Power systems, digital twin, data quality, prognostics and health management

The energy transition is accelerating the digitalization of power systems, now densely instrumented, connected and increasingly mirrored by digital twins. Yet a twin is only as trustworthy as the data that feed it, and that data is rarely characterized beyond cleaning. Power-system data are heterogeneous and multi-scale, affected by noise, missing values, sensor drift and distribution shifts, and scarce precisely where it matters most: rare and faulty operating conditions.

This session positions data characterization as the foundation of digital twins. The central question is not how much data is available, but whether it can support reliable learning. We invite contributions on measuring data quality and its impact on model performance, assessing representativeness, and quantifying uncertainty. Degradation of sensors and devices, and its separation from model drift, is a key theme, as are out-of-distribution detection and physics-constrained synthetic data generation. We also welcome work on validating virtual representations against experimental data. Physics-based, data-driven and hybrid approaches are equally encouraged.

The goal is to bridge power electronics, smart grids, industrial artificial intelligence (AI), and prognostics and health management, and to align digital-twin development with emerging requirements for trustworthy, certifiable hybrid physics-AI approaches.

• Data quality assessment and characterization for power electronics and power system measurements
• Multi-scale, multi-source data fusion (electrical, thermal, mechanical and environmental signals)
• Synthetic data generation, accelerated ageing tests and data augmentation
• Verification, validation and benchmarking of digital twins against experimental data
• Digital twin architectures for power converters, drives, energy storage and grids
• Data pre-processing: denoising, missing data, sensor drift, feature selection and model-order reduction
• Prognostics and health management (PHM) of power electronic components and energy storage