[Militaum]

I've been really struggling with the simulation of a MOV(Metal Oxide Varistor) , that has a transient volumetric heat generation (W/m³) , with convective and radiative heat transfer to ambient air. The component is modelled as a filleted circular disk (see attachment), surrounded by a cylindrical enclosure. Due to the axysimmetrical conditions, a 2D axysimmetric approach was used, in order to be able to handle a more refined and detailed mesh. My goal is to obtain the transient profile of the temperature at the upper surface of the component, and it is supposed to match the experimental and analytical results( attached below), but so far they diverged quite significantly.


The mesh is composed of triangle elements, of a standard size of 5 x 10^-5 m, with an edge refinement for the upper and lower surface of the MOV ( element size of 1x 10^-5 m). Resulting in a mesh with 3.5 million elements , approximately.

A pressure-based formulation is being considered, with Energy and Viscous Laminar models ON , and Radiation OFF. As far as BC are concerned, the lower wall is kept at 53ºC(326K),the side wall is treated as a velocity inlet ( v=0.001 m/s) and the top wall is treated as a pressure outlet ( 0 gauge pressure). The upper surface of the component is treated as a wall that radiates with and emissivity of 0.71(determined experimentally) to an external temperature of 23ºC(296K). As for the lower surface, it has the same BC type and emissivity, but it radiates to an external temperature of 53ºC(326 K).

A transient volumetric heat source profile is imported through a .txt inside fluent.

An implicit approach was selected, with coupled scheme for pressure-velocity coupling, second order upwind for energy and momentum equations and second order implicit transient formulation.

The initial temperature of the component is 53ºC(326K) and of the surrounding air is 23ºC(296K). The timestep size is 5x10^-5s , with a total number of 100 time steps. A maximum number of iterations per timestep of 50 was established. Unfortunately, I'm not able to increase the timestep size because of the volumetric heat source curve behavior(attached below).

I've tried messing around with the Flow Courant Number, since I noticed by reading threads here in the forum that this can influence convergence significantly. But I had no luck in doing so, the temperature behavior of the component was the same. It rises around 1 K after 1 ms and stays at that value for the rest of the time interval.

Any ideas or inputs are valuable, since I've tried a lot of different approaches to this problem.

[deniz13]

Have you find the solution
We want to demostrate Electric field on the surface of surge arrester using ANSYS.
How can we create varistor model in ANSYS?