Balancing Safety and Efficiency in Aircraft Health Diagnosis: A Task Decomposition Framework with Heterogeneous Long-Micro Scale Cascading and Knowledge Distillation-based Interpretability

2026-03-24

AI Summary

This work addresses the challenges of data uncertainty, task heterogeneity, and computational inefficiency in whole-aircraft health diagnostics for general aviation. Existing end-to-end approaches struggle to jointly model global context and local features while incurring high training costs under severe class imbalance. To overcome these limitations, the authors propose a Diagnostic Decoupling Framework (DDF), which explicitly decomposes the diagnostic task into anomaly detection and fault classification subtasks for the first time. The framework employs ConvTokMHSA for long-range global screening and MMK Net for fine-grained local diagnosis, complemented by a knowledge distillation–based saliency extraction layer to enhance interpretability. Evaluated on the real-world NGAFID dataset, DDF improves multi-class weighted penalty metrics (MCWPM) by 4–8% over baseline methods while significantly reducing training time, thereby achieving a unified balance of efficiency, adaptability, and physical interpretability.

Technology Category

• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Machine Learning: Calibration & Uncertainty Quantification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Whole-aircraft diagnosis for general aviation faces threefold challenges: data uncertainty, task heterogeneity, and computational inefficiency. Existing end-to-end approaches uniformly model health discrimination and fault characterization, overlooking intrinsic receptive field conflicts between global context modeling and local feature extraction, while incurring prohibitive training costs under severe class imbalance. To address these, this study proposes the Diagnosis Decomposition Framework (DDF), explicitly decoupling diagnosis into Anomaly Detection (AD) and Fault Classification (FC) subtasks via the Long-Micro Scale Diagnostician (LMSD). Employing a "long-range global screening and micro-scale local precise diagnosis" strategy, LMSD utilizes Convolutional Tokenizer with Multi-Head Self-Attention (ConvTokMHSA) for global operational pattern discrimination and Multi-Micro Kernel Network (MMK Net) for local fault feature extraction. Decoupled training separates "large-sample lightweight" and "small-sample complex" optimization pathways, significantly reducing computational overhead. Concurrently, Keyness Extraction Layer (KEL) via knowledge distillation furnishes physically traceable explanations for two-stage decisions, materializing interpretability-by-design. Experiments on the NGAFID real-world aviation dataset demonstrate approximately 4-8% improvement in Multi-Class Weighted Penalty Metric (MCWPM) over baselines with substantially reduced training time, validating comprehensive advantages in task adaptability, interpretability, and efficiency. This provides a deployable methodology for general aviation health management.

Problem

• data uncertainty
• task heterogeneity
• computational inefficiency
• class imbalance
• aircraft health diagnosis

Innovation

• Task Decomposition
• Long-Micro Scale Cascading
• Knowledge Distillation
• Interpretability-by-Design
• Heterogeneous Diagnosis