Wind turbine blade damage detection using artificial intelligence (Blade-AI)

DNV will launch a Joint Industry Project which aims at developing a standard approach for the evaluation of AI assisted drone-based blade inspections.

SHARE:

Challenge

The wind energy industry is increasingly seeking to minimize the operational and maintenance costs of Wind turbines. Recent technological advancements have enabled the automation of wind turbine inspections while newly developed data-driven techniques allow for the processing of very large data sets to assist decision making. This has led to the implementation of Automatic Image recognition techniques combined with machine learning algorithms for the processing of drone-based inspection data.  

The question that arises is, how can it be ensured that such automated tools, will operate as required, meeting an acceptable level of reliability?

Solution

DNV will launch a joint industry project which aims at defining an approach for the third-party validation of the techniques used for the automatic processing of blade inspection data. Additionally, the project will outline the technical requirements for the acquisition and handling of inspection data, develop a system for the classification of findings based on the detecting and recognition capabilities of the algorithms, and define a reporting format related to inspection findings.

Value

Establish a third-party validation of drone assisted wind turbine inspections and the techniques used for the automatic processing of the generated inspection data.

Benefits

Through the project a consensus will be reached regarding:

• The technical and data requirements related to blade inspections
• The validation metrics and evaluation approach of machine learning algorithms for wind turbine blade inspections
• The classification of inspection findings
• The reporting and sharing format of relevant information