A Satellite Incipient Fault Detection Method Based on Decomposed Kullback-Leibler Divergence.

Area of Science:

• Satellite engineering
• Fault detection and diagnostics
• Signal processing

Background:

• Early fault detection is crucial for satellite service availability and continuity.
• Existing methods using local optimum projection vector and Kullback-Leibler (KL) divergence are effective but computationally intensive.
• High time complexity hinders real-time application of advanced fault detection techniques.

Purpose of the Study:

• To develop a computationally efficient method for detecting incipient faults in satellite systems.
• To reduce the time complexity associated with Kullback-Leibler (KL) divergence in fault detection models.
• To enhance the sensitivity of fault detection to subtle anomalies in satellite operational data.

Main Methods:

• Decomposition of the Kullback-Leibler (KL) divergence within the optimization model.
• Application of the generalized Rayleigh quotient property to reduce computational complexity.
• Establishment of two distribution models (for subfunctions F1(w) and F3(w)) to detect anomalies in mean and covariance.
• Validation using numerical simulations and a real satellite fault case.

Main Results:

• The proposed method significantly reduces computational complexity compared to traditional approaches.
• Demonstrated high sensitivity in detecting slight anomalous behavior indicative of incipient faults.
• Effectiveness validated through both simulated and real-world satellite fault scenarios.
• Achieved a balance between computational efficiency and detection accuracy.

Conclusions:

• The novel approach offers a practical and efficient solution for incipient fault detection in satellite systems.
• Reduced time complexity makes the method suitable for real-time monitoring and enhances satellite service reliability.
• The method's high sensitivity to minor deviations improves early warning capabilities for potential satellite failures.