Windage, churning and pocketing power losses of gears: different modeling approaches for different goals

Energy efficiency represents one of the most relevant trends in many fields, including the sector of power transmissions and gears, which are involved whenever power has to transmitted and transformed. For instance, in the automotive industry, gearboxes can contribute to the overall efficiency of the system and promote lower fuel consumption and emissions, both allowing an optimization of the whole system and reducing their own power losses. In many circumstances a better efficiency corresponds to lower operating temperatures and to a higher reliability of the systems, which can be related to the final profit, like in industrial applications, or even to the success, like for instance in motorsport racing. Improving the efficiency is therefore a main issue also for the gearbox manufacturers, and the availability of methods and tools to forecast the behavior with respect to lubrication and power losses since the beginning of the design phase strongly contributes to the goal. In the years, many empirical models were derived from experimental tests and have represented the only available tool for such purpose, but today, thanks to the recent developments in the computer science, numerical approaches allow a more accurate modeling of the physics behind the power dissipation and also allow a description of the oil flow inside a gearbox, which is fundamental with respect to the reliability of the components of the transmission. Both approaches, either derived from experimental tests or based on numerical simulations, have advantages and drawbacks. For each single case and problem, depending on the specific condition, the most appropriate model is not always the same. In this paper a review of the different available tools is proposed, describing critically the properties of the single approaches in order to understand when each of them should be preferred. The review also includes the latest developments by the authors, which have not been previously published yet.