Corporation de l'École Polytechnique de Montréal

Published purpose

Ceramic high temperature superconductors (HTS) are now routinely fabricated in the form of long lengths of tapes (several km), making it possible to envision large-scale superconducting applications. One issue concerns the physical properties of those HTS tapes, which can vary over length scales shorter than 1 millimeter, and lead to destructive hot spots in some circumstances. A strategy to find a solution to this problem at the design stage is to simulate the complete transient electrothermal behavior of long lengths of HTS tapes in real operating conditions while retaining their microscopic features. However, time transient simulations requires tremendous calculation times. AI techniques has proven useful for speeding up numeric simulations, but there are still open problems from the point of how to best use AI approaches due to the diversity and complexity of the physical systems object of the simulation. Because of its specificities, this particular application poses several challenges to the introduction of AI techniques, which makes it suitable from the point of view of both core AI research and material science. The AI solutions would benefit similar problems in other domains and the improvements in the design of superconductive tapes will find applications in many important areas, where high field magnets are required.