A Spatial Point Process-based Approach for Dropout Events Modeling in High-Temperature Superconductor Manufacturing

Recently, high-temperature superconductor (HTS) tapes have shown promising properties of high critical current, which are prerequisites for applications in high-field magnets. However, due to the unstable growth conditions in the HTS manufacturing process, dropout events, which refer to the significant drops of critical current at certain locations of HTS tapes, occur frequently resulting in the non-uniform performance of produced tapes. To produce HTS tapes with large scale, high yield and uniform performance, it is important to develop novel data analytic approaches focusing on modeling the dropout events and their associated process parameters. In this study, we develop a spatial point process-based framework that integrates the Granger causality-based feature selection, principle component analysis-based data fusion, and dropout events modeling based on a nonhomogeneous Poisson process. The proposed method is applied and evaluated on real data from HTS tapes, and the important process parameters associated with dropout occurrences are identified, e.g., substrate temperatures.