Efficient approach to simulate EM loads on massive structures in ITER machine

Operation of the ITER machine is associated with high electromagnetic (EM) loads. An essential contributor to EM loads is eddy currents induced in passive conductive structures. Reasoning from the ITER construction, a modelling technique has been developed and applied in computations to efficiently predict anticipated loads. The technique allows us to avoid building a global 3D finite-element (FE) model that requires meshing of the conducting structures and their vacuum environment into 3D solid elements that leads to high computational cost. The key features of the proposed technique are: (i) the use of an existing shell model for the system "vacuum vessel (VV), cryostat, and thermal shields (TS)" implementing the magnetic shell approach. A solution is obtained in terms of a single-component, in this case, vector electric potential taken within the conducting shells of the "VV + cryostat + TS" system. (ii) EM loads on in-vessel conducting structures are simulated with the use of local FE models. The local models use either the 3D solid body or shell approximations. Reasoning from the simulation efficiency, the local boundary conditions are put with respect to the total field or an external field. The use of an integral-differential formulation and special procedures ensures smooth and accurate simulated distributions of fields from current sources of any geometry. The local FE models have been developed and applied for EM analyses of a variety of the ITER components including the diagnostic systems, divertor, test blanket modules, cryopumps, blanket modules. (iii) Two integration algorithms can be applied to an ordinary differential equation system (ODES) describing a discrete problem. First, a direct integration of ODES can be performed in accordance with operating scenarios (variations of field sources). Second, complex variations of field sources can be decomposed for each source into individual components via a set of basic (influence) functions. A generalized solution is obtained as a superposition of individual solutions. (iv) The use of a combination of different computer codes implementing the shell models and 3D solid-body models. The codes and developed models were validated and approved, particularly, in the course of an ITER-initiated extensive benchmark to support of the blanket modules design. (C) 2013 Elsevier B.V. All rights reserved.