Abstract

The model order reduction of thermoacoustic systems, consisting of an acoustic and a flow-flame interaction model, is investigated. The presented approach focuses on only reducing the pure acoustic model and a posteriori coupling to the flow-flame model. In this study, a simplistic 1D network tool has been used to model the acoustics and a low-pass transfer function represents the flow-flame interaction. Three different linear reduction algorithms, namely modal model order reduction, truncated balance realization (TBR) and the iterative rational Krylov algorithm (IRKA) are employed to obtain the reduced acoustic model. After coupling these to the flow-flame interaction model, the quality of the reduction is assessed by comparing the most unstable eigenvalue of the reduced coupled thermoacoustic system to the full coupled thermoacoustic system. The results show good convergence for the TBR and IRKA algorithms with increasing reduced model order, while the modal reduction may even decrease in quality when additional modes are included in the reduced acoustic system.