This paper presents an Artificial Intelligence based methodology to detect typical manufacturing errors in external gear machines based on a hybrid approach by means of data-driven and model-based approach using Multics-HYGESim. In particular, the study is finalized to detect manufacturing gear errors such as conicity, concentricity, and axial runout defects in those machines, which can affect their hydraulic performance, reliability, and noise emission. To validate the effectiveness of the proposed method, a specific experimental test campaign has been carried out on eight physical units considering realistic operating conditions during a standard end-of-line test, demonstrating the robustness and efficacy of the proposed method in detecting flawed components in the electric motor-driven external gear pump. This study offers valuable insights into the development of advanced condition monitoring techniques for fluid power components based on different machine learning classifiers combined in an ensemble learning algorithm, which can contribute to improving system reliability, efficiency, and safety. Besides, the numerical simulations can ultimately help manufacturers to investigate these interactions and optimize their pump design accordingly. The findings demonstrate that the proposed methodology achieves above 92% of accuracy in the tested operating conditions, offering insights into the development of future condition monitoring techniques particularly in fluid power systems.

Detection of Typical Manufacturing Errors in External Gear Machines Using Numerical Simulation and Data Driven Machine Learning

Frosina, Emma;
2023-01-01

Abstract

This paper presents an Artificial Intelligence based methodology to detect typical manufacturing errors in external gear machines based on a hybrid approach by means of data-driven and model-based approach using Multics-HYGESim. In particular, the study is finalized to detect manufacturing gear errors such as conicity, concentricity, and axial runout defects in those machines, which can affect their hydraulic performance, reliability, and noise emission. To validate the effectiveness of the proposed method, a specific experimental test campaign has been carried out on eight physical units considering realistic operating conditions during a standard end-of-line test, demonstrating the robustness and efficacy of the proposed method in detecting flawed components in the electric motor-driven external gear pump. This study offers valuable insights into the development of advanced condition monitoring techniques for fluid power components based on different machine learning classifiers combined in an ensemble learning algorithm, which can contribute to improving system reliability, efficiency, and safety. Besides, the numerical simulations can ultimately help manufacturers to investigate these interactions and optimize their pump design accordingly. The findings demonstrate that the proposed methodology achieves above 92% of accuracy in the tested operating conditions, offering insights into the development of future condition monitoring techniques particularly in fluid power systems.
2023
978-0-7918-8743-1
flaw detection, end-of-line test, external gear machine, Multics-HYGESim, machine learning
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/20.500.12070/62159
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact