Automatic WindTurbine Infrastructure Inspection

Powered by IBM Watson AI

Results:


0%

Submit Data

Submit data from any device, no matter the size!

Analyze

The data is processed through a custom tensorflow-based Machine learning model

Results

The results are exported to popular formats such as JSON, CSV

Action Flow

Using image recognition and machine learning algorithms, BrainChild Innovation can detect minor and major damage to a wind turbine’s blade dataset faster and more thoroughly than current methods used by engineers. What takes an engineer up to an hour to review, we can review the same data in less than a couple minutes to provide various metrics of each blade, turbine and wind farm. The results from the data collected are exported to various files, including JSON & CSV available on a traditional computer or interfaced with a Virtual Reality or Augmented Reality headset, which allows engineers to quickly locate, analyze, mark, and share the dataset.

Results

Model accuracy results vary based on the quality of data which is trained and tested. A minimum of 85% damage prediction accuracy is required for the approach to be feasible, which is why our models traditionally achieve an accuracy rating of 91% or higher. The final prototype should pass all the end-to-end tests which includes correctly identifying the defect and notifying the user.

.carousel-indicators li { background-color: #999; background-color: rgba(70,70,70,.25); } .carousel-indicators .active { background-color: #444; }