A Multi-Objective Macro-Geometry Spur Gear Optimisation Process to Improve Weight, Efficiency and NVH Performance Utilising Neural Networks

Dynamic gear simulations can be expensive. To overcome this a team of researchers and academics have investigated using peak-to-peak Static Transmission Error (STE) as a surrogate measure for NVH. This resulting paper demonstrates that minimising peak-to-peak STE leads to substantial reductions in Dynamic Transmission Error (DTE), which is closely related to gear vibration and noise.

The paper proposes a multi-objective optimisation framework for spur gear macro-geometry that simultaneously improves:

• Weight
• Mechanical efficiency (reduced power losses)
• NVH (Noise, Vibration and Harshness)

Unlike many previous studies that focus on micro-geometry modifications (such as profile relief), this work demonstrates that substantial NVH improvements can be achieved through macro-geometry optimisation alone, preserving manufacturing simplicity and gear interchangeability.

The paper shares findings from across seven case studies, where the optimisation achieved:

• >35% reduction in RMS Dynamic Transmission Error (improved NVH)
• >40% reduction in power losses
• simultaneous optimisation of efficiency and vibration without micro-geometry modifications

These are significant improvements considering only macro-geometry changes were used.

Thanks to the authors for granting us permission to link to the paper on Science Direct’s website.
Read it here: https://www.sciencedirect.com/science/article/pii/S0094114X26000728?via%3Dihub

Authors: Christos Kalligeros, Christos Papalexis, Georgios Kostopoulos, Klearchos Terpos, Panteleimon Tzouganakis, Konstantinos Kostas, Dunant Halim, Jian Yang, Christos Spitas, Antonios Tsolakis, Emmanouil Sakaridis, Vasilios Spitas