Nowadays, condition monitoring (CM) techniques have prepared facility to optimize maintenance systems. CM includes some methods that vibration test is the most important method for rotational machines. CM of slow speed machines is challenging to the various industries, because they are often the most important machines in the production lines. On the other hand, CM of them is not achieved by the vibration test. In the recent years, several studies are carried out by acoustic emission (AE) techniques to measure and analyze physical wave produced by slow speed machines. Since the AE wave is created by variation of the tension and the microscopic strain in material or on its surface, some defects like friction and impact produce AE waves independent of rotational speed. So, we have an applied method for fault detection of slow speed machines. This research focuses on the fault detection of rolling element bearings, because they are used in the vast majority of slow speed machines and most sensitive parts to be damaged. Finally, AE techniques are able to detect creation and progress of their fault. In this research, the AE phenomenon and propagation principles, producer objects in machinery, data measurement, indication and processing methods, fault detection of rolling element bearings and artificial neural network are studied. Experimental tests are then done. In the first step a test rig is subjected to AE test, after that, two slow speed machines in Mobarakeh Steel Complex are tested and analyzed. In this thesis, it is cleared that if a slow speed is periodically tested and analyzed by using AE techniques and gathering data is processed by a suitable wavelet transform then statistical parameters, specifically kurtosis, are calculated and their trends are observed, we can realize fault progress of the machine and detect type of the defects with the use of envelop function. It is also observed if an artificial neural network is correctly designed for a machine, prediction of the failure intensity is achievable by using AE data. Additionally, friction and impact phenomenon in the damaged slow speed bearings are studied. Previous research done with the use of test rig, shows friction falls between 50 kHz to 100 kHz. Here this is ascertained for an industrial machine. Relation to impact phenomenon, it is cleared that natural frequency of AE sensor is able to be excited and dominant frequency of wave happens in the range of natural frequency of the sensor. Keywords: Condition monitoring, Acoustic emission, Slow speed machines, Wavelet transform, Rolling element bearing, Friction, Impact, Kurtosis
