CAPE CANAVERAL, Fla. – Even as the wounded space shuttle Endeavour brought its seven astronauts safely home Tuesday, NASA is looking ahead to three more launches at risk for the same kind of damage.

There is a striking parallel with the 2003 Columbia disaster in the space agency’s failure to anticipate the harm from breaking ice or insulating foam – this time from a new area of the shuttle’s fuel tank.

The 3«-inch-long gouge in Endeavour’s belly did not put the astronauts at risk. And as soon as the damaged tiles are popped off, engineers will know whether repairs are needed to the underlying aluminum structure. The gash seemed to weather the return flight well, NASA said.

But for the early part of Endeavour’s 13-day mission there was an eerie sense of deja vu.

Back before Columbia flew its last mission four years ago, NASA knew it had a foam problem with its fuel tanks but never imagined a piece of the airy insulation could severely wound a space shuttle.

The result: Columbia shattered during re-entry to Earth’s atmosphere, just five days before engineers were to propose possible repairs.

This time, NASA knew it had a foam problem with brackets on its fuel tanks but never imagined a stray piece would ricochet off the tank and smash into the shuttle.

The result: Endeavour was gouged.

Retired Navy Adm. Harold Gehman Jr., who headed the 2003 Columbia investigation, was reluctant to comment this week on the bracket problem. He said he didn’t have enough information.

But he observed: “You have to assume things are going to happen and you have to mitigate the consequences, that’s what our report was all about.”

Endeavour’s gash, although deep, was too small for scorching atmospheric gases to penetrate and cause serious damage, mission managers said during the flight. It was also on the belly, a more benign area than the nose or wings, which are subjected to much higher heat. The platesize hole that brought down Columbia pierced the left wing.

After checking out Endeavour on the runway Tuesday, officials said there was no apparent charring to the exposed felt fabric, the last barrier before the shuttle’s aluminum frame.

But now NASA finds itself playing catch-up. It’s analyzing a variety of temporary bracket solutions, which may or may not be in place before the next shuttle flight in late October.

Making the brackets with titanium, which would require far less foam insulation than the aluminum version, is the permanent solution ordered after the problem first cropped up last summer. But that won’t happen until next spring.

Engineers are considering a variety of short-term options: shaving some foam from the brackets or possibly applying an oil to the foam to reduce condensation and the buildup of ice.

Because the bracket problem has intensified for the launches since Columbia, engineers theorize it might be due to the one-hour earlier start of fueling – a new rule intended to provide more time for ice checks. That extra hour that the super-cold fuel is in the tank could be allowing more undetected ice to form, which then can cause the neighboring foam to pop off.

NASA Administrator Michael Griffin said after Endeavour’s landing that he will need to be satisfied that any change “is necessary and, in fact, beneficial” before ordering modifications. The last thing the space agency wants to do is to change something and make it worse.

NASA still doesn’t know whether the debris that smacked Endeavour was foam, ice or a combination of both. Whatever it was, it broke off the bracket, fell nearly 25 feet onto a strut lower on the external fuel tank, then shot into Endeavour’s belly.

The four-by-four-inch piece of debris, just a third of an ounce, was traveling more than 200 mph when it hit the strut and 150 mph when it bounced into Endeavour. Engineers were surprised the fragment didn’t shatter when it hit the strut, which holds the tank to the shuttle’s belly. That’s why they suspect ice may have been attached to flying foam.

“We have been looking at this area for some time,” said LeRoy Cain, a launch manager. Even though NASA factored the bracket debris into its risk assessment, Cain said managers didn’t consider the ricochet possibility. That risk assessment is now being re-evaluated.

Shuttle program manager Wayne Hale called it an unanticipated “unlucky bounce.”

“When we returned to flight, we knew that we had not eliminated all the risks,” he said. “In fact, at the end of the day, we will never eliminate all the risk.”

Gehman agreed that launch debris can never be entirely eliminated.

Each fuel tank is covered with 4,000 pounds of foam and each bracket has only ounces of foam on it, NASA’s space operations chief Bill Gerstenmaier said. “We’ve really taken this huge complex problem … and we’ve really shrunk it down to just very, very few areas, essentially ounce-size pieces of foam that we need to go work with.”

A NASA veteran who now chairs the mission management team, John Shannon, is quick to point out that once the problem was discovered, the difference in the way the Columbia and Endeavour flights were handled was “night and day.”