PSS/E power flow blows up, more often, because of a large phase angle difference between two buses which is seen to be a large power flow by the program.
Example, you added a new bus and connected it to an existing bus. The existing bus would have a phase angle calculated by the program but the new bus would have a default of zero phase angle, if you did not change it. Thus when power flow is called, the program sees this as a large mismatch and can't converge the solution.
So, when adding new buses, see to it that you change the phase angle equal to the phase angle of the bus where you connect it.
Also, the program will provide hints where the solution when berserk thus its good to look at the progress window.
In the US, different Transmission Owners would use different solution parameters (PAR adjustment, transformer tap/shunt compensation adjustment, interchange power flow control, FNSL or FDNS) thus when the power flow exchanged hands, it is a must that you know what are the solution options if not, the solution blows up when the right options are not applied.
I'd agree with @waltterval that there are lots of reasons other than incorrect data