Investigation and Prediction of Vibration Attenuation in Rotor Motor MRD System Under Variable Current Condition Using Artificial Neural Networks (ANNs)

This research investigates the application of artificial neural networks (ANNs) to predict fault characteristics in rotating machinery through vibration analysis under variable current conditions. The study utilizes a magnetorheological fluid (MRF) damper, integrated into a vibrational analysis setup, to modulate system damping properties. Arus Oil is considered for the study of vibration attenuation by varying the supply current. A predictive model is developed by applying the ANN model to the experimental data collected by changing supply current for 0.3 A, 0.4 A, and 0.8 A. The ANN-based predictive approach demonstrated high accuracy, with predicted vibration characteristics aligning closely with experimental data. This validation was achieved through graphical comparisons, underscoring the reliability of the model in identifying fault-related patterns in unmeasured operational conditions. The study highlights how varying magnetic fields influence the MRF’s rheological properties, altering the damping force and enabling dynamic control over vibration characteristics. The range of unstable speed at resonance is observed. It is also observed that as supply current is increased, the vibration amplitude decreases. The vibration amplitude for 0.1 A current is obtained by using the proposed ANN model, where the experimental data is unknown. The accuracy between the actual and predictive amplitude has very less difference. The findings illustrate how data-driven methodologies, combined with advanced material technologies like MRFs, can transform maintenance strategies from reactive to predictive, ensuring improved reliability, operational efficiency, and sustainability in industrial applications. The approach lays the groundwork for further innovation in fault diagnosis, predictive maintenance, and smart engineering systems.