Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ Recolector de Cienci...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
UCrea
Master thesis . 2021
License: CC BY NC ND
Data sources: UCrea
versions View all 3 versions
addClaim

Mantenimiento predictivo y análisis espectral: de Fourier al aprendizaje automático

Predictive Maintenance and Spectral Analysis: From Fourier to Machine Learning
Authors: Sáinz-Pardo Díaz, Judith;

Mantenimiento predictivo y análisis espectral: de Fourier al aprendizaje automático

Abstract

RESUMEN: El mantenimiento predictivo es un conjunto de técnicas que analizan parámetros físicos para prevenir fallos en los equipos. El presente estudio explora las técnicas más avanzadas para abordar los principales problemas que plantea el mantenimiento predictivo, a saber, la detección y clasificación de los fallos, y la predicción del tiempo hasta el fallo. Estas técnicas se aplican a series temporales obtenidas a partir de sensores colocados en los rodamientos. En primer lugar, se revisan las técnicas clásicas de detección de fallos, basadas en técnicas de análisis espectral mediante las transformadas de Fourier y de Hilbert-Huang. Se realiza una revisión de las técnicas contemporáneas de aprendizaje automático para la detección de fallos, y se presentan resultados experimentales sobre series temporales obtenidas a partir de varias bases de datos de código abierto de rodamientos. A continuación, se ofrece una visión general del problema de la predicción del tiempo hasta el fallo, que se aborda a través del contexto de la detección de anomalías. Se presentan varias técnicas de aprendizaje automático para abordar este problema, y se incluye una comparación de los clasificadores obtenidos basada en una amplia gama de resultados experimentales. El presente estudio sienta las bases de investigación para un proyecto industrial que pretende operar con datos reales. Todo el código desarrollado se ha escrito en Python y se ha distribuido a través de un repositorio de código abierto

ABSTRACT: Predictive maintenance is a set of techniques that analyzes physical parameters to prevent equipment failure. The present study explores state-of-the-art techniques to address the main problems involved in predictive maintenance, namely the detection and classification of failures, and the timeto-failure prediction. These techniques are applied to time series obtained from sensors placed on bearings. First, the classical techniques for fault detection are reviewed, based on spectral analysis techniques using the Fourier and the Hilbert-Huang transforms. A review is performed of contemporary machine-learning techniques for failure detection, and experimental results are presented on time series obtained from several open-source databases of bearing data. Next, an overview is given of the time-to-failure prediction problem, which is approached through the context of anomaly detection. Several machine-learning techniques are presented to tackle this problem, and a comparison of the obtained classifiers is included based on a wide range of experimental results. The present study lays the research groundwork for an industrial project that aims to operate on live data. All developed code was written in Python and distributed through an open-source repository.

Máster en Ciencia de Datos

Country
Spain
Related Organizations
Keywords

Detección de fallos, Rodamiento, Bearing, Machine learning, Mantenimiento predictivo, Predictive maintenance, Anomaly detection, Detecciónn de anomalías, Fault detection, Aprendizaje automático

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    0
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Average
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
    OpenAIRE UsageCounts
    Usage byUsageCounts
    visibility views 203
    download downloads 109
  • 203
    views
    109
    downloads
    Powered byOpenAIRE UsageCounts
Powered by OpenAIRE graph
Found an issue? Give us feedback
visibility
download
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
views
OpenAIRE UsageCountsViews provided by UsageCounts
downloads
OpenAIRE UsageCountsDownloads provided by UsageCounts
0
Average
Average
Average
203
109
Green