摘要：The quality of maintenance has an important influence on firm performance. Detection and monitoring of the progression of the defects of machines can reduce maintenance costs by minimizing the number of unplanned shutdowns and ensuing economic losses. The state of a machine depends largely on the condition of rotating elements that compose it and especially bearings. The vibration measurements are used as indicators of the health status of machines and timeliness of maintenance. The objective of this work is to establish a system based on artificial neural networks to know precisely the position of an incipient defect on the bearing elements and quantify its severity, using indicators from vibration signals.
