CAPE CANAVERAL, Fla. - Boeing's Delta 4 rockets are being retrofitted with new pressure valves to alleviate bubbling in liquid oxygen fuel lines that possibly occurred in the fleet's maiden liftoff in 2002 and investigators determined caused engine trouble during the first Heavy booster's December test flight.

The so-called cavitation is a phenomenon when super-cold cryogenic liquid oxygen changes to vapor bubbles within a rocket's feedlines running from fuel tanks to engines.

Rocket designers never caught the potential for such a problem in the Delta 4's first stage until after the Heavy's demonstration launch that suffered early extinctions of its three main engines during ascent. The shortened firings resulted in the rocket failing to achieve the proper altitude to deploy a pair of student-built nanosat experiments and a massive satellite mockup.

The inquiry into the incident formally concluded with corrective actions to fix both the Delta 4-Medium and -Heavy rocket versions to prevent a repeat on future launches.

Government and industry investigators said the bubbling started at the entrance of the liquid oxygen feedline where a filtration screen, the line's elbow bend and an internal gimbal strut altered the fuel flow. This changed the fluid's speed and decreased pressure as the liquid oxygen streamed from the tank.

The Delta 4-Heavy features three Common Booster Core stages mounted together, each identical with a liquid oxygen tank on top, the large liquid hydrogen fuel tank accounting for three-quarters of the stage's length and a Rocketdyne RS-68 engine at the bottom.

All three CBCs on the Heavy demo flight saw cavitation. The bubbles made their way five feet downstream from the liquid oxygen tanks to internal sensors used to tell the engines when the fuel supply is expended and command the shutoff sequence. The bubbles fooled the sensors into thinking the tanks were going dry and triggered the engine cutoff despite plenty of liquid oxygen still aboard.

"This feedline design has been present in all previous Delta 4 flights, but the unique combination of vehicle acceleration, liquid level in the tank and propellant flow rate for the Heavy mission reduced the fluid pressure enough to enable the creation of gaseous oxygen at this location as the tanks emptied," investigators determined.

"Further draining of the liquid oxygen tank worsened the conditions at the feedline inlet, causing the cavitation effect to extend further down the feedline. A pocket of gaseous oxygen continued to enlarge until it reached the Engine Cut-Off (ECO) sensors and caused the ECO sensors to momentarily indicate dry. This ECO sensor dry indication was sensed by the flight computer, which initiated the sequence to throttle-down and shut off the main engines as it is programmed to do.

"In reality, flight data showed that sufficient propellant remained in the tank to complete the planned first stage burn time."

Despite extensive analytical work to understand the rocket's performance before the Heavy flight and even the first Delta 4-Medium in 2002, engineers did not look at the engine cutoff sensor area for possible cavitation.

"Based on a lot of previous experience we look very hard at cavitation at the inlet to the RS-68 engine, as that is the traditional place where cavitation might cause a problem. We did not look at this issue up at the top of the feedline where the ECO sensor is," Dan Collins, Boeing's Vice President of Expendable Launch Systems, said in a news conference Friday. The media briefing followed a remarkably open and forthcoming flow of information throughout the investigation that began immediately after the December 21 launch.

"We are as much at fault as Boeing in terms of not looking at this. Cavitation at the engines is a routine phenomenon. For whatever reason we did not think to look 100 feet up the LOX feedline," said Ken Holden, general manager of the EELV division at the federally-funded engineering support firm Aerospace Corp.

"The nuances of this are such that it would not occur on each and every mission. It took a particular set of circumstances to bring this together."

As part of the Heavy's investigation, engineers reviewed data from the earlier three Medium missions to see if bubbling occurred on those launches. The inaugural flight possibly experienced cavitation, however it did not trigger an engine shutdown.

"We have gone back and looked very hard at this phenomenon. There is a possibility that it may have occurred on one of the earlier flights," Collins said.

"We did not have the special instrumentation that was present on the Heavy demo, though, to know that for sure. There was absolutely no premature (engine cutoff), no signs of this at all. And when I say cavitation, it is possible that a very small vapor bubble well away from the engine cutoff sensors may have developed. There was absolutely no interaction with the cavitation and the engine cutoff sensors."

Hardware and computer software changes are being ordered to increase the pressure in the liquid oxygen tank to counteract the pressure losses in the upper portion of the feedline, thereby removing the potential for bubbles forming. The earlier Delta 4 flights permitted the pressure to drop during the rocket's ascent.

To raise the tank pressure later in the launch phase, the currently pressure relief valve will be replaced with one having a higher pressure. Also, flight software will be modified to provide commands needed to increase the tank pressure later in the launch.

Other alterations to onboard software includes changes to guard against the engine cutoff sensors being tricked. The time in which the rocket's computer brain will begin accepting "dry" signals from the fuel lines will be moved later, the Air Force said.

"It's really a one-time fix that we will then carry through all vehicles going forward. We will have those completed over the next several months," Collins said.

The next Delta 4 mission, using a Medium vehicle, does not require the fixes because its flight profile is not deemed susceptible to cavitation. That May launch will carry the next civilian geostationary weather satellite into orbit for the U.S. government from Cape Canaveral.

The first rocket to receive the changes is stacked on the Space Launch Complex-6 pad at Vandenberg Air Force Base in California for a planned August 30 liftoff with a classified national security satellite payload.

Schedules call for the next Heavy mission to occur on October 28 to deploy a missile-launch detection satellite, called DSP-23, from Cape Canaveral. Collins said the fixes won't delay the launch plans.

"All in all, the Heavy demo mission was successful in providing a wealth of vehicle design, environment and performance validation data. We are now taking all our knowledge and understanding from this Demo mission and applying it as we prepare for our operational missions," Collins said.

"Straightforward fixes have been identified that will allow us to safely fly the Delta 4 fleet without risk of repeating the cavitation that led to premature engine shutdown," said Col. John Insprucker, Evolved Expendable Launch Vehicle system program director.

"The team findings do not detract from the successes we experienced on the Delta 4 mission. We successfully checked out and launched a Heavy rocket from the Boeing launch pad that also saw its first Heavy flight. We successfully flew three liquid-propelled booster cores side-by-side and cleanly separated the strap ons. We successfully flew the upper stage through the long three-burn profile required of a geosynchronous mission. We successfully separated the Demosat satellite.

"The ability of the Delta 4 to fly an end-to-end mission, despite the premature engine shutdown, made this an exemplary test flight," Insprucker said.