The Nuclear Regulatory Commission yesterday repudiated the findings of the government's most widely accepted report on the safety of nuclear reactors, a study concluding that there is essentially no chance of an atomic power plant disaster.

After a second review of the 1975 Rasmussen report, as the study is known, the NRC concluded that it was misleading and unreliable. "The commission does not regard as reliable the Reactor Safety Study's numerical estimate of the overall risk of reactor accident," the five-member commission said.

The NRC also called for a staff analysis of its nuclear licensing decisions to determine whether they were affected by an overreliance on the Rasmussen report, and said it would decide "where regulatory modifications are appropriate."

The NRC oversees nuclear licensing and safety.

Commissioned by the NRC's predecessor, the Atomic Energy Commission, the Rasmussen report concluded that the probability of an individual's being killed in a nuclear power plant disaster was about the same as his being killed by a meteorite. It has been used by the nuclear power industry and the government as an assurance that nuclear power is safe.

Yesterday's NRC endorsement of a review headed by Dr. Harold Lewis of the University of California is expected to broaden the national debate in some quarters over the safety of nuclear power.

In a carefully worded statement, the NRC came to three major conclusions:

It withdrew its endorsement of the Rasmussen study's executive summary, the widely circulated conclusion that nuclear power plants were not likely to have major accidents.

It said that the Rasmussen report's peer review process in the scientific community had been "inadequate" and said that the NRC would take corrective steps in future analysis of the probability of nuclear accidents.

It agreed with the findings of the review headed by Lewis that the probabilities of disaster should not be used "uncritically." Lewis' study of the Rasmussen report found that, if anything, the estimates of the probability of an accident were "understated."

One senior NRC official, summarizing the commission's decision on the 1975 study directed by Dr. Norman Rasmussen, a Massachusetts Institute of Technology professor, said: "It means there isn't any estimate of nuclear reactor safety; the bottom line from the Rasmussen report has evaporated."

Rasmussen's findings, officially known as "WASH-1400," were reconsidered by the NRC in response to a 1977 letter from Rep. Morris Udall (D-Ariz.), chairman of the House Interior Committee. Yesterday Udall applauded the NRC's acceptance of Lewis' review, saying, "Nuclear proponents have for years used the [Rasmussen] study to assure the public nuclear power is safe. The commission has now made clear that the report is useless for this purpose."

The Lewis study, however, described the Rasmussen report as "a conscientious and honest effort" and said that its methodology -- based on the analysis of statistical probabilities -- was "sound." Lewis' seven-member review group found, however, that the Rasmussen report's executive summary had been used improperly as a promotional document, not fully reflecting its findings.

A spokesman for the Atomic Industrial Forum, Carl Goldstein, said yesterday that in the view of the nuclear power industry the NRC decision "is the first of two shoes dropping." Future studies of reactor safety, Goldstein said, "Can go up or down, but the chance is so horrendously small that there will be an accident in real life, it doesn't make a difference."