The nuclear fusion devices using the principle of magnetic plasma confinement such as the ITER tokamak are going to consist of a variety of highly heterogeneous, nuclear-radiation-sensitive components. The compactness of the ITER tokamak makes it difficult to rely on large safety margins. Under these circumstances the use of reasonably heterogeneous, highly precise models for the nuclear analysis is going to be unavoidable. Techniques have been developed to create these models based as directly as possible on computer-aided design (CAD) specifications, thereby retaining fidelity and speeding up the process. Inevitably, some adaption of the CAD model is necessary as part of the conversion process.

This paper describes the approach to the production of the models for nuclear analysis for ITER developed by the neutronics group in the ITER Organization. Algorithmization of the CAD-based modeling for MCNP code has been undertaken.