Answers are grounded in this paper and cite their sources.
The practical deployment of reacting flow modeling in industrial design processes is largely reliant on computational efficiency and accuracy. The Flamelet-Generated Manifold (FGM) approach is one of the most popular modeling methods developed to address these requirements. The FGM model separates the reaction zone from the bulk flow and simplifies the reaction space by defining reaction scalars as functions of fewer parameters, such as the mixture fraction, progress variable, and variances. However, the single-mixture-fraction-based FGM model is not applicable to dual-fuel applications due to inherent formulation limitations. This study presents an extension of the FGM model to accommodate dual-fuel applications (i.e. three streams: two fuel streams and one oxidizer stream) and to replace the Probability Density Function (PDF) table with an Artificial Neural Network (ANN) model. This modification significantly reduces the memory requirements for the tabulated data, enabling higher manifold refinements across the controlling parameter space. The proposed model is used to simulate the KAIST multi-tube combustor operating under varying fuel-composition conditions. The model results are compared with experimental data and the detailed chemistry simulations to assess the accuracy and limitations of the proposed method. Similar predictions are obtained between the PDF and ANN-based approaches, demonstrating the effectiveness of the proposed method in efficiently handling complex industrial cases with higher resolution.
No takes yet. Share an insight, caveat, or question.
Shrivastava et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: