Energy Digest
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Summary
Technical Papers & Research
AI-curated academic research for power system engineers
Other 2 papers
The paper establishes a connection between threshold-based event encoding and discretized learning systems, identifying a dimensionless ratio governing information retention, and derives an information-sufficiency boundary using normal-condition data. It also formulates a data-driven criterion that selects the sparsest threshold retaining a prescribed fraction of peak discriminative information. The proposed design improves power-system fault classification performance by identifying sparse operating thresholds.
A physics-constrained autoencoder is proposed to learn interpretable health representations in turbofan engines, enabling component-level analysis and diagnostic value. The model incorporates physical equations in the decoder architecture and learns from a loss function that promotes the representation of predefined physical quantities. The methodology achieves comparable reconstruction performance to fully data-driven models and demonstrates its effectiveness in tracking degradation and producing a health index with high correlation to ground truth.
Was this digest helpful?