Selectel and ITMO JV develops multi-agent system technologies for business

Bayesian thermal digital twin for space habitat under impact

A Bayesian thermal digital twin for space habitats was developed to model impact-induced thermal anomalies. It uses a thermal resistance-capacitance network with physics-based activation functions for model selection and health estimation, enhancing resilience of life support systems.

Data-driven transient forecasting for generator digital twin decision support

A calibrated surrogate model for forecasting generator transients was developed to support digital twin decision-making. The model evaluates planned load commands before application using event-conditioned Hankel-DMDc and split-conformal calibration for prediction bands, helping operators ensure voltage and frequency remain within limits.

Seed-Anchored Graph Rendering for QA on Power-Grid CIM/CGMES Models