Processing of massive Rutherford Back-scattering Spectrometry data by artificial neural networks

Guimarães, Renato da S.; Silva, Tiago Fiorini da; Rodrigues, C. L.; Tabacniks, M.H.; Bach, Simon; Burwitz, V. V.; Hiret, P.; Mayer, M.

doi:10.1016/j.nimb.2021.02.010

Cited by 9 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To obtain the accuracy and precision of the low- and high-temperature domain ANN analysis, ten ANNs are trained independently using the same training set and employed for data analysis. For each dual-spectrum input, this analysis with the independently trained ANNs results in the mean value and standard deviation of each ANN output 30 .…”

Section: Resultsmentioning

confidence: 99%

“…The user bias can be minimized by applying a machine learning approach. The latter mainly involves artificial neural networks (ANN), which relate a single RBS spectrum to the corresponding compositional depth profile 12 , 29 , 30 . Analogously to the transmission and processing of electric signals in a biological network of neurons, the architecture of a multilayer perceptron ANN consists of one layer of input nodes and one layer of output nodes, separated by one or more hidden layers.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Enhanced accuracy through machine learning-based simultaneous evaluation: a case study of RBS analysis of multinary materials

Magchiels,

Claessens,

Meersschaut

et al. 2024

Sci Rep

View full text Add to dashboard Cite

We address the high accuracy and precision demands for analyzing large in situ or in operando spectral data sets. A dual-input artificial neural network (ANN) algorithm enables the compositional and depth-sensitive analysis of multinary materials by simultaneously evaluating spectra collected under multiple experimental conditions. To validate the developed algorithm, a case study was conducted analyzing complex Rutherford backscattering spectrometry (RBS) spectra collected in two scattering geometries. The dual-input ANN analysis excelled in providing a systematic analysis and precise results, showcasing its robustness in handling complex data and minimizing user bias. A comprehensive comparison with human supervision analysis and conventional single-input ANN analysis revealed a reduced susceptibility of the dual-input ANN analysis to inaccurately known setup parameters, a common challenge in material characterization. The developed multi-input approach can be extended to a wide range of analytical techniques, in which the combined analysis of measurements performed under different experimental conditions is beneficial for disentangling details of the material properties.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Enhanced accuracy through machine learning-based simultaneous evaluation: a case study of RBS analysis of multinary materials

Magchiels,

Claessens,

Meersschaut

et al. 2024

Sci Rep

View full text Add to dashboard Cite

show abstract

“…As a case study, we produced activation maps for the predictions of roughnesses and thicknesses of both the carbon and molybdenum layers, for the ANN trained to analyze RBS spectra of the internal walls of the W7x fusion reactor [5]. We aim at verifying the model stability, thus enhancing the reliability to use ANN to process RBS spectra.…”

Section: Resultsmentioning

confidence: 99%

“…Thus, an ANN was trained to speed up and automate the analysis. Therefore, after the training, the neural network evaluates the RBS spectrum of the samples and predicts the elementary concentrations and the layer structure [5]. Moreover, it was possible to compare the performances of manual analysis, automatic fit, and ANN evaluation.…”

Section: Introductionmentioning

confidence: 99%

What do artificial neural networks learn? A study for analysis of RBS spectra

Oliveira

Silva

2022

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

The massive analysis of Rutherford Backscattering Spectrometry (RBS) data is complex. When the data is processed manually, it requires a long time of an experienced person. Artificial Neural Networks (ANN) can analyze, speed up, and automate data processing. In fact, after training, the ANN processes one RBS spectrum in a fraction of a second with the advantage of keeping the consistency over the whole set of spectra. Our group used ANN to process a large set of RBS spectra from the inner walls of the vacuum chamber of the W7-X fusion reactor. In this work, we used a perturbation-based method to study the local explanations of the neural network predictions. In this method, we apply small perturbations to the inputs. Then, the outputs’ variations are evaluated. Thereby activation maps were created to visualize how sensitive the ANN is to perturbations. The activation maps enable the identification of the parts of the spectrum the neural network is getting information to make predictions. Therefore, we can better understand the behavior of the machine learning model and verify if the neural network learned the features of the spectra similarly to humans.

show abstract

“…(a) About 500 spectra from pre-exposure measurements were fitted using automated fits. (b) Two different artificial neural networks (ANNs) were trained [28]. The model ANN 2 consisted of a carbon substrate and two layers: a Mo interlayer and a top carbon layer.…”

Section: Erosion/deposition Marker Layersmentioning

confidence: 99%

Carbon erosion/deposition on the divertor of W7-X during the operational period OP 1.2b

et al. 2022

Self Cite

View full text Add to dashboard Cite

Carbon net erosion/deposition at the Test Divertor Unit (TDU) of Wendelstein 7-X (W7-X) were measured for the discharge period OP 1.2b in the year 2018 using 18 specially prepared target elements in all 10 TDUs. These had lengths between 30 and 60 cm and were coated with marker layers for erosion/deposition investigations of 5-10 µm carbon on top of about 300 nm molybdenum. The marker layer thicknesses were measured by elastic backscattering spectrometry (EBS) before and after plasma exposure using 2.5 MeV protons; the surface morphology was investigated using scanning electron microscopy (SEM) and focused ion beam cross-sectioning (FIB), the surface roughness was determined using a 2-dimensional optical profiler. Plasma-exposed surfaces were considerably smoother than unexposed surfaces with decreased mean roughness and a shift of the inclination angle distribution towards lower values. The erosion on the 10 TDUs was unequal within a factor of about two. During the discharge period in total 20.4±5.7 g carbon were eroded from the 10 TDUs. Adjacent to the strike line some deposition of carbon is observed. Compared to the discharge period OP 1.2a in the year 2017, the net carbon erosion rate dropped by a factor of 5-6 due to regular boronizations, which reduced the oxygen (and subsequently also the carbon) content in the plasma by 1-2 orders of magnitude. The significance of erosion/deposition processes for long-pulse discharges is discussed.

show abstract

Processing of massive Rutherford Back-scattering Spectrometry data by artificial neural networks

Cited by 9 publications

References 14 publications

Enhanced accuracy through machine learning-based simultaneous evaluation: a case study of RBS analysis of multinary materials

Enhanced accuracy through machine learning-based simultaneous evaluation: a case study of RBS analysis of multinary materials

What do artificial neural networks learn? A study for analysis of RBS spectra

Carbon erosion/deposition on the divertor of W7-X during the operational period OP 1.2b

Contact Info

Product

Resources

About