Avionic Main Fuel Pump Simulation and Fault-Diagnosis Benchmark

2026-04-23

AI Summary

This study addresses the scarcity of training data for fault diagnosis in safety-critical cyber-physical systems such as aviation, which stems from data privacy constraints and partial observability. To this end, the authors develop a high-fidelity, physics-based co-simulation model of an aircraft main fuel pump using MATLAB/Simulink Simscape Fluids and, for the first time, publicly release a time-series simulation benchmark dataset annotated with both healthy and fault modes. By integrating an RNN-VAE for unsupervised anomaly detection and a SOM-VAE for operational condition discretization, the experimental results effectively distinguish between healthy and faulty states. The findings validate the feasibility and practical utility of the proposed dataset in supporting data-driven fault diagnosis research.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models

Application Category

• Security and Privacy: Data transparency and provenance
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Web Mining and Content Analysis: Web data generation and simulation

Abstract

In many cyber-physical systems, especially in critical applications such as aeroplanes, data to train anomaly detection and diagnosis algorithms is lacking due to data protection issues and partial observability. To combat this inherent lack of data, we introduce a high-fidelity, physics-informed co-simulation of a common aircraft main-fuel-pump system modelled in \textsc{MATLAB/Simulink Simscape Fluids}. We also describe its generated time-series data with health and fault mode annotations. To show feasibility of our benchmark, we apply an unsupervised Recurrent Variational Autoencoder (RNN-VAE) for anomaly detection and a SOM-VAE for operating mode discretization, trained to separate healthy and faulty conditions.

Problem

• cyber-physical systems
• anomaly detection
• fault diagnosis
• data scarcity
• avionic fuel pump

Innovation

• physics-informed co-simulation
• fault-diagnosis benchmark
• anomaly detection
• RNN-VAE
• SOM-VAE