Ultrasonic guided wave estimation of minimum remaining wall thickness using Gaussian process regression

Tabatabaeipour, Morteza; Tzaferis, Konstantinos; McMillan, Ross; Jackson, William E.; Dobie, Gordon; Edwards, R. S.; Trushkevych, Oksana; Gachagan, Anthony

doi:10.1016/j.matdes.2022.110990

Cited by 8 publications

(6 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The second approach considered in the context of this study is the GPR method used to establish the data mapping from an input x-vector to an output y-vector based on a set of a training data denoted 𝐷 = (𝑥 𝑖 , 𝑦 𝑖 ) 𝑖=𝟏 𝑛 , where n is the number of data points. So, a noisy GPR model is expressed as [34][35]:…”

Section: B Gaussian Process Regression (Gpr) Modelmentioning

confidence: 99%

Faults prediction and monitoring of complex processes using an ensemble of machine learning regression models: application to the Tennessee Eastman Process

Rezgui

Rezki²,

Kerrouchi³

2023

Preprint

View full text Add to dashboard Cite

Modern industrial installations are a source of big amount of data, these data serve as a means of monitoring and control and can also be used for the prediction of the parameters characterizing the supervised process and thus the anomalies. For this, there are several machine learning regression models that can be considered in order to select the best prediction tool. Our study consists in selecting the best means of prediction of the different operating parameters of the Tennessee Eastman (TEP) process. Indeed, this consists in selecting the best model that can ensure an efficient and cost-effective prognosis and predictive monitoring system. Four regression models were considered during our comparative study: the Support Vector Regression (SVR), Gaussian Process Regression (GPR), Decision Tree Regression (DTR) and Ensemble of Learners approaches for Regression (ELR). As evaluation criteria, we chose the MSE regression error by re-substitution, the MSE regression loss for the cross-validation kernel, the optimization of the hyper-parameters and the training time.

show abstract

Section: B Gaussian Process Regression (Gpr) Modelmentioning

confidence: 99%

Faults prediction and monitoring of complex processes using an ensemble of machine learning regression models: application to the Tennessee Eastman Process

Rezgui

Rezki²,

Kerrouchi³

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…To increase the amount of information in the map produced by the system thickness estimation of the minimum remaining wall thickness using Gaussian process regression machine learning [22]. This increases the amount of information on the map, which is better for the inspector when choosing the high detail method of inspection to evaluate an area of interest, and it also increases the certainty of the type of feature under inspection.…”

Section: Minimum Remaining Wall Thickness Estimation Using Guided Wavesmentioning

confidence: 99%

Crawler-Based Automated Non-Contact Ultrasonic Inspection of Large Structural Assets

McMillan

Tabatabaeipour

Tzaferis

et al. 2022

2022 49th Annual Review of Progress in Quantitative Nondestructive Evaluation

View full text Add to dashboard Cite

This paper presents an update on the progress of developing a crawler-based automated non-contact ultrasonic inspection system for the evaluation of large structural assets. The system presented is a significant improvement on current robotic NDT crawlers and aims to greatly reduce the time of inspection by creating an internal feature map of the subject in a Simultaneous Localisation And Mapping (SLAM) style method instead of using a lawnmower scanning style where all areas are scanned regardless if they contain features or are featureless. This map will be generated through rapid automated path planning and scanning and will show the location of potential areas of interest, where then, the appropriate method of inspection can be used for a high detailed evaluation. Current and ongoing work presented is as follows; the use of guided waves as the sensory input of an occupancy grid map; evaluating guided wave modes to find the mode most appropriate for this system; minimum thickness estimation using machine learning; improving the transducer setup using a unidirectional transmitter.

show abstract

“…In the presented work, the HT is used to derive the phase relation between an emitter and receiver via emitter-receiver signal combinations by using a linear sensor array consisting of several piezoelectric elements [15]. In [16], the IP characteristics of shear horizontal waves are used for the detection of various defects in an aluminum specimen. The shift in the IP of the shear horizontal waves shows different patterns for different defects.…”

Section: Introductionmentioning

confidence: 99%

The Guided Ultrasonic Wave Oscillation Phase Relation between the Surfaces of Plate-like Structures of Different Material Settings

et al. 2023

View full text Add to dashboard Cite

Lamb waves occur in thin-walled structures in two wave modes—the symmetric and the antisymmetric mode. Their oscillation on the structures‘ surfaces is either in phase (symmetric) or shifted by a phase angle of π (antisymmetric). In this work, a method is developed by which to compare the surfaces’ oscillation phase relation to answer the question of whether fiber metal laminates show the same surface oscillation behavior as described for metals. The evaluation of time signals regarding the instantaneous phase angle is performed by using the continuous wavelet transformation and the short-time Fourier transformation. Numerical simulations utilizing the finite element method provide time signals from the top and bottom surface of different thin-walled structures of different material settings and configurations. The numerically obtained time signals are evaluated by the developed methods with respect to the oscillation phase. Subsequently, the oscillation phase is evaluated experimentally for the wave propagation in a fiber metal laminate. It is shown that the method based on the continuous wavelet transformation is suitable for the evaluation of oscillation phase relations in time signals. Additionally, it is proven that fiber metal laminates show only two phase relations, which indicates the occurrence of Lamb waves.

show abstract

Ultrasonic guided wave estimation of minimum remaining wall thickness using Gaussian process regression

Cited by 8 publications

References 40 publications

Faults prediction and monitoring of complex processes using an ensemble of machine learning regression models: application to the Tennessee Eastman Process

Faults prediction and monitoring of complex processes using an ensemble of machine learning regression models: application to the Tennessee Eastman Process

Crawler-Based Automated Non-Contact Ultrasonic Inspection of Large Structural Assets

The Guided Ultrasonic Wave Oscillation Phase Relation between the Surfaces of Plate-like Structures of Different Material Settings

Contact Info

Product

Resources

About