Dheeraj Sharan Singh

Optimization of symbolic feature extraction for pattern classification

The concept of symbolic dynamics has been used in recent literature for feature extraction from time series data for pattern classification. The two primary steps of this technique are partitioning of time series to optimally generate symbol sequences and subsequently modeling of state machines from such symbol sequences. The latter step has been widely investigated and reported in the literature. However, for optimal feature extraction, the first step needs to be further explored. The paper addresses this issue and proposes a data partitioning procedure to extract low-dimensional features from time series while optimizing the class separability. The proposed procedure has been validated on two examples: (i) parameter identification in a Duffing system and (ii) classification of fatigue damage in mechanical structures, made of polycrystalline alloys. In each case, the classification performance of the proposed data partitioning method is compared with those of two other classical data partitioning methods, namely uniform partitioning (UP) and maximum entropy partitioning (MEP).

Symbolic dynamic analysis of surface deformation during fatigue crack initiation

This rapid communication addresses early detection of fatigue damage evolution in polycrystalline alloys, based on the observation of surface deformation (e.g. roughness, linings and incisions). This method is well suited for calibration of other model-based and experimental tools for damage analysis and prediction in the fatigue crack initiation phase. To this end, the existing theory of symbolic dynamics-based feature extraction from time-series data is extended to the analysis of two-dimensional surface images. The resulting algorithms are experimentally validated on a fatigue-testing machine and a surface interferometer in the laboratory environment. The experiments have been conducted for analysis of statistical changes in the surface profiles due to gradual evolution of deformation in specimens, made of the 2024-T6 aluminum alloy.

Semantic sensor fusion for fault diagnosis in aircraft gas turbine engines

Data-driven fault diagnosis of a complex system such as an aircraft gas turbine engine requires interpretation of multi-sensor information to assure enhanced performance. This paper proposes feature-level sensor information fusion in the framework of symbolic dynamic filtering. This hierarchical approach involves construction of composite patterns consisting of: (i) atomic patterns extracted from single sensor data and (ii) relational patterns that represent the cross-dependencies among different sensor data. The underlying theories are presented along with necessary assumptions and the proposed method is validated on the NASA C-MAPSS simulation model of aircraft gas turbine engines.

Fractal analysis of crack initiation in polycrystalline alloys using surface interferometry

Microstructural degradation is a predominant source of damage in polycrystalline alloys that are commonly used in diverse applications. For early diagnosis and prognosis of failures, it is essential to understand the mechanisms of damage growth specifically in the crack initiation phase, which is still an intriguing phenomenon for scientists due to sensing inaccuracies and modeling uncertainties. Measurements of gradually evolving deformations on the material surface during crack initiation provide early warnings of forthcoming widespread damage. In this paper, a surface interferometer is used to generate 3-D surface profiles of polycrystalline alloy specimens under oscillating load. The concepts of fractal geometry are used to quantify the changes in the 3-D surface profiles as early indicators of damage evolution in the crack initiation phase.

Ultrasonic measurement of crack opening load for life-extending control of mechanical structures

This paper develops and validates a technique for real-time measurement of crack tip opening load using ultrasonic sensors and its application to life-extending control of mechanical structures. To experimentally validate the proposed measurement technique, fatigue tests have been conducted in the laboratory environment on center-notched 7075-T6 aluminium alloy specimens. The energy of reflected ultrasonic waves is used to detect crack closure and opening phenomena that lead to real-time measurement of crack opening load. Experimental results are compared with the Newmans crack opening stress model under constant amplitude cyclic loading. A life-extending control scheme is proposed by taking advantage of the real-time information on fatigue damage in mechanical structures.

In-situ fatigue damage monitoring using symbolic dynamic filtering of ultrasonic signals

Abstract This article presents a data-driven method of pattern identification for in-situ monitoring of fatigue damage in polycrystalline alloys that are commonly used in aerospace structures. The concept is built upon analytic signal space partitioning of ultrasonic data sequences for symbolic dynamic filtering of the underlying information. The statistical patterns of evolving damage are generated for real-time monitoring of the possible structural degradation under fatigue load. The proposed method is capable of detecting small anomalies (i.e. deviations from the nominal condition) in the material microstructure and thereby generating early warnings on damage initiation. The damage monitoring algorithm has been validated on time series data of ultrasonic sensors from a fatigue test apparatus, where the behavioural pattern changes accrue because of the evolving fatigue damage in polycrystalline alloys.

Online recursive estimation of remaining life using ultrasonic measurements

This article addresses diagnosis and prognosis of evolving fatigue crack damage in polycrystalline alloy structures. It presents a statistically inspired recursive method for in situ estimation of the remaining useful life in machinery components, based on real-time measurements. The underlying algorithm is built upon (a) symbolic dynamic filtering of (online) ultrasonic sensor data and (b) Karhunen–Loève decomposition of optical measurements for (off-line) construction of a stochastic model of fatigue crack propagation. The proposed method has been experimentally validated on a computer-instrumented and computer-controlled fatigue test apparatus for estimation of crack damage and prediction of the remaining useful life in test specimens, made of 7075-T6 aluminum alloys.

Optimal partitioning of ultrasonic data for fatigue damage detection

This paper presents an analytical tool for online fatigue damage detection in polycrystalline alloys that are commonly used in mechanical structures. The underlying theory is built upon symbolic dynamic filtering (SDF) that optimally partitions time series data for feature extraction and pattern classification. The proposed method has been experimentally validated on a fatigue test apparatus that is equipped with ultrasonics sensors and a traveling optical microscope for fatigue damage detection.

Early detection of fatigue damage using escort distributions of ultrasonic data sequences

This paper presents real-time detection of fatigue damage in mechanical structures using ultrasonic sensing methodology. The data-driven pattern identification method for anomaly detection is based on the tools derived from statistical mechanics and symbolic dynamics. The concept of escort distributions has been used to identify the behavioral patterns changes in complex systems due to gradual evolution of anomalies. The real-time information of evolving fatigue damage provides early warnings of forthcoming catastrophic failures. The anomaly detection method has been experimentally validated on poly-crystalline alloys using ultrasonic data generated from a fatigue damage testing apparatus.