Our CompMech Group Member, Kayalvizhi Lakshmanan, is going to present her results in a talk entitled “Predictive Maintenance of an external gear pump using Machine Learning algorithms, Support Vector Machine and Multilayer Perceptron” at the WCCM-ECCOMAS CONGRESS will take place from January 11 to 15, 2021.
For any manufacturing industries, performing predictive maintenance is essential to prevent unexpected failures that may cause huge disruption in the production. In recent years, AI-based techniques are widely used to perform predictive maintenance. In this work, due to the absence of an experimental dataset of the external gear pump, machine learning with a high-fidelity computational fluid dynamic model are combined to perform predictive maintenance.
Under these premises, her work carried out under the supervision of Prof. Ferdinando Auricchio and Prof. Antonio Gil (Swansea University), and in collaboration with Industrial partner Dr. Fabrizio Tessicini (F-Lab), focusing on finding the best possible machine learning algorithm to perform predictive maintenance of an external gear pump.
Preliminary results of this work have been published in the conference proceedings of UKACM, 1-3 April 2020:
- K. Lakshmanan, A.J. Gil, F. Auricchio, and F. Tessicini. “A fault diagnosis methodology for an external gear pump with the use of Machine Learning classification algorithms: Support Vector Machine and Multilayer Perceptron”, DOI: 10.17028/rd.lboro.12097668.v1
Fig. 1: Quantification analysis of prognosis (prediction of remaining useful life): Test mean square error (MSE) of pressure and flow rate dataset with respect to the number of samples used to train the machine learning algorithms, Multilayer Perceptron (MLP) and Support vector Machine (SVM).
December 1st, 2020