Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
This study develops a reinforcement learning (RL) framework to optimize liquid-fueled reactor designs for predicting lean blowout (LBO) in gas turbine combustors, addressing limitations of traditional heuristic-based clustering methods. By employing a multi-stage clustering-classification strategy, the framework generates micro-clusters and refines them through an actor-critic RL agent to enhance prediction accuracy and computational efficiency. Validation with a complex Jet-A mechanism reveals that the RL approach significantly improves predictive fidelity and captures LBO trends more effectively than conventional methods, while also accelerating the modeling process.
Reinforcement learning can transform LBO predictions in gas turbine combustors, achieving higher accuracy and faster computations than traditional methods.
This study introduces a reinforcement learning (RL) framework for generating optimal liquid-fueled reactors to improve lean blowout (LBO) predictions in gas turbine combustors. Existing approaches for determining cluster boundaries rely on manual heuristics or distance-based metrics in the input space. In contrast, the proposed method is goal-oriented, explicitly accounting for the target metric (e.g., LBO prediction accuracy) during cluster formation. The framework employs a multi-stage clustering--classification strategy: an initial clustering step (e.g., $k$-means clustering) generates a large set of homogeneous micro-clusters, followed by an actor-critic RL agent that merges them into optimal reactor zones. The validation study, performed using a Jet-A mechanism (119 species, 841 reactions), shows the RL framework offers improved predictive fidelity compared to $k$-means and captures the correct LBO trends, while achieving substantial speedups relative to the high-fidelity computational model. Overall, the RL-driven approach demonstrates strong potential as a computationally efficient reduced-order modeling technique that can complement high-fidelity simulations for rapid design-space exploration.