Digital processing of speech signals

A novel feature extracted from a nonlinear time series is presented within the context of vibration-based damage detection in a system An eight-degree-of-freedom spring-mass-damper `structure' is considered with damage incurred by a stiffness degradation in one spring The system is excited with a chaotic input, and by tuning the Lyapunov exponents of the chaotic excitation to the dominant eigenvalue of the structure the dimensionality of the entire system is effectively controlled Both the input and output are viewed in state space as geometric objects, and the effect of the damage is shown to alter the geometric properties of the corresponding attractors at a local level, which may be captured in construction of the feature The utility of the feature is compared with that of a number of modal-based features and shown to be superior in resolving capability and in robustness

Vibration-based damage assessment utilizing state space geometry changes: local attractor variance ratio

A multivariate, attractor-based approach to structural health monitoring

The field of vibration based structural health monitoring involves extracting a ‘feature’ which robustly quantifies damage induced changes to the structure in the presence of ambient variation, that is, changes in ambient temperature, varying moisture levels, etc In this paper, we present an attractor-based feature derived from the field of nonlinear time-series analysis Emphasis is placed on the use of chaos for the purposes of system interrogation The structure is excited with the output of a chaotic oscillator providing a deterministic (low-dimensional) input Use is made of the Kaplan–Yorke conjecture in order to ‘tune’ the Lyapunov exponents of the driving signal so that varying degrees of damage in the structure will alter the state space properties of the response attractor The average local attractor variance ratio (ALAVR) is suggested as one possible means of quantifying the state space changes Finite element results are presented for a thin aluminum cantilever beam subject to increasing damage, as specified by weld line separation, at the clamped end Comparisons of the ALAVR to two modal features are evaluated through the use of a performance metric

Structural Health Monitoring Through Chaotic Interrogation

On the inverse power laws for accelerated random fatigue testing

This paper explores the application of statistical pattern recognition and machine learning techniques to vibration-based damage detection. First, the damage detection process is described in terms of a problem in statistical pattern recognition. Next, a specific example of a statistical-pattern-recognition-based damage detection process using a linear discriminant operator, ''Fisher's Discriminant'', is applied to the problem of identifying structural damage in a physical system. Accelerometer time histories are recorded from sensors attached to the system as that system is excited using a measured input. Linear Prediction Coding (LPC) coefficients are utilized to convert the accelerometer time-series data into multi-dimensional samples representing the resonances of the system during a brief segment of the time series. Fisher's discriminant is then used to find the linear projection of the LPC data distributions that best separates data from undamaged and damaged systems. The method i s applied to data from concrete bridge columns as the columns are progressively damaged. For this case, the method captures a clear distinction between undamaged and damaged vibration profiles. Further, the method assigns a probability of damage that can be used to rank systems in order of priority for inspection.

/pdf/damage-identification-with-linear-discriminant-operators-1v00a9v3au.pdf

Damage Identification with Linear Discriminant Operators

Accelerated vibration testing seeks to compress long service exposures to vibration into a reduced length laboratory test by increasing the amplitude and/or frequency of the applied inputs during the laboratory test relative to the amplitude and/or frequency experienced during service This testing procedure provides an important tool that can reduce testing time associated with a new design and reduce time to market This paper will summarize current methods that are employed to develop accelerated testing criteria and will highlight the attributes and limitations of these methods Typically there are two ways of accelerating vibration testing The first method involves testing at fewer cycles but at higher amplitude levels; and the second method involves testing at higher frequencies (rates) A combination of the two is also an option Development of an accelerated test based on either of these methods requires a priori knowledge of the controlling failure mechanisms The review will begin with a discussion of Miner's Rule for developing accelerated testing criteria This rule, which is based on a linear damage accumulation assumption, was first proposed in the 1940's for fatigue failures of ductile metals loaded repetitively in bending Confounding factors associated with developing accelerated testing criteria for nonlinear vibration response (eg, rattling of components) will be illustrated with an example

T. A. Duffey

Papers

Damage Identification with Linear Discriminant Operators

A review of methods for developing accelerated testing criteria