Nuclear reactors are designed to respond in a completely safe manner to a wide range of possible transients arising from all reasonably conceivable initiators. These initiators will include failure of important valves or pumps and rupture of any part of the piping. In a PWR, main coolant pipes are some 1,000 mm in diameter and in safety analysis, the unlikely event of a guillotine rupture of one of these pipes is considered. The consequence could be the rapid ejection of most of the water from the primary circuit, which would flash into steam and may also be radioactive. The containment structure is designed to capture this steam and to limit containment pressure, which may be a value consistent with the design strength of the containment. Most water reactors have this type of containment. The issues in gas-cooled reactors (which have a single-phase coolant) and sodium-cooled reactors (which are not pressurized) are rather different, and are not included in this discussion.
As reactor safety rationale developed, it has become necessary to consider so-called 'Severe Accidents,' which result from the extremely unlikely failure of safety systems to cool the reactor fuel adequately following some accident initiator. Containment has been shown to have a strong influence on the subsequent course of events and could well limit the consequences of such an accident. Designs have progressed by considering whether there are some extra features which could be added to further reduce the consequences of a severe accident, without having a large additional effect on reactor cost. Examples of such additions are provisions for cooling and solidification of any molten core material which might penetrate the reactor vessel, or for the release of excessive internal pressure through filters. (See also Blowdown.)
According to the International Atomic Energy Agency (IAEA) Safety Guide, the containment system consists of the containment structure and associated subsystems including:
the containment structure and extensions, such as external passive fluid retaining boundaries which, together, form an envelope around the reactor coolant system;
the active features of the containment isolation system, which, in general, provide closure of openings in the containment envelope upon demand;
energy management features (collectively referring to the functions of pressure suppression, containment atmosphere pressure and temperature reduction and containment heat removal after a postulated accident);
radionuclide confinement features, which are provided to reduce the release of radionuclides to the external environment, after their release to the containment volume;
combustible gas control features, which limit the build up of combustible gases, such as hydrogen, in the containment envelope and prevent uncontrolled combustion or detonation of these gases;
other prevention or mitigation features for impulsive loads produced by severe accidents (such as ex-vessel steam explosions).
The main functions of the containment system are:
to prevent uncontrolled large releases to the environment in all those plant conditions which need to be taken into account, and to minimize controlled releases;
to maintain the system's structural integrity to assure the required leak-tightness and the necessary support to systems and components;
to allow the removal of decay heat and the cool-down of the reactor to safe shutdown conditions;
to prevent radioactive releases to the environment as a consequence of external events of natural and man-induced origin;
to provide a biological shield for operating personnel and the public.
Additionally, the containment system must allow access for operating and maintenance staff and equipment.
There are a large number of different designs of containment, but they can be divided into a few broad headings. The first major category is Dry Containments and Pressure Suppression Containments. In addition, containment systems may be either single or dual; in the latter case, a second structure is provided to catch any gases which may leak from the primary structure and also, in some cases, to provide extra protection against aircraft crashes.
Containment systems need to have provisions for removal of radioactive decay heat in the event of more extreme transients and severe accidents. In most existing reactors, these systems require the operation of pumps and heat exchangers and hence, require guaranteed electrical supplies. In many advanced designs being considered, there is a move away from dependence on such supplies towards reliance on natural processes, such as natural convection in which gravity is the motive force. The specifications for such systems allow for operation without operator intervention for at least 72 hours. Most reactors with these containment designs claim a cost reduction due to simplification of emergency systems.
The following brief descriptions indicate the most important types of containment in existence or being designed (1994).
The principle of dry containment is to provide a structure which is large enough and strong enough to contain without rupture all the contents of the primary coolant system which have flashed to an equilibrium pressure. Design pressure is in the range of 0.2 MN/m2 and the vessel is either spherical or cylindrical in steel or concrete, with a steel liner for leak tightness. An outer structure permits an interspace between the two structures, which is maintained at subatmospheric pressure by pumps which discharge through filters. Any activity escaping from the primary vessel is trapped on the filters. The In-containment Refuelling Water Storage Tank (IRWST) provides water for residual heat removal and for cooling of molten core material in the remote event of such an occurrence.
In a pressure suppression system, much of the vapor emitted from the ruptured primary system is condensed by passing these through a water tank before reaching the main volume of the containment. The primary reactor vessel is situated in the so-called dry-well. The water is in the wet-well and there are ducting arrangements to channel the vapor between the two parts of the containment. The advantage is a smaller and less expensive containment than a dry containment. It is particularly well-suited to BWRs since the maximum pipe sizes are smaller than in PWRs and depressurization through a ruptured pipe takes place more slowly, giving time for condensation in the pressure suppression system to occur without an unacceptable pressure transient. There have been several different designs embodying the same principle. Designs for PWRs also exist; for example, the PWR Bubbling Condenser Containment has tanks of water half way up the containment building through which steam from the reactor space has to pass.
This is a form of pressure suppression containment used in some of the early PWRs. Banks of ice are maintained at a high level inside the main containment building. Steam from a ruptured pipe is constrained to pass through these banks of ice to condense as much as possible and so reduce the ultimate pressure. A variant is to replace the ice by banks of gravel which form a completely passive, but limited, heat sink.
This has been developed for the Canadian design of Pressure Tube, Heavy Water Reactors. See CANDU Nuclear Power Reactor. The main reactor building is maintained at subatmospheric pressure. Attached to this building by a duct is another building maintained as a pressure less than atmospheric. In the event of a pressure rise in the main building, valves open to connect the vacuum building to the reactor building which, combined with a spray system in the reactor building, ensures that the pressure does not exceed the design pressure.
In multiunit stations, several reactors are connected to a single vacuum building.
Many concepts of advanced reactor design have adopted the principle that reactors should safely contain the consequences of accidents considered in the design without need of electrical supplies for at least 72 hours. This implies that the provisions for removing residual heat during this period must rely solely on natural processes, such as heat conduction and natural convection. The containment atmosphere can be cooled by conduction through a metal containment structure. The amount of heat lost in this way, and hence the power of the reactor which can be operated, depends on the area of the containment structure and the temperature which can be tolerated within it. Heat transfer can be enhanced by spraying the outside of the structure with water from a high-level tank and by inducing an enhanced flow of air through a secondary enclosing structure and a chimney effect. Many variants of this principle are under development (1994).
The above cooling system has a single steel containment only, and some safety authorities regard it as inadequate based on integrity grounds. A dual containment is not possible. An alternative is to install a heat exchanger system operating between the containment atmosphere and an external heat sink, such as a cooling tower. The components have to be configured so as to allow it to operate by natural convection.
Containment systems have been designed to alleviate the consequences of so-called design basis accidents, such as the rupture of any pipework. The Three Mile Island accident, coupled with the development of Probabilistic Safety Assessment (PSA), have led to the consideration of a greater range of extremely-improbable accident sequences, such as overheating, failure and possibly melting of the fuel. These would require consideration of a set of physical processes concerned with the emission and dispersion of radioactive material from the fuel and with the progression of molten fuel through the reactor core structure, the reactor pressure vessel and its interaction with concrete and other structural materials.
Much of the radioactive material is emitted in the form of aerosols, in which the different volatile fission products interact with each other and with their environment. They get carried along in the flow of steam and gases and are deposited in the reactor pipes and within the containment building. It is important to know how much of the material remains in suspension to allow for the possibility of its discharge to the external environment through any breaches in the integrity of the containment.
If there were to be a complete failure of the emergency core cooling systems, there is the possibility that some of the core material may melt due to the continued generation of decay heat from the radioactive decay of radioactive fission products and actinides. All volatile species will be released and other materials may be caught up in the aerosols. The molten material itself may penetrate through the core structure and may reach and melt through the reactor vessel. Some cooling processes do still occur, such as natural convection cooling by steam and other gases generated. There is also significant conduction through the core materials. It is not necessarily a foregone conclusion that the vessel will be penetrated, but if it is, the molten core material will fall onto the concrete of the reactor cavity and will interact with it, producing further aerosols and hydrogen.
Considering these processes, the objective has been to assess the degree of protection afforded by existing containment designs. A large number of computer codes have been developed to consider these complex and linkend processes. They are of two types; Mechanistic codes attempt to treat all the processes using the basic laws of physics in as exact a way as possible. These codes are expensive to run. Much faster codes are needed for assessment purposes, where a large number of accident sequences have to be developed. These codes make use of empirical data from experiments which have been carried out to study all aspects of the phenomena. These studies have shown that existing containment designs offer a large degree of protection against severe accidents.
In some new containment designs, the objective is to include additional systems in the containment to further mitigate the consequences of these severe accidents. One example is the provision of containment venting through a filter. This avoids the possibility of rupture of the containment through excessive build-up of pressure in an accident, but retains radioactive species in the filter. Such filters—using gravel beds in the filter—have been back-fitted to existing reactors in some countries. Another example is the provision of a core catcher in the reactor cavity underneath the reactor vessel. This causes the molten core to take up a predetermined geometry if it penetrates the reactor vessel and provides a cooling system to stabilize it.