End-user: Laboratoire National des Champs Magnétiques Intenses (LNCMI)

The LNCMI is a French large scale facility [12] also part of the European Magnetic Field Laboratory (EMFL), enabling researchers to perform experiments in the highest possible magnetic field (up to 35 T static field provided by water cooled resistive magnets connected with a 24 MW power supply). Magnets are accessible to the international scientific community through project calls twice a year. Studies range from solid physics to applied superconductivity and magneto-science. In strong international competition driven by NHMFL (USA), and with the emergence of magnet labs in China, the LNCMI has entered the race for higher magnetic field. To keep up in this context, magnet technologies have to be pushed to their limits, both in terms of materials (active research is carried out to have materials - either resistive or superconductor - with improved mechanical and electrical properties) and of design methods.

From an engineering point of view designing, such high field magnet reaches the limits of our current methodology and the models upon which it relies. In particular it raises questions about the model precision, from a pure numerical point of view and from the model itself: is the physics considered sufficient to correctly represent the observed phenomena. On top of that, to guarantee the requested homogeneity it is mandatory to account for geometrical uncertainties, slight plays and mechanical clearances. Moreover material properties and operating parameters uncertainties should be accounted for to assure a robust design.

HL-31-cadgeom.png H1H8_B3D.png

Figure 5. Left: View of a PolyHelix Magnet Insert (a quarter of the structure has been removed to give better insight of the considered geometry). Right: Magnetic Field B produced by a typical PolyHelix magnet Insert. On the right plots of specific B components on the low pressure side of the magnet.

The HIFIMAGNET pilot application [13,14] has been developed in the frame of a collaboration between LNCMI and Institut de Recherche Mathématique Avancée (IRMA) from Unistra to address these questions.  HIFIMAGNET consists of: (i) a set of numerical models ranging from 2D to 3D, including more and more physics, and (ii) a framework for sensitivity analysis and uncertainty quantification with respect to material properties, operational parameters and geometry, that aims to complement LNCMI standard design. This framework relies on Feel++ Reduced Basis Method facility.

The pilot consists will be evaluated through 3 test cases. The first two tests are designed to show the TRL level of Feel++/HiFiMagnet. The 3rd test case is more challenging, and demonstrates the potential of HiFiMagnet for robust design optimization.

Test Case 1: An existing polyhelix magnet will be modelled in operation, at low and high field using a 3D non-linear multi-physics model.  The simulations will be validated against experimental results, more precisely with magnetic field measurements in the zone of interest for researchers. Depending on the availability of other measurements, such as the temperature at the low-pressure side of the magnets, which are currently under development more validations can be performed on the cooling model and the global thermal behaviour of the magnets.


Figure 6. Left: Experimental Setup for Low field Measurements on a Workbench. Right: Experimental Setup for Low field Measurements in Situ.


Figure 7. Left Magnet Field profile along circles of increasing radius at a given altitude for a test magnet of 2 helices: comparison between measurements and different numerical models.

Test Case 2: The commissioning of a magnet is performed when a new magnet is first set into operation. It consists in measuring the resistance of each helix or pair of helices, as a function of the total current and the mean temperature of the coolant. This data is then used to control and monitor the magnet. A deviation of more than 3% from the expected resistance is a signal for power shutdown. This test case will involve 3D non-linear multi-physics simulations, and also a reduced model for parametric studies. The result will be compared with experimental data measured for an existing magnet.

Test Case 3: The goal of this test case is to design or re-design part of an existing magnet to reach a more homogeneous magnetic field within a small volume around the magnet centre. This kind of magnet is of special interest for the community of RMN researcher, and could be a booster for some applications.