Early Detection of Thermoacoustic Combustion Instability Using a Methodology Combining Complex Networks and Machine Learning
TL;DR: A feature space consisting of the principal component plane estimated from the probability distribution of the transition patterns, which is obtained by a support vector machine, allows the early detection of thermoacoustic combustion instability.
Abstract: Early detection of thermoacoustic instabilities is of interest to both applied physicists and engineers, to avoid resonance leading to self-destruction of gas-based engines and turbines. This study shows how a combination of complex-network physics and machine learning can be used to detect a precursor of thermoacoustic instabilities, which can help to prevent the onset of a potentially destructive combustion-driven instability.
Citations
More filters
112 citations
TL;DR: In this article, the authors discuss various prognosis and mitigation strategies for thermo-acoustic instability based on complex system theory in turbulent combustors, where the authors view the thermoacoustic system in a turbulent combustor as a complex system and the dynamics exhibited by the system is perceived as emergent behaviors of this complex system.
Abstract: Thermoacoustic instability in turbulent combustors is a nonlinear phenomenon resulting from the interaction between acoustics, hydrodynamics, and the unsteady flame Over the years, there have been many attempts toward understanding, prognosis, and mitigation of thermoacoustic instabilities Traditionally, a linear framework has been used to study thermoacoustic instability In recent times, researchers have been focusing on the nonlinear dynamics related to the onset of thermoacoustic instability In this context, the thermoacoustic system in a turbulent combustor is viewed as a complex system, and the dynamics exhibited by the system is perceived as emergent behaviors of this complex system In this paper, we discuss these recent developments and their contributions toward the understanding of this complex phenomenon Furthermore, we discuss various prognosis and mitigation strategies for thermoacoustic instability based on complex system theory
88 citations
TL;DR: A review of data sources, data-driven techniques, and concepts for combustion machine learning can be found in this article , focusing on interpretability, uncertainty quantification, robustness, consistency, creation and curation of benchmark data, and the augmentation of ML methods with prior combustion domain knowledge.
47 citations
TL;DR: The present review can boost future network-based research on turbulent and vortical flows, promoting the establishment of complex networks as a widespread tool for turbulence analysis.
Abstract: Turbulent and vortical flows are ubiquitous and their characterization is crucial for the understanding of several natural and industrial processes. Among different techniques to study spatio-temporal flow fields, complex networks represent a recent and promising tool to deal with the large amount of data on turbulent flows and shed light on their physical mechanisms. The aim of this review is to bring together the main findings achieved so far from the application of network-based techniques to study turbulent and vortical flows. A critical discussion on the potentialities and limitations of the network approach is provided, thus giving an ordered portray of the current diversified literature. The present review can boost future network-based research on turbulent and vortical flows, promoting the establishment of complex networks as a widespread tool for turbulence analysis.
39 citations
TL;DR: An experimental study on early detection of thermoacoustic combustion oscillations using a method combining statistical complexity and machine learning, including the characterization of intermittent combustionscillations.
Abstract: We conduct an experimental study on early detection of thermoacoustic combustion oscillations using a method combining statistical complexity and machine learning, including the characterization of intermittent combustion oscillations. Abrupt switching from aperiodic small-amplitude oscillations to periodic large-amplitude oscillations and vice versa appears in pressure fluctuations. The dynamic behavior of aperiodic small-amplitude pressure fluctuations represents chaos. The complexity-entropy causality plane effectively captures the subtle changes in the combustion state during a transition to well-developed combustion oscillations. The feature space of the complexity-entropy causality plane, which is obtained by a support vector machine, has potential use for detecting a precursor of combustion oscillations.
27 citations
References
More filters
Book•
25 Mar 2010
TL;DR: This book brings together for the first time the most important breakthroughs in each of these fields and presents them in a coherent fashion, highlighting the strong interconnections between work in different areas.
Abstract: The scientific study of networks, including computer networks, social networks, and biological networks, has received an enormous amount of interest in the last few years. The rise of the Internet and the wide availability of inexpensive computers have made it possible to gather and analyze network data on a large scale, and the development of a variety of new theoretical tools has allowed us to extract new knowledge from many different kinds of networks.The study of networks is broadly interdisciplinary and important developments have occurred in many fields, including mathematics, physics, computer and information sciences, biology, and the social sciences. This book brings together for the first time the most important breakthroughs in each of these fields and presents them in a coherent fashion, highlighting the strong interconnections between work in different areas. Subjects covered include the measurement and structure of networks in many branches of science, methods for analyzing network data, including methods developed in physics, statistics, and sociology, the fundamentals of graph theory, computer algorithms, and spectral methods, mathematical models of networks, including random graph models and generative models, and theories of dynamical processes taking place on networks.
10,567 citations
TL;DR: How the abstract learning theory established conditions for generalization which are more general than those discussed in classical statistical paradigms are demonstrated and how the understanding of these conditions inspired new algorithmic approaches to function estimation problems are demonstrated.
Abstract: Statistical learning theory was introduced in the late 1960's. Until the 1990's it was a purely theoretical analysis of the problem of function estimation from a given collection of data. In the middle of the 1990's new types of learning algorithms (called support vector machines) based on the developed theory were proposed. This made statistical learning theory not only a tool for the theoretical analysis but also a tool for creating practical algorithms for estimating multidimensional functions. This article presents a very general overview of statistical learning theory including both theoretical and algorithmic aspects of the theory. The goal of this overview is to demonstrate how the abstract learning theory established conditions for generalization which are more general than those discussed in classical statistical paradigms and how the understanding of these conditions inspired new algorithmic approaches to function estimation problems.
5,370 citations
"Early Detection of Thermoacoustic C..." refers background in this paper
...The SVM allows the combustion state to be clearly classified into three regimes on the feature space: combustion noise (blue), thermoacoustic combustion instability (red), and the transition from combustion noise to thermoacoustic combustion instability (yellow)....
[...]
...support vector machine (SVM) [15,16], which is a class of machine learning to recognize patterns, is a binary classi-...
[...]
...The standard support vector machine (SVM) [15,16], which is a class of machine learning to recognize patterns, is a binary classifier predicting the boundaries of classes and has received considerable attention in data science for the past few decades [17]....
[...]
...The SVM [17] searches for an optimal separating hyperplane in the feature space using the given training data set {xi, yi}, i = 1, 2, . . . , n, where xi is the input vector and yi is the class label taking a value of +1 or −1....
[...]
...Figure 6(A) shows the feature space consisting of S1 and S2 obtained by the SVM....
[...]
TL;DR: The method introduces complexity parameters for time series based on comparison of neighboring values and shows that its complexity behaves similar to Lyapunov exponents, and is particularly useful in the presence of dynamical or observational noise.
Abstract: We introduce complexity parameters for time series based on comparison of neighboring values. The definition directly applies to arbitrary real-world data. For some well-known chaotic dynamical systems it is shown that our complexity behaves similar to Lyapunov exponents, and is particularly useful in the presence of dynamical or observational noise. The advantages of our method are its simplicity, extremely fast calculation, robustness, and invariance with respect to nonlinear monotonous transformations.
3,433 citations
"Early Detection of Thermoacoustic C..." refers methods in this paper
...In this study, we construct the ordinal partition transition networks consisting of order patterns (permutation patterns [26,27])....
[...]
TL;DR: The aim of this work is to provide the readers with the know how for the application of recurrence plot based methods in their own field of research, and detail the analysis of data and indicate possible difficulties and pitfalls.
2,993 citations
"Early Detection of Thermoacoustic C..." refers methods in this paper
...Here, Jr is the joint probability of recurrence plots [23]; Rr (= ∑ m,n Rm,n/N 2) is the recurrence rate, M = [(1/N 2) ∑N m,n Jm,n]/Rr, Jm,n = [ − ||p(tm) − p(tn)||] [ − ||WOH∗(tm) − WOH∗(tn)||], is the Heaviside function, p(t) = {p ′(t), p ′(t + τ), ....
[...]
...In this study, we first adopt the synchronization index SI , defined as the product of the joint probability of recurrence plots Jr [23] and the phase synchronization parameter rpq, to study the significant change in the dynamics from combustion noise to thermoacoustic combustion instability with increasing equivalence ratio:...
[...]
TL;DR: A comprehensive review of the advances made over the past two decades in this area is provided in this article, where various swirl injector configurations and related flow characteristics, including vortex breakdown, precessing vortex core, large-scale coherent structures, and liquid fuel atomization and spray formation are discussed.
1,048 citations