Underground sited. This scale was chosen so that the physical size allowed it to be factory manufactured and transported to the site,which is a significant potential cost reducer. . . .

- John Rawls, Chief Scientist at General Atomics

From the dawn of the Nuclear era, Edward Teller was deeply concerned about reactor safety. Teller favored underground placement of civilian nuclear power plants primarily for safety reasons. Indeed in his last paper Teller advocated underground placement of Molten Salt Reactors, although arguably MSRs could be designed to be safe enough to make further safety measures unnecessary. A further justification of underground placement of MSRs, would be that it would be consistent with low manufacturing costs and rapid reactor deployment.
Underground deployment of nuclear tractors offers a number of advantages including,
Higher Resistance to...
– Terrorist attack
– Aircraft impacts
– Proliferation
– Sabotage and vandalism
– Conventional warfare effects
Underground sites offer superior protection against the effects of severe weather events and some potentially protection even from the effects of earthquakes. Underground sites also offer superior protection against fission product release in the event of a serious reactor accidents. Studies of underground siting conducted during the 1970's reported that underground siting would cost more than traditional reactor siting, but these studies assumed the use of conventional nuclear technology and that the entire nuclear facility would be located underground. From the viewpoint of safety and security it is only necessary to house reactors underground. Turbines and generators, as well as other Nuclear Power Plant related facilities can be located above ground without any disadvantages if the cost of underground facilities placement become a matter of concern. In addition Generation IV reactors are generally more compact than conventional reactors. There are other ways to limit underground housing costs. For example salt formations offer unique advantages for nuclear reactor housing, with low cost excavation. Existing underground salt mines offer unique placement advantages. In addition to existing salt mines, many old mines and natural caverns offer potential underground siting for reactors. Studies of underground placement of nuclear facilities made during the 1970's assumed that reactors would be placed 300 feet or more beneath the surface, but reactor manufacturer Babcock and Willcox intend to place their small mPower Reactor just below the surface.

Underground placement of small, compact Generation IV nuclear power plants would be inexpensive, and underground placement is often featured in many small Modular nuclear designs including the B&W mPower Reactor. A recent report to the American Nuclear Society by Mark S. Campagn and Walter Sawruk and titled, "PHYSICAL SECURITY FOR SMALL MODULAR REACTORS" states,
Rely on government response for SMR facilities with vital assets underground or otherwise well protected. Shallow burial or a hardened structural design provides excellent protection against large explosive weapons and aircraft impact as well as an excellent means of enhancing security system effectiveness against sabotage. Application of the traditional multilayered defensive approach of detection, deterrence, delay, and defeat can be used effectively for physical protection of SMRs. Detection, deterrence, and delay concepts must be integrated into the early design phase of the facility in order to provide sufficient lead time for government response.
A few years ago three University of Tennessee Nuclear Engineering Graduate students, William A Casino, Kirk Sorensen, and Christopher A Whitener wrote a paper titled "A Small Mobile Molten Salt Reactor (SM-MSR) For Underdeveloped Countries and Remote Locations." The paper won first prize in an American Nuclear Society reactor design contest. This design exercise focused on a reactor small enough to be transportable by truck, yet large enough to be transportable by truck. The design is highly suggestive although it turns out to be a little big to be truck transported. The reactor was designed to produce 100 MWe, with an active core region that weighed 216 tons (about 200 metric tons). This is too heavy to be easily transported by truck, but it might be possible to shave that weight down significantly. More than half of the core weight is contributed by core graphite (about 147 of 215 tons). Thus a method of inserting core graphite into the core at the destination site, would offer considerable advantage if this could be accomplished quickly and at low cost. The use of graphite pebbles would be consistent with these goals. This would lower of the weight of the core moduel to 68 US tons, which would certainly be manageable by either truck or train. Further the primary heat exchange and connecting pipes are included in the core module, and this might be considered a flaw in the design.
There were a number of flaws in the SM-MSR design, which was after all a student design exercise. Although a core dump tank was included in the SM-MSR design, no colling system was included. However, a passive cooling system for the disposal of fission product decay heat is possible with an underground MSR. Air can be drawn into the underground chamber and heated by the dump tank exterior, and then the heated air could rise through a chimney. The rising heated air, would, of course, draw more air into the underground chamber by lowering its air pressure, thus creating a passive decay heat cooling system.
Casino, Sorensen, and Whitener noted that,
One site specific limitation is that the primary containment module as proposed is to be placed into a silo to be trenched into the earth. This silo needs to be approximately 28 meters in depth and be approximately 25 square meters in area. The water table in most locations will likely occur above this level, and the SMMSR containment module shall be constructed to withstand moisture impingement on the outer surface. Other corrosive elements in the water need to be checked for.
The reactor silo would not requite a significant amount of excavating, and thus could be dug quickly. As has already pointed our, building a silo from scratch might not always be required. Silos built for cold war guided missles, and well as a variety of underground mines might be useful, although preexisting underground structures would not be the only solution to small reactor siting the problem. Rapid drilling of a silo could advance at rates of as much as 10 meters a day. With prefacricated silo liners, site preperation might require no more than a week. Thus the re-use of coal fired power plants sits to house clusters of small baseload reactors could easily include underground housing of a number of reactors.
Underground housing of small reactors appears to be practical and it is credible to argue that Underground reactor housing can lower nuclear costs, and dramatically shorten reactor construction times. In addition underground housing can increase nuclear safety and offer significant protection of reactors from aircraft and other forms of terrorist attacks.