Designers of the Three Mile Island 2 nuclear power plant all but washed their hands yesterday of responsibility for the March 28 near-disaster there, saying instead that plant operators need a lot more training.
John MacMillan, vice president of Babcock & Wilcox Co.'s nuclear power generation division, avoided any direct discussion of blame for the incident at the facility near Middletown, Pa. Instead, in a presentation to the Advisory Committee on Reactor Safeguards, a group of technical experts who advise the Nuclear Regulatory Commission, MacMillan assured them that "B&W systems can be operated safely for a spectrum of equipment failures, including those experienced at Three Mile Island 2" (TMI).
The company, he said, had already met with representatives of the other B&W designed nuclear power plants to discuss the incident and had begun retraining plant operators on an accident simulator that reproduces the early elements of the TMI event.
"Based on information currently available to B&W, the equipment in both the primary and secondary plants, with the exception of the pilot-operated [automatic] pressurizer relief valve, performed as "designed," MacMillan said.
He noted, however, that the firm would "consider . . . potential design improvements" that would give a "more positive indication" of whether that valve was open or shut. The valve stuck open during the TMI incident, allowing water that was intended to cool the overheated reactor to flow out of the core and into the basement of the containment vessel.
MacMillan stressed that the operator did not notice the stuck valve for more than two hours. "In our view the significance of this [valve] factor is not only the failure of the valve to reseat but, more importantly, the time which elapsed between the failure to reseat and the recognition that this had occurred."
In the 150 other times that the valve opened during events at other B&W plants, he said, it failed to close only twice - once because of corrosion and once because of a missing electrical relay.
MacMillan also indicated that his company would consider a design change to stop the automatic activation of a sump pump that drove the radioactive overflow water from the TMI plant's basement to an auxiliary building, even though that pump was designed and licensed to perform exactly as it did. The NRC had blamed the automatic pump for part of the release of radioactivity to the environment at TMI.
A pressurizer gauge that earlier had been blamed for misleading an operator into prematurely turning off a series of pumps was, MacMillan said, "not significantly in error" during the accident. It showed a water level that was off by only one foot, more or less, out of 33 feet, he said.
At the same time, he said, other indicators of reactor pressure were available and the operator action "should not have been based on the single parameter of pressurizer level. Another design change is contemplated that would provide a more graphic picture of the pressure level in the reactor, he said.
"The severity of the TMI 2 incident warrants timely actions to encourage proper operator performance in these events," MacMillan said. He called for tighter monitoring of safety system controls and stronger emphasis on training operators to focus on keeping the reactor core cool.
Although a B&W task force will assess the long-term implications of the incident, MacMillan said, "we do not believe that major re-analysis is necessary for these near and longer-term design decisions."
Earlier yesterday, an NRC official endorsed the basic safety of B&W plants. "We find no apparent major design deficiencies on Babcock & Wilcox plants," said Robert Tedesco, assistant director for reactor safety in the NRC's office of nuclear reactor regulation.
Under questioning by the 12 committee members present, MacMillan said B&W plants had experienced about 60 incidents in which the flow of water to theturbines suddenly stopped, as it did at TMI. However, he said, feed water systems are not part of the equipment designed and sold by B&W but rather are contracted for individually by each power plant from various suppliers. "We set criteria for temperature flow. . . . we work with the customer to analyze these problems. In most cases the corrective measures involve changes that are not in B&W-supplied equipment," he said.
He added he did not yet know why the TMI 2 feed water pump had failed nor had the B&W simulator ever predicted the TMI 2 sequence of events.
"While the TMI 2 incident was serious, we believe that there are constructive lessons to be learned and that timely, responsive actions have been taken to assure the safe operation of B&W reactors," MacMillan said.
At 8:30 a.m. today, committee members will resume deliberating what recommendations for technical changes, if any, they will make to the NRC.