Using Artificial Intelligence to Determine the Type of Rotary Machine Fault

Keywords: Vibrodiagnostics, Neuron Network, Classification Learner, Machine Learning, Matlab, Industry 4.0, Classification Method, Static Unbalance, Dynamic Unbalance

Abstract

The article deals with the possibility of using machine learning in vibrodiagnostics to determine the type of fault of rotating machine. The data source is real measured data from the vibrodiagnostic model. This model allows simulation of some types of faults. The data is then processed and reduced for the use of the Matlab Classication learner app, which creates a model for recognizing faults. The model is ultimately tested on new samples of data. The aim of the article is to verify the ability to recognize similarly rotary machine faults from real measurements in the time domain.

References

Broch, J.T.: Mechanical Vibration and Shock Measurements, Naerum, Denmark (1984). ISBN 8787355361

Blata, J., Juraszek, J.: Metody technicke diagnostiky, teorie a praxe [in Czech]. REPRONIS, s.r.o.: Ostrava (2013). ISBN 978-80-248-2997-5

Castellani, M., Lalchandani, R.: An experimental study on competitive coevolution of MLP classiers. Mendel 23(1), 41–48 (2017)

Ren, Y., Suganthan, P.N., Srikanth, N., Amaratunga, G.: Random vector functional link neural network for short-term wind power ramp forecasting. Mendel 21(1), 77–87 (2015)

Zuth, D.: Using HIL simulation and genetic algorithms for controller tuning. Mendel 22(1), 25–30 (2016)

Zuth, D., Vdolecek, F.: Meren vibrac ve vibrodiagnostice [in Czech]. Automa 16(1), 32–36 (2010)

Sedenka, D., Blata, J., Kasiar, L., Heisig, L., Zarsky, V.: Deployment of technical diagnostics during commissioning of small pumped storage Hydropower plant. International Multidisciplinary Scientic Geo-Conference Surveying Geology and Mining Ecology Management, SGEM, 17(42), 175–182 (2017)

Vdolecek, F.: Terminologie v oboru nejistot mereni [in Czech]. Akustika, 16(1), 40–42 (2012)

ADXL335 Datasheet and Product Info – Analog Devices http://www.analog.com/en/products/sensors-mems/accelerometers/adxl335.html. [Online; accessed 20-Apr-2018]

USB-6009 – National Instruments http://www.ni.com/cs-cz/support/model.usb-6009.html. [Online; accessed 20-Apr-2018]

Statistics and Machine Learning Toolbox – MathWorks – Makers of MATLAB and Simulink https://ch.mathworks.com/help/stats/. [Online; accessed 20-Apr-2017]

Train models to classify data using supervised machine learning – MathWorks – Makers of MATLAB and Simulink https://www.mathworks.com/help/stats/classicationlearner-app.html. [Online; accessed20-Apr-2017]

Published
2018-12-21
How to Cite
[1]
ZuthD. and MaradaT. 2018. Using Artificial Intelligence to Determine the Type of Rotary Machine Fault. MENDEL. 24, 2 (Dec. 2018), 49–54. DOI:https://doi.org/10.13164/mendel.2018.2.049.
Section
Articles