ENHANCING AIR BLOWER HEALTH MONITORING: A COMBINATION OF EXPERIMENTAL ANALYSIS AND MACHINE LEARNING TECHNIQUES

Abstract

Undesirable vibrations in equipment activities result in energy wastage and undesired noise. This work quantifies the Peak and Root Mean Square (RMS) values to identify machinery vibrational features, important for anomaly identification. The essay outlines a methodical strategy for diagnosing defects via experimental analysis, offering in-depth understanding of the challenges involved in maintaining and monitoring industrial blowers and motors. Comparing fixed frame arrangement to flexible design shows less vibrations in the flexible setup. Machine learning techniques have proven to be useful in automated fault detection through the analysis of frequency vs magnitude graphs. During the training and validation stages, the models consistently show advancements in accuracy and decrease in loss, ultimately reaching detection rates that are nearly flawless. Combining traditional analysis with machine learning improves the reliability of defect diagnosis.