Ground connection in simulating three phase PFC

Hi everyone,

I’m simulating a three‑phase PFC in ORCAD PSPice. The real circuit also includes an input LC filter, which I’ve omitted here for simplicity, but the behavior I’m seeing is always the same. In my current setup, the Pspice ground reference is connected to the negative node of the rectifier bridge, while the three AC sources have their star point left floating. During simulation, I observe that it is the star point of the generators that “moves” with respect to ground, while the negative node of the bridge remains fixed because it is tied to the reference. Which is of course normal for my simulation. 

I’m wondering whether this behavior is correct. Can this simply be considered a consequence of how the simulator assigns the ground reference, or am I introducing a significant error compared to the real‑world behavior of the circuit? In other words, is it acceptable that the generators’ star point moves while the rectifier negative stays fixed, or could this lead to unrealistic results?

Another thing I don’t fully understand: if I try the opposite approach—connecting the ground to the star point of the generator and leaving the bridge negative floating (using a very large resistor like 5 MΩ instead of the direct connection)—the simulation immediately diverges, leaving me with no results. Is there a way to stabilize the simulation in this configuration, or is it inherently problematic?

My main question is: how can I simulate a circuit in which I connect the ground to the star point of the generator and leave the bridge negative floating? 

My simulation profile is: time transient form 0 to 180ms, starting to save data from 160ms. this time to allow the PFC to stabilyze.

Thanks in advance to anyone who can help clarify this.