Ball valves fail because the seat "extrudes" when we close the valve on 150# saturated steam service.  When closing the valve, the ball lip graps part of the seat and deforms seat, not letting valve close completly.
Worcester has a technical paper TP12D "Failure analysis of Ball Valves".  Your local distributor should have it on a CD-ROM and he can zap you a copy. It has photos of various seat failure modes and it pretty useful.
Basically if the seat is extruding your seat material becomes too soft for the pressure at the working temperature.  If your operators are throttling with the ball valve in a partially open condition the assymetrically-supported seat and local high velocity will aggravate the problem.  Ball valves for throttling are not the same as off-the-shelf commodity ball valves. 
Just guessing, but you are probably using virgin PTFE seats.  Loaded or glass-filled seats will be better, or you could go to PEEK or even metal seats. Some tradenames for reinforced PTFE seats are Polyfill and High-Per-Fill. The  Metal seats should not be needed for 150 psi saturated steam, but it's an option. 
You might also want to look at whether the line upstream of the valve is appropriately steam-trapped, and if the trap is functioning.  If the valve is getting liquid condensate, the liquid could be flashing in the valve and causing 2-phase flow with velocity as high as sonic.  That's not good for the valve, either, and can cause additional damage besides just sucking the seats out. 

MORE NEWS