A Structured Neural ODE Approach for Real Time Evaluation of AC Losses in 3D Superconducting Tapes

2025-10-16

AI Summary

Real-time evaluation of AC losses in high-temperature superconductors (HTS) remains challenging due to the prohibitive computational cost of full-order electromagnetic simulations, stemming from strong material nonlinearity. Traditional model reduction techniques—such as Proper Orthogonal Decomposition with Discrete Empirical Interpolation Method (POD-DEIM)—are intrusive and rely heavily on numerous interpolation points, limiting their applicability. Method: This work proposes a novel non-intrusive framework that integrates POD-DEIM with the Integral Equation Method (IEM) and introduces a structured Neural Ordinary Differential Equation (Neural ODE) architecture. The Neural ODE directly learns the nonlinear dynamics in the low-dimensional reduced subspace, eliminating explicit interpolation and avoiding modifications to the underlying physical model. Results: Experimental validation on 3D HTS tape electromagnetic response prediction demonstrates substantial improvements in both accuracy and generalization over conventional POD-DEIM, alongside superior computational efficiency—enabling practical real-time HTS simulation.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Hardware-aware ML
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Energy management for devices in mobile Web and WoT environments
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

Efficient modeling of High Temperature Superconductors (HTS) is crucial for real-time quench monitoring; however, full-order electromagnetic simulations remain prohibitively costly due to the strong nonlinearities. Conventional reduced-order methods, such as the Proper Orthogonal Decomposition (POD) and Discrete Empirical Interpolation Method (DEIM), alleviate this cost but are limited by intrusive implementation and by the need for many interpolation points. This work investigates reduced-order strategies for Integral Equation Method (IEM) of HTS systems. We present the first application of POD-DEIM to IEM-based HTS models, and introduce a Structured Neural Ordinary Differential Equation (Neural ODE) approach that learns nonlinear dynamics directly in the reduced space. Benchmark results show that the Neural ODE outperforms POD-DEIM in both efficiency and accuracy, highlighting its potential for real-time superconducting simulations.

Problem

• Real-time AC loss evaluation in 3D superconducting tapes
• Reducing computational cost of electromagnetic HTS simulations
• Developing efficient reduced-order models for nonlinear dynamics

Innovation

• Structured Neural ODE learns nonlinear dynamics
• POD-DEIM applied to Integral Equation Method
• Reduced-order approach enables real-time simulation