Optimizing gear transmission efficiency through advanced machine learning and analytical techniques

This research explores the reduction of frictional power losses in spur geared transmissions, a pivotal aspect impacting the efficiency of mechanical systems. The study scrutinizes factors contributing to these losses, such as sliding and rolling velocities, normal loads at the tooth contact zone, oil temperature, and surface roughness of the gears. By integrating advanced analytical, numerical, and machine learning (ML) techniques—namely sensitivity analysis and neural networks—within the modeFRONTIER optimization framework, we enhance the methodology for identifying and implementing optimal design and operational parameters. This integration not only aids in effectively diminishing frictional losses but also leverages large-scale datasets for model training, improving the accuracy of predictions concerning loss patterns and their interactions. Experimental validations confirm the predictions of the models, demonstrating significant improvements in system efficiency by reducing power losses through optimization. This study employs ML models to accelerate analysis and ensure reliability in the face of varying operational conditions. The application of ML tools using modeFRONTIER represents a significant advancement in gear transmission performance optimization, pioneering a methodical approach that could be adapted across various mechanical engineering applications to improve robustness and reliability.