Much research in recent years has focused on using empirical machine learning approaches to extract useful insights on the structure-property relationships of superconductor material. Notably, these approaches are bringing extreme benefits when superconductivity data often come from costly and arduously experimental work. However, this assessment cannot be based solely on an open black-box machine learning, which is not fully interpretable, because it can be counterintuitive to understand why the model may give an appropriate response to a set of input data for superconductivity characteristic analyses, e.g., critical temperature. The purpose of this study is to describe and examine an alternative approach for predicting the superconducting transition temperature Tc from SuperCon database obtained by Japan's National Institute for Materials Science. We address a generative machine-learning framework called Variational Bayesian Neural Network using superconductors chemical elements and formula to predict Tc. In such a context, the importance of the paper in focus is twofold. First, to improve the interpretability, we adopt a variational inference to approximate the distribution in latent parameter space for the generative model. It statistically captures the mutual correlation of superconductor compounds and; then, gives the estimation for the Tc. Second, a stochastic optimization algorithm, which embraces a statistical inference named Monte Carlo sampler, is utilized to optimally approximate the proposed inference model, ultimately determine and evaluate the predictive performance. As a result, in comparison with the standard evaluation metrics, the results are promising and also agree with the existing models prevalent in the field. The R 2 value obtained is very close to the best model (0.94), whereas a considerable improvement is seen in the RMSE value (3.83 K). Notably, the proposed model is known as the first of its kind for predicting a superconductor's Tc.Index Terms-Critical transition temperature, machine learning, Bayesian neural network, variational inference, stochastic optimization algorithm, high temperature superconducting (HTS).

No abstract

In this paper, the general electromagnetic design process of a 10-MW-class high-temperature superconducting (HTS) synchronous generator that is intended to be utilized for large scale offshore wind generator is discussed. This paper presents three-dimensional (3D) electromagnetic design proposal and electrical characteristic analysis results of a 10-MW-class HTS synchronous generator for wind power. For more detailed design by reducing the errors of a two-dimensional (2D) design owing to leakage flux in air-gap, we redesign and analyze the 2D conceptual electromagnetic design model of the HTS synchronous generator using 3D finite element analysis (FEA) software. Then electrical characteristics which include the no-load and full-load voltage of generator, harmonic contents of these two load conditions, voltage regulation and losses of generator are analyzed by commercial 3D FEA software.

A cooling system is an essential part of hightemperature superconducting (HTS) rotating machine manufacturing. Moreover, thermal behavior is a crucial parameter of the cooling system that shows unique characteristics of superconductivity below a specific temperature to maintain a superconducting state. Therefore, many experiments have been performed to investigate new reliable cryogenic cooling systems for a large-scale HTS rotating machine. The motivation for the study is our recent development of the cryogenic cooling system using thermal trigger switches; it effectively minimizes non-operational downtime of the HTS machine in cases of power supply or cryocooler failure. This paper focuses on two main targets. First, the thermal design of the cooling system for the 10 MW-class HTS rotating machine is enhanced. Second, the performance of the cooling system is observed for various cryogens to investigate the feasibility of using solid cryogen.

No abstract