Experimental investigation of efficiency of worm gears and modeling of power loss through artificial neural networks

Karabacak Y. E., Baş H.

Measurement: Journal of the International Measurement Confederation, vol.202, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 202
  • Publication Date: 2022
  • Doi Number: 10.1016/j.measurement.2022.111756
  • Journal Name: Measurement: Journal of the International Measurement Confederation
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, INSPEC
  • Keywords: Worm gearbox, Power loss, Efficiency, Friction, Gear fault, Artificial neural network, Data -driven model, REGRESSION-MODELS, PREDICTION, OPTIMIZATION, CAPABILITIES
  • Karadeniz Technical University Affiliated: Yes


© 2022 Elsevier LtdIn this study, an experimental system that can operate at different speeds and loading rates was developed for the efficiency calculations of worm gears (WGs), and measurements were made accordingly. A comprehensive efficiency analysis has been made for WGs based on power loss calculations under different operating conditions. The efficiencies obtained from the healthy gearbox were also compared with three different gearboxes with different failures (pitting, wear and breakage). The negative effect of the faults on the efficiency that may occur in the gearbox has been investigated comparatively. A widely used type of machine learning algorithms, called artificial neural network (ANN), is applied for data-driven modeling of the power loss in WG under different operating conditions. The input and output power of the gearbox are utilized to design an ANN-based model for the power loss.