Backflow when using "outlflow" ?
 

is it possible occur backflow when using "outflow" as outlet ?

It is possible because outflow condition means that the flow field in the boundary are extrapolated from within the domain while maintaining overall mass balance. So the backflow would occur if the boundary cut some recirculation zone.

If the back flow are significant, then the result is not physically meaningful because the flow near the boundary is not parabolic. You should relocate the outflow boundary, or change to other type of boundary.

I earlier had the same problem, reversed outflow. I haven't figured out this for sure, but check if you have enabled shell conduction. It may create too thick layer.

thanks a lot for your suggetion. I simulate downdraft gasification process. in the beginning, I use "pressure outlet" as outlet but occur backflow. So i change the outlet become "outflow", but still occur backflow. however, it is less than before.

Would you like to suggest me how to prevent occur backflow?

Try turn on the "radial equilibrium pressure distribution" if the flow exhibits strong swirling when exiting the domain.

You may enlarge your computational domain by adding an artificial domain (sponge region) to the outlet boundary. The flow at the outlet of the sponge region will be relative uniform so that any boundary condition should be reasonable. A low resolution mesh in the sponge region is acceptable so it won't add much to the computation cost.

Ok, thanks a lot. but, How to add an articial domain to outlet boundary?

image that a large tank is attached to the outlet of your original computation domain.
now the computation domain consists of the original one and the tank.

the shape of tank could be arbitrarily chosen.