Physics-Informed Hybrid Anomaly Detection and Adaptive Power Reconfiguration for Spacecraft Electrical Power Systems

The spacecraft electrical power system (EPS) is the most failure-critical on-orbit subsystem, while its connecting cable network is an often-overlooked source of degradation. Most methods monitor the two separately and decouple detection from corrective power management. This paper proposes a physics-informed hybrid framework jointly monitoring the EPS power link and cable network and closing the loop with adaptive reconfiguration. A simulator generates 36-channel telemetry over eight operating modes, in which a harness thermal model produces wire-temperature, contact hot-spot, and voltage-drop residuals invisible to power-only sensing. A dual-residual detector fuses a mode-conditioned physics residual with an autoencoder reconstruction residual via an OR rule and three-sample persistence. On 102 episodes spanning 17 fault types (11 power-side, 6 cable-side), it reaches a recall of 0.98 (100/102) with zero false alarms on held-out nominal data, matching an autoencoder-only detector, whereas the physics-only and rule baselines reach 0.51 and 0.47; a random forest attributes the fault class (accuracy 0.465, macro-F1 0.448). Under a 28% under-supply fault, minimal-load MPC reconfiguration retains 0.720 of load power versus 0.636 for rule-based shedding and near-zero in open loop. Joint harness-aware monitoring with minimal reconfiguration offers a simulation-validated basis for autonomous EPS health management, pending cross-simulator and hardware-in-the-loop confirmation.

Authors

Institutions

Publication Details

Journal
Applied Sciences
Published
2026-11-16
DOI
https://doi.org/10.3390/app162211362
Primary Topic
Fault Detection and Control Systems
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Physics-Informed Hybrid Anomaly Detection and Adaptive Power Reconfiguration for Spacecraft Electrical Power Systems

Jian Li, Ye Wang, Ze Zheng, Lili Yuan et al.
2 links
Applied Sciences
Fault Detection and Control Systems
article

Physics-Informed Hybrid Anomaly Detection and Adaptive Power Reconfiguration for Spacecraft Electrical Power Systems

Jian Li, Ye Wang, Ze Zheng, Lili Yuan, Kang Li
article en

Abstract

The spacecraft electrical power system (EPS) is the most failure-critical on-orbit subsystem, while its connecting cable network is an often-overlooked source of degradation. Most methods monitor the two separately and decouple detection from corrective power management. This paper proposes a physics-informed hybrid framework jointly monitoring the EPS power link and cable network and closing the loop with adaptive reconfiguration. A simulator generates 36-channel telemetry over eight operating modes, in which a harness thermal model produces wire-temperature, contact hot-spot, and voltage-drop residuals invisible to power-only sensing. A dual-residual detector fuses a mode-conditioned physics residual with an autoencoder reconstruction residual via an OR rule and three-sample persistence. On 102 episodes spanning 17 fault types (11 power-side, 6 cable-side), it reaches a recall of 0.98 (100/102) with zero false alarms on held-out nominal data, matching an autoencoder-only detector, whereas the physics-only and rule baselines reach 0.51 and 0.47; a random forest attributes the fault class (accuracy 0.465, macro-F1 0.448). Under a 28% under-supply fault, minimal-load MPC reconfiguration retains 0.720 of load power versus 0.636 for rule-based shedding and near-zero in open loop. Joint harness-aware monitoring with minimal reconfiguration offers a simulation-validated basis for autonomous EPS health management, pending cross-simulator and hardware-in-the-loop confirmation.

Applied SciencesVol. 16(22)
China Academy of Space Technology (CN)
Industry, innovation and infrastructure
Fault Detection and Control Systems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon