Utilities are digitally transforming infrastructure management programs through aerial power line inspection with AI/ML Image Processing

Electric utilities in power generation, transmission and distribution are turning to aerial power line inspection combined with AI/ML image processing to increase the frequency, scope and accuracy of utilities infrastructure inspections and vegetation management programs. Discovering, prioritizing and repairing degraded transmission and distribution assets or vegetation encroachment issues more quickly has a direct impact on the bottom line by improving operational performance, SAIDI/SAIFI, safety, and energy transmission capacity while reducing systemic risk from wildfires and periods of extraordinary system demand.

Manual utility infrastructure inspection methods are time-proven, but they do not easily scale. Inspecting a single transmission line using manual methods may involve maintenance inspectors gathering data via multiple remote visits, which can be in a variety of formats, such as images collected via Pan-Tilt-Zoom (PTZ) camera, helicopter-mounted video cameras, mobile phones, drawings, or even handwritten notes. Despite best efforts, there could still be missed preventative or corrective maintenance actions due to the number of variables involved in manual processes, as well as issues properly recording the location of each specific photo. Traditional power line inspection methods are also expensive, requiring a variety of data collection methods and supporting personnel to gather the required data for further analysis, classification and prioritization. Finding and hiring skilled utility inspectors is also becoming increasingly difficult, with 84% of energy employers experiencing difficulty in hiring qualified workers, and an estimated 40% of current electric utility workers eligible for retirement by 2030. [2020 U.S. Energy and Employment Report, National Association of State Energy Officials (NASEO)]

The Rise of Aerial Power Line Inspections

Power utilities have long grappled with the logistical challenges of inspecting and maintaining the vast networks of power lines that crisscross their service territories. These electrical grids traverse everything from dense urban environments to remote, rugged terrains, making routine inspections a challenging, resource-intensive task. These utility inspections have traditionally been conducted by maintenance inspector ground crews with PTZ cameras augmented with imagery collected via manned aircraft, methods that are not only costly and time-intensive but can also expose workers to potential environmental and electrical hazards.

Recent advances in artificial intelligence and machine learning (AI/ML), combined with drone-based data collection techniques, have allowed utilities to dramatically improve the accuracy and capacity of their infrastructure inspection programs with existing resources.

drone inspection services
Aerial drones enable inspection of infrastructure located in heavily wooded areas that would be difficult to accomplish with ground-based methods.  

Continued developments in drone inspection software have made aerial power line inspection a practical and cost-effective option for energy utilities. Advancements in battery technology now allow power line drones to fly longer and further, reducing the need for battery change outs to inspect distribution and transmission lines. Further, improvements in drone platforms and control software have made drones a practical option for congested line inspections, as many platforms are not only easier to pilot, but also include automated obstacle avoidance technology and improved sensor technology. This enables inspection of areas with a higher concentration of power lines, as well as allows capturing high quality inspection imagery from further distances for particularly challenging areas.

Aerial drones offer several advantages that are catching the attention of both large Investor Owned Utilities (IOUs) and smaller rural electrical co-ops (RECs) & Public Utility Districts (PUDs):

Safety

By keeping inspectors on the ground, drone power line inspections minimize the risk of accidents associated with climbing structures or maneuvering through difficult terrain while conducting power line drone inspections.

Cost-effectiveness

By reducing the reliance on aerial data capture via helicopter and airplane, and decreasing the amount of time maintenance inspectors need to spend in the field manually capturing power line inspection imagery, drone utility inspections offer a more economical alternative to traditional inspection methods. Beyond Visual Line of Sight (BVLOS) drone operations promise even greater operational and cost efficiencies for conducting transmission line inspections, which often span hundreds of miles across remote territories, and can be effectively conducted using fixed wing drones for data capture beyond the 1.5KM limit for standard VLOS operations. 

Efficiency

Drones can cover large areas quickly, capturing detailed images that can be reviewed and analyzed faster than ever before. The aerial imagery captured by aerial power line inspection is also typically superior, as some of the inspection defects are impossible to view from below (pole tops) or are frequently obstructed by vegetation, power lines, poles and crossarms.

Enhanced Asset Insights

The data captured by drones is not limited to imagery. Equipped with advanced sensors, these UAVs can collect a range of data from thermal imaging to detailed topographical information via LiDAR (Light Detection and Ranging). When paired with AI and machine learning algorithms, this data can be processed to identify potential issues before they become critical, such as overheating components, structural vulnerabilities, wear and tear on the lines, and vegetation management program encroachments.

Asset Health Monitoring

While utilities have traditionally used field images for locating and documenting asset defects, images of assets without defects were often either discarded or otherwise not used. With the advent of automated drone data management, it is now possible to collect and retain aerial power line inspection imagery of all assets regardless of condition, associate those images with specific assets and utilize that data to provide a comprehensive view of infrastructure health over time. By combining this imagery with AI/ML processing for defect detection, utilities can monitor asset health by geography or feeder through dashboards. This detailed asset data can be used to prioritize repair actions based on diagnosed repair severity to improve SAIFI and SAIDI and system resilience.

Optelos drone inspection services

Case Study

“Automating Grid Analytics Through Drone Inspection and Computer Vision AI”

Read about the partnership between Optelos and ComEd

comed thumb

Automating Aerial Power Line Inspections with AI/ML

While the quest for operational efficiency is driving drone inspection adoption, the true innovation lies in what happens after the data capture with AI/ML data processing. To realize the full benefits of aerial power line inspection, computer vision AI/ML algorithms are utilized to automatically analyze the massive quantity of raw visual image data to classify assets and identify predetermined defect conditions. This dramatically improves the efficiency of inspection programs by flagging images with detected issues for further MI (maintenance inspector) review, saving time and reducing inspector fatigue.

Optelos Qualified Aerial Data Capture patch
Computer vision AI algorithms classify and detect failure conditions by asset type, accelerating image processing to provide fast, accurate and reliable asset insights. 

Integration of AI and ML in Data Analysis

At the heart of an automated inspection system is a powerful AI that analyzes the high-resolution imagery captured by aerial drones. The AI algorithms are trained to classify the asset under inspection (e.g., pole tops, cross arms, cutouts, insulators, lightening arrestors) and detect defect conditions that indicate wear, damage, or other issues expected to impact system performance. This could be a broken crossarm, a rotted pole top or a cracked insulator, for example.

Enhancing Inspection Capacity and Accuracy

By automating the analysis of this data, utilities are able to significantly enhance their inspection capacity utilizing the same resources. Optelos customers have seen a 3X increase in inspection capacity by combining aerial inspection with AI/ML image processing. What once took weeks can now be accomplished in days, with a consistent level of detail and accuracy that doesn’t vary due to work distractions, fatigue, staffing levels or personal experience level; it consistently applies the same rigorous standards to every image and ensures that defects are detected and flagged with repetitive precision.

Improving Inspection Accuracy

The reduction of human error is a key benefit of aerial inspections augmented with AI/ML processing. While human inspectors are skilled, they are also fallible, and the monotony of reviewing thousands of images can lead to errors, estimated at 17% in manual inspection programs (Colin G. Drury, Inspection of Sheet Materials — Model and Data, Department of Industrial Engineering, State University of New York at Buffalo, Amherst, New York, June 1, 1975). AI in a cloud environment, however, maintains a constant performance level with near instantaneous processing of thousands of images, which can lead to earlier detection of issues, thereby enhancing risk management and preventing failures before they occur.

“While human inspectors are skilled, they are also fallible, and the monotony of reviewing thousands of images can lead to errors, estimated at 17% in manual inspection programs”

Colin G. Drury, SUNY Buffalo, Amherst, NY

Scalability and Inspection Resource Efficiency

Airborne inspection with computer vision AI processing also addresses the challenge of scalability. As utility electrical grids expand and the demand for the frequency of inspections increases, the strain on MI resources grows. The automation of inspections means that utility companies can maintain, or even increase, their inspection capacity without a proportional increase in resources. This scalability is crucial for utilities striving to meet growing regulatory requirements for powerline inspections for wildfire mitigation planning and other oversight requirements.

Continuous Improvement through AI Retraining

Machine learning is another critical element of inspection automation. When the AI model is retrained with additional annotated imagery showing new environmental conditions and equipment, it learns, improving its algorithms and becoming more adept at classifying assets and identifying failure conditions. This enables the system to become more efficient over time, adapting to new patterns of wear, new component types and evolving changes to the transmission network.

Knowledge Retention

According to the 2020 U.S. Energy and Employment Report by National Association of State Energy Officials (NASEO), finding and hiring skilled utility inspectors is also becoming increasingly difficult. According to the 2020 report, 84% of energy employers experiencing difficulty in hiring qualified workers, and an estimated 40% of current electric utility workers eligible for retirement by 2030. This makes it imperative to begin capturing the knowledge of these experienced workers before they retire, empowering the next generation of workers with new tools to multiply their efficiency in order to keep up with asset inspection demands.  

According to the 2020 U.S. Energy and Employment Report by National Association of State Energy Officials (NASEO), 84% of energy employers experiencing difficulty in hiring qualified workers, and an estimated 40% of current electric utility workers eligible for retirement by 2030.

Advanced Vegetation Management through Drone Technology

Vegetation management is a critical component of maintaining powerline integrity and service reliability. Overgrown trees and brush can cause outages, fires, and costly damage to the grid. Traditional methods of vegetation management often involve ground-based surveying and manual assessments—a labor-intensive process with limitations in accuracy and frequency – and aerial imagery collection via airplane and helicopter. Utilities are increasingly adopting drone image collection for vegetation monitoring programs due to the ability to complete this task more quickly, effectively and at a lower cost.

Leveraging LiDAR and AI for Precision Mapping

Utilities are utilizing drones equipped with Light Detection and Ranging (LiDAR) sensors to create detailed 3D maps of powerline corridors. LiDAR systems can rapidly collect vast amounts of data, creating 3D models with centimeter-level accuracy. This technology provides precise measurements of vegetation proximity to powerlines, enabling a more accurate assessment of risk areas. The integration of AI algorithms further processes this data, identifying vegetation encroachment and predicting growth patterns to preemptively address potential issues.

Operational Efficiency and Cost Reduction

The use of aerial drones for vegetation management significantly reduces the operational costs associated with manual vegetation inspections. By covering more ground in less time, drones allow for more frequent and thorough inspections. The data collected is not only more accurate but also provides a comprehensive overview that manual inspections cannot match, leading to better-informed decision-making and resource allocation.

Enhancing Safety and Compliance

Safety is a paramount concern in vegetation management. Aerial drones/UAVs minimize the need for workers to navigate hazardous terrain or work in close proximity to live power lines. This not only enhances the safety of the workforce but also ensures a higher level of compliance with health and safety regulations.

vegetation management program with LiDAR scans
LiDAR scans can ve effectively used to quickly inspect for vegetation encroachment in powerline corridors. These scan can be analyzed with computer vision AI algorithms to automatically detect current or projected vegetation encroachment. In this scan, the floor was set at 15′, eliminating ground noise and improving powerline visibility. 

Conclusion

By combining aerial power line inspection with computer vision AI/ML processing, utilities are digitally transforming their asset inspection and vegetation management programs. These evolving technologies are providing electric utilities with the ability to monitor their distributed assets more frequently and methodically, providing insights to deliver better grid reliability and resilience as demands on infrastructure grow.  The use of drones for aerial power line inspections and vegetation management is not merely a technological novelty but an operational necessity that enhances safety, efficiency, inspection capacity and provides proactive maintenance capabilities.

Optelos has emerged as a technological leader in the digital transformation of utilities inspections. With utility customers on three continents, Optelos has implemented inspection automation programs around the world in diverse environmental conditions. This hands-on experience provides us with the ability to deliver an operational  inspection platform to meet your exacting requirements. Our SaaS solution, known as Optelos Enterprise Asset Advisor, enables the swift adoption of these transformative new technologies into existing inspection workflows, allowing you to realize the benefits of drone powerline inspection automation quickly.

The advances in aerial power line inspection and drone data management with AI/ML processing technologies are not just changing how utilities conduct inspections—they are redefining what is possible. Optelos provides a comprehensive suite of tools and expertise to empower utilities to not only meet the demands of today, but to innovate for the challenges of tomorrow. If you would like to learn how Optelos can empower your inspection program, please contact us for further information or a demonstration tailored to your specific goals.