Identification of Unbalance Severity through Frequency Response Function and Artificial Neural Networks

Rotating equipment is the main production process player. The new industrial revolution aims to increase the running time and efficiency for the rotating equipment to maximize the profit and reduce maintenance costs through avoid or early predict and correct occurred faults like unbalance, looseness, misalignment, etc.. Mass unbalance represents 35% of common problems in rotating equipment by control unbalance problem a void the equipment deterioration and power losses. These lead to open a wide research in rotor dynamic challenges fields. This study proposed a new method to identify unbalance severity in the system under study with statistical features and the amplitude frequency domain. The generated ANN validate by the statistical features with the amplitude values which is used in the ANN test. Practical works on overhang rotor bearing system at 25 Hz and measure the output unbalance conditions of the rotor. Different features in statistical and spectrum are generated from the vibration and the neural network is created. It is found that the statistical features results are better than the frequency domain amplitudes. ANNs are used to identify the unbalance severity and the method will help in predict faults and arrange best maintenance action