Pradeep Ramuhalli

Electromagnetic NDE signal inversion by function-approximation neural networks

In the magnetic flux leakage (MFL) method of nondestructive testing commonly used to inspect ferromagnetic materials, a crucial problem is signal inversion, wherein the defect profiles must be recovered from measured signals. This paper proposes a neural-network-based inversion algorithm to solve the problem. Neural networks (radial-basis function and wavelet-basis function) are first trained to approximate the mapping from the signal to the defect space. The trained networks are then used iteratively in the algorithm to estimate the profile, given the measurement signal. The paper presents the results of applying the algorithm to simulated MFL data.

Integrated Imaging and Vision Techniques for Industrial Inspection

This pioneering text/reference presents a detailed focus on the use of machine vision techniques in industrial inspection applications. An internationally renowned selection of experts provide insights on a range of inspection tasks, drawn from their cutting-edge work in academia and industry, covering practical issues of vision system integration for real-world applications. Topics and features:presents a comprehensive review of state-of-the-art hardware and software tools for machine vision, and the evolution of algorithms for industrial inspection; includes in-depth descriptions of advanced inspection methodologies and machine vision technologies for specific needs; discusses the latest developments and future trends in imaging and vision techniques for industrial inspection tasks; provides a focus on imaging and vision system integration, implementation, and optimization; describes the pitfalls and barriers to developing successful inspection systems for smooth and efficient manufacturing process.

Neural network-based inversion algorithms in magnetic flux leakage nondestructive evaluation

Magnetic flux leakage (MFL) methods are commonly used in the nondestructive evaluation (NDE) of ferromagnetic materials. An important problem in MFL NDE is the determination of flaw parameters such as the flaw length, depth, and shape (profile) from the measured values of the flux density B. Commonly used methods use a forward model in a loop to determine B for a given set of flaw parameters. This approach iteratively adjusts the flaw parameters to minimize the error between the measured and predicted values of B. This article proposes the use of neural networks as forward models. The proposed approach uses two neural networks in feedback configuration—a forward network and an inverse network. The second network is used to predict the profile given the measured value of B, and acts to constrain the solution space. Results of applying these methods to MFL data obtained from a two-dimensional finite-element model, with rectangular flaws of various dimensions, are presented.

Finite-element neural networks for solving differential equations

The solution of partial differential equations (PDE) arises in a wide variety of engineering problems. Solutions to most practical problems use numerical analysis techniques such as finite-element or finite-difference methods. The drawbacks of these approaches include computational costs associated with the modeling of complex geometries. This paper proposes a finite-element neural network (FENN) obtained by embedding a finite-element model in a neural network architecture that enables fast and accurate solution of the forward problem. Results of applying the FENN to several simple electromagnetic forward and inverse problems are presented. Initial results indicate that the FENN performance as a forward model is comparable to that of the conventional finite-element method (FEM). The FENN can also be used in an iterative approach to solve inverse problems associated with the PDE. Results showing the ability of the FENN to solve the inverse problem given the measured signal are also presented. The parallel nature of the FENN also makes it an attractive solution for parallel implementation in hardware and software.

A Recursive Bayesian Estimation Method for Solving Electromagnetic Nondestructive Evaluation Inverse Problems

Estimating flaw profiles from measurements is a typical inverse problem in electromagnetic nondestructive evaluation (NDE). This paper proposes a novel state-space approach for solving such inverse problems. The approach is robust in the presence of measurement noise. It formulates the inverse problem as a tracking problem with state and measurement equations. The state-space model resembles the classical discrete-time tracking problem. The model allows recursive Bayesian nonlinear filters based on sequential Monte Carlo methods to be applied in conjunction with numerical models that represent the measurement process (i.e., solution of the forward problem). We apply our approach to simulated eddy-current and magnetic flux leakage NDE measurements (with and without measurement noise) from known flaw shapes, and the results indicate the feasibility and robustness of the proposed method.

Prognostics and life beyond 60 years for nuclear power plants

Safe, secure, reliable, and sustainable energy supply is vital for advanced and industrialized life styles. To meet growing energy demand there is interest in longer-term operation for the existing nuclear power plant fleet and enhancing capabilities in new build. There is increasing use of condition-based maintenance for active components and growing interest in deploying on-line monitoring instead of periodic in-service inspection for passive systems. Opportunities exist to move beyond monitoring and diagnosis based on pattern recognition and anomaly detection to prognostics with the ability to provide an estimate of remaining useful life. The adoption of digital I&C systems provides a framework within which added functionality including on-line monitoring can be deployed, and used to maintain and even potentially enhance safety, while at the same time improving planning and reducing both operations and maintenance costs.

Particle-Filter-Based Multisensor Fusion for Solving Low-Frequency Electromagnetic NDE Inverse Problems

Flaw profile characterization from nondestructive evaluation (NDE) measurements is a typical inverse problem. A novel transformation of this inverse problem into a tracking problem and subsequent application of a sequential Monte Carlo method called particle filtering has been proposed by the authors in an earlier publication. In this paper, the problem of flaw characterization from multisensor data is considered. The NDE inverse problem is posed as a statistical inverse problem, and particle filtering is modified to handle data from multiple measurement modes. The measurement modes are assumed to be independent of each other with principal component analysis used to legitimize the assumption of independence. The proposed particle-filter-based data fusion algorithm is applied to experimental low-frequency NDE data to investigate its feasibility.

Combining multiple nondestructive inspection images with a generalized additive model

In this paper, multiple nondestructive inspection (NDI) images are combined with a generalized additive model to achieve a more precise and reliable assessment of hidden corrosion in aircraft lap joints. Two inspection techniques are considered in this study. One is the conventional multi-frequency eddy current testing technique and the other is the pulsed eddy current technique. To characterize the thickness loss or equivalently to achieve a quantitative measure of corrosion, multiple NDI images are fused to produce a thickness map that reflected the amount of corrosion damage. These results are further compared with corresponding digital x-ray thickness maps, which are obtained by mapping the remaining thickness after the specimen is dissembled and all the corrosion products are cleaned. Experimental results demonstrate that the proposed algorithms outperform the traditional calibration method aligned with a single testing approach.

Towards a theory of autonomous reconstitution of compromised cyber-systems

Effective reconstitution approaches for cyber systems are needed to keep critical infrastructure operational in the face of an intelligent adversary. The reconstitution response, including recovery and adaptation, may require significant reconfiguration of the system at all levels to render the cyber-system resilient to ongoing and future attacks or faults while maintaining continuity of operations. A theoretical basis for optimal dynamic reconstitution is needed to address the challenge of ensuring that dynamic reconstitution is optimal with respect to resilience metrics, and is being developed and evaluated in this project. Such a framework provides the technical basis for evaluating cyber-defense and reconstitution approaches. This paper describes a preliminary framework that may be used to develop and evaluate concepts for effective autonomous reconstitution of compromised cyber systems.

Multichannel signal processing methods for ultrasonic nondestructive evaluation

Ultrasonic inspection methods are commonly used in the nondestructive evaluation of welds to detect flaws in the weld region. An important characteristic of ultrasonic inspection is the ability to identify the type of discontinuity that gives rise to a particular signal. Standard techniques rely on differences in individual A-scans to classify the signals. This paper proposes an ultrasonic signal classification technique based on the information in a group of signals. The approach is based on a 2-dimensional transform and principal component analysis, for generating a reduced dimensional feature vector for classification. Results of applying the technique to data obtained from the inspection of welds are presented.