Wyniki wyszukiwania - BazTech

1

Zastosowanie konwolucyjnej sieci neuronowej do identyfikacji struktury przepływu ciecz-gaz w rurociągu z wykorzystaniem histogramów sygnałów z detektorów scyntylacyjnych

Ochał Piotr, Hanus Robert, Zych Marcin

Przegląd Elektrotechniczny

|

2022

|

R. 98, nr 11

185--189

PL

Znajomość struktury przepływów dwufazowych w rurociągach jest niezbędna dla oceny prawidłowego przebiegu wielu procesów przemysłowych. W niniejszej pracy zastosowano Konwolucyjną Sieć Neuronową (CNN) do analizy histogramów sygnałów uzyskanych dla przepływu ciecz-gaz z wykorzystaniem absorpcji promieniowania gamma. Eksperymenty przeprowadzono na laboratoryjnej instalacji hydraulicznej wyposażonej w radiometryczny układ pomiarowy. W pracy zbadano cztery typy przepływu: rzutowy, tłokowy, tłokowo-pęcherzykowy i pęcherzykowy. Stwierdzono, że sieć CNN poprawnie rozpoznaje strukturę przepływu w ponad 90% przypadków.

EN

Knowledge of the two-phase flow structure is essential for the proper conduct of industrial processes. In this work, the Convolutional Neural Network (CNN) is applied for analysis of histograms of signals obtained for liquid-gas flow by use gamma-ray absorption. The experiments were carried out on the laboratory hydraulic installation fitted with radiometric measurement system. Four types of flow regimes as plug, slug, bubble, and transitional plug – bubble were studied in this work. It was found that the CNN network correctly recognize the flow structure in more than 90% of cases.

2

Evaluation of liquid-gas flow in pipeline using gamma-ray absorption technique and advanced signal processing

Hanus Robert, Zych Marcin, Mosorov Volodymyr, Golijanek-Jędrzejczyk Anna, Jaszczur Marek, Andruszkiewicz Artur

Metrology and Measurement Systems

|

2021

|

Vol. 28, nr 1

145--159

EN

Liquid-gas flows in pipelines appear in many industrial processes, e.g. in the nuclear, mining, and oil industry. The gamma-absorption technique is one of the methods that can be successfully applied to study such flows. This paper presents the use of the gamma-absorption method to determine the water-air flow parameters in a horizontal pipeline. Three flow types were studied in this work: plug, transitional plug-bubble, and bubble one. In the research, a radiometric set consisting of two Am-241 sources and two NaI(TI) scintillation detectors have been applied. Based on the analysis of the signals from both scintillation detectors, the gas phase velocity was calculated using the cross-correlation method (CCM). The signal from one detector was used to determine the void fraction and to recognise the flow regime. In the latter case, a Multi-Layer Perceptron-type artificial neural network (ANN) was applied. To reduce the number of signal features, the principal component analysis (PCA) was used. The expanded uncertainties of gas velocity and void fraction obtained for the flow types studied in this paper did not exceed 4.3% and 7.4% respectively. All three types of analyzed flows were recognised with 100% accuracy. Results of the experiments confirm the usefulness of the gamma-ray absorption method in combination with radiometric signal analysis by CCM and ANN with PCA for comprehensive analysis of liquid-gas flow in the pipeline.

3

Wpływ oczyszczenia danych pomiarowych na wyznaczenie wartości przepływu masowego za pomocą kryzy

Golijanek-Jędrzejczyk Anna, Hanus Robert, Zych Marcin, Mrowiec Andrzej

Przegląd Elektrotechniczny

|

2019

|

R. 95, nr 11

56--58

PL

W artykule przedstawiono analizę oczyszczenia eksperymentalnych wyników pomiaru przepływu masowego za pomocą kryzy z systematycznych oddziaływań. Zaprezentowano także wpływ przeprowadzenia takiej procedury na zmianę wartości estymaty menzurandu przepływu oraz wartości niepewności typu A. Wyniki uzyskanych analiz pozwalają potwierdzić zasadność kontrolowania składowych systematycznych w wynikach pomiaru przepływu cieczy i ich eliminacji przed przeprowadzeniem procedury szacowania niepewności pomiaru.

EN

The article presents an analysis of the purification of experimental mass flow measurement results from systematic interactions. The influence of performing such a procedure on changing the value of the flow estimation and the uncertainty of type A is presented. The results of the obtained analyzes allow to confirm the validity of controlling the systematic components in the results of liquid flow measurement and their elimination before conducting the procedure of estimating the uncertainty of measurement.

4

Improved recognition of rock formation on the basis of well logging and laboratory experiments results using factor analysis

Puskarczyk Edyta, Jarzyna Jadwiga A., Wawrzyniak-Guz Kamila, Krakowska Paulina I., Zych Marcin

Acta Geophysica

|

2019

|

Vol. 67, no. 6

1809--1822

EN

Several data sets from the Silurian and Ordovician formations from three wells on the shore of Baltic Basin in Northern Poland prepared on the basis of well logging data and results of their comprehensive interpretation were used in factor analysis. The goal of statistical analysis was structure recognition of data and proper selection of parameters to limit the number of variables in study. The top priority of research was recognition of specific features of claystone/mudstone formations predisposing them to be potential shale gas deposits. The identified data scheme based on data from one well, was then applied to: 1) well 2 and well 3 separately, 2) combined data from three wells, 3) depth intervals treated as sweet spots, i.e., formations of high hydrocarbon potential. Numbers of samples from well logging were proportional to number of laboratory data from individual formations. The extended data set comprising all available log samples in explored formations was also prepared. Outcomes from standard (Triple Combo—natural gamma log, resistivity log, neutron log and bulk density log and Quad Combo—with addition of sonic log and spectral gamma log) and sophisticated (GEM™—Elemental Analysis Tool, Wave Sonic and Nuclear Magnetic Resonance—NMR) logs were the basis for data sets. Finally, laboratory data set of huge amount of variables from elemental, mineralogical, geochemical and petrophysical laboratory experiments was built and verified in FA to select the most informative components. Conclusions on the data set size, number of factors and type of variables were drawn.