Advancements in Machine Learning-Based Condition Monitoring for Crack Detection in Windmill Blades: A Comprehensive Review

Abstract

Globally, the amount of wind turbines used to produce sustainable, renewable power is always increasing. Achieving dependable and easily accessible performance requires integrating innovative real-time condition monitoring technology. Ensuring the efficacy of wind power generation while maintaining its ability to generate revenue is fundamental. Machine learning (ML) has emerged as a crucial method for monitoring the condition of wind power systems in the past several years. This research study offers a comprehensive and current overview of contemporary condition monitoring technology employed in wind turbines for the purpose of detecting and predicting failures. Emphasizing machine learning algorithms for identifying significant faults and failure modes, preprocessing methods, and evaluation metrics, the review evaluates several references to determine past, present, and future developments in this field of study. Most of the analyzed references come from recent papers, reports, and journal articles that are freely available online.