Proceedings of The Institution of Mechanical Engineers Part G-journal of Aerospace Engineering, 2008
Degradation monitoring is of paramount importance to safety and reliability of aircraft operation... more Degradation monitoring is of paramount importance to safety and reliability of aircraft operations and also for timely maintenance of its critical components. This two-part paper formulates and validates a novel methodology of degradation monitoring of aircraft gas turbine engines with emphasis on detection and isolation of incipient faults. In a complex system with multiple interconnected components (e.g. an aircraft engine), fault isolation becomes a crucial task because of possible input-output and feedback interactions among the individual components.
The paper addresses data-driven statistical pattern identification in complex dynamical systems, ... more The paper addresses data-driven statistical pattern identification in complex dynamical systems, where the concept is built upon thermodynamic formalism of symbolic data sequences in the setting of lattice spin systems. The transfer matrix approach has been used for generation of pattern vectors from time series data of observed parameters. Efficacy of pattern identification is demonstrated for early detection of anomalies (i.e., deviations from the nominal pattern) on an experimental apparatus of nonlinear active electronic circuits.
The paper addresses the issue of online diagnosis and prognosis of emerging faults in human-engin... more The paper addresses the issue of online diagnosis and prognosis of emerging faults in human-engineered complex systems. Specifically, the paper reports a dynamic data-driven analytical tool for early detection of incipient faults and real-time estimation of remaining useful fatigue life in polycrystalline alloys. The algorithms for fatigue life estimation rely on time series data analysis of ultrasonic signals and are built upon the principles of symbolic dynamics, information theory and statistical pattern recognition. The proposed method is experimentally validated by using 7075-T6 aluminium alloy specimens on a special-purpose fatigue test apparatus that is equipped with ultrasonic flaw detectors and an optical travelling microscope. The real-time information, derived by the proposed method, is useful for mitigation of widespread fatigue damage and is potentially applicable to life extending and resilient control of mechanical structures.
This paper presents a novel analytical tool for early detection of fatigue damage in polycrystall... more This paper presents a novel analytical tool for early detection of fatigue damage in polycrystalline alloys that are commonly used in mechanical structures. Time series data of ultrasonic sensors have been used for anomaly detection in the statistical behaviour of structural materials, where the analysis is based on the principles of symbolic dynamics and automata theory. The performance of the proposed method has been evaluated relative to existing pattern recognition tools, such as neural networks and principal component analysis, for detection of small changes in the statistical characteristics of the observed data sequences. This concept is experimentally validated on a special-purpose test apparatus for 7075-T6 aluminium alloy specimens, where the anomalies accrue from small fatigue crack growth. r
Real-time data-driven pattern classification requires extraction of relevant features from the ob... more Real-time data-driven pattern classification requires extraction of relevant features from the observed time series as low-dimensional and yet information-rich representations of the underlying dynamics. These low-dimensional features facilitate in situ decision-making in diverse applications, such as computer vision, structural health monitoring, and robotics. Wavelet transforms of time series have been widely used for feature extraction owing to their time-frequency localization properties. In this regard, this paper presents a symbolic dynamics-based method to model surface images, generated by wavelet coefficients in the scale-shift space. These symbolic dynamics-based models (e.g., probabilistic finite state automata (PFSA)) capture the relevant information, embedded in the sensor data, from the associated Perron-Frobenius operators (i.e., the state-transition probability matrices). The proposed method of pattern classification has been experimentally validated on laboratory apparatuses for two different applications: (i) early detection of evolving damage in polycrystalline alloy structures, and (ii) classification of mobile robots and their motion profiles.
... signals for fatigue damage monitoring in polycrystalline alloys Shalabh Gupta, Asok Ray and E... more ... signals for fatigue damage monitoring in polycrystalline alloys Shalabh Gupta, Asok Ray and Eric Keller Mechanical Engineering Department, The Pennsylvania State University, University Park, PA 16802, USA E-mail: [email protected], [email protected] and [email protected] ...
This paper presents information-theoretic analysis of time-series data to detect slowly evolving ... more This paper presents information-theoretic analysis of time-series data to detect slowly evolving anomalies (i.e., deviations from a nominal operating condition) in dynamical systems. A measure for anomaly detection is formulated based on the concepts derived from information theory and statistical thermodynamics. The underlying algorithm is first tested on a low-dimensional complex dynamical system with a known structure-the Duffing oscillator with slowly changing dissipation. Then, the anomaly detection tool is experimentally validated on test specimens of 7075-T6 aluminum alloy under cyclic loading. The results are presented for both cases and the efficacy of the proposed method is thus demonstrated for systems of known and unknown structures.
Proceedings of The Institution of Mechanical Engineers Part G-journal of Aerospace Engineering, 2008
The first part of this two-part paper, which is a companion paper, has developed a novel concept ... more The first part of this two-part paper, which is a companion paper, has developed a novel concept of fault detection and isolation (FDI) in aircraft gas turbine engines. The FDI algorithms are built upon the statistical pattern recognition method of symbolic dynamic filtering (SDF) that is especially suited for real-time detection and isolation of slowly evolving anomalies in engine components, in addition to abrupt faults. The FDI methodology is based on the analysis of time series data of available sensors and/or analytically derived variables in the gas path dynamics.
Proceedings of The Institution of Mechanical Engineers Part I-journal of Systems and Control Engineering, 2006
This paper presents symbolic time series analysis of observable process variables for anomaly det... more This paper presents symbolic time series analysis of observable process variables for anomaly detection in thermal pulse combustors. The anomaly detection method has been tested on the time series data of pressure oscillations, generated from a non-linear dynamic model of a generic thermal pulse combustor. Results are presented to exemplify early detection of combustion instability due to reduction of friction coefficient in the tailpipe, which eventually leads to flame extinction.
The paper presents an analytical tool for early detection and online monitoring of fatigue damage... more The paper presents an analytical tool for early detection and online monitoring of fatigue damage in polycrystalline alloys that are commonly used in mechanical structures of human-engineered complex systems. Real-time fatigue damage monitoring algorithms rely on time series analysis of ultrasonic signals that are sensitive to micro-structural changes occurring inside the material during the early stages of fatigue damage; the core concept of signal analysis is built upon the principles of Symbolic Dynamics, Statistical Pattern Recognition and Information Theory. The analytical tool of statistical pattern analysis has been experimentally validated on a special-purpose test apparatus that is equipped with ultrasonic flaw detection sensors and a travelling optical microscope. The paper reports fatigue damage monitoring of 7075-T6 alloy specimens, where the experiments have been conducted under load-controlled constant amplitude sinusoidal loadings for low-cycle and high-cycle fatigue.
This paper addresses online monitoring of fatigue damage in polycrystalline alloy structures base... more This paper addresses online monitoring of fatigue damage in polycrystalline alloy structures based on statistical pattern analysis of ultrasonic sensor signals. The real-time data-driven method for fatigue damage monitoring is based on the concepts derived from statistical mechanics, symbolic dynamics and statistical pattern identification. The underlying concept is detection and identification of small changes in statistical patterns of ultrasonic data streams due to gradual evolution of anomalies (i.e., deviations from the nominal behavior) in material structures. The statistical patterns in terms of the escort distributions from statistical mechanics are derived from symbol sequences that, in turn, are generated from ultrasonic sensors installed on the structures under stress cycles. The resulting information of evolving fatigue damage would provide early warnings of forthcoming failures, possibly, due to widespread crack propagation. The damage monitoring method has been validated by laboratory experimentation in real time on a computer-controlled fatigue damage testing apparatus which is equipped with a variety of measuring instruments including an optical travelling microscope and an ultrasonic flaw detector.
This paper examines the efficacy of a novel method for anomaly detection in mechanical systems, w... more This paper examines the efficacy of a novel method for anomaly detection in mechanical systems, which makes use of a hidden Markov model, derived from the time-series data of pertinent measurement(s). The core concept of the anomaly detection method is symbolic time-series analysis that is built upon the principles of Automata Theory, Information Theory, and Pattern Recognition. The performance of this method is compared with that of other existing pattern-recognition techniques from the perspective of early detection of small fatigue cracks in ductile alloy structures. The experimental apparatus, on which the anomaly detection method is tested, is a multi-degree-of-freedom mass-beam structure excited by oscillatory motion of two electromagnetic shakers. The evolution of fatigue crack damage at one or more failure sites are detected from symbolic time-series analysis of displacement sensor signals.
Gradual development of anomalies (i.e., deviations from the nominal condition) may alter the quas... more Gradual development of anomalies (i.e., deviations from the nominal condition) may alter the quasi-static behavior patterns of human-engineered complex systems. This chapter presents a recently reported technique of pattern recognition, called Symbolic Dynamic Filtering (SDF), for early detection and prognosis of such changes in behavior patterns due to slowly evolving anomalies that may be benign or malignant. The underlying concept of SDF is built upon the principles of Symbolic Dynamics, Information Theory, and Statistical Signal Processing, where time series data from selected sensor(s) in the fast time scale of the process dynamics are analyzed at discrete epochs in the slow time scale of anomaly evolution. The key idea here is early detection and identification of (possible) changes in quasi-static statistical patterns of the dynamical system behavior. An important feature of this pattern recognition technique is extraction of the relevant statistics by conversion of the time series data into a symbol sequence by appropriate coarse-graining of the imbedded information. As an alternative to the currently practiced method of phase-space partitioning in the time domain, a new concept of partitioning is introduced for symbol generation, based on wavelet analysis of the time series data. This chapter also discusses various aspects of the wavelet-based partitioning tool, such as selection of the wavelet basis, noise mitigation, and robustness to spurious disturbances. The partitioning scheme is built upon the principle of maximum entropy such that the regions of the data space with more information are partitioned finer and those with sparse information are partitioned coarser. The algorithms of SDF are constructed to solve two problems: (i) Forward problem of Pattern Recognition for (offline) characterization of the anomalous behavior, relative to the nominal behavior; and (ii) Inverse problem of Pattern * Identification for (online) estimation of parametric or non-parametric changes based on the knowledge assimilated in the forward problem and the observed time series data of quasi-stationary process response.
This paper presents a statistical mechanics concept for identification of behavioral patterns in ... more This paper presents a statistical mechanics concept for identification of behavioral patterns in complex systems based on measurements (e.g., time series data) of macroscopically observable parameters and their operational characteristics. The tools of statistical mechanics, which provide a link between the microscopic (i.e., detailed) and macroscopic (i.e., aggregated) properties of a complex system are used to capture the emerging information and to identify the quasi-stationary evolution of behavioral patterns. The underlying theory is built upon thermodynamic formalism of symbol sequences in the setting of a generalized Ising model (GIM) of lattice-spin systems. In this context, transfer matrix analysis facilitates construction of pattern vectors from observed sequences. The proposed concept is experimentally validated on a richly instrumented laboratory apparatus that is operated under oscillating load for identification of evolving microstructural changes in polycrystalline alloys.
Proceedings of The Institution of Mechanical Engineers Part G-journal of Aerospace Engineering, 2008
Degradation monitoring is of paramount importance to safety and reliability of aircraft operation... more Degradation monitoring is of paramount importance to safety and reliability of aircraft operations and also for timely maintenance of its critical components. This two-part paper formulates and validates a novel methodology of degradation monitoring of aircraft gas turbine engines with emphasis on detection and isolation of incipient faults. In a complex system with multiple interconnected components (e.g. an aircraft engine), fault isolation becomes a crucial task because of possible input-output and feedback interactions among the individual components.
The paper addresses data-driven statistical pattern identification in complex dynamical systems, ... more The paper addresses data-driven statistical pattern identification in complex dynamical systems, where the concept is built upon thermodynamic formalism of symbolic data sequences in the setting of lattice spin systems. The transfer matrix approach has been used for generation of pattern vectors from time series data of observed parameters. Efficacy of pattern identification is demonstrated for early detection of anomalies (i.e., deviations from the nominal pattern) on an experimental apparatus of nonlinear active electronic circuits.
The paper addresses the issue of online diagnosis and prognosis of emerging faults in human-engin... more The paper addresses the issue of online diagnosis and prognosis of emerging faults in human-engineered complex systems. Specifically, the paper reports a dynamic data-driven analytical tool for early detection of incipient faults and real-time estimation of remaining useful fatigue life in polycrystalline alloys. The algorithms for fatigue life estimation rely on time series data analysis of ultrasonic signals and are built upon the principles of symbolic dynamics, information theory and statistical pattern recognition. The proposed method is experimentally validated by using 7075-T6 aluminium alloy specimens on a special-purpose fatigue test apparatus that is equipped with ultrasonic flaw detectors and an optical travelling microscope. The real-time information, derived by the proposed method, is useful for mitigation of widespread fatigue damage and is potentially applicable to life extending and resilient control of mechanical structures.
This paper presents a novel analytical tool for early detection of fatigue damage in polycrystall... more This paper presents a novel analytical tool for early detection of fatigue damage in polycrystalline alloys that are commonly used in mechanical structures. Time series data of ultrasonic sensors have been used for anomaly detection in the statistical behaviour of structural materials, where the analysis is based on the principles of symbolic dynamics and automata theory. The performance of the proposed method has been evaluated relative to existing pattern recognition tools, such as neural networks and principal component analysis, for detection of small changes in the statistical characteristics of the observed data sequences. This concept is experimentally validated on a special-purpose test apparatus for 7075-T6 aluminium alloy specimens, where the anomalies accrue from small fatigue crack growth. r
Real-time data-driven pattern classification requires extraction of relevant features from the ob... more Real-time data-driven pattern classification requires extraction of relevant features from the observed time series as low-dimensional and yet information-rich representations of the underlying dynamics. These low-dimensional features facilitate in situ decision-making in diverse applications, such as computer vision, structural health monitoring, and robotics. Wavelet transforms of time series have been widely used for feature extraction owing to their time-frequency localization properties. In this regard, this paper presents a symbolic dynamics-based method to model surface images, generated by wavelet coefficients in the scale-shift space. These symbolic dynamics-based models (e.g., probabilistic finite state automata (PFSA)) capture the relevant information, embedded in the sensor data, from the associated Perron-Frobenius operators (i.e., the state-transition probability matrices). The proposed method of pattern classification has been experimentally validated on laboratory apparatuses for two different applications: (i) early detection of evolving damage in polycrystalline alloy structures, and (ii) classification of mobile robots and their motion profiles.
... signals for fatigue damage monitoring in polycrystalline alloys Shalabh Gupta, Asok Ray and E... more ... signals for fatigue damage monitoring in polycrystalline alloys Shalabh Gupta, Asok Ray and Eric Keller Mechanical Engineering Department, The Pennsylvania State University, University Park, PA 16802, USA E-mail: [email protected], [email protected] and [email protected] ...
This paper presents information-theoretic analysis of time-series data to detect slowly evolving ... more This paper presents information-theoretic analysis of time-series data to detect slowly evolving anomalies (i.e., deviations from a nominal operating condition) in dynamical systems. A measure for anomaly detection is formulated based on the concepts derived from information theory and statistical thermodynamics. The underlying algorithm is first tested on a low-dimensional complex dynamical system with a known structure-the Duffing oscillator with slowly changing dissipation. Then, the anomaly detection tool is experimentally validated on test specimens of 7075-T6 aluminum alloy under cyclic loading. The results are presented for both cases and the efficacy of the proposed method is thus demonstrated for systems of known and unknown structures.
Proceedings of The Institution of Mechanical Engineers Part G-journal of Aerospace Engineering, 2008
The first part of this two-part paper, which is a companion paper, has developed a novel concept ... more The first part of this two-part paper, which is a companion paper, has developed a novel concept of fault detection and isolation (FDI) in aircraft gas turbine engines. The FDI algorithms are built upon the statistical pattern recognition method of symbolic dynamic filtering (SDF) that is especially suited for real-time detection and isolation of slowly evolving anomalies in engine components, in addition to abrupt faults. The FDI methodology is based on the analysis of time series data of available sensors and/or analytically derived variables in the gas path dynamics.
Proceedings of The Institution of Mechanical Engineers Part I-journal of Systems and Control Engineering, 2006
This paper presents symbolic time series analysis of observable process variables for anomaly det... more This paper presents symbolic time series analysis of observable process variables for anomaly detection in thermal pulse combustors. The anomaly detection method has been tested on the time series data of pressure oscillations, generated from a non-linear dynamic model of a generic thermal pulse combustor. Results are presented to exemplify early detection of combustion instability due to reduction of friction coefficient in the tailpipe, which eventually leads to flame extinction.
The paper presents an analytical tool for early detection and online monitoring of fatigue damage... more The paper presents an analytical tool for early detection and online monitoring of fatigue damage in polycrystalline alloys that are commonly used in mechanical structures of human-engineered complex systems. Real-time fatigue damage monitoring algorithms rely on time series analysis of ultrasonic signals that are sensitive to micro-structural changes occurring inside the material during the early stages of fatigue damage; the core concept of signal analysis is built upon the principles of Symbolic Dynamics, Statistical Pattern Recognition and Information Theory. The analytical tool of statistical pattern analysis has been experimentally validated on a special-purpose test apparatus that is equipped with ultrasonic flaw detection sensors and a travelling optical microscope. The paper reports fatigue damage monitoring of 7075-T6 alloy specimens, where the experiments have been conducted under load-controlled constant amplitude sinusoidal loadings for low-cycle and high-cycle fatigue.
This paper addresses online monitoring of fatigue damage in polycrystalline alloy structures base... more This paper addresses online monitoring of fatigue damage in polycrystalline alloy structures based on statistical pattern analysis of ultrasonic sensor signals. The real-time data-driven method for fatigue damage monitoring is based on the concepts derived from statistical mechanics, symbolic dynamics and statistical pattern identification. The underlying concept is detection and identification of small changes in statistical patterns of ultrasonic data streams due to gradual evolution of anomalies (i.e., deviations from the nominal behavior) in material structures. The statistical patterns in terms of the escort distributions from statistical mechanics are derived from symbol sequences that, in turn, are generated from ultrasonic sensors installed on the structures under stress cycles. The resulting information of evolving fatigue damage would provide early warnings of forthcoming failures, possibly, due to widespread crack propagation. The damage monitoring method has been validated by laboratory experimentation in real time on a computer-controlled fatigue damage testing apparatus which is equipped with a variety of measuring instruments including an optical travelling microscope and an ultrasonic flaw detector.
This paper examines the efficacy of a novel method for anomaly detection in mechanical systems, w... more This paper examines the efficacy of a novel method for anomaly detection in mechanical systems, which makes use of a hidden Markov model, derived from the time-series data of pertinent measurement(s). The core concept of the anomaly detection method is symbolic time-series analysis that is built upon the principles of Automata Theory, Information Theory, and Pattern Recognition. The performance of this method is compared with that of other existing pattern-recognition techniques from the perspective of early detection of small fatigue cracks in ductile alloy structures. The experimental apparatus, on which the anomaly detection method is tested, is a multi-degree-of-freedom mass-beam structure excited by oscillatory motion of two electromagnetic shakers. The evolution of fatigue crack damage at one or more failure sites are detected from symbolic time-series analysis of displacement sensor signals.
Gradual development of anomalies (i.e., deviations from the nominal condition) may alter the quas... more Gradual development of anomalies (i.e., deviations from the nominal condition) may alter the quasi-static behavior patterns of human-engineered complex systems. This chapter presents a recently reported technique of pattern recognition, called Symbolic Dynamic Filtering (SDF), for early detection and prognosis of such changes in behavior patterns due to slowly evolving anomalies that may be benign or malignant. The underlying concept of SDF is built upon the principles of Symbolic Dynamics, Information Theory, and Statistical Signal Processing, where time series data from selected sensor(s) in the fast time scale of the process dynamics are analyzed at discrete epochs in the slow time scale of anomaly evolution. The key idea here is early detection and identification of (possible) changes in quasi-static statistical patterns of the dynamical system behavior. An important feature of this pattern recognition technique is extraction of the relevant statistics by conversion of the time series data into a symbol sequence by appropriate coarse-graining of the imbedded information. As an alternative to the currently practiced method of phase-space partitioning in the time domain, a new concept of partitioning is introduced for symbol generation, based on wavelet analysis of the time series data. This chapter also discusses various aspects of the wavelet-based partitioning tool, such as selection of the wavelet basis, noise mitigation, and robustness to spurious disturbances. The partitioning scheme is built upon the principle of maximum entropy such that the regions of the data space with more information are partitioned finer and those with sparse information are partitioned coarser. The algorithms of SDF are constructed to solve two problems: (i) Forward problem of Pattern Recognition for (offline) characterization of the anomalous behavior, relative to the nominal behavior; and (ii) Inverse problem of Pattern * Identification for (online) estimation of parametric or non-parametric changes based on the knowledge assimilated in the forward problem and the observed time series data of quasi-stationary process response.
This paper presents a statistical mechanics concept for identification of behavioral patterns in ... more This paper presents a statistical mechanics concept for identification of behavioral patterns in complex systems based on measurements (e.g., time series data) of macroscopically observable parameters and their operational characteristics. The tools of statistical mechanics, which provide a link between the microscopic (i.e., detailed) and macroscopic (i.e., aggregated) properties of a complex system are used to capture the emerging information and to identify the quasi-stationary evolution of behavioral patterns. The underlying theory is built upon thermodynamic formalism of symbol sequences in the setting of a generalized Ising model (GIM) of lattice-spin systems. In this context, transfer matrix analysis facilitates construction of pattern vectors from observed sequences. The proposed concept is experimentally validated on a richly instrumented laboratory apparatus that is operated under oscillating load for identification of evolving microstructural changes in polycrystalline alloys.
Uploads
Papers by Shalabh Gupta