Blade fault localization with the use of vibration signals through artificial neural network: a data-driven approach

• Main Authors: Keng, Ngui Wai, Mohd Salman, Leong, Mohd Ibrahim, Shapiai, Hee, Lim Meng
• Format: Article
• Language: English
• Published: UPM Press 2023
• Subjects: T Technology (General); TA Engineering (General). Civil engineering (General); TJ Mechanical engineering and machinery; TL Motor vehicles. Aeronautics. Astronautics
• Online Access: http://umpir.ump.edu.my/id/eprint/38233/1/Blade%20fault%20localization%20with%20the%20use%20of%20vibration%20signals%20through%20artificial%20neural%20network_.pdf; http://umpir.ump.edu.my/id/eprint/38233/; https://doi.org/10.47836/pjst.31.1.04; https://doi.org/10.47836/pjst.31.1.04

Turbines are significant for extracting energy for petrochemical plants, power generation, and aerospace industries. However, it has been reported that turbine-blade failures are the most common causes of machinery breakdown. Therefore, numerous analyses have been performed to formulate techniques for detecting and classifying the fault of the turbine blade. Nevertheless, the blade fault localization method, performed to locate the faulty parts, is equally important for plant operation and maintenance. Therefore, this study will propose a blade fault localization method centered on time-frequency feature extraction and a machine learning approach. The purpose is to locate the faulty parts of the turbine blade. In addition, experimental research is carried out to simulate various blade faults. It includes blade rubbing, blade parts loss, and twisted blade. An artificial neural network model was developed to localize blade fault through the extracted features with newly proposed and selected features. The classification results indicated that the proposed feature set and feature selection method could be used for blade fault localization. It can be seen from the classification rate for blade faultiness localization.