hi all: If the fluid want to be fully develpoed,the length of the channel would be very large,both the number of cell.If the length of the channel is not large enough,will it cause the solution divergence?By the way,how to make the fluid become fully developed fast?

Use a cyclic boundary condition at inlet & outlet of the channel. Impose the mass flux (or alternatively the pressure drop). That should give you a fully developed turbulent flow.