D.-S. Huang, L. Heutte, and M. Loog (Eds.): ICIC 2007, LNAI 4682, pp. 952–960, 2007. © Springer-Verlag Berlin Heidelberg 2007 Bearing Diagnosis Using Time-Domain Features and Decision Tree Hong-Hee Lee, Ngoc-Tu Nguyen, and Jeong-Min Kwon School of Electrical Engineering, University of Ulsan, Ulsan, Korea hhlee@mail.ulsan.ac.kr, nntu@hcmut.edu.vn, ruee006@mail.ulsan.ac.kr Abstract. Bearing fault detection with the aid of the vibration signals is presented. In this paper, time-domain features are extracted to indicate bearing fault, which collected from tri-axial vibration signal. Decision tree is chosen as an effective diagnostic tool to obtain bearing status. The paper also introduces principal component analysis (PCA) algorithm to reduce training data dimension and remove irrelevant data. Both original data and PCA-based data are used to train C4.5 decision tree models. Then, the result of PCA-based decision tree is compared with normal decision tree to get the best performance of classification process. Keywords: bearing diagnosis, decision tree, vibration, principal component analysis. 1 Introduction Bearing defects are the most popular type of machinery fault. Nowadays, most of diagnostic methods are based on measurement of vibration, acoustic noise, stator currents, or temperature. The vibration measurement is commonly method used in industry, because it is relatively cheaper and more reliable than the others. Vibration measurement methods can be based on time-domain, frequency-domain vibration signals, or both of them. Frequency domain bearing diagnosis method often monitors the fundamental frequencies generated by the defective bearing: rotating frequency, fundamental train frequency, ball pass frequency of the outer race, ball pass frequency of the inner race, ball spin frequency and their harmonics. Meanwhile time-domain bearing diagnosis method is using simple processing to analyze the time waveform characteristics. Recently, the time signal analysis method for fault diagnosis has been introduced in many researches, such as proximal support vector machine [1], artificial neural network [2] for bearing diagnosis. Applications of decision tree [3], support vector machine [4] for motor diagnosis, etc. have been shown as the effective way on machine fault diagnosis field. However, there are still few projects of decision tree in bearing condition monitoring in particular. Therefore, this paper presents a bearing fault detection method by developing a decision tree, which is based on C4.5 algorithm. Compare to other methods such as neural network, fuzzy system, etc., decision tree has construction that users can understand easily, and have very fast learning rate.