Work

Detección y Diagnóstico de Fallas en Rodamientos con Velocidad Variable Aplicando Aprendizaje Profundo

Year: 2020
Type: dissertation
Abstract: Detect possible anomalies in rotating equipment bearings, minimize the risk of loss of lost \nprofits, repair costs and process safety impairment as a consequence of a failure. Bearings \nare an essential component in any rotating equipment, their load capacity and reliability are \nimportant to the overall performance of the equipment. \nMost methods of detection and diagnosis of bearing failures support their models at a certain \nspeed of rotation and do not have the ability to quantify damage in real time, but in reality \nrotating equipment operated at variable speed. In this operational scenario, the accuracy of \nthese detection methods is low and complex parameter adjustments are required. \nThis thesis develop a software for the detection and diagnosis of failures in rotating \nequipment bearings, applying Deep Learning based on a Convolutional Neural Network on a \nContinuous Wavelet Transform Scalogram and Decision Tree, which serves as support \ntechnical staff for timely decision making in operational situations. This innovative proposal \nhighlights the combination of computational perception methods, artificial cognitive systems, \nand machine learning to detect faults and damage severity of a bearing, under different \nrotational speed variations, simulating real operating conditions, adding greater complexity to \nthe characterization and classification of faults. The results of the evaluation and prediction of \nthe model were satisfactorily meeting the stated objectives. \nThe application of artificial cognitive systems techniques is relevant to solve these types of \nproblems due to the high capacity for identifying characteristics and classification that they \nrequire, and it also opens the door for the use of learning transfer theories where the \nproposed solutions are extrapolated. with experimental data on similar problems present in \nindustrial settings.
Language: Spanish
Open Access status: green
OpenAlex ID: W3097737383
Cites: 40
Cited by: 1
Related to: 10