I know that Goober, karajorma, or possibly someone else ran a rather comprehensive Colossus-vs-Sathanas scenario a while back, starting each of the ships in various positions. Even in the setup that was least favorable to the Colossus, it still decimated the Lucifer while suffering only relatively minor damage.
As for the issue of shield penetration itself, I've always theorized that it wasn't only the amount of power applied, but how it was distributed. Remember, a Harbinger most likely deals damage to a relatively wide area of a target's hull; the fact that one bomb can take out multiple closely-spaced subsystems is proof enough of that. In contrast, look at a beam cannon; it's delivering a comparable amount of power to a relatively small area of hull. Perhaps it's this characteristic that allows beam weaponry to pierce right through shielding, while ordinary matter weapons are distributed across them. There could also be some electromagnetic or other physical characteristic of shielding in general (and the Lucifer's in particular) that renders it vulnerable to however FS2's beam cannons actually work.
The bottom line is, I just can't see the GTVA sinking 20 years and God knows how much money into such a massive project as the Colossus without being absolutely certain that it would be able to more than handle itself in a fight against the Lucifer. Between Alpha 1's scans in "Playing Judas," the information obtained from the Ancient remains in Altair, and whatever data was confiscated from the GTI at the end of Silent Threat, I'm fairly certain that Command had more than enough data available on the Lucifer's shields to ensure the Colossus's success against them.
