Pumping hydrocarbons (aka fuel)
HP pumps, be they gear (positive displacement / lobe) or centrifical, behave very strangley under many conditions that are just marginally outside of the nominal state. Any one variable can have significant impact to pumping performance. (Cavitation pitting is the clue!!!)
Variables include:
Turbulant vs laminar flow at LP side (Reynolds Number).
S.G. variation at temp deltas (pumps are pre set for nominals NOT all temps).
Clearances and gaps between rotator and stator faces are critical, too big equals deminished performance, too tight equals tip gassing (cavitation).
Contamination (even at < 25 micron level) plays havoc with the flow transition state across the face of the impeller. (Again cavitation).
Cavitation pitting is NOT an instant catastrophic occurance it is a chronic progressive development.
I am sure that many engine HP pumps exhibit similar pitting but I believe that we have a concoction of variables that culminate in an event of the HP pumps creating an air (gas) saturated "fluid" on the HP side that is inadequate to provide the combustion energy required for thrust.
Please, if the AAIB lurk here, look outside of aviation for similar examples of HP hydocarbon pumps behaving very strangly outside of the norm.