Is has to do with the relative diameters of the seat-to-ball sealing surface and seat ring-to-body outer seal (usually O-ring), what is called the "piston effect" of the seats.
For a trunnion ball valve, with the ball closed, the force that presses the seat sealing surface against the ball is generated by the annular area of the seat ring (the distance outside the bore). With the ball closed, you should be able to open the body vents because the pressure is sealed off.
In the event of leakage past the upstream seat, you will get pressure in the body cavity. For double piston seats, the sealing surface diameter is larger than the O-ring diameter, so pressure in the body cavity goes around and pushes the downstream seat back against the ball and you have a secondary seal. Single piston seats have the O-ring larger than the sealing surface and thus cavity pressure will push the seat off the ball and you will get leakage downstream of the ball.
Double piston seats sound like the way to go, right? Well, they require body pressure vents/equalizers to release pressure inside the body whereas single piston seats vent excess body pressure back into the flow when the valve is open. Single piston seats also allow "double block & bleed" for checking the seals and maintenance. Also single piston seats generate much larger sealing force for a given pipe size and valve geometry, so require less material (money) to make.
There may be some other things that I've never had explained to me.
TR1NTX has it about right. The big thig is that 'standard' seats in a trunnion mounted valve will relieve the cavity pressure to the downstream side. With Double Piston Effect (DPE) seats the pressure in the cavity increases the sealing force of both seats onto the ball. This is a double edged sword as the downside is that you need a secondary device attached to the body cavity to prevent over pressurisation (especially if you experience a change of state from liquid to gas) within the body cavity.
You can, of course, mix the seats: have the upstream seat a 'standard' and the downstream seat as DPE. This means that once the cavity pressure exceeds the UPSTREAM pressure PLUS the spring force of the seat spring the cavity pressure would relieve UPSTREAM. It does mean that the valve will only work as intended in one direction and you need to be confident that the upstream pressure will never be full class rated if a secondary pressure relief device is not used - I am not sure what standards, if any, address this installation.
All the explanation are right and complete.
Just want to add that in both configuration, single and double piston, the valve (trunnion) is always double block and bleed.
API 6D rules the construction of these valves.
Looking at valves for a pig trap and thinking DPE (one side only)for pig trap isolation valve giving double block isolation. Thinking similar for pig trap bypass valve but maybe DPE both sides as pressure coming from both sides - any ideas?
