International Journal of Engineering & Technology IJET-IJENS Vol: 11 No: 04 97 118604-5454 IJET-IJENS @ August 2011 IJENS I J E N S Abstract—Research in damage of gear and gear pairs using vibration signals is still very attractive, because vibration signals from a gear pairs are complex in nature and not easy to interpret. Predicting gear pairs defects by analyzing changes in vibration signal of gears pairs in operation is a very reliable method. Therefore, a suitable vibration signal processing technique is necessary to extract defect information usually covered under noise of other gear pairs dynamic factors. This paper presents the results of an evaluation of vibration analysis techniques as a method for the gear and gear pairs condition assessment. The origin of vibration in gear pairs and useful definition of damage identification techniques are presented. The detection and assessment capability of some of the most effective vibration techniques are discussed and experimentally compared, concerning a multistage industrial gearbox. In particular, the results of estimated vibration signal processing techniques are compared. Advantages and disadvantages of estimated techniques having in mind specific limitation have been shown. Further research in damage identification of gear pairs have been pointed out. Index Term— spiral bevel gear, gear pairs, gear damage, vibration signal processing I. INTRODUCTION Gear pairs in gearboxes normally generate vibration and corresponding vibration signal could be used as reference characteristics when the gear is in good mechanical condition. If defects occur to one of the gears during operation, the faulty gearbox would result in serious damage. Changes in vibration signals are often an indication that the gear pair meshing condition is changing. Measuring vibration gearbox with the help of accelerometer mounting on the gearbox housing is one of the best methods for gear pair damage assessment. Therefore, condition monitoring of the gearbox system during its operation is crucial to prevent the system from malfunction that could cause damage or entire system shutdown [1]. R. B. Author is with the Coal Mines Kreka, Open pit mine Dubrave, Mechanical Structure and Vibration Research Division, Tuzla, Bosnia & Herzegovina, rusmir.bajric@kreka.ba Up to now condition monitoring and damage identification of industrial gearboxes has received significant attention by researchers engaged in multidisciplinary activities. The rapid progress in materials technology, intelligent sensor technology, signal processing and information technologies brings new solutions to solve a variety of problems associated with failures of industrial gearboxes in a real operational environment accurately and efficiently [2]. There have been many investigations carried out to monitor and assess of industrial gearboxes using different techniques. These methods are well-established industrial practices and among them vibration signal processing technique is well known. However, since the vibration signals measured from gearboxes are non- stationary and transient in nature, when damage occurs it is even more interesting to carry out investigation. All those technique have some limitations and cannot be applied in all conditions, i.e. some types of failure cannot be detected by simple vibration methods. Hence it is more desirable to investigate possibilities of some of those methods. The simple spectral analysis is generally unable to detect gear damage at an early stage; for this reason, many researchers have proposed the application of other vibration assessment techniques for the early detection of damage symptoms. The aim of this paper, on the basis of experimental results is to evaluate and compare detection and diagnostic capabilities of some of vibration signal processing techniques. II. THE SOURCES OF GEAR VIBRATION Gears are very widely used in machines to transmit power from one shaft to another, usually with a change in speed and torque. The study of the dynamic behavior of gearboxes has received moderate attention in literature, but due the great dynamics complexity of gears, it remains an insufficiently understood area. In practice out of balance and bearing forces are active in gear dynamics, but also the geometry of the gear profile has a crucial effect on the vibration behavior. In general flexural vibration will be more important then torsional vibration because flexural vibrations are transferred directly to the housing via the bearings [1]. In practice, the situation is not so ideal, as the teeth deform under load, introducing a Review of Vibration Signal Processing Techniques Towards Gear Pairs Damage Identification Rusmir Bajrić 1 , Denijal Sprečić 2 , Ninoslav Zuber 3 1 Coal Mines Kreka, Open pit mine Dubrave, Tuzla, Bosnia & Herzegovina 2 Faculty of Mechanical Engineering, University of Tuzla, Tuzla, Bosnia & Herzegovina 3 Faculty of Technical Sciences, University of Novi Sad, Novi Sad, Serbia