I have done it with limited success by creating the pits in the terrain model. It will really only be successful provided the pipe capacity is the limiting factor, not pit inlet capacity. The raw terrain model was created using a very small grid, 0.5 metre, however, in future I would use a 0.25 metre grid. The problem is that the terrain Hec creates can not have vertical drops, hence the sloping side needs to be as steep as possible. The top of the pits have two weirs, one along three sides of the pit, the other along the four side. If you need to limit the overland flow into the pits, the weirs can be raised accordingly. The pipes are culverts connecting the pits.
The 2D flow area grids in the model I am currently building are 1×1 metre in one area and 2×2 metre in the other area. It is a small dam breach study in an urban area, hence the small grids. The outlet from the basin is by grated pits with the pipes discharging into the downstream pipe system. When inserting the weirs for each pit, make the min and max cell size adjacent the weir 1 metre, even in the 2 metre flow area.
There has been a little instability at some of the pits, however, I have minimised this by adopting a timestep of 0.3 sec.
This works for me, but I would not be using this method for design.