Fuel "excessive aeration" theory might be explained, in part, by a significant difference in more recent Boeings (post TWA800). Recent Boeings use centre tank fuel scavenge jet pumps with motive power provided by main tank fuel from the main tank booster pumps. Moreover in 777s such jet pumps remain on, once main tanks reduced to half full, for remainder of flight; whereas in 737s (other than more recent NGs) they are powered on for just 20 minutes (and link only to main tank No1) before a shut-off valve removes motive power.
Accordingly, here is another slightly different, ridiculously remote Swiss cheese line-up:
1. Jet A-1 in main tanks was saturated with air, either when uplifted and/or as result of prior centre fuel scavenge tank jet pump operation on fuel remaining from previous sectors. (Kerosene has much greater propensity to become saturated with air and retain it even at altitude compared to say water.)
2. Centre fuel scavenge tank jet pumps operated to entrain and dissolve air into fuel in each main tank thereby supersaturating fuel in each main tank during last hour of flight (by then last of centre tank fuel has been scavenged). Increasing pressure on descent helped increase the level of saturation.
3. LP main tank booster pumps were unaffected by supersaturated fuel (i.e. no cavitation) and no LOW PRESS alert as a result.
4. Cavitation induced in HP pump as flow increased by order of magnitude on finals.
5. Relatively cold fuel made cavitation much more pronounced (see
http://naca.central.cranfield.ac.uk/...rc/cp/1128.pdf) and, once started, cavitation worsened preventing increase in flow rate as more and more air was released into system. Some flow was maintained, just insufficient.
6. Suction feed as alternate to HP pumps was ineffective and/or became ineffective to increase fuel flow to level demanded. Perhaps due to:
(a) inherent weakness of suction feed (for inadequacy of suction feed in certain flight conditions, see article on UAL flight 767 which in climb at altitude suffered rollback when boost pumps turned off and suction feed alone turned out to be insufficient -
http://findarticles.com/p/articles/m..._73578123/pg_4).
(b) secondary effect of release of air at HP pumps as a result of cavitation impeding operation of suction feed bypass.
(c) reduced flow at suction feed inlet from main tanks due to (i) FOD blockage on right hand side at least and/or (ii) perhaps (if sufficient newly melted water, derived from ice formed in surge tanks or centre tank, could have entered and frozen in main tanks) ice particles.