Already three days late, the space shuttle Discovery and its five-man crew today raced away from Earth just ahead of a torrential rainstorm, then quickly ran into a shower of problems in space that has caused a reordering of the mission's schedule.

Plans called for launching one satellite each of the next three days. However, a schedule change led to deployment of two of the satellites today.

But Australia's first communications satellite almost didn't make it out of Discovery's cargo bay today, after a piece of the satellite and the shuttle's robot arm apparently collided.

When the satellite, called "Aussat," was finally released, "We just breathed a sigh of relief down here," astronaut Robert C. Springer told the crew from the Mission Control Center in Houston. "Outstanding," replied Discovery commander Joe H. Engle when Springer told him the satellite's on-board motor was driving it to a permanent position east of Australia over the equator. By evening the crew was forced by the earlier accident and rescheduling to launch a second satellite within hours of leaving Florida.

The trouble began when the time came to test the Aussat's sun shield as it sat in the cargo bay. The crew opened it, then saw it give a strange jerk when it was ordered closed. The shield had jammed, half-open.

A half-open sun shield meant the satellite could not be launched and also meant the satellite's electronics would overheat and leave the craft useless.

Flight directors gave the satellite no more than seven hours of life when they decided to pry open the shield with the shuttle's 50-foot robot arm, to try to deploy the satellite right away instead of waiting for Wednesday.

Using the arm from inside the shuttle cabin, astronaut John M. (Mike) Lounge pushed it against the clam-shaped screen to force it open.

If the arm had failed, the next move would have been a spacewalk by astronauts William F. Fisher and James D. van Hoften to try to pry the shield open with their hands, a last-ditch attempt that would have cost two days and could have threatened a daring plan late in the mission by Fisher and van Hoften to cling to the shuttle, capture a passing Navy satellite dead in orbit since last April, manually stop its rotation, then fix it.

Ironically, the robot arm used to pry open the sun shield may have caused it to jam.

Flight director Gary Coen said: "We think the camera on the arm's elbow may have collided with the sun shield. The arm and the satellite are both on the port side of the cargo bay, with less than six inches of clearance normally between them. We don't know why but there may have been an impact."

Less than five hours after the Australian satellite was fixed and launched, Discovery's crew deployed a satellite for the American Satellite Co., a company half-owned by Fairchild Industries of Chantilly, Va.

The crew had planned to deploy this satellite Wednesday, and the unplanned deployment of the Australian craft marked the first time a shuttle crew deployed two satellites in one day.

"Good work," astronaut Springer said from Houston. "A new world's record for satellite deploys in one day."

The day began for Discovery and her crew on a note that could only be described as bleak. Less than an hour before liftoff, the skies around Cape Canaveral were black with clouds. A tropical depression that had moved in during the night from Cuba had sucked up energy from the Gulf Stream and was dumping more rain on central Florida than meteorologists had forecast.

"We could see large holes in the system 50 or 100 miles across so we prayed for the breaks that would let us thread the needle when the time came," launch director Robert Sieck said later at the Kennedy Space Center. "We had two things in our favor: There was no lightning in the cloud cover, and most of the rain was south of the launch pad, not right over it." Launch rules prohibit a liftoff into a rainstorm and into cloud cover containing lightning. Rain can strike the spaceliner like buckshot as the shuttle gathers speed, damaging its protective tiles so badly they may not stand up to the heat of reentry when the shuttle returns to Earth. Lightning can short-circuit the shuttle's guidance and navigation machinery just when it needs it the most during assent.

Launch directors gambled that rain would not be falling on the pad at liftoff time, which was delayed three minutes to make sure the gamble was the right one. Even as a heavy rain fell on the press site three miles to the west, Discovery roared away from the pad through a hole in the clouds and sped into the air 10 minutes before the pad was pelted with rain.

"That cloud is black!" Engle yelled as the craft rose through a heavy cloud just over the launch pad.

"We may have pushed the rules to the limit but we did not violate those rules," Sieck said. "Technically, we were within our limits."

While the rules may not have been broken at liftoff, there was some question that rules on an abort may have been stretched. A steady rain fell on the shuttle landing strip four miles from the launch pad before, during and after liftoff. Abort rules say that rain cannot impede an attempt by the crew to land their spaceliner back at Kennedy if they must abort their flight.

"I believe that technically we were within our limits because visibility was good," Sieck said. "Realistically, that part of it was marginal."