The author of the Rasmussen Report, which was until two weeks ago the definitive work on the safety of nuclear power, remains convinced that his findings are accurate and that nuclear power is the safest there is.

Norman C. Rasmussen, head of the nuclear engineering department at Massachusetts Institute of Technology, led the blue-ribbon panel of scientists that produced WASH-1400, the Reactor Safety Study, in October 1975. Its complex calculations led to many an industry assurance that death in a nuclear addident was as likely as being hit on the head by a falling meteorite.

But the Nuclear Regulatory Commission, which oversees reactor safety and licensing, repudiated the report in part on Jan. 19. Rasmussen is a little bitter over the way he has been treated since.

"An hour before they made (the repudiation) public, the public relations office called and read me the press release," Rasmussen said. "I guess I don't want to be quoted on what I said back."

At 51, Rasmussen is the iron-jawed, pipe-chomping stereotype of nuclear power's defenders. A veteran of countless advisory commissions, panels, study groups and research teams, he already is preparing graphs and charts to defend his basic findings later this month at a congressional hearing.

"The report has been misued by supporters as well as opponents of nuclear power," he said. "It's not right to say there are no risks, but the risks are low compared to other energy producers. I absolutely defend that statement."

The NRC's action on Rasmussen's report was triggered by a reassessment prepared at the NRC's request by Harold W. Lewis, a physicist at the University of California at Santa Barbara, and six other nuclear specialists.

The Lewis group lavishly praised Rasmussen's methods of analyzing the dangers involved in running a nuclear power plant, but it criticized some of his mathematics. He made some incorrect assumptions, occasionally used inconsistent or skimpy data, and made some calculations with a "degree of arbitrariness [that] boggles the mind," the Lewis group said.

The net result was that although much of the Rasmussen report was very useful, it "understand" the width of the range of probability for a major nuclear accident, such as the melting of a reactor's radioactive core.

"We are unable to define whether the overall probability of a core melt given in WASH-1400 is high or low, but we are certain that the error bands are understated. We cannot say by how much," the Lewis analysis said.

Seeing that, the Nuclear Regulatory Commission ordered its staff to stop using Rasmussen's numbers as the sole basis for licensing or regulatory decisions, to use them only with Lewis' criticisms in mind, and to review their use in past decisions to see if any actions require change.

But the press release announcing these change, Rasmussen complained, said boldly that the NRC "does not regard as reliable" Rasmussen's overall risk evaluation, and that repudiation was what got picked up by the media and by anti-nuclear critics. "The word 'unreliable' was never used either by Lewis or the NRC," Rasmussen said, "only in that press release."

In the interview, Rasmussen insisted his risk assessment probably was not off by much, and in another interview, Lewis generally agreed with him.

Rasmussen's report said that one could assume a core meltdown once in every 20,000 reactor-years, give or take 16,000 reactor-years either way.That would mean a core meltdown would be likely after about 4,000 years of reactor operation. Since the United States already has accumulated 659 reactor-years, any large increase in the range of probability would bring the numbers into the near future.

"In other works, if the probabilities were much more, we should have had one by now," Lewis said. "Further, if we were on the verge of a core melt there would have been a history of near-misses, and there hasn't been even one."

Most experts, he said, including those critical of nuclear power, now regard a serious fire at the Brown's Ferry, Ala., reactor in 1975 as "a major accident" worthy of minute study but not as "a near miss." This, Lewis said, is because there were very few elements in that incident that were random in the sense that they could have happened otherwise so as to lead to a meltdown.

In any case, Rasmussen said, he and others had made recalculations in ways more consistent with Lewis' critique and had not found any major differences. He acknowledged several of Lewis' other points.

The Rasmussen report's executive summary, which substituted for the foot-thick full report in most public use, was "not as good as it could have been," Rasmussen said. Lewis' group said the summary "masqueraded as a summary of the report" and had been "badly misused."

Lewis criticized Rasmussen's handling of the risks from earthquakes, tornadoes and other unpredictable disasters, and his evaluation of the health dangers expectable over the long run from a core meltdown. Rasmussen acknowledged that those figures might have been improved, but argued that nothing significant would change if they were.

Some of Lewis' critique suggested that nuclear accidents might be less probable than Rasmussen had figured, rather than more. Rasmussen neglected to factor in human creative problem-solving ability, Lewis noted, but only figured in human stupidity.

For example, one scenario Rasmussen had examined showed a reactor overheating for 29 hours to meltdown while dozens of reactor personnel did nothing, "presumably sitting around playing bridge," Lewis said.

Lewis reserved his most biting criticism for the "inscrutability" of the Rasmussen study. "Anyone who has tried to work with the thing... comes away drinking heavily," he told the Nuclear Regulatory Commission last fall.

At this, Rasmussen shrugged. "Any study involving 60 scientists is going to be hard to read," he said. CAPTION: Picture, "It's not right to say there are no risks, but the risks are low compared to other energy producers. I absolutely defend that statement."

Norman C. Rasmussen