Faced with a nuclear crisis they did not know how to solve, federal officials considered breaking a pipe, blowing out radiation seals or starting a fire inside the Three Mile Island reactor to create an accident they thought they could handle.

Transcripts of the Nuclear Regulatory Commission's closed-door deliberations on the second and third days after the March 28 accident near Harrisburg. Pa., make it clear that the regulators had never contemplated the malfunctions and mistakes that occurred at the Three Mile Island power plant.

"We have an accident that we have never been designed to accommodate," Roger Mattson, NRC director of reactor systems safety, told the commissioners. Accordingly, Mattson suggested, it might make sense to trigger a further accident to "get into a mode for which all these systems were designed and we could cope with."

The idea of provoking new accidents came up several times in the wandering, confused discussions as the five commissioners and their technical experts groped for a way to resolve the danger in the power plant 10 miles south of Harrisburg. The meetings were taped and transcribed; transcripts of the sessions were released this week by a House subcommittee.

One suggestion, proposed by the plant's designers on Friday, March 30, two days after the accident, was described by Mattson:

"Start up all the reactor pumps, burn them out, blow the seals and hope they cause a loss of coolant water accident that way, which would depressurize the system rapidly."

But Mattson, who expressed strong distrust of the competence of the companies the NRC had licensed to build and operate the reactor, said this idea "doesn't hold a lot of promise." It was rejected.

A few hours later, the commissioners were discussing how much simpler things would be if the problems included a ruptured water pipe, the kind of accident that backup systems could handle.

"You know," suggested Chairman Joseph M. Hendrie, "What we need at the moment . . ."

"Is a good pipe break," offered Commissioner Peter Bradford, completing the thought.

Hendrie even had a scenario for how the "good pipe break" might occur. He hoped that one of the heavy electric motors driving reactor control rods "would just fall off the head of the damn vessel and give us a nice 6-inch diameter small-break loss-of-coolant accident."

The commissioners then carried on a rambling discussion of Hendrie's idea, with some indicating it had promise and some suggesting it might exacerbate the major problem of the moment, the hydrogen bubble in the reactor. The bubble was driving out cooling water, raising the risk of an explosion or a melting of the uranium fuel.

The "good pipe break" notion came up again late on Saturday, March 31, when Mattson called from the reactor site to discuss methods of forcing a control rod motor to fall-in effect, making Hendrie's hope come true.

"We think we've got a way we can break the control rod," Mattson said.

"We've got people looking at the way to fail a control rod drive on purpose and provide a crack. Unfortunately, the only way you can do that that we know is to heat it; in other words, you want to start a fire."

Without deciding whether they wanted to start a fire, the commissioners moved to another subject.

The efforts to set up one of the specific problems covered in their emergency plans reflected the general aura of uncertainty that pervades the six days of NRC deliberations covered by the Three Mile Island transcripts.

The transcripts, uncorrected drafts of tape recordings of the commissioners' meetings and many of their telephone conversations over speaker phones, were made public by a House subcommittee which obtained them from the NRC on threat of a subpoena.

The transcripts, which provide an unusually vivid behind-the-scenes picture of how a government agency deals with a crisis, provoked reactions in Washington yesterday that generally ranged from dismay to outrage.

Presidential press secretary Jody Powell, whose repeated consultations with the NRC on public relations aspects of the problem are included in the transcripts, was surprised to learn that he was in the drama.

"You mean my telephone coversations were taped?" Powell said yesterday. "Isn't that against the law? Hendrie didn't tell me."

Powell was anxious to refute any suggestion that NRC members were acting at his urging when they worked to make their news statements on the incident reasssuring to the public.

"I'm not denying we were very much involved, I said that at the time," Powell said. "But my concern was not to be either reassuring or not reassuring, but to be accurate and understandable."

Of the various NRC officials involved only Commissioner Victor Gilinsky would discuss the transcripts yesterday. He said the situation involved, "an operational mode that went way beyond what we could deal with, or what I could deal with at least."

Gilinsky said, "I think our people did very well up there . . . but we really have to go over this record, really pore over it."

One expression of outrage came from consumer advocate Ralph Nader. He said in an interview that the transcripts displayed "negligence, recklessness, and ill concern" for public safety on the part of NRC Chairman Hendrie. He demanded that Hendrie resign.

"For political reasons the mass evacuation that that should have been carried out was not because it would have shown 150 million people watching on TV a picture of half a million people fleeing from a potential disaster," Nader said. "That picture would have terminated the nuclear industry right there and then."

Others expressed concern over the NRC's preoccupation with unfavorable publicity. "They were far more interested in public relations than in public safety," said Rep. Bob Carr (D-Mich.), a member of two House nuclear engergy subcommittees.

"It's clear the decision not to evacuate people was based on political considerations to protect the nuclear power industry and the president's image and not the public," said Dr. Ernest Sternglass, a University of Pittsburgh physician, who was criticized in the transcripts.

"It's clear to anyone who has followed the situation at Three Mile Island that the NRC, as Chairman Hendrie said, was flying blind," said David Masselli, energy policy director for the Friends of the Earth, an environmental group. "What is new from the transcripts is the extent to which the NRC and the White House was engaged in news manipulation."

The rejection of recommendations to evacuate the Middletown area by the two top federal officers on the scene, Masselli charged, "again shows there are very strong pronuclear elements in the administration who see this as nothing more than a public relations problem and want to minimize the dangers."

With much of Congress away for the Easter recess, there was only a ripple of reaction to the transcripts from Capitol Mill.

Sen. Gary Hart (D-Colo.), chairman of the Senate subcommittee on nuclear regulation, called the transcripts "both distressing and alarming." They showed, he said, a far greater degree of concern and uncertainly over plant dangers than the NRC had indicated to senators in a series of private briefings after the March 28 accident.

Hart and others said that the behind-the-scenes look the transcripts provided raises serious doubts about NRC's ability to regulate the nuclear power industry.

The commission was "totally unprepared for the kind of crisis which occured," he said. "It was unable to verify the accuracy of information it was receiving from the site and to sort out conflicting information from the utility and state officials."

"All of this indicates a very serious gap in the NRC's capacity in crisis management situations," Hart added.

"The behavior of the operator at Three Mile Island demonstrated an astonishing lack of competence," said Rep. Toby Moffett (D-Conn.) "The accident presents a strong case for federal intervention in the event of another major accident."

The only moderately kind word for the NRC from Capitol Hill came from a freshman Republican, Sen. Alan K. Simpson (Wyo.), who said, "I admire individual members of the NRC and especially Chairman Hendrie."

Simpson, a member of Hart's sub-committee, however, said he was "startled and disappointed" that the commission didn't have more accurate information to work with-a problem he said Congress and the press shared.

The NRC issued one correction yesterday of the transcripts. Where staff official Harold Denton is quoted as saying "the important thing . . .is to get a start rather than sitting here waiting to die," the agency said Denton said "decide" not "die."

In addition to its discussions about causing new accidents, the NRC spent much of the second and third days after the accident anguishing over whether to call general evacuation around the site.

The officals lacked two key pieces of information needed to decide the question: how serious the danger at the plant and the geographic and demographic make-up of the surrouding area.

At one point, an NRC official reported that a complete evacuation of a five-mile radius from the plant could be completed in one hour. Gilinsky replied, "I guess I find that hard to believe. I guess I just don't think these sorts of things work out."

The commissioners then got to talking about the main roads near the site and the proximity of Harrisburg. Unable to answer either question for certain, Hendrie said. "I wish we had a map here someplace," a remark that was followed by laughter, according to the transcript.

Another moment the commissioners apparently found humorous came when Hendrie was discussing how much radioactive iodine had escaped into the air above the plant.

"Well, when you breach your containment," Hendrie says, "I don't know, we've probably got 15, 20 percent of the core iodine out there which is-you know, that will do for starters.

The transcript then reads: "(Laughter)." CAPTION: Picture, VICTOR GILINSKY