Supervised process monitoring and fault diagnosis based on machine learning methods

2019 ◽  
Vol 102 (5-8) ◽  
pp. 2321-2337 ◽  
Author(s):  
Hajer Lahdhiri ◽  
Maroua Said ◽  
Khaoula Ben Abdellafou ◽  
Okba Taouali ◽  
Mohamed Faouzi Harkat
Keyword(s):  
Machine Learning ◽  
Fault Diagnosis ◽  
Process Monitoring ◽  
Learning Methods ◽  
Machine Learning Methods

An experimental methodology to evaluate machine learning methods for fault diagnosis based on vibration signals

2020 ◽  
pp. 114022
Author(s):  
Thomas Walter Rauber ◽  
Antonio Luiz da Silva Loca ◽  
Francisco de Assis Boldt ◽  
Alexandre Loureiros Rodrigues ◽  
Flávio Miguel Varejão
Keyword(s):  
Machine Learning ◽  
Fault Diagnosis ◽  
Experimental Methodology ◽  
Learning Methods ◽  
Vibration Signals ◽  
Machine Learning Methods

Fault diagnosis of ball bearings using machine learning methods

2011 ◽  
Vol 38 (3) ◽  
pp. 1876-1886 ◽  
Author(s):  
P.K. Kankar ◽  
Satish C. Sharma ◽  
S.P. Harsha
Keyword(s):  
Machine Learning ◽  
Fault Diagnosis ◽  
Ball Bearings ◽  
Learning Methods ◽  
Machine Learning Methods

Einzelteilüberwachung in Beschleunigungsphasen*/Single item processes containing acceleration phases – A comparison between conventional regression tools and machine learning

2019 ◽  
Vol 109 (05) ◽  
pp. 347-351
Author(s):  
B. Denkena ◽  
B. Bergmann ◽  
M. Witt ◽  
Keyword(s):  
Machine Learning ◽  
Process Monitoring ◽  
Input Signal ◽  
Regression Models ◽  
Learning Methods ◽  
Single Item ◽  
Machine Learning Methods ◽  
Model Based ◽  
Conventional Regression

Die modellbasierte Prozessüberwachung für die Einzelteilfertigung scheitert häufig an Prozessphasen, welche Beschleunigungsanteile enthalten. Der Beitrag widmet sich der Kompensation von Beschleunigungs- und Reibanteilen im Messsignal durch kennfeldbasierte Ansätze sowie Methoden des maschinellen Lernens. Für die anschließende Modellierung des Gesamtsignals zur Prozessüberwachung werden Regressionsansätze mit Methoden des maschinellen Lernens verglichen, welche sich als besonders geeignet zeigen.   Model-based process monitoring often fails to adequately assess process segments that exhibit acceleration. This work evaluates different compensation approaches for non-process-specific parts of the measured signals. For a subsequent modelling of the complete input signal enabling the monitoring of single item processes, conventional regression models are compared to machine learning methods. The latter prove to be very well suited for these requirements.

Sign in / Sign up

Export Citation Format

Share Document