Robotics-as-a-Service & Technology Deployment

Using Drones and Machine Learning for Offshore Wind Farm Inspection

Project Detail

In September 2019, ULC Robotics completed its second annual inspection of the five offshore wind turbine foundations located off the coast of Rhode Island. Automated flight operations and machine learning-enabled ULC’s UAS pilots to perform the inspection of all five jacket foundations, capturing and processing more than one thousand images to satisfy the requirements of the customer.

machine learning offshore wind

Client : Orsted, Keystone Engineering

Location : Block Island Wind Farm

Project Date : Fall 2019

Aircraft : ULC Custom-Developed Hexacopter UAS

Payload : 42MP High-Resolution DSLR, HD Video Camera

Total Flight Time : 90 minutes


Conventional methods of inspecting the offshore wind foundations:

  • Expose workers to risk, climbing at heights above sea level
  • Are labor-intensive, and can require 6 hours of inspection time per foundation
  • Do not capture all data points, due to inaccessible areas or inconsistent photography

Additionally, traditional methods of data analysis:

  • Can take substantial time to manually process, review and sort all data
  • Can result in inaccuracies, as manual data review is just 95-97% accurate

ULC proposed a detailed flight plan using automated operations to deliver comprehensive inspection data to detail the condition of all key inspection points.

Project Results

Using our custom-developed hexacopter UAV outfitted with a high-resolution 42MP DSLR camera and HD video camera, ULC’s Aerial Services team provided detailed images and reporting on all inspection points as outlined by Orsted.

  • High-resolution images of all key inspection points across each of the five foundations
  • 100% data capture by ULC, with 360° insight enabled by aerial views, in approximately 1.5 hours versus 6 hours per foundation manually

In addition to automated flight and data capture, ULC Robotics developed and implemented a machine learning application to rapidly process all images captured during the inspection process.

  • ULC’s offshore wind machine learning model can be used to rapidly process the data
  • Raised accuracy of data analysis to over 99% through machine learning
  • Provided the client with an interactive cloud-based portal for detailed data review
  • Enabled better analytics and predictive models for comparative analysis when reviewing previously captured data

Drones and machine learning offshore wind inspection

Don't Miss Out on Our Technology and Innovation Updates:

Sign Up for Our Email Newsletter