Experimental data-based reduced-order model for analysis and prediction of flame transition in gas turbine combustors

Abstract

In lean premixed combustors, flame stabilization is an important operational concern that can affect efficiency, robustness and pollutant formation. The focus of this paper is on flame lift-off and re-attachment to the nozzle of a swirl combustor. Using time-resolved experimental measurements, a data-driven approach known as cluster-based reduced order modeling (CROM) is employed to 1) isolate key flow patterns and their sequence during the flame transitions, and 2) formulate a forecasting model to predict the flame instability. The flow patterns isolated by the CROM methodology confirm some of the experimental conclusions about the flame transition mechanism. In particular, CROM highlights the key role of the precessing vortex core (PVC) in the flame detachment process in an unsupervised manner. For the attachment process, strong flow recirculation far from the nozzle appears to drive the flame upstream, thus initiating re-attachment. Different data-types (velocity field, OH concentration) were processed by the modeling tool, and the predictive capabilities of these different models are also compared. It was found that the swirling velocity possesses the best predictive properties, which gives a supplemental argument for the role of the PVC in causing the flame transition. The model is tested against unseen data and successfully predicts the probability of flame transition (both detachment and attachment) when trained with swirling velocity with minimal user input. The model trained with OH-PLIF data was only successful at predicting the flame attachment, which implies that different physical mechanisms are present for different types of flame transition. Overall, these aspects show the great potential of data-driven methods, particularly probabilistic forecasting techniques, in analyzing and predicting large-scale features in complex turbulent combustion problems.

Influence of flow field dynamics on soot evolution in an aero-engine model combustor

“…Specifically, this work employs a maximumoverlap discrete packet wavelet transform (MODPWT) in the MRPOD to isolate PVC from low-frequency dynamics (LFD), such that their influences on the flow field as well as on soot volume fraction (f v ) can be separately evaluated. Beside POD-based methods the cluster-based reduced-order modeling (CROM) method [8,9] with its probabilistic nature, may also prove appropriate at treating the nonlinear LFD. mary injector consists of two concentric, swirled airflows, into which ethylene is injected co-axially via 64 rectangular channels.…”

Data-driven analysis of relight variability of jet fuels induced by turbulence

“…The clustering process is done in an unsupervised manner using K-means clustering [46][47][48], which has been successfully applied in a variety of fluid applications [49][50][51][52]. The K-means algorithm takes as an input a number of clusters and subsequently groups the realizations (or snapshots) into these many clusters.…”

“…In the following results, the distance chosen for the clustering is the L 2 norm computed in physical space. While this choice has been sufficient in other applications [49][50][51], it is not necessarily ideal. Other distance measures based on image recognition concepts were investigated [58], though these led to the same conclusions as the ones presented in Sec.…”

Reconstruction of cross-modal visual features from acoustic pressure time series in combustion systems

“…Majority of the efforts has been spent in characterizing the combustor pressure signal and detecting temporal precursors by various reduced order modeling, time series analysis (Gotoda et al (2011); Murayama et al ( 2018)), and signal processing (Nair and Sujith (2014); Sen et al (2018Sen et al ( , 2016; Barwey et al (2019a)) or machine learning approaches (Sarkar et al (2015b)).…”

“…The common methods for coherent structure detection are rotational symmetry(Ramanan and Chakravarthy (2019)), proper orthogonal decomposition (POD)(Berkooz et al (1993); LoCurto et al (2018)) (similar to principal component analysis(Bishop (2006))) and dynamic mode decomposition (DMD)(Schmid (2010);Ghosal et al (2016)), which utilize spectral theory to derive spatial coherent structure modes. Although these methods have widespread applications in understanding spatiotemporal flow characteristics, it is shown that they demonstrate high parametric sensitivity for instability detection purposes and their applicability does not extend to noisy line-of-sight videos at par with more realistic data collection scenario.Using artificial neural networks and clustering methods, studies(Barwey et al (2019a(Barwey et al ( ,b, 2020;Cui et al (2020)) have been performed in the domain of combustion systems, especially on fluid mechanical aspects of combustion and identifying mapping between flow-field and heat release rate markers. Researchers have also shown early instability detection capability by combining complex networks and machine learning(Kobayashi et al (2019)) and have used Bayesian machine learning for combustion diagnostics(Sengupta et al (2020)).…”