To estimate more quantitatively the fueling problem, let us consider the typical luminosity and power of an active nucleus, that can be of the order or higher than 1046 erg/s. If we assume a mass-to-energy conversion efficiency ~ 10% (L = dM/dt c2 ), then the mass accretion rate dM/dt should be:
dM/dt ~ 1.7 (0.1/) (L/1046 erg/s) M/yr
If the duty cycle of the AGN is of the order of 108 yr, then a total mass up to 2 108 M should be available. It is a significant fraction of the gas content of a typical galaxy, like the Milky Way! The time-scale to drive such a large mass to the center is likely to be larger than 1 Gyr.
For the mass to infall to the center, it must lose its angular momentum. Could this be due to viscous torques? In a geometrically thin accretion disk, one can consider the gas subsonic viscosity, where the viscous stress is modelled proportional () to the internal pressure, with a factor < 1. This can only gather in 1 Gyr the gas within 4 pc typically (e.g. Shlosman et al 1989, Phinney 1994). This shows that viscous torques will not couple the large-scale galaxy to the nucleus, only the very nuclear regions could play a role through viscous torques.