# Modelling a pulley in ANSYS workbench

Tags:
1. Feb 9, 2012

### LukeJS

Hi,

I’m trying to do an FEA analysis of a single rubber belt-style CVT pulley in ANSYS workbench. I’ve created a Solidworks assembly consisting of the pulley, key and a shaft with a keyway.

To begin with I’m starting with a Static Structural simulation. (I realise this will not give me centrifugal stresses.) I’ve added a fixed support to the segment of the pulley in contact with the belt (outlined in the Solidworks model as a split line projected on to the pulley face). A cylindrical support is added to the shaft, with everything fixed except the tangential component. A moment is applied to the shaft, and a compression force is applied to the back of the pulley in the direction of the belt so that the pulley is compressed against the belt. All the contact regions are defined as frictional (coefficient of friction I’m just taking as 0.5 while I try and work out how I’m going to do this – I’ll figure out more accurate values later).

The other thought I’ve had is to fix the shaft completely and instead apply a force to the belt contact area by creating a cylindrical coordinate system about the centre of the shaft applying the force to the y axis of that. The problem here is I really don’t know how ANSYS is applying the force, or how it’s interpreting what I’ve told it to do. In fact the pulley will be subject to a maximum moment from the belt of about 80 Nm.

Regarding the centrifugal forces – the maximum rpm won’t exceed 150, and I plan to get an estimate of the stresses by hand - if I stick with a static simulation - (e.g. using this: http://www.codecogs.com/reference/engineering/materials/rotating_discs_and_cylinders.php).

Incidentally I’m only really concerned with stresses in the pulley – not so much the key or shaft.

Also I’m not averse to doing a dynamic/transient simulation, but as you can probably tell I’m a novice at this and I’m beginning to think I’ve already bitten off slightly more than I can chew – the learning curve is pretty steep. That said if someone could give me some pointers to start off setting up a transient simulation of this type I’m game to give it a go.

The goal of this is primarily to allow me to compare designs – but I’d like to have some confidence in the results I’m getting. Unfortunately building an experiment to verify the model is not an option at the moment.

I’ve attached a couple of pictures to try and illustrate the setup.

Thanks,

Luke

File size:
78.9 KB
Views:
988
File size:
78.7 KB
Views:
702
File size:
67.4 KB
Views:
1,565