 A study is presented to diagnosis of misalignment fault using feature extraction and selection technique and support vector machine (SVM) classifier. The timedomain vibration signals of a compressor with normal and misalignment conditions in driven end (DE) are gained for feature extraction. The features are extracted by using the statistical and vibration parameters. Then stepwise backward selection is applied for selecting the significant features. The selected features are used as inputs to the classifier for twoclass (normal or misalignment) identification. The roles of stepwise backward selection technique and SVM classifier are investigated. The results indicate the potential of the proposed intelligent method in misalignment fault diagnosis of the compressor.  Fault diagnosis, Misalignment fault, Feature selection, Support vector machine. I. INTRODUCTION ONDITION monitoring of rotating machinery is important in terms of system maintenance and process automation [1]. Reliability has always been an important aspect in the assessment of industrial products. By development of technology, cost of timebased preventive maintenance increased thus, new approaches in maintenance such as conditionbased maintenance (CBM) developed. Machine condition monitoring has long been accepted as one of the most effective and costefficient approaches to avoid catastrophic failures of machines [2,22]. Precise and high accuracy assessment of machinery condition results in fewer stoppages and better product quality and reduces maintenance costs for plants. Thus, they can optimize workforce and implement more efficient operations [319]. Manuscript received October 9, 2007: Revised version received March 4, 2008.. This work was supported by University of Tehran. Hojat Ahmadi is an associate professor in Department of Agricultural Machinery Engineering, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran (email: hjahmadi@ut.ac.ir). Ashkan Moosavian is MSc student in Department of Mechanical Engineering of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran. (Corresponding author, phone: +989375468339; fax: +982166593099; email: a.moosavian@ut.ac.ir). Meghdad Khazaee is MSc student in Department of Mechanical Engineering of Agricultural Machinery, Faculty of Agricultural Engineering and Technology, University of Tehran, Karaj, Iran (email: khazaee.meghdad@ut.ac.ir). Most of machinery used in the modern world operates by means of rotary components which can develop failures. The monitoring of the operative conditions of rotating machinery provides a great economic improvement by decreasing maintenance costs, as well as improving the safety level. So, it is essential to analyze the external information so as to evaluate the internal components state which, generally, are inaccessible without disassemble the machine [21]. Fault diagnosis improves the reliability and availability of an existing system. Since various failures degrade relatively slowly, there is potential for fault diagnosis at an early step. This avoids the sudden, total system failure which can have serious consequences. Fault diagnosis provides more information about the nature or localization of the failure. This information can be used to minimize downtime and to schedule adequate maintenance proceeding. In recent years, online fault diagnostic systems have been gaining considerable amount of business potential. The need for automating industrial processes and reducing the cost maintenance has simulated the research and extension of faster and robust fault diagnosis. Attempts have been created towards classification of the most common type of rotating machinery problem [23]. Vibration analysis is one of the main techniques used to the nondestructive diagnosis and identification of various defects in rotary machines. Vibration analysis provides early information about progressing malfunctions for future monitoring purpose. Rotating machineries are used considerably utilized in the manufacturing of industrial products. Shafts as a key rotating motion transmission component, plays a critical role in industrial applications. Therefore, attracts research interests in condition monitoring and fault diagnosis of this equipment [20]. Importance of shafts and bearings in condition monitoring of the machine is undeniable, thus, processing and analysis of acoustic and vibration signals of the shafts and bearings is the common way of extracting reliable representative of the machine condition. Misalignment faults are one of the foremost causes of failures in rotating machinery. One of the major effects of misalignment between rotors in drive system is the production of rotor preload in a specific radial direction. Consequently, misalignment produces a constant radial force, which pushes An Appropriate Approach for Misalignment Fault Diagnosis Based on Feature Selection and Least Square Support Vector Machine H. Ahmadi, A. Moosavian, and M. Khazaee C INTERNATIONAL JOURNAL OF MECHANICS Issue 2, Volume 6, 2012 97