Technology Category
Machine Learning: Calibration & Uncertainty QuantificationReasoning under Uncertainty: Uncertainty RepresentationsKnowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
Fuzzy Decisions on Fluid Instabilities: Autoencoder-Based Reconstruction meets Rule-Based Anomaly Classification
🤖 AI Summary
Addressing the dual challenges of scarce labeled data and complex flow-field structures in shockwave classification for shadowgraph imaging, this paper proposes an unsupervised, interpretable classification framework integrating a β-VAE autoencoder with a fuzzy inference system. Methodologically, the β-VAE learns low-dimensional flow-field representations and reconstructs anomaly maps, enhanced by spatial contextual constraints to improve structural sensitivity; concurrently, a fuzzy rule engine maps latent-space features to physically interpretable shock criteria, ensuring mechanical consistency of diagnostic outcomes. Experiments demonstrate accurate identification of coherent shock structures without labeled data and high-precision anomaly classification across canonical flow instability scenarios. This work establishes the first paradigm for shock diagnostics in experimental fluid dynamics that simultaneously satisfies three critical requirements: unsupervised learning, human-interpretable reasoning, and physical consistency—enabling real-time, robust intelligent monitoring of industrial flow fields.
Technology Category
Application Category
📝 Abstract
Shockwave classification in shadowgraph imaging is challenging due to limited labeled data and complex flow structures. This study presents a hybrid framework that combines unsupervised autoencoder models with a fuzzy inference system to generate and interpret anomaly maps. Among the evaluated methods, the hybrid $β$-VAE autoencoder with a fuzzy rule-based system most effectively captured coherent shock features, integrating spatial context to enhance anomaly classification. The resulting approach enables interpretable, unsupervised classification of flow disruptions and lays the groundwork for real-time, physics-informed diagnostics in experimental and industrial fluid applications.
Problem
Research questions and friction points this paper is trying to address.
Classify shockwaves in shadowgraph images with limited labeled data
Combine autoencoder and fuzzy inference for anomaly detection
Enable interpretable unsupervised classification for fluid flow disruptions
Innovation
Methods, ideas, or system contributions that make the work stand out.
Hybrid autoencoder with fuzzy inference system
Unsupervised anomaly classification using β-VAE
Physics-informed diagnostics for fluid applications
🔎 Similar Papers
No similar papers found.
