Early Detection of Combustion Instability by Neural-Symbolic Analysis on Hi-Speed Video

This paper proposes a neural-symbolic framework for analyzing a large volume of sequential hi-speed images of combustion flame for early detection of instability that is extremely critical for engine health monitoring and prognostics. The proposed hierarchical approach involves extracting low-dimensional semantic features from images using deep Convolutional Neural Networks (CNN) followed by capturing the temporal evolution of the extracted features using Symbolic Time Series Analysis (STSA). Furthermore, the semantic nature of the CNN features enables expert-guided data exploration that can lead to better understanding of the underlying physics. Extensive experimental data have been collected in a swirl-stabilized dump combustor at various operating conditions for validation.

[1]  A. Hussain.  Coherent structures—reality and myth , 1983 .

[2]  Asok Ray,et al.  Dynamic Data-Driven Prediction of Lean Blowout in a Swirl-Stabilized Combustor , 2015 .

[3]  Y-Lan Boureau,et al.  Learning Convolutional Feature Hierarchies for Visual Recognition , 2010, NIPS.

[4]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[5]  S. Sarkar,et al.  Investigation of Combustion Instability in a Swirl-Stabilized Combustor Using Symbolic Time Series Analysis , 2014 .

[6]  Asok Ray,et al.  Review and comparative evaluation of symbolic dynamic filtering for detection of anomaly patterns , 2009, 2008 American Control Conference.

[7]  Asok Ray,et al.  Symbolic time series analysis via wavelet-based partitioning , 2006, Signal Process..

[8]  Kushal Mukherjee,et al.  State splitting and state merging in probabilistic finite state automata , 2011, Proceedings of the 2011 American Control Conference.

[9]  Kushal Mukherjee,et al.  Symbolic Dynamic Analysis of Transient Time Series for Fault Detection in Gas Turbine Engines , 2013 .

[10]  J. Janicka,et al.  Experimental characterization of onset of acoustic instability in a nonpremixed half-dump combustor. , 2007, The Journal of the Acoustical Society of America.

[11]  Asok Ray,et al.  Early detection of lean blow out (LBO) via generalized D-Markov machine construction , 2014, 2014 American Control Conference.

[12]  Asok Ray,et al.  Symbolic dynamic analysis of complex systems for anomaly detection , 2004, Signal Process..

[13]  Eitan M. Gurari,et al.  Introduction to the theory of computation , 1989 .

[14]  Luc De Raedt,et al.  Neural-Symbolic Learning and Reasoning: Contributions and Challenges , 2015, AAAI Spring Symposia.

[15]  Christopher M. Bishop,et al.  Pattern Recognition and Machine Learning (Information Science and Statistics) , 2006 .

[16]  Asok Ray,et al.  Early Detection of Combustion Instability from Hi-speed Flame Images via Deep Learning and Symbolic Time Series Analysis , 2015, Annual Conference of the PHM Society.

[17]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[18]  Abhishek Srivastav,et al.  Maximally Bijective Discretization for data-driven modeling of complex systems , 2013, 2013 American Control Conference.

[19]  P. Holmes,et al.  The Proper Orthogonal Decomposition in the Analysis of Turbulent Flows , 1993 .

[20]  P. Schmid,et al.  Dynamic mode decomposition of numerical and experimental data , 2008, Journal of Fluid Mechanics.

[21]  C. Finney,et al.  A review of symbolic analysis of experimental data , 2003 .

[22]  Asok Ray,et al.  Performance comparison of feature extraction algorithms for target detection and classification , 2013, Pattern Recognit. Lett..

[23]  Kushal Mukherjee,et al.  State splitting and merging in probabilistic finite state automata for signal representation and analysis , 2014, Signal Process..