The plant currently being designed is projected to produce 114 MWe, making it only some 10% the size of typical conventional PWR power reactors currently in operation. Consequently the entire nuclear generating facility can be correspondingly smaller than a conventional nuclear plant.

The reactor under design is of the high temperature gas reactor (HTGR) type, which means that its core is gas-cooled and not water-cooled. Helium is used in a closed loop Brayton cycle to power a gas turbine. In addition, this HTGR is of the "pebble bed" type, which means that the fuel is cricket ball (baseball) sized spheres of graphite containing coated uranium dioxide particles.

The reactor system is being designed in a modular fashion so that additional units can be added easily, as demand so dictates. Hence the name of the Pebble Bed Modular Reactor (PBMR). The system can accommodate ten modules to one control room.

The natural physics of the reactor system is such that it exhibits passive safety, which means that in the event of any fault occurring during reactor operations, the system, at worst, will come to a standstill and merely dissipate heat on a decreasing curve without any core failure. This means that the PBMR is inherently safe from the highly publicized public "fear factors" of reactor core meltdown and radiation release which in the public mind are associated with conventional water-cooled reactors.

At the end of its design life, fuel will be kept on site for 40 years after the plant has been shut down. At this point it can be removed and the entire PBMR can be demolished, and the site returned to greenfield status. In fact, the current costing of the project has assumed this option, and has also included the cost of long term storage of depleted fuel.

It is instructive to consider the South African circumstances that gave rise to the decision by the electricity utility Eskom to investigate the feasibility of constructing PBMRs.

About six years ago, only 30% of the South African population had access to electricity. Today the figure has risen to 60%, thanks to a concerted effort by Eskom to extend electrification (Ref 1). By comparison, in Africa south of the equator, only 10% of persons had access to electricity six years ago, and today it is still only 10%. In other words the African challenge is great (Ref 2). It is also interesting to note that South Africa produces and consumes over half of the electricity of the entire continent of Africa (Ref 3), (Ref 4).

Both domestically and in the rest of Africa, South Africa is seen as the regional economic powerhouse, and as such is assumed to have some moral obligation to engage in the economic uplifting of the whole region (Ref 5).

Currently 93.5% of South Africa’s electricity is coal-fired, with one large PWR nuclear station near Cape Town providing 4.5%; a further 1.5% is hydro. There are essentially no more hydro sites that could deliver significant amounts of power, and furthermore for some years Eskom has been of the opinion that it is not an option to build another large-scale nuclear plant. However, South Africa is rich in uranium.

Most of South Africa’s coal-fired electricity is generated by large scale plants built near the pit-heads of two large coal fields, both far inland on the eastern side of the country (Figure 1).

South Africa is a large country, and the distance between Cape Town and Pretoria is the same as the distance from London to Berlin. This implies long power lines from the coal fields or long fuel delivery routes between the inland coal fields and coastal centers of electricity consumption, if the electricity were to be generated by coal in the future. The economic equations are stacked against these options.

Transporting huge quantities of coal over very long distances is just not an option, so South Africa needs to place power generation units closer to the sites that need it, to avoid having to construct very long (1000 km) high voltage transmission lines.

At the present time there is a world move against fossil-fuel power stations due to public concerns about CO₂ emissions leading to possible global warming. The United Nations sponsored conference on global warming held in Kyoto, Japan, in December 1997 gave rise to the Kyoto Protocol which sets limits on countries’ emissions of CO₂. This is in turn bringing pressure to bear on the development of fossil-fuel power stations that may produce "excessive" CO₂ (Ref 6).

Overall this general scenario led Eskom to a decision to investigate the development of a small modular pebble bed nuclear powered plant, the PBMR.

The PBMR consists of a vertical steel pressure vessel 18 m in height, and lined with 60 cm of graphite (Figure 2).

During normal operation, the pressure vessel contains a fuel load of 440 000 balls, each 60 mm in diameter. This load consists of 310 000 fuel balls containing uranium dioxide particles encased in graphite and silicon carbide, as well as an additional 130 000 pure graphite balls that serve the function of an additional nuclear moderator. The uranium dioxide particles are less than half a millimeter in size, and there are 15 000 of them in one fuel ball, totaling 9 g of uranium, which means that total uranium in one fuel load is 2.79 tonnes.

A reactor will use 10 to 15 total fuel loads in its design lifetime. The actual calculated figure, for continuous full power operation, is 13.8 fuel loads.

To remove the heat generated by the nuclear reaction, helium gas at 540°C is passed into the pressure vessel at the top. It passes between the fuel balls, and then leaves the bottom of the pressure vessel at 900°C. The hot gas then passes through a conventional gas turbine system to drive electrical generators (Figure 3). The conventional section of the turbine gas cycle can be water-cooled or air-cooled which is an important consideration, since this factor provides the plant with a great variety of site options.

The team developing the PBMR is hoping to obtain a decision before the end of 1999 as to whether or not Eskom intends to proceed with the construction of a demonstration unit. If so, the time scale for construction of this first unit is indicated in Table 1. A typical production PBMR module will take 24 months to construct.

The staff needed to operate such a reactor do not need to be particularly highly skilled, and furthermore relatively few staff are required. This clearly is an important factor in considering the operating costs of such a plant.

Fuel will be manufactured in South Africa, and there is currently a development project underway as a joint venture between the PBMR team and the South African Atomic Energy Corporation.

The fuel consists of 60 mm diameter balls of graphite that contain coated uranium dioxide particles evenly distributed throughout, as shown in Figure 4. The graphite is a mixture of 75% natural graphite and 25% synthetic graphite. The small uranium dioxide spheres are each coated with a layer of porous carbon, then high density pyrolytic carbon, silicon carbide, and then another layer of pyrolytic carbon. This is known as TRISO fuel.

A function of the porous carbon is to absorb any mechanical deformation that the uranium dioxide may undergo during its lifetime, thereby enhancing the integrity of the silicon carbide containment and so ensuring that radioactive daughter products cannot escape from the particle. The inner layer of pyrolytic carbon serves the dual function of providing structural support to the silicon carbide, and also forms a barrier against the migration of some fission products. The weight of a single fuel ball is 202g, including 9 g of uranium.

The start-up fuel is enriched to a level of 4%, and the equilibrium fuel has an enrichment of 8%.

As has been explained above, each fuel particle serves as its own inherent containment boundary. Any fuel containment failure which might occur on PBMR fuel would be of small concern since there is no possibility of any type of massive simultaneous failure, as is possible in a core meltdown scenario in a conventional PWR or BWR.

When South Africa embarked on this project in 1993, there was already a considerable body of knowledge in existence in the world relating to HTGRs (see Table 2).

Eskom was therefore able to form a collection of partnerships and coalitions with various organizations that had worked on a variety of operating plants and advanced designs over the years (see Table 3). The PBMR concept is thus not an unproven new technology. In fact, the contrary is true, and Eskom has been able to collect together the "best of all worlds" and then carry out its own design on the basis of a sound body of proven knowledge and experience.

The PBMR is ideal for South Africa’s future needs, but it is also clearly ideal for other countries, both in the developing world and in the first world.

In the case of the developing world, it is easier to operate a PBMR than a coal-fired power station with twenty-first century specifications. The PBMR does not need a particularly highly skilled team of operators, and the team is small. A year’s supply of fuel can easily be stored on site, which is important in countries where road and rail supply routes can be highly unreliable. Imagine the constant transportation of coal! PBMRs can also be sited anywhere, being independent of water-cooling. In developing countries they can be sited such that they have their own localized grid, and do not need to be integrated into a major national grid.

As far as first world countries are concerned, PBMRs will be ideal. They can either be deployed individually in places where small amounts of power are required, or can be clustered on one site in modular fashion. Units can easily be added incrementally as required. First world countries could also use PBMRs to replace existing plants that come to the end of their operating lives. At that point the option could be exercised either to build a large scale modular cluster of reactors on one site, or alternatively to distribute the reactors to individual sites and close down the original large site.

The low enriched uranium fuel is half millimeter sized particles of uranium dioxide encased in graphite and silicon carbide, which in turn is encased in the main graphite ball (Ref 6). Not only is it expensive and difficult to process such spent fuel, but the proportion of the fuel burnt is very high.

Fuel balls are recycled through the reactor some 10 times. They are extracted, analyzed to determine the amount of fissionable material left, and then either returned to the reactor or sent to depleted fuel storage.

Unlike conventional large fuel elements, the fuel balls cannot jam in reactor core housings, and so they can be burnt to depletion without any fear of mechanical deformation causing handling problems in the core. There is therefore minimal fissile material to extract from depleted fuel.

There is no low or medium level waste produced by a PBMR, except for gloves, jackets and similar items that might be used by workers in a designated radiation area.

Depleted fuel balls are planned to be stored on site for many years. However, they can either be removed to a repository, or left on site for decades, depending on the operator’s policy. The current PBMR design assumes that depleted fuel balls will be stored on site for the duration of the plant’s lifetime, and then for a further 40 years after plant shutdown and decommissioning.

The nuclear licensing process is currently underway in South Africa, and is being carried out by the Council for Nuclear Safety (CNS) in accordance with long established norms. The CNS is an independent body formed by statute.

The author would like to acknowledge the valuable input in the preparation of this paper of members of the PBMR team, including Dave Nicholls, Wynn Roscoe and Thabang Makubire, as well as Ken Brown of the Atomic Energy Corporation. He would also like to thank Sue Cooke and Sakkie du Plessis of Eskom for assistance with research and preparation of the diagrams.