I'm finally at a computer and can type instead of struggle with the dang phone...
it could be the HPFP harness... most often, though, the harness shorting pops a fuse- no?
the part about your issue that is confounding is that it does it around the same interval- or around two hours into run time.. that, imHo, helps you track it down to either an electrical issue (heat and expansion on a feeble area of wiring NOT specifically fuse protected), or a flow issue in the pump or lines.. if it were mechanical, it should be present no matter how long the engine has been running.. that is just my opinion, though, and maybe misguided..
going with the flow issue, there are a bunch of places where it could happen..
regardless of the code given (P0194), I have a hard time believing the HPFP works and then all of the sudden decides NOT to.. that thing is either good to go, or it isn't..
that leaves the HMFC and the fuel cooler, the lines between, and the source.. and it simply sounds like a vacuum issue..
but, then there is this, too:
the FRP/ICP sensor is the only sensor that figures fuel pressure- it can't be triangulated with other sensors.. it should have some sort of redundancy, but I reckon that ain't doing us no good to know right now.. what happens with that thing is this:
when the key turns, in less than half a second a Built In Test (BIT) is ran, and all key sensors are given a diagnostic check.. either they check out, or they don't.. if they don't, a CEL is ignited..
question- if you clear the code, does it come back on before you start the engine?
the way the FRP works is with swinging voltages.. IIRC, it is looking for a 5vdc swing to determine FP.. it's pretty sensitive to voltage, as these things can range from 20kpsi to nearing 30kpsi during normal operations..
if the FRP fails, and I know this from experience so it isn't just conjecture: the PCM resorts to a different operating logic.. since it can't determine what the fuel rail pressure is, it can't determine how to inject in terms of duty cycle (how long injector is open).. it will still run, but the PCM limits how much power it applies to the drivetrain, and it eliminates functions such as cruise control.. it does this, by my own figuring, by operating on a pre defined script that calculates between throttle position and response.. if you start out slow, it will drive okay- if you apply a lot of throttle, it will buck-n-fight you.. all of this is if it fails outright.. usually indicated in a P0192/93 code..
there is another area on these rigs where chaffing of harnesses is a problem.. it is NOT fuse protected, either- or, at least, not directly
fuse protected.. it is the harness leading to the FRP sensor itself..
that bundle is under the passenger valve cover, and is evident as it connects to the FRP.. it bundles with some other wires while still under the vc's.. it comes out of the front of the head in a round bundle before rejoining yet another bundle.. point being: if you remove your vc's and see an evident chafe, you've likely found your huckleberry.. if you don't see it, and replace the sensor and shorter harness and the issue still persist, it could be in the harness exterior to the heads, which exits the heads, and twists back toward the firewall between and below the battery tray and the air box.. my guess is that it is the shorter harness leading to the FRP sensor, though..
my figuring, fwiw, is either a delivery issue that is between your tank and your HPFP, or a short on the FRP sensor harness under the valve covers.. If you go buy a new FRP sensor, make sure and get a new harness, too..