[xh110120]

Hellow everyone,
Here is a question confusing me for several days. I want to use the temperature gradient mesh adaption to do the combustion calculation (refer to the fig. 1). In such method, I want to refine the region that the temperature gradient is large (flame front) and coarsen the region that the temperature gradient is small. This is a dynamic mesh adaption method which will be automatic executed after every 10 time steps.(transient calculation) When I use the serial calculation, this method function well (refer to fig. 2). However, when I use the parallel calculation, the meshes can only be refined and cannot be coarsened (refer to fig. 3). What's more, the refined region is not limited to where the temperature gradient is large, on the contrary, the whole high temperature region is refined. I don't know why and I want this adaption method can also function well under the parallel calculation, just like that under the serial calculation. Could you please help me? Thank you!

Fig. 1 Temperature contour

Fig. 2 Dynamic mesh adaption under serial calculation

Fig. 3 Dynamic calculation under parallel calculation

[xh110120]

I've tried several times. At last, I get the solution but I don't know why. Before, I used the Curvature as the method. Now I change it into Gradient. In this way, the dynamic mesh adaption can function well under parallel calculations. Though the problem has been solved, I don't know the reason.