Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Smart Fault Detection in Nanosatellite Electrical Power System

Alireza Rezaee, Niloofar Nobahari, Amin Asgarifar, Farshid Hajati

TL;DR

This work addresses fault detection in nanosatellite electrical power systems operating without ADCS in LEO, focusing on faults in the photovoltaic subsystem, MPPT/DC-DC converters, and battery regulation. It proposes a residual-based neural-network framework, using inputs tied to solar radiation and PV temperature to model outputs and detect faults, complemented by PCA, KNN, and decision-tree classifiers for fault categorization. The photovoltaic faults are classified with >99% accuracy via an MLP, and system-wide fault classification reaches 100% accuracy when the first moment of load current is added as a feature, with PCA also performing at a high level. The results indicate a viable, computation-efficient fault-detection approach suitable for nanosatellite power systems, potentially enhancing reliability and mission success in constrained environments.

Abstract

This paper presents a new detection method of faults at Nanosatellites' electrical power without an Attitude Determination Control Subsystem (ADCS) at the LEO orbit. Each part of this system is at risk of fault due to pressure tolerance, launcher pressure, and environmental circumstances. Common faults are line to line fault and open circuit for the photovoltaic subsystem, short circuit and open circuit IGBT at DC to DC converter, and regulator fault of the ground battery. The system is simulated without fault based on a neural network using solar radiation and solar panel's surface temperature as input data and current and load as outputs. Finally, using the neural network classifier, different faults are diagnosed by pattern and type of fault. For fault classification, other machine learning methods are also used, such as PCA classification, decision tree, and KNN.

Smart Fault Detection in Nanosatellite Electrical Power System

TL;DR

This work addresses fault detection in nanosatellite electrical power systems operating without ADCS in LEO, focusing on faults in the photovoltaic subsystem, MPPT/DC-DC converters, and battery regulation. It proposes a residual-based neural-network framework, using inputs tied to solar radiation and PV temperature to model outputs and detect faults, complemented by PCA, KNN, and decision-tree classifiers for fault categorization. The photovoltaic faults are classified with >99% accuracy via an MLP, and system-wide fault classification reaches 100% accuracy when the first moment of load current is added as a feature, with PCA also performing at a high level. The results indicate a viable, computation-efficient fault-detection approach suitable for nanosatellite power systems, potentially enhancing reliability and mission success in constrained environments.

Abstract

This paper presents a new detection method of faults at Nanosatellites' electrical power without an Attitude Determination Control Subsystem (ADCS) at the LEO orbit. Each part of this system is at risk of fault due to pressure tolerance, launcher pressure, and environmental circumstances. Common faults are line to line fault and open circuit for the photovoltaic subsystem, short circuit and open circuit IGBT at DC to DC converter, and regulator fault of the ground battery. The system is simulated without fault based on a neural network using solar radiation and solar panel's surface temperature as input data and current and load as outputs. Finally, using the neural network classifier, different faults are diagnosed by pattern and type of fault. For fault classification, other machine learning methods are also used, such as PCA classification, decision tree, and KNN.
Paper Structure (12 sections, 12 equations, 10 figures, 5 tables)

This paper contains 12 sections, 12 equations, 10 figures, 5 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Materials and Methods
  4. Electrical power system modeling
  5. Faulty electrical power system modeling
  6. Photovoltaic subsystem faults calcification with neural network MLP
  7. Electrical power system fault classification with neural network MLP
  8. Fault classification with another intelligence method
  9. KNN method
  10. Decision tree
  11. PCA method
  12. Conclusion

Figures (10)

  • Figure 1: Proposed fault diagnosis model.
  • Figure 2: System output coherent and its estimated amount, the output electrical power system without fault and its estimated amount diagram, estimated fault histogram and amount of fault square average, and the average of fault square based on samples.
  • Figure 3: Current of photovoltaic system output coherent and its estimated amount, output diagram and its estimated amount, estimated fault histogram, and amount of fault square average and average of fault square based on samples.
  • Figure 4: Voltage of photovoltaic system output coherent and its estimated amount, output diagram and its estimated amount, estimated fault histogram, and amount of fault square average and an average of fault square based on samples.
  • Figure 5: Neural network classifier confusion matrix MLP for identifying fault of a photovoltaic system.
  • ...and 5 more figures