A puff of radiation twice as strong as anything previously revealed came out of the Three Mile Island nuclear power plant stacks during the March accident, but a report on it was lost within the Nuclear Regulatory Commission, the commissioners were told yesterday.

The revelation stunned some high-level NRC decision makers, who said privately that it was outrageous to learn only now of such an event.

Victor Stello, director of inspection and enforcement at NRC, noted the emission in a two-inch-thick report on the Pennsylvania incident that is sure to add fuel to the ongoing controversy over nuclear power.

On the one hand, the report said the incident was "clearly preventable" if only plant design, equipment and procedures "had been permitted to function or be carried out as planned."

Stello said he is considering charging the owners and operators of TMI with 35 violations of federal procedures.

On the other hand, the report blamed a "mindset" of the operators, apparently established by their training in those very same procedures, for what it said was their failure to control the situation better.

The radiation emission, Stello maintained, was dispersed by the wind so that there was no danger to the public.

It was recorded by helicopter at 3,000 millirems per hour at 2:10 p.m. March 29, the day after the accident began, at a point 15 feet above the plant vent stacks, Stello's report said. This was more than twice the 1,200 millirem-per-hour level noted at 8:01 a.m. March 30 at a spot 130 feet above the Unit 2 reactor building.

The 1,200 milleren level, the highest then made public, caused Gov. Richard Thornburgh to suggest that children and pregnant women leave the Three Mile Island area.

The NBC's Region 1 headquaeters in King of Prussian, Pa., was notified of the 3,000-millirem reading shortly after it happened, according to Michael Slobodien, a radiation specialist in that office.

He said in an interview that the information was telephoned to the NRC's headquarters in Bethesda, but that so much was going on at that moment that any of a dozen people might have taken the report. It was never heard again.

Stello said that didn't surprise him because communications between Harrisburg and Bethesda were terrible that day. "The exposure received by the public was really not much," he said.

"The two [radiation ] readings are really not comparable" because of the varying distances from the radiation source, he continued.

Stello's report also noted that operators first learned of temperatures above 2,000 degrees Fahrenheit inside the reactor core-high enough to damage the fuel cladding between 8 and 9 a.m. the first day, but did not report them to the NRC. "They just didn't believe the readings were real," he said.

Even if they had believed the readings, they had no standard procedures to follow for dealing with high temperatures, Stello said. Similarly, the NCR was not notified immediately about a high-pressure pulse now thought to have been a small hydrogen explosion or burn in the containment building.

Asked whether these reporting delays violated NRC rules, Stello said the regulations did not cover the reporting of details within a larger event.

The four-month investigation, Stello said, found no indication of sabotage at any stage of the incident.It is still not clear why two auxiliary water pumps were blocked off before the accident, he said.

The study also indicated that no iodine was released from the building, contrary to early reports, and that the principal source of radiation that escaped into the atmosphere was leakage of evaporated gases from the reactor water purification system. It had previously been thought that spillage from the basement sump pumps caused the radiation release.

Operators had a "mindset" that caused them to focus on a misleading water level gauge and not on other, more serious indications of trouble, the report said. "had the operators allowed the emergency core cooling system to perform its intended function, damage to the core would most likely have been prevented," the study asserted.

Herman Dieckamp, president of General Public Utilities, the holding company that owns Three Mile Island, criticized Stello's report for ignoring creative operator action in the face of unprecedented conditions. It also left our actions that the Nrc or the reactor supplier, the Babcock & Wilcox Co., might have taken to prevent the accident, Dieckamp said.