Press P. S. Saputra, Yusri Syam Akil, Rifqi Firmansyah
Induction motors are the predominant component in factory drive systems when compared to other motor types. The widespread popularity of induction motors can be attributed to several advantages, such as affordability, low maintenance requirements, high efficiency, solid construction, and user-friendly operation. Despite these merits, as induction motors operate for extended periods, they are susceptible to damage, particularly concerning their bearings. Bearing issues, responsible for up to 50% of overall damage, surpass other forms of damage. Some causes of the bearing damage, misalignment during motor installation with a load and it causes vibrations during motor operation. This research aims to identify and categorize misalignment phenomena based on the extent of damage, specifically focusing on scenarios with 1mm and 1.5mm misalignments. Vibration data from both normally aligned and misaligned motors will be recorded using a vibration sensor. Subsequently, the vibration signals will undergo transformation using Daubechis wavelet transform. The resulting high-frequency signals will be extracted based on signal characteristics such as number, and range, and energy. The effectiveness of the Fuzzy C-means method in classifying the data will be assessed, with performance evaluations conducted after testing the dataset. To gauge its performance, the Fuzzy C-means method will be compared with the K-Means method in terms of their ability to accurately classify the conditions of the induction motor. The findings indicate that employing either Fuzzy C-means or K-Means in conjunction with the first-level Daubechis wavelet transform yields the best classification accuracy, reaching an impressive 95.83%. © 2025 Author(s).
Electrical Engineering Department, Universitas Muhammadiyah Gresik, Gresik, Indonesia; Electrical Engineering Department, Universitas Hasanuddin, Gowa, Indonesia; Electrical Engineering, Universitas Negeri Surabaya, Surabaya, Indonesia
