This collaboration on ITER began with Conceptual Design Activities (CDA) in April 1988 and was successfully completed in December 1990.

The CDA took place in Garching, hosted by the Max Planck Insitut für Plasmaphsik (IPP), for periods of several months, and was attended by experts from the design teams of national next step experiments. The overall direction of the CDA and supervision of its execution was undertaken by the ITER Council (IC), chaired by John Clarke (USA). Advice was given by the ITER Science and Technical Advisory Committee (ISTAC), chaired by Boris Kadomtsev (USSR). The ITER Management Committee (IMC) - Romano Toschi (EU), Ken Tomabechi (Japan), Yuri Sokolov (USSR) and John Gilleland (USA) were responsible for day-to-day execution.

Design Groups for the CDA

There were two work phases: definition and design. The aim was:

• to define the technical characterisitics of the machine and produce its conceptual design;
• to carry out validating R&D in a coordinated fashion;
• to define future R&D needed as well as schedule, cost and human resource estimates for its realisation;
• to carry out a safety and environmental analysis of the dsign and define the construction site technical characteristics.

This work followed the schedule shown below. Around 50 professional staff met together for about 6 months each year. The work was carried out jointly by all the Parties in each area, in order to be sure of building consensus. Each Party carried out technology R&D worth about US$10M/year over the three years of the CDA, in addition to their ongoing experimental programme.

Schedule of the CDA

Although the overall programmatic objective of ITER was clear - to demonstrate the scientific and technological feasibility of fusion power - this needed interpreting into a set of technical objectives. These were established before the start of the CDA in 1988. In summary the main aim of the device was to be to achieve ignition and to guarantee an inductive pulse length of a few hundred seconds to allow plasma density and temperature profiles to relax to their equilibrium state. The plasma current thought to be needed would exceed 20 MA, and would be lower for the steady state operation that would be the ultimate goal for nuclear component testing. This testing was considered to require an average 14 MeV neutron wall loading of about 1 MW/m², and the design should allow for irradiation for up to 3 operating years (i.e a fluence of up to 3 MWam²). A tritium-breeding blanket, possibly not of reactor-relevant design to keep it simpler, would be needed in the high availability (average 25%) technology phase. During that phase one aim would be to operate at very high availability for two weeks.

There were intense discussions on the range of desirable parameters. At the time it was being realised that high plasma currents might be needed to ensure ignition, under the then current understanding of the scaling law for plasma energy confinement. Tight aspect ratio (ration of major to minor radius of the plasma torus) devices were cheaper to build but were always pushing magnetic field limits on the coils and were tight on the ability to drive long plasma pulses inductively. Large aspect ratio devices looked to be better, but the physics database for confinement was very weak, and there were questions on the ease of access for remote maintenance. The kind of parametric survey graph derived is shown below.

Schematic representation of the parameter space for the choice of the ITER design. R = major radius PFUS = fusion power Pn = neutron wall load The names refer to contours for specific confinement time scaling laws. Devices in the top left corner are the cheapest, subject to magnet design limitations. But devices must be chosen sufficiently towards the bottom right to be able to reach high Q or ignition. The type of device chosen depends on the confinement time scaling law that best describes the plasma physics. This graph, produced in 1988, is still valid today

Eventually the parameters shown below were chosen.

The design particularly of the tokamak was then elaborated in some detail (see figure). Estimates were made of the performance capacity of ancillary equipment, and the dimensions and characteristics of the buildings. Safety analysis was carried out, and a preliminary estimate of costs was derived. The last indicated that ITER would cost $4.9B to construct: $1.7B for the tokamak, $1.4B for tokamak auxiliaries, $0.8B for buildings and plant auxiliaries, $0.3B for transport and assembly, with a reserve of $0.7B for cost contingencies. A further $0.8B would be needed for 2900 pmy (professional man-years) estimated to be needed for the construction, with a team of 300 professionals, and a further $0.3B would be needed for R&D during construction. The necessary design and R&D, to be carried out during the EDA, was also elaborated and costed. Design costs both for a central team of up to 180 professionals, and in the Parties home teams, was estimated in total at 1200 pmy ($0.25B). Technology R&D both for basic technology and for specific engineering (prototypes, etc.) was estimated to cost $0.75B.

The ITER CDA Tokamak Design

The schedule foresaw selection of the EDA site by April 1991, an EDA of 6 years, with construction site decision by the start of 1996, and a start of the construction phase of 8 years at the beginning of 1997.

The conceptual design was documented in a series of reports published by the IAEA.

The CDA confirmed the Parties' common view of the overall programmatic and technical objectives for a next step machine. In technical terms, the collaboration started to orientate the fusion research efforts of the various Parties towards a common goal. In addition, the process of the CDA and the successful addressing of organisational and human issues gave the Parties confidence that the project could move to its next stage of engineering design activities as an international project under the terms of an inter-governmental collaborative agreement.