Kalman Prototypical Networks for Few-shot Fault Detection in Combined Cycle Gas Turbines

2026-06-25

AI Summary

This study addresses the challenge of effectively detecting faults in combined-cycle gas turbines under conditions of scarce labeled fault data by proposing a metric learning–based few-shot fault diagnosis method termed the Kalman Prototypical Network (KPN). KPN uniquely integrates Kalman filtering into prototypical networks, modeling class prototypes as latent states of a stochastic dynamic system and employing a dynamic update mechanism to reduce embedding variance and enhance prototype stability. Evaluated on high-fidelity Modelica simulation data for a simulated leakage fault detection task, KPN significantly outperforms existing few-shot learning approaches—including Matching Networks, Relation Networks, and MAML—in both accuracy and training stability, thereby improving model convergence and generalization capability.

Abstract

Combined-cycle gas turbines (CCGTs) play a key role in modern power generation, offering both high efficiency and reduced environmental impact. However, their complex thermo-fluid and mechanical interactions complicate fault detection, particularly when labeled fault data are scarce. In this paper, we introduce the Kalman Prototypical Network (KPN), a metric-based few-shot learning (FSL) framework specifically tailored for CCGT fault diagnosis. We model the evolution of class prototypes as latent stochastic states in a dynamic system to reduce episodic variance and improve robustness in embedding representation. Synthetic data sets generated with a high-fidelity Modelica-based dynamic simulation of an offshore CCGT system were used, simulating both normal operation and progressive leak faults under transient conditions. Application of the proposed framework on simulated leak fault detection tasks demonstrate that KPN outperforms conventional FSL methods such as Matching Networks, Relation Networks, and MAML in both accuracy and stability under varying support and query configurations. The proposed framework significantly improves training convergence and generalization by stabilizing class representations, making it well-suited for real-world CCGT fault detection where labeled data is limited.

Problem

• fault detection
• combined cycle gas turbines
• few-shot learning
• labeled data scarcity
• complex system diagnostics

Innovation

• Kalman Prototypical Network
• few-shot learning
• fault detection
• dynamic system modeling
• prototype stabilization