The final “defense in depth” barrier to the release ofradioactive materials to the environment is the containment building or system itself. The TMI-2 accident demonstrated the importance of the reactor containment in converting a very severe accident in the reactor itself into one that had very little public health im­pact. There has been much study of the integrity of containments, including ex­perimental research under simulated accident conditions, particularly for P^^s.

The main forms of reactor containment that have been employed are as follows:

• Large prestressed or reinforced concrete shells that are designed to withstand internal pressures of 3—4 bars above atmospheric.

• Spherical steel vessels (as used in German reactors) that are similar in concept to the concrete vessels and withstand about the same pressure.

• Steel or concrete vessels in which ice is used to condense any steam released from the reactor system (so-called ice-condenser plants). Here the design pressure can be lower, but such concepts are less popular than they used to be since the ice-condenser is of little help in containing, say, a hydrogen explosion.

• Pressure-suppression containments in which the system is arranged so that any steam escaping from the reactor circuit will bypass through vent tubes into a pool of cold water where it is condensed.

Advanced containment systems often involve a double wall containment with a steel or prestressed concrete inner wall and a reinforced concrete outer wall. A subatmospheric pressure is maintained in the interwall space. Alterna­tively it is possible to combine a pressure-suppression system inside a conven­tional dry-well containment.

The likelihood of an accident’s leading to a breach in the containment is low. As exemplified by the case of TMI-2, in the majority of severe accidents the containment will fulfill its function. Challenges, however, can come from over­heating or overpressurization, hydrogen explosion, or missile impacts. These could result in structural failure or damage to the liner or a penetration resulting in a high rate of leakage. In addition, failure to isolate the containment during an accident could allow the transfer of radioactivity to other parts of the plant or to the environment. The timing of any failure is also relevant. The longer the containment remains intact, the greater the opportunity to take action to protect the public from any release.

In the previous section we saw that if molten fuel reaches the reactor vessel lower head, then this may fail. If this failure occurs rapidly with the primary sys­tem still at high pressure, the molten fuel will be ejected into the reactor cavity and from there it can move into the containment building. Rapid heating and pressurization of the containment will result from

• molten debris particles heating the containment atmosphere

• chemical reactions between the debris and water-steam leading to additional heating

• hydrogen, produced by chemical reactions, burning or detonating

Tests have been carried out at the Sandia National Laboratories using one — tenth-scale containments modeling the geometric details of actual nuclear power plants. Iron/aluminium/chromium thermite was used to simulate the molten core. This was ejected by high-pressure steam from the scale pressure vessel bottom head. Pressures and temperatures inside the containment were measured. Water that might flood the reactor cavity or be on the containment basemat was present in some experiments.

The results demonstrated that heating of the containment is less if the debris is contained below the main operating deck. Water in the reactor cavity reduces the pressurization due to the steam released by quenching the melt. Water on the basemat has little or no effect. Hydrogen produced by the cavity interaction and dispersal processes burned with a diffusion flame in the upper dome and contributed about half to the pressurization. Any preexisting hydrogen burned slowly and had little effect. These experiments and modeling calculations sug­gest that a wide range of variables influence containment heating

• Primary circuit pressure prior to vessel failure

• Mass of molten core in pressure vessel lower head

• Temperature and composition of the molten core

• Particle-size distribution and entrapment of debris

• Impingement onto surfaces and freezing

• Chemical reaction of debris with steam

• Quenching by reactor cavity water

• Formation, transport, and burning of hydrogen

From a variety of studies it can be judged that containment integrity will be at risk only if a large fraction of the core is ejected and a hydrogen detonation occurs. In these circumstances, the containment may be subjected to continu­ous static pressures of 10-15 bars as well as a transient pressure pulse. How­ever, a major mitigating feature would be depressurization of the primary circuit prior to molten core ejection so as to limit debris dispersion. Depressurization to 15-25 bars is effective in this way (see Section 6.2.4).

Even if the containment survives the early containment heating and pressur­ization, there are still challenges to its integrity that occur later. These include

• hydrogen combustion

• gradual overpressurization

• basemat melt-through

Hydrogen formation and combustion are described in Section 6.3.3. Slow overpressurization may occur if there is no heat removal or venting of the con­tainment. Moreover, interaction of the core debris with concrete produces copi­ous amounts of noncondensible gases such as carbon monoxide and carbon dioxide to add to the pressurization.

The interaction between the molten core and the concrete depends on a number of factors including the presence of water. If the cavity is initially dry and the core debris forms a deep bed, then extensive interaction may occur. Flooding the debris may effectively cool the melt. Some comments about the coolability of debris beds are made in Section 6.3.2.

There remains a finite possibility that the molten core materials may attack the containment basemat. This scenario is discussed in Section 6.3.4. Com­plete melt-through of the basemat would take several days, but the conse­quences of failure are relatively small compared with the failure of the containment above ground.

Finally, the integrity of the containment building may be compromised by the failure to isolate the building. As explained in Chapter 4, once the emer­gency core cooling system in the P’^TC is initiated, the containment is isolated. In practice, a number of essential services must still be provided to the reactor (emergency feedwater, etc.), and this provides a number of routes for release from the containment. For example, rupture of the decay heat removal system outside the containment, coupled with failure of an isolation valve, could give a route out from the containment that bypasses the building itself. Proper se­curing of personnel and equipment airlocks is also essential. Two specific acci­dent sequences are important in this context:

The so-called interfacing systems LOCA in which important check (or nonre­turn) valves fail and low-pressure piping connected to the reactor coolant sys­tem fails outside the containment. This provides a bypass to the enviroment.

Failure of steam generator tubes during the course of an accident again may permit bypass to the enviroment via the secondary side steam relief valves