The Three Mile Island Accident The Early History of the TMI Accident

The Three Mile Island (TMI) accident occurred in one of two similar reactors at the Three Mile Island site in Pennsylvania.7 The accident was in the second

4 The number of fatalities is in question, given the uncertainties surrounding the effects of radiation at low doses and dose rates, but in the discussion of these accidents, we quote numbers based on the adoption of the linearity hypothesis (see Section 4.3).

5 This summary is based largely on an IAEA report prepared shortly after the accident [10].

6 There was no explosion, but criticality continued with a low power output for about 20 h, stabilized by thermal expansion of the fluid and the formation of bubbles. The chain reaction was terminated by draining water from a cooling jacket surrounding the tank, which reduced reflection of neutrons back into the tank, and as a precaution by injecting a boric acid solution into the tank.

7 Extensive studies were carried out after the accident. One, referred to later as the "Kemeny report," was by a commission appointed by President Carter and

Tmi Diagram

Fig. 15.1. Schematic of the TMI-2 facility, including reactor building and turbine building. Piping goes, from right to left in the diagram, through the containment building wall to the auxiliary building (not shown); piping also goes, from left to right, through the turbine building wall to the condensate storage tank and cooling tower (not shown). (From Ref. [11, pp. 86-87].)

Fig. 15.1. Schematic of the TMI-2 facility, including reactor building and turbine building. Piping goes, from right to left in the diagram, through the containment building wall to the auxiliary building (not shown); piping also goes, from left to right, through the turbine building wall to the condensate storage tank and cooling tower (not shown). (From Ref. [11, pp. 86-87].)

unit, known as TMI-2. It was a 906-MWe pressurized water reactor built by Babcock and Wilcox, the smallest (in terms of number of units completed) of three U.S. manufacturers of PWRs. It had first received a license to operate at low power in February 1978 and was in routine operation at full power by the end of 1978. A schematic of the TMI-2 facility is shown in Figure 15.1 [11, pp. 86-87].

The accident started with a failure of the cooling system of TMI-2 in the early morning of March 28, 1979. The initial problem was an interruption in the flow of water to the secondary side of the steam generator. This water is the so-called feedwater. In the secondary loop, feedwater enters the steam chaired by John Kemeny, the president of Dartmouth College [11]. The second, the Rogovin Report, was by a special inquiry group instituted by the Nuclear Regulatory Commission and chaired by Mitchell Rogovin, a partner in an independent Washington law firm. The description here is drawn largely from the Kemeny Report [11] and Part 2 of Volume II of the Rogovin Report [12], as well as a further review article [13].

generator, and steam emerges to drive the turbine. The steam is condensed in a second heat exchanger (the condenser), and water is returned to the steam generator after passing through a "polisher," in which dissolved impurities are removed. The flow of water between the condenser and steam generator is maintained by the condensate pump and the main feedwater pump (see Figure 15.1).

The chain of events that led to the accident appears to have been initiated by work done to clean the polishers. In a sequence that has not been conclusively established, this operation may have caused one or more of the valves in the condensate polisher system to close, automatically shutting off (tripping) one of the condensate pumps. The tripping of the condensate pump, whatever the cause, in turn, tripped the main feedwater pumps.8 This failure caused the emergency feedwater pumps to start automatically, in order to maintain the flow of water to the steam generator. Maintenance of feedwater flow is essential to cool the water from the reactor that flows through the primary side of the steam generator.

Up to this point, everything was "normal," in the sense that reactors are designed to handle occasional equipment failures; protection then comes from backup systems. However, the block valves in the emergency feedwater lines (there were two) were closed; according to proper operating procedures, they were supposed to be open. Indicator lights in the control room showed the closed status, but the operators at first did not notice this. Thus, no water was being fed to the secondary side of the steam generator because the pumps for the main supply were off and valves in the emergency line were closed. With no flow of water, the pressure in the steam generator rose and in response, the pilot-operated relief valve (PORV) on the "hot" side of the steam generator heat exchanger opened. The pressure excursion also caused the reactor to trip, with automatic insertion of the control rods. With the reactor turned off and the PORV open, the pressure dropped. The PORV should then have automatically closed.

At this point, there were additional equipment and design failures. The PORV did not close properly, but the control panel indicator light displayed the status of the control power to the valve (namely that it was supposedly closed), not the actual status of the valve (namely that it was open). Thus, the operators had to cope with unusual conditions in the cooling system without knowing the actual status of the valves in it. In particular, the PORV remained open for almost 2.5 h, causing a very large loss of needed cooling water.

Within 2 min after the start of the accident, the steam generators boiled dry because they had no feedwater source and there was a substantial heat output from the reactor core due to radioactive decay. Overall, the conditions of the cooling system were both unusual and confused, with the operators

8 Figure 15.1 does not show the redundancy in the system. There were two main feedwater pumps and three emergency feedwater pumps.

not having correct information or sufficient training to recognize the nature of the evolving anomalies and cope with them. They did recognize that there were serious problems, and by 4:45 AM supervisory personnel began to arrive at TMI, only three-quarters of an hour after the start of the accident. By 6:22 AM, the PORV was closed, but the problems were not over. At 7:00 AM a "site emergency" was declared because there had been some release of radioactivity.

Was this article helpful?

0 0

Post a comment