Active Learning Enhanced Surrogate Modeling of Jet Engines in JuliaSim

2025-01-13

AI Summary

High-fidelity surrogate modeling of turbofan engines remains challenging due to prohibitive computational costs and insufficient accuracy of conventional data-driven approaches. Method: This paper proposes an active learning–driven surrogate modeling framework, the first to integrate uncertainty-aware active learning into the JuliaSim platform. It establishes an error-guided, adaptive sampling and iterative model refinement pipeline that synergistically combines physics-informed modeling, Gaussian process regression (GPR), and adaptive sample augmentation. Contribution/Results: The method significantly improves initial design solution quality and accelerates optimization convergence. Experiments demonstrate a 0.1% relative error in key performance metrics across the full operational envelope—representing an order-of-magnitude accuracy gain over uniform or brute-force sampling—and substantially reduce the number of optimization iterations. This work establishes a scalable, high-fidelity, physics–data hybrid modeling paradigm for rapid design and optimization of complex propulsion systems.

Technology Category

Search and Optimization: Sampling/Simulation-based SearchMachine Learning: Active LearningReasoning under Uncertainty: Stochastic Optimization

Application Category

User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systemsSearch and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Surrogate models are effective tools for accelerated design of complex systems. The result of a design optimization procedure using surrogate models can be used to initialize an optimization routine using the full order system. High accuracy of the surrogate model can be advantageous for fast convergence. In this work, we present an active learning approach to produce a very high accuracy surrogate model of a turbofan jet engine, that demonstrates 0.1% relative error for all quantities of interest. We contrast this with a surrogate model produced using a more traditional brute-force data generation approach.

Problem

• Jet Engine Modeling
• Accuracy Improvement
• Complex System Optimization

Innovation

• Intelligent Learning Method
• Jet Engine Simulation
• Ultra-Accurate Surrogate Model