The Three Mile Island accident

Summary description of the Three Mile Island no.2 Plant 

Three Mile Island on the Susquehanna River is located about 16 km SE of Harrisburg Pa, USA. It is a flat island with a surface of several square kilometres. Some years ago it was chosen as the site for a nuclear power station with two units named TMI-1 and TMI-2. Each unit has its own reactor and turbine-generator group for the conversion of steam into electric energy. The two units could supply 1700 MW to the grid, sufficient for the needs of 300 000 families (based on the average consumption of a US family). 

The power station was the joint property of the Pennsylvania Electric Company, the Jersey Central Power Light Company and the Metropolitan Edison Company. The three companies were part of a holding , the General Public Utilities Corporation 

The nuclear part of the plant (i.e. the reactor and its auxiliary systems - the nuclear island ) had been supplied by the Babcock Wilcox company. The architect engineer, Bums Roe, had built the remainder of plant. 

The vessel (1) contains the reactor core (2) in which the control rods can be inserted from above (3). The cooling system is formed by two circuits (in the figure only one is represented), each one provided with two recirculation pumps (4) and with one steam generator (5). The steam produced in the secondary side of the generator is routed to the turbine (6) and converted to water again in the condenser. The condensate returns to the steam generators through 

---

Several accidents in nuclear facilities have been extensively analyzed and reported. The three most widely publicized accidents were at Windscale (now known as Sellafield), United Kingdom, in 1957 Three Mile Island, Pennsylvania, in 1979 and Chernobyl, Ukraine, in 1986 (UNSCEAR 1988 Severn and Bar 1991 Eisler 1995). From the accident at Windscale about 750 trillion (T)Bq 22 TBq Cs, 3 TBq Sr, and 0.33 TBq °Sr were released and twice the amount of noble gases that were released at Chernobyl, but 2000 times less and Cs. From the Three Mile Island accident, about 2% as much noble gases and 50,000 times less than from the Chernobyl accident were released. The most abundant released radionuclides at Three Mile Island were Xe, Xe, and but the collective dose equivalent to the population during the first post-accident days was <1% of the dose accumulated from natural background radiation in a year. 

The Three Mile Island accident, rated as one of the top media events of the century, was a crowning blow in the battle. The media consistently portrayed the accident as a close call on a public health disaster, and continue to do so to this day, although none of the studies done after the accident gives any reason to believe that to be the case. As demonstrated in those studies, the containment building would have prevented release of large amounts of the radioactivity regardless of what might have hap-... 

In the five years since the first volume was published, there has been increased interest in the chemistry within gas lasers and the chemistry induced by laser radiation, the kinetics and photochemistry within fusion and industrial plasmas, as well as in the normal and perturbed lower and upper atmosphere. And. since the Three Mile Island accident there has been renewed interest in radiation damage to living and nonliving things. This state of affairs has not only precipitated a variety of spectroscopic studies, but has also brought more attention to the nonspectroscopic aspects of excited state production and the interaction of excited species. The latter topic was stressed in the earlier volume and the emphasis is retained here. 

The release of 131I and other fission products in reactor accidents has been considered in the previous chapter. In the Windscale accident, the temperature in the fire zone reached an estimated 1300°C and 8 tonne of uranium metal melted. Over 25% of the 1311 in the melted fuel escaped to atmosphere. In the Chernobyl accident, the fuel was U02, the temperature exceeded 2000°C, and about 25% of the total reactor inventory of 131I was released to atmosphere, as vapour or particulate aerosol. In the Three Mile Island accident, 131I remained almost completely in the reactor coolant. The activities of 131I released in reactor accidents, including that at Chernobyl, have totalled much less than the activities released from weapons tests (Table 2.3). 

As anticipated, there have been occasional equipment failures involving reactors, but the safety systems have been sufficiently redundant so that one or more have always worked. Even in the Three Mile Island accident in 1979, the safety systems worked as designed. Much of the damage resulted from operator actions to override the safety systems. As concluded in the Reactor Safety Report, the limitations of the operator created and then seriously aggravated the Three Mile Island incident. Nevertheless, the features of the containment system prevented significant exposure to the plant personnel or any off-site individual, this despite failure of the barriers provided by the fuel cladding and the primary coolant system. 

It should also be remembered that most evolving technologies, whether boilers during the 19th century, airplanes in this century, or nuclear plants, entail some accidents from which lessons are learned. Both the Three Mile Island accident, from which only limited radioactivity escaped to the environment, and the Chernobyl disaster, have led to the introduction of new safety features in nuclear reactors, in plant operating procedures, and in regulations. 

Only two earlier reactor accidents caused significant releases or radionuclides the one at Windscale (United Kingdom) in October 1957 and the other at Three Mile Island (United States) in March 1979 (UNSCEAR-1982). While it is very difficult to estimate the fraction of the Windscale radionuclide core inventory that was released to the atmosphere, it has been estimated that the accident released twice the amount of noble gases that was released at Chernobyl, but 2,000 times less and Cs (DOE-1987). The Three Mile Island accident released approximately 2% as much noble gases and 0.00002% as much l as the Chernobyl accident. 

Systems analysis techniques are used only to a limited extent in the process industries for two reasons. First, such techniques are not generally effective at predicting human behavior (e.g., WASH 1400 did not anticipate the Three Mile Island accident). Yet human performance is a very important component of safety performance in the process industries. Second, the use of PRA methodologies is generally time-consuming and expensive— particularly when used in the chemical industries where there is so much difference from facility to facility. 

This, the Windscale accident, was for many years the most serious reactor accident it had more radiological consequences than the Three Mile Island accident of 1979. Clearly the accident had the salutary effect of making designers and operators safety conscious in the UK. 

With the Three Mile Island accident, much publicity was given to the potential explosion hazard associated with a hydrogen bubble which had formed above the nuclear core. 

From the late 1950s to the Three Mile Island accident... 

The decision to create DID in the plants was taken at the start of nuclear energy development which indicates a remarkable farsightedness, as subsequent history has demonstrated that it has been the best defence against the uncertainties of the technology and the mistakes initially made (see, for example, the Three Mile Island accident). Obviously, in the first period of nuclear energy, many protests were made against this waste of resources which consisted in the construction of costly barriers (e.g. the containment) without, according to some, a real need for them. 

Whereas the dangers of the above example never materialized, the behaviour of the reactor pressure vessel during the Three Mile Island accident is exceptional. It withstood the outpouring of about twenty tons of molten core on its bottom, in conditions of highly deteriorated internal cooling. This behaviour indicated to the technical experts the presence of a powerful and up to then neglected barrier in the Defence in Depth, which is now utilized in a planned way as a potential asset. 

In the following, both experience data (from the Three Mile Island accident) and deterministic considerations justified by the existing knowledge are shown. At least a picture of the important factors for the decrease of the vessel damage probability and indications on the still necessary research will be obtained. In particular, the importance of the prevention of severe accidents will be clearly demonstrated. 

The fact that in a small LOCA with the rupture on the top of the pressurizer it is possible to have the pressurizer almost full of water and the vessel almost empty was demonstrated in a couple of transients in Belgium and in the USA, but in a more dramatic way in the Three Mile Island Accident. 

As it can be seen, both the risk objectives and the risk analyses on existing plants are reassuring, but, it is frequently asked, how reliable these analyses are How much the inevitable uncertainties on data and methods can influence the results Is it possible that some accident sequence has been forgotten in performing a probabilistic analysis All the available information, including the analyses made before the Three Mile Island accident and the sequence of events in the accident itself, indicate that a corrective... 

The Three Mile Island Accident started 36 seconds after 4.a.m. TMI-2 had already met problems with the feedwater purification system 18 months before the accident. During this time, however, no effective measures were taken to guarantee the needed safety of operation of this equipment. 

A technical superintendent at TMI-2 who arrived on the plant at 03 45, subsequently said I had the perception that we were in a very unusual situation, since I had never seen the pressurizer level increase and stay at a high value and, at the same time, the pressure staying low. They [the pressure and the level] had always behaved in the same way . As a consequence of the described evaluation errors the primary circuit continued to lose water for hours and in addition the automatic core cooling system, correctly activated, could not perform its function of fuel integrity protection. It is now known that if the block valve had been closed after one and half or two hours or if the operation of the HPI only had not been arrested, even without the closure of the valve, the Three Mile Island accident would have been no more than a modest nuisance of operation. For completeness of information it has to be added that the possibility of an accident of the type of TMI-2 had been foreseen by some experts. If these foresights had been confirmed by in-depth theoretical studies and possibly by experimental tests, their results, duly made known to interested people, would have enabled the TMI-2 operators to correctly diagnose the fault and react correctly. 

---