Title: Total organic carbon from well logging - statistical approach, Polish shale gas formation case study
Authors: Jadwiga A. Jarzyna; Paulina I. Krakowska; Edyta Puskarczyk; Kamila Wawrzyniak-Guz; Marcin Zych
Addresses: Department of Geophysics, Faculty of Geology Geophysics and Environmental Protection, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Krakow, Poland ' Department of Geophysics, Faculty of Geology Geophysics and Environmental Protection, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Krakow, Poland ' Department of Geophysics, Faculty of Geology Geophysics and Environmental Protection, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Krakow, Poland ' Department of Geophysics, Faculty of Geology Geophysics and Environmental Protection, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Krakow, Poland ' Department of Geophysics, Faculty of Geology Geophysics and Environmental Protection, AGH University of Science and Technology, Al. Mickiewicza 30, 30-059 Krakow, Poland
Abstract: Advanced statistical methods - artificial neural networks (ANN) and support vector machines (SVM) were used to calculate total organic carbon (TOC) on the basis of well logging. Data from three wells from the Silurian and Ordovician formations in the Baltic Basin from the north Poland were used. In learning procedures TOC data as Rock-Eval pyrolysis results were applied. Research was undertaken in four steps aiming to determine the best ANN in aspect of number and type of input variables. Regarding increase of cases number used in learning process data from two wells were combined and next, determined ANN and SVR were used to predict TOC. There were also made tests of number of input variables (results of standard and sophisticated well logs and laboratory data). Results obtained using standard logs when number of available cases for learning was big enough did not gave way to results based on very many input variables. [Received: October 29, 2016; Accepted: October 21, 2017]
Keywords: total organic carbon; TOC; Polish shale gas formations; Baltic Basin; artificial neural networks; ANN; support vector machines; SVM.
DOI: 10.1504/IJOGCT.2019.102784
International Journal of Oil, Gas and Coal Technology, 2019 Vol.22 No.2, pp.140 - 162
Received: 29 Oct 2016
Accepted: 21 Oct 2017
Published online: 07 Oct 2019 *