pacific Gas & Electric estimates more than $500 million on vegetation management in 2019 and the expenditure is likely to reach $1 billion.  Despite such substantial spending, traditional approaches (endorsed by regulators) – cycle-based trimmings where vegetation is cut at fixed time intervals seem, inadequate.

Improved data collection, low cost sensors combined with advanced analytics is helping utilities detect and respond more effectively and perform predictive maintenance of grid infrastructure prior to its catastrophic failure.

 

Recent Challenges:

Those familiar with destructive wildfires of 2017 and 2018 in California concur that a new approach to vegetation management is an absolute necessity now. California Public Utilities Commission (CPUC) records 9000 fires (2017) responsible for 43 deaths and burning of 1.2 million acres.  They further state that California’s wildfires were the result of ‘utility ignitions’ covering even vegetation in contact with power lines.   It is not only California that needs improving vegetation management, Federal Energy Regulatory Commission opines that tree contact with transmission lines is the leading cause of power outages including blackout in Northeast US and Canada that impacted 50 million people.

The impact of vegetation assumes greater value when distribution lines are taken into account.  Electric Power Research Institute reported that 92% of weather related outages were caused by vegetation.  Regulators and utilities who rely on System Average Interruption Duration Index (SAIDI) and System Average Interruption Frequency Index SAIFI) for utility’s safety and reliability performance reflect imperfect metrics.  In the light of rise in extreme storms, utilities face increased pressure from regulators and customers to achieve overall resiliency and reduce prolonged outages.  Unfortunately, these outcomes are not fully captured by SAIDI and SAIFI metrics. In addition, both of these do not cover extreme weather events.

More effective vegetation management is obvious to embrace customer-centric business even for those who have not suffered prolonged outages.  Some of the distribution utilities have invested in hundreds of deployments for customer and grid analytics to meet changing customer habits.

 

What does the future look like?

This prompts a proactive approach to power outages as a clear differentiator among utilities.  This facilitates preventing them happening and communicating about the cause, anticipated duration and repair of any outages.

Vegetation Management poses operational and maintenance challenges with both internal and external pressure to increase efficiency in the business.  Currently, it is being carried out manually for vegetation trimming and removal.  There is no doubt that efficient vegetation management improves safety, reliability and performance during major storms which could not only be fatal but may cause property and environmental damage.  Accenture provides a solution that help utilities to integrate this new and emerging technology to unlock efficiency, productivity and safety.

Utilities approach has altogether been different on vegetation management for distribution and transmission infrastructure.  North American Electricity Reliability Corporation, in the wake of 2003 blackout mandated that utilities with high voltage transmission lines of 200kV and higher, that power lines rights-of-way be inspected annually and a vegetation management plan in line with its standards be adopted.  In addition, strategy to ensure that vegetation do not interfere lines, transformers and other critical infrastructure must be submitted to state regulators for approval.  The realization that the current cycle-based approach to vegetation management seems more efficient prompted research into investigating alternatives.

 

Modern approach:

The increased use of satellite imagery, light detection and ranging (LIDAR), land surveying method that uses pulsed laser light from helicopters and fixed-wing airplanes for a 3D representation be deployed.  Drones equipped with LIDAR would perhaps make it simpler.  In view of limitations that these applications have, predictive analytics are applied to massive data sets from variety of sources including historical and current weather conditions including soil moisture levels.  In this regard, IBM vegetation management offers a sophisticated and scalable insight into state of vegetation.

IBM Vegetation Management built on IBM PAIRS Geoscope provides data scientists’ quick access to data that can be used to develop models, enhance and enable applications.  The wealth of data can be used within a utility’s geographical system (GIS) allowing them to leverage the existing GIS asset mapping with an overly of vegetation data to understand real-time picture of their service territory.

The next phase focuses on techniques to determine species of trees and identifying areas of rapid growth for a clearer understanding on which areas of vegetation may pose a threat to transmission and distribution network.  IBM Vegetation Management also triggers an alert when vegetation is encroaching utility’s buffer zone prompting deployment of crews to trim bushes and trees for the problem arises.  The enhanced monitoring by IBM Vegetation Management allows refining trimming cycles and reduces costs as well.

Regulatory pressures, Increases customer & worker expectations, increased volume & complexity of work, Pressure to improve efficiency & quality of work and proliferation of digital technologies are a few challenges facing this sector.  The Vegetation Management (VM) solution facilitates utilities integrate emerging technologies towards unlocking efficiency, productivity and improve safety.  Advanced analytics for scenario analysis to evaluate efficiency vs. reliability or risk with different VM cycles happens to be Pre-initiation.  Identification of work that needs to be attended at the circuit level with least manual intervention and automation kick starts initiation. Optimizing VM crew utilization towards integrated work management for resource management and work schedule, coordination and execution fall under execution.  Automating QA/QC via mobile allowing crews to capture field data through ML/AI to identify potential ‘go backs’ account for close out.  Analytics could be assessed to rebalance the supplier contract risk and arrange performance expectations of sourcing ends the chain.

 

New Model:

While the problem is not the same with reference to vegetation interfering electrical outages, the approach would be different although the basics remain the same.  Machine Learning (ML) has initiated a model called ‘Outage Prediction Model (OPM)’ to serve simulation capable of quantifying the damage reduction.  Experimenting with one of the grid territories, uses a random forest model with variables such as vegetation management, land cover, drought, elevation, weather and electrical infrastructure to predict outages for each unit of the power distribution network.  The model accounting for considerable number of storms, showed that the annual reduction in trouble spots due to enhanced tree trimming is a good matching between increased trouble spot reduction and increased vegetation management.  In addition, it further indicated improved outage prediction while including the vegetation management data as a variable.  The results would be helpful for the utilities /municipalities and regulators in optimizing vegetation management and a broader grid resilience enhancement.

 

Conclusion:

A K Shyam  373 Canarabank Layout  First Main 1st Cross Vidyaranyapura Post BENGALURU – 560 097, Karnataka, India. E-mail: [email protected]