Utveckla en FE-modell för att studera effekten av presspassning på en hypoid växelsats

Sammanfattning: Gears are an integral part of devices ranging from simple wrist watches to complex systems like automotive and wave energy converters. They play a very important role in the transfer of torque. There are several types of gears to be chosen from, depending on the application. This thesis work deals exclusively with hypoid gears. A hypoid gear is a type of bevel gear, that transmits torque between two non-parallel shafts. It is similar to a spiral bevel gear, except that the pinion axis can be oset. It is this ability to offset, that renders hypoidgears as highly sought-after gears for automotive applications. It is crucial that these hypoid gears are designed efficiently and that their life before failure is predictable, to a reasonable degree of accuracy. To enable such predictions, this thesis makes an effort to build a finite element (FE) model. With the developed FE model, a study of the effect of press-fit on the ring gear's root bending stresses, that is present in a hypoid gear set of a car's Rear-Drive Unit (RDU), was carried out. The FE model was built on three different software - Transmission 3D (T3D), Hypoid Face Milled (HFM) and MSC Marc (Marc). The effect of press-fit on root bending stresses of the ring gear was first analyzed on T3D. To determine if the press-fit was correctly induced, the model was rebuilt and analyzed on Marc. HFM was used to determine the effect of inclusion of different components on the root bending stresses. Additionally,the HFM model was also replicated on Marc and a press-fit of 100m was induced. This was done to see if modelling the press-fit on HFM gave similar results to that of T3D and if using HFM in conjunction with Marclead to a better modelling procedure. It was found that the maximum root bending stresses increased linearly with increasing press-fit dimension. It was also found that the inclusion of different parts does not cause a significant increase in root bending stresses, except for the inclusion of differential cage. The effect of press-fit could not be quantified despite knowing that it affects the root bending stresses. When the same analysis done on HFM or T3D was done on Marc, there wasa 10% increase in stresses at highly stressed zone in Marc model of T3D; there was 10-15% increase in stresse sat highly stressed zone in the HFM model of Marc. Hence, quantification remained an impediment. It was not possible to quantify the error that occurred during the migration of analysis from HFM or T3D to Marc. However, potential causes of these errors could be the error in computation of bearing forces and difference in the definition of contacts in the software. Owing to large computational time and limited working period, all the analysis on Marc was carried out for one position and for the first time step of the gear mesh. If the errors in migration of analysis from one software to another could be quantified, then this modelling can be used to estimate the contribution of press-fit to rootbending stresses on the ring-gear. Although the contact pattern comparison between the virtual models andthe physical test suggests that HFM is a more trustworthy software, it is recommended to conduct some straingauge measurements on different ring gear teeth and then compare the results with those of the virtual model.