First of all, such porous carbon structures require considerably large system sizes, which is why you very likely need to use classical potentials to model the system.
You could follow the procedure outlined in the paper
http://homepages.rpiscrews.us/~shiy2/publications/jcp.porous.pdfAlthough we do not have the exact potential of this paper implemented, you could use one of our Tersoff or Brenner potentials for Carbon.
Then, you essentially need to create a random arrangement of carbon atoms at the desired density.
You can do this e.g. by taking a diamond cell, repeating it to obtain the desired number of atoms, then increasing the lattice constants to obtain the desired atomic density (while keeping the fractional coordinates constant). As a new plugin in ATK-2015, you can also use the new Amorphous prebuilder plugin after downloading the SCAITools AddOn.
This initial structure has to be annealed at very high temperatures (in the paper they use 21 000 K) in a NVT thermostat e.g. Langevin or NVT Nose Hoover Chain to randomize the atomic arrangement. Such high temperatures require a very small time step e.g. 0.1 fs or smaller.
Then you can basically follow the instructions in the paper, by cooling the system to lower temperatures. In ATK-2015 this can easily be achieved in the NVT Nose Hoover Chain thermostat by choosing an initial temperature (e.g. 21 000 K and a final temperature, e.g. 1000 K). In ATK-2014 you need to run a series of constant temperature simulations, e.g. with the Langevin thermostat, as described in the amorphous systems tutorial.
Choose a sufficiently large number of steps, ideally resulting in several nanoseconds simulation time, to achieve a relatively slow quenching. Remember also to set the log interval parameter to something like 1000 or even larger, otherwise your trajectory file will become huge and the simulation very slow.
The atoms will move around wildly due to the high temperatures, so in order to be able to inspect your structures properly you need to send the final structure of the MD simulation from the Movie Tool to the Builder and wrap all atoms back into the cell via 'Bulk Tools > Wrap' .