AD ecooler Electromagnetic models Mechanical considerations Hugo Bajas

AD e-cooler Electromagnetic models Mechanical considerations Hugo Bajas 13 -jan-2021

Let us focus on the mechanics today 12/13/2021 H. BAJAS | AD e-cooler modelling 2

Lorentz forces retrieved from EM modelling Integration of the Biot-Savart gives only an average value for the force… ZUT! 1 7 13 A solenoid is balanced in terms of forces in the X and Y directions Along the coil axis, a solenoid is slightly in compression. Unfortunately, as modelled, the hoop stress is not an OUTPUT ! Need to mesh the conductor 12/13/2021 H. BAJAS | AD e-cooler modelling 3

Outline • OPERA allows coupled physics computations (EM-Thermo-mechanics)! Let us do it! • The material properties (Young Modulus, Poisson coeff. , density) entered as INPUT. • The Lorentz forces from EM computation (Biot-Savart integration) can be transferred into a mechanical model. • The stress-strain state of the various materials can be retrieved from the mechanical simulation. • In order to compute the s-e state of the coils, these have to be meshed. (The iron is already mesh). • The way the e-cooler model is built (face extrusion + morph) allows to mesh the conductor quite regularly. • Let us focus on a simple model for the drift solenoid. 12/13/2021 H. BAJAS | AD e-cooler modelling 4

Coupled Physics Simple model of solenoid to start the analysis with. flanges yoke coils 12/13/2021 H. BAJAS | AD e-cooler modelling 5

ELENA e-cooler EM simulation: main outcome Thanks to the control of the mesh, the coil can be meshed easily. (another would be to use BRICKS to create the coil) 12/13/2021 H. BAJAS | AD e-cooler modelling 6

Von Mises stress and Displacement vectors: COILS The coils hoop stress tends to “explode” outward the coils (very few of the stress in the axial direction). 12/13/2021 H. BAJAS | AD e-cooler modelling 7

Von Mises stress and Displacement vectors: YOKE The iron shell is in tension as well (hoop stress). 12/13/2021 H. BAJAS | AD e-cooler modelling 8

Von Mises stress and Displacement vectors Giving the present dimension, the applied stress are quite low. 12/13/2021 H. BAJAS | AD e-cooler modelling 9

Von Mises stress and Displacement vectors Giving the present dimension, the applied stress are quite low. 12/13/2021 H. BAJAS | AD e-cooler modelling 10

Von Mises stress and Displacement vectors Giving the present dimension, the applied stress are quite low. The axial forces are quite impacted by the shape and distance 12/13/2021 H. BAJAS | AD e-cooler modelling 11

Von Mises stress and Displacement vectors The axial forces are quite impacted by the shape and distance Aperture diameter Distance to the flange 12/13/2021 H. BAJAS | AD e-cooler modelling 12

Need to study other options for optimization 12/13/2021 H. BAJAS | AD e-cooler modelling 13

Conclusion • The Lorentz forces as output from the EM model cannot be exploited as is: • Need to discretize the conductor (block solenoid) or • Need to mesh the conductor • Coupled simulations have been successfully run (for the first time yesterday) for simple case (only the dirft solenoids). • Encouraging results are obtained but very preliminary. • The stress depends on: • the pin points of the structure (now, all d. o. f blocked on the flanges surface); i. e. where are the supporting points • the location of the iron (flange shape, distance to coil) • Obisously the size of the copper conductor (now quite big) lower the stress; conservative? • The biggest force in clearly in the Expansion solenoid (k. N)… work in progress. • The work on optimization of the field uniformity is still on-going. 12/13/2021 H. BAJAS | AD e-cooler modelling 14

Lorentz forces along z for the e-cooler AD without expansion solenoid. 12/13/2021 H. BAJAS | AD e-cooler modelling 15

Lorentz forces along z for the whole e-cooler AD 12/13/2021 H. BAJAS | AD e-cooler modelling 16