Data based approach for magnet design

Magnets play an important role in particle accelerators used for various nuclear physics experiments, most notably at CERN. The magnets ensure that the particles follow the desired path during operation. Therefore it becomes important to have the necessary Magnetic Flux Density prole, i.e. the variation of Magnetic Flux Density along the considered coordinates. Although it is straight forward to obtain the Magnetic Flux Density profile given the parameters of the magnet, it still remains a challenge to obtain the parameter from the given Magnetic Flux Density profile. Such study can help the engineers get some idea of design parameters such as the length and thickness while designing the magnet. Machine learning algorithms have been used here to detect patterns between the design parameters of the magnet and its respective Magnetic field values, and further try to see the accuracy to which the parameter can be predicted using the data about the magnetic field