The utilization of signal processing techniques in nondestructive testing, especially in ultrasonics, is widespread. Signal averaging, matched filtering, frequency spectrum analysis, neural nets, and autoregressive analysis have all been used to analyze ultrasonic signals. The Wavelet Transform (WT) is the most recent technique for processing signals with time-varying spectra. Interest in wavelets and their potential applications has resulted in an explosion of papers; some have called the wavelets the most significant mathematical event of the past decade. In this work, the Wavelet Transform is utilized to improve ultrasonic flaw detection in noisy signals as an alternative to the Split-Spectrum Processing (SSP) technique. In SSP, the frequency spectrum of the signal is split using overlapping Gaussian passband filters with different central frequencies and fixed absolute bandwidth. A similar approach is utilized in the WT, but in this case the relative bandwidth is constant, resulting in a filter bank with a self-adjusting window structure that can display the temporal variation of the signal's spectral components with varying resolutions. This property of the WT is extremely useful for detecting flaw echoes embedded in background noise. The detection of ultrasonic pulses using the wavelet transform is described and numerical results show good detection even for signal-to-noise ratios (SNR) of -15 dB. The improvement in detection was experimentally verified using steel samples with simulated flaws.

Signal detection and noise suppression using a wavelet transform signal processor: application to ultrasonic flaw detection

Improving the time-resolution and signal-to-noise ratio of ultrasonic NDE signals

Nondestructive evaluation (NDE) techniques are used widely in many manufacturing and operational areas of industry. NDE is a highly statistical science that has developed rapidly over the past decade with relatively little input from individuals with forma1 training in statistics. This review article describes some of the many methods and applications of NDE, including eddy-current methods for detecting fatigue in jet-engine turbine disks, ultrasonic methods for detecting flaws in forged titanium, and radiographic methods for detecting out-of-control conditions in a casting process. We will also outline the important potential for increased use of modern methods of statistics to improve the practice of nondestructive evaluation. Areas of statistics important in NDE include data analysis and signal processing, decision theory, screening and calibration, the assessment of measurement capability, and the use of designed experiments for improvement of measurement/detection processes.

Applications of statistical methods to nondestructive evaluation

Ultrasonic guided waves offer the possibility of inspecting large areas of structures from a small number of sensor positions. However, inspection of complex structures is difficult as the reflections from different features overlap. Estimating the number and amplitude of the wave packets in ultrasonic time traces is crucial for the development of a guided wave inspection system, in order to detect and locate damage. Deconvolution has been extensively used in geophysical applications to resolve overlapping echoes in the recorded signals. The main objective of this work was to evaluate the applicability of the deconvolution approach for enhancing the resolution of ultrasonic time traces in structural health monitoring (SHM). Numerical simulations on strongly overlapping signals were carried out to evaluate the performance of the two techniques that have been considered: (i) Wiener filter, (ii) single most likely replacement. It was shown that the relatively narrow bandwidth of the input signals and phase s...

Potential and Limitations of a Deconvolution Approach for Guided Wave Structural Health Monitoring

Further improvement of temporal resolution of seismic data by autoregressive (AR) spectral extrapolation

Flaw signals measured in ultrasonic testing include the effects of the measurements system and are corrupted by noise The measurement system response is both bandlimited and frequency dependent within the bandwidth, resulting in measured signals which are blurred and distorted estimates of actual flaw signatures The Wiener filter can be used to estimate the flaw's scattering amplitude by removing the effect of the measurement system in the presence of noise A method is presented for implementing an optimal form of the Wiener filter that requires only estimates of the noise distribution parameters The theoretical error for scattering amplitude estimation, assuming various levels of available prior information, is analyzed Three estimation techniques, one a maximum-likelihood based method and the other two residual-sum-of-squares methods, are formulated and tested The results demonstrate that any of the three approaches could be used to optimally implement the alternative form of the Wiener filter with limited prior information >

Flaw signature estimation in ultrasonic nondestructive evaluation using the Wiener filter with limited prior information

In the evolution of 6G and Future Networks, a dynamic, flexible, learning-based security architecture will be essential with the ability to handle both current and evolving cybersecurity threats. This is specially critical with future networks' increased reliance on distributed learning-based approaches for operation. To address this challenge, a distributed learning framework must provide security and trust in an integrated fashion. In contrast to existing approach such as federated learning (FL), that update parameters of a shared model, this work proposes an architecture that is capable of integrating advanced learning with real-time digital forensics, e.g. monitoring compute and storage resources. With real-time monitoring, it is possible to develop a learning-based, real-time Zero-Trust Architecture (ZTA) to achieve the high levels of security. The proposed architecture, serves as a framework to enable and spur innovation, where new machine learning based techniques can be developed for enhanced real-time, adaptive and proactive security, thus, embedding future networks' security with learning-based ZTA elements.

A Learning-Based Zero-Trust Architecture for 6G and Future Networks

In the evolution of artificial Intelligence (AI) and machine learning (ML), reasoning, knowledge representation, planning, learning, natural language processing, perception, and the ability to move and manipulate objects have been widely used. These features enable the creation of intelligent mechanisms for decision support to overcome the limits of human knowledge processing. In addition, ML algorithms enable applications to draw conclusions and make predictions based on existing data without human supervision, leading to quick, near-optimal solutions even in problems with high dimensionality. Hence, autonomy is a key aspect of current and future AI/ML algorithms. This chapter focuses on the development and implementation of AI/ML technologies for 5G and future networks. The objective is to illustrate how these technologies can be smoothly migrated into 5G systems to increase their performance and to decrease their cost. To that end, this chapter presents the Drivers, Needs, Challenges, Enablers, and Potential Solutions identified for the AI/ML field as applicable to future networks over three-, five-, and ten-year horizons. AI/ML applications for 5G are wide and diverse. In this document, some of the key areas are described which includes networking, securing, cloud computing and others. Over time, this white paper will evolve to encompass even more areas where AI/ML technologies can improve future network performance objectives.

M.A. Enright

Papers

Flaw signature estimation in ultrasonic nondestructive evaluation using the Wiener filter with limited prior information

A Learning-Based Zero-Trust Architecture for 6G and Future Networks

Artificial Intelligence And Machine Learning