A Comprehensive Review on Blade Damage Detection and Prediction

Abstract

High-speed rotating blades are widely used in modern industry, such as aviation gas turbine blades and wind turbine blades. Different degrees of damage can occur under extreme and long-term working conditions, such as surface damage, mechanical damage, creep and crack. In severe cases, it will cause blade fracture and cause inestimable loss. Thus, it is much necessary to identify blade damages as early as possible. By now, many studies have been done to investigate blade damage detection and prediction. The aim of this paper is to give a comprehensive review on existing blade damage detection and prediction. Firstly, common types of blade damages are summarized and the corresponding causes and consequences are analyzed. Then different measurement techniques of blade conditions are compared, the corresponding signal preprocessing and feature extraction algorithms are presented. Next, several classical machine learning-based blade detection and prediction methods are presented and compared. Also, the state of the art of deep learning-based methods and their applications in blade detection and prediction is reviewed. The powerful automatic feature extraction capability of deep learning has become a new research direction in the field of damage detection. This paper introduces related methods in detail and gives prospects at the end, providing a comprehensive perspective for this field.